General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.
Users may download and print one copy of any publication from the public portal for the purpose of private study or research.
You may not further distribute the material or use it for any profit-making activity or commercial gain
You may freely distribute the URL identifying the publication in the public portal If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from orbit.dtu.dk on: Jan 19, 2021
Identifying simple and cost-effective gear solutions for an effective implementation ofthe new EU Common Fisheries Policy (CFP)
Melli, Valentina
Publication date:2019
Document VersionPublisher's PDF, also known as Version of record
Link back to DTU Orbit
Citation (APA):Melli, V. (2019). Identifying simple and cost-effective gear solutions for an effective implementation of the newEU Common Fisheries Policy (CFP). Technical University of Denmark.
Identifying simple and cost-effective gear solutions for an effective implementation of the new EU Common Fisheries Policy (CFP)
PhD Thesis
By Valentina Melli
DTU Aqua National Institute of Aquatic Resources
Identifying simple and cost-effective gear solutions for
an effective implementation of the new EU Common
Fisheries Policy (CFP)
Ph.D. Thesis, 2018
Valentina Melli
Technical University of Denmark
National Institute of Aquatic Resources
Section for Management Systems - Fisheries Technology, Hirtshals, Denmark
”It is not the strongest of the species that survives, nor the most intelligent, it is
the one that is most adaptable to change” - Charles Darwin
2
3
Preface
The present thesis is submitted in partial fulfilment of the requirements for obtaining a
Doctor of Philosophy (Ph.D.) degree. The thesis consists of a review and four supporting
papers. Three papers are published and one is a manuscript ready for publication.
I wish to express my sincere gratitude to my supervisors: Dr Ludvig A. Krag, Dr Junita D.
Karlsen and Prof Henrik Gislason from the Technical University of Denmark – National
Institute of Aquatic Resources (DTU Aqua). Their support and guidance have been
invaluable for me and this dissertation. Further, I would like to thank Dr Bent Herrmann
from SINTEF for the endless hours spent developing analytical facilities for my analyses.
I also owe a debt of gratitude to:
- all my highly appreciated colleagues in the Fisheries Technology group for many
inspiring talk and discussions, and for being always there whenever I needed;
- the amazingly skilled DTU Aqua Fisheries technicians for measuring thousands
and thousands of fish and Nephrops with me, helping to deal with any technical
challenge during the experimental trials and most of all, for being always
interested and passionate about the work;
- Helle Andersen for helping me extract the data, with precious care and precision;
- my colleague and friend Tiago Malta, for sharing with me every very step of these
three years. I couldn’t have asked for a better person to share the office with;
- my dear partner Marco for following me in this adventure and all over the world.
You made me brave.
Finally, I would like to acknowledge the financial support granted by the European
Maritime and Fisheries Fund and the Ministry of Environment and Food of Denmark that
made the research described in the supporting papers possible. Projects: FlexSelect –
Scaring lines, an innovative and flexible solution for the Nephrops fishery (Grant
Agreement No 33113-I-16-068) and Vision - Development of an optimal and flexible
selective system for trawl by use of new technology and underexploited fish behaviour
(Grant Agreement No 33113-I-16-015).
Hirtshals, December 2018
Valentina Melli
4
5
Table of contents
Preface ................................................................................................................. 3
Dansk resumé (abstract in Danish) ....................................................................... 9
Abstract ............................................................................................................... 11
1. Introduction ..................................................................................................... 13
2. The outcomes of a discard ban: a global overview ......................................... 17
3. Case study fishery: why the Nephrops trawl fishery? ...................................... 19
3.1 Fishing dynamics ....................................................................................... 20
3.1.1 Nephrops catchability .......................................................................... 21
3.1.2 Fish catchability ................................................................................... 22
3.2 Bycatch reduction measures before the landing obligation........................ 24
3.3 Bycatch reduction under the landing obligation ......................................... 24
4. Gear modifications for the Nephrops-directed mixed trawl fishery .................. 27
4.1 Anterior modifications ................................................................................ 28
4.1.1 Doors, sweeps and bridles .................................................................. 29
4.1.2 Counter-herding and anterior fish excluder devices ............................ 29
4.1.3 Headline height ................................................................................... 29
4.1.4 Topless or cutaway trawl ..................................................................... 30
4.1.5 Netting tapering ................................................................................... 31
4.2 Posterior modifications .............................................................................. 33
4.2.1 Trawl body mesh size .......................................................................... 33
4.2.2 Horizontal separator panel .................................................................. 34
4.2.3 Grids .................................................................................................... 35
4.2.4 Square mesh panels (SMP) ................................................................ 35
4.2.5 Sieve panels ........................................................................................ 36
4.2.6 Horizontally divided codends ............................................................... 36
4.2.7 Codend configuration .......................................................................... 37
5. Flexible gear modifications ............................................................................. 39
5.1 Flexible anterior modifications ................................................................... 40
5.2 Flexible posterior modifications ................................................................. 42
6. Towards haul-by-haul control over selectivity ................................................. 45
7. Conclusions and future work ........................................................................... 47
Reference List ..................................................................................................... 49
6
7
List of papers
Paper I: Melli, V., Karlsen, J. D., Feekings, J. P., Herrmann, B., Krag, L.
A., 2018. FLEXSELECT: counter-herding device to reduce
bycatch in crustacean trawl fisheries. Canadian Journal of
Fisheries and Aquatic Sciences, 75: 850–860. https://doi:
10.1139/cjfas-2017-0226
Paper II: Melli, V., Krag, L.A., Herrmann, B., Karlsen, J.D., 2018.
Investigating fish behavioural responses to LED lights in trawls
and potential applications for bycatch reduction in the Nephrops-
directed fishery. ICES Journal of Marine Science, 75: 1682–1692.
https://doi.org/10.1093/icesjms/fsy048
Paper III: Melli, V., Krag L.A., Herrmann, B., Karlsen J.D., 2019. Can active
behaviour stimulators improve fish separation from Nephrops
(Nephrops norvegicus) in a horizontally divided trawl codend?
Fisheries Research. https://doi.org/10.1016/j.fishres.2018.11.027
Paper IV: Melli, V., Herrmann, B., Karlsen J.D., Feeking, J.P., Krag L.A.
Predicting optimal combinations of bycatch reduction devices in
fishing gears: a meta-analytical approach. Manuscript.
8
9
Dansk resumé (abstract in Danish)
Med den nye Fælles fiskeripolitik (FFP), sigtede den Europæiske Union (EU)
mod at fjerne discarden af kommercielle arter, gennem indførelsen af en
forpligtelse til at lande hele fangsten (”landings forpligtelsen”). Denne fangst-
baserede tilgang, hvor alle størrelser af regulerede arter skal landes og
modregnes kvoten, er designet til at tilskynde fiskere til at undgå uønsket fangst.
For at kunne imødekomme dette, er der behov for et større udvalg af redskaber,
der matcher ændringer både i forekomsten af arter og størrelser men også i
skipperens aktuelle ønske om en fangstsammensætning, der kan maksimere
profitten inden for kvoten. Fleksible redskabsløsninger, som det enkelte fartøj kan
ændre fra slæb til slæb for at tilpasse deres størrelses- og arts-selektion, kan føre
til en effektiv implementering af FFP’en samtidig med at fiskeriets økonomiske
bæredygtighed opretholdes. Disse løsninger er specielt vigtige i flerartsfiskerier,
hvor ”choke”-arter kan begrænse udnyttelsen af mere produktive bestande. Dette
studie fokuserer derfor på det danske flerartsfiskeri efter Jomfruhummer
(Nephrops norvegicus), der har en af de højeste rater af uønsket fangst og derfor
ventes at blive stærkt påvirket af den nye FFP. Vi undersøgte fleksible
redskabsmodifikationer som kunne støtte alternative fangststrategier, såsom
reduktionen af bifangst af fisk både over og under det kommercielle mindstemål,
eller en selektiv tilbageholdelse af kun den mest værdifulde bifangst.
Afhandlingen består af en sammenfatning og 4 artikler.
Artikel 1 indeholder udvikling og afprøvning af det første studie af skræmmeliner i
jomfruhummerfiskeriet. Skræmmeliner er en fleksibel anordning der let kan
monteres og afmonteres foran selve trawlen og er designet til at lede fisk ud af
trawlsporet. Effektiviteten af skræmmelinerne varierede med fiskenes art og
størrelse men ikke med tidspunktet på dagen. Resultaterne viste at fangst af
potentielt uønskede fiskearter kunne undgås uden at fangsten af jomfruhummer
blev påvirket.
Artikel 2 fokuserer på et lovende design, den horisontalt delte fangstpose, som
kan føre til en fleksible opdeling af fangsten. Vi prøvede at anvende visuel
stimulering til at forbedre artsopdelingen og sammenlignede dette med et
baseline-redskab. Ved brug af ”Light Emitting Diodes” (LED) undersøgte vi om
enten positiv eller negativ phototaxis, kunne bruges til at øge den vertikale
separering af fisk fra jomfruhummer.
Artikel 3 fortsætter undersøgelsen af den horisontalt delte fangstpose men ved
brug af andre typer af adfærdsstimulering. Vi undersøgte om, og i hvilken
udstrækning, det er muligt til at forbedre den vertikale opdeling af fisk fra
jomfruhummer, ved at tilføje stimulatorer designet til at aktivere fisks
undvigelsesadfærd. Vi undersøgte to slags adfærdsstimulatorer: et gardin af
kæder ved indgangen til den nedre fangstpose og en serie af flydeliner foran den
opdelte fangstpose.
10
Artikel 4 beskriver en meta-analytisk tilgang til at forudsige størrelsesselektionen
af et redskab med flere forskellige bifangst-reducerende anordninger samt til at
sammenligne deres effektivitet under forskellige fangstscenarier. Vi brugte denne
teoretiske tilgang på trawlfiskeriet efter jomfruhummer for at identificere den mest
egnede kombination af bifangstanordninger og den alternative fangststrategi som
denne kombination ville kunne understøtte. Denne meta-analytiske tilgang kan
accelerere processen med at identificere optimal brug af fleksible
redskabsløsninger og dermed udvide fiskerens muligheder i håndteringen af EU
landingsforpligtelsen.
11
Abstract
With the new Common Fisheries Policy (CFP) the European Union (EU) aimed at
eliminating the discard of commercial species, introducing the obligation to land
all catches (“landing obligation”). This catch-based approach, where all sizes of
regulated species have to be landed and counted against quota, is designed to
encourage fishermen to minimize unwanted catches. Therefore, fishermen are
now in need of fishing gear options to cope with the variability in unwanted
catches and maximize their profit within the allowed catch limits. Flexible and
specialized gear solutions, which can be used on a haul-by-haul basis to adjust
size and species selectivity, can lead to an effective implementation of the CFP
while maintaining the economic viability of the fishery. These solutions are
particularly urgent in mixed trawl fisheries, where “choke” species can limit the
exploitation of more productive stocks. Therefore, this study focused on the
Danish Nephrops (Nephrops norvegicus) directed mixed trawl fishery, one of the
economically most important fisheries in Europe. This multispecies fishery has
one of the highest rates of unwanted catches, and is expected to be strongly
affected by a fully implemented and controlled landing obligation. We investigated
flexible gear modifications that could support alternative harvest strategies, such
as the reduction of both undersized and commercial sized fish bycatch or the
retention of only the most valuable bycatch species and sizes. The thesis
consists of a review and four papers.
Paper I contains the development and test of the first counter-herding device for
Nephrops-directed trawl fisheries. This flexible anterior modification, easily
mountable and de-mountable on the gear at a haul-by-haul level, was designed
to lead fish out of the trawl path. Its efficiency varied across species and sizes,
but was consistent regardless of diel period. The results showed a major
reduction of catches of potentially unwanted fish species, in particular haddock
(Melanogrammus aeglefinus) and whiting (Merlangius merlangus), with no effect
on Nephrops catches.
Paper II focuses on a horizontally divided trawl codend, which could lead to a
flexible separation of the catch in different compartments of the trawl. We
attempted to use visual stimulation to improve species separation. Using Light
12
Emitting Diodes (LED), we investigated if either positive or negative phototaxis
could be used to improve fish vertical separation from Nephrops. The results
showed significant changes in vertical separation but no clear species-specific
phototactic response. Moreover, overall LED lights increased the proportion of
individuals entering the lower compartment together with Nephrops.
Paper III continues the research on the horizontally divided trawl codend, but
applying other types of behavioural stimulators. We investigated if and to which
extent it is possible to improve the vertical separation of fish from Nephrops by
adding active stimulators designed to exploit fish avoidance behaviour. We tested
two types of behaviour stimulators: a chain curtain at the entrance of the lower
compartment and a set of rising float-lines ahead of the point of separation. The
results showed that species separation can be partially improved by the
stimulators, but the effect may not be sufficient to justify the additional complexity
in design with respect to the baseline.
Paper IV describes a meta-analytical approach to predict the size-selectivity of a
gear with a combination of Bycatch Reduction Devices (BRDs) and to compare
their performance under different catch scenarios. We applied this theoretical
approach to the Nephrops-directed trawl fishery, to identify the most pertinent
BRDs combinations and the alternative harvest strategies that they could
support. By including the results obtained in the previous papers, as well as
relevant BRDs available in literature, we predicted the selectivity of up to 100
possible combinations. Their performance was investigated for the target
species, Nephrops, and two bycatch species, cod (Gadus morhua) and haddock.
This meta-analytical approach can accelerate the process of identifying optimal
uses of flexible gear solutions, broadening fishermen’s options when coping with
the EU landing obligation.
13
1. Introduction
Mixed trawl fisheries are characterized by complex combinations of harvested
species, varying in productivity (i.e. abundance and state of the stock) and
economic value. Although theoretically all the commercial species caught are
target, in practice, within mixed fishery, individual catch goals vary according to
multiple economic, sociological, technical and legislative factors (Salomon et al.,
2014). In particular, the management framework strongly influence which fraction
of the catch is wanted and which one is unwanted. For example, measures such
as quotas and minimum landing sizes can increase the unwanted fraction
(Graham et al., 2007; Feekings et al., 2012). These unwanted catches, also
known as “bycatch”, include undersized and/or damaged target species, valuable
species whose quota is not available or has been exhausted, and low valued or
non-commercial species (Kelleher, 2005). In most mixed fisheries, the unwanted
fraction can equal or even exceed the wanted one (Hall and Mainprize, 2005;
Kelleher, 2005). Unless otherwise regulated, these unwanted catches are
discarded at sea, often dead or injured (Evans et al., 1994). Because discarding
unwanted catches is generally considered as an ecological and economical
waste of marine resources, most mixed fisheries around the world are managed
through mandatory technical measures (Kennelly, 2007). These technical
measures, termed Bycatch Reduction Devices (BRDs; Kennelly, 2007), are
designed to restrict the amount of unwanted catches, by modifying the gear
selectivity. The term “selectivity” expresses the ability of a gear to retain the
individuals encountered, on the basis of factors such as species, size and
behaviour (MacLennan, 1992). For trawl gears, the selectivity is mostly
determined by the characteristics of the codend, i.e. the final part of the gear
where the catch accumulates (Glass, 2000). In particular, regulation of the
selectivity of trawl gears has historically focused on mesh size and shape (Glass,
2000; Herrmann et al., 2009). Nevertheless, BRDs can be introduced in the
codend or ahead of it to select out undersized individuals and/or unwanted
species. However, depending on the management framework in place, the BRDs
can compromise the economic viability of the fishery and limit its capacity to cope
with spatial and temporal variability in catches.
14
In the European Union (EU), the management system consists of total allowable
catches (TACs) for the most valuable marine commercial species, determined
annually on the basis of estimated Maximum Sustainable Yields, i.e. “the
maximum annual catch which on average can be taken from an exploited stock
without deteriorating its productivity” (Salomon and Holm-Müller, 2013). As most
fishing grounds in the EU are shared among different Member States, TACs are
assigned per fishing area and then divided among the countries. Moreover, to
prevent unknown and unrecorded discarding of unwanted catches from
compromising the efficiency of the quota system (Punt et al., 2006; Crowder and
Murawski, 1998), the new EU Common Fishery Policy (CFP) introduced a
landing obligation for important harvested stocks (EU, 2013; 2016). The so-called
“discard ban” compels fishermen to land all catches, both wanted and unwanted.
Consequently, unwanted catches now count against fishermen’s quotas.
Moreover, this creates additional costs for the industry due to the processing of
the unwanted fraction of the catch (Hall et al., 2000; Hall and Mainprize, 2005).
Both sorting time and handling costs will likely increase as a bigger part of the
catch has to be separated and stored; on a limited storage space this could force
fishermen to increase the number of journeys to the harbour. Moreover, in mixed
fisheries, whenever the quota for one species is exhausted, and the catch of that
species cannot be avoided, fishing activities have to stop. These “choke” species
can potentially lead to the under-exploitation of more productive stocks, with
consequences on the economy of the fishery (Ulrich et al., 2011; Baudron and
Fernandes, 2015). Therefore, one of the main expected outcomes of a discard
ban is to strongly incentivize fishermen to couple selectivity with economy (Hall
and Mainprize, 2005; Graham et al., 2007). Indeed, in the frame of a landing
obligation, it is in fishermen’s interest to avoid or reduce the amount of unwanted
catches by improving the selectivity of the fishing gear, for example adopting
BRDs.
To be implemented effectively, with less undesirable economic impacts on the
industry, a landing obligation needs to be combined with flexible technical
regulations to increase fishermen’s ability to adjust the selectivity of their gears
(EU, 2016; Eliasen et al., 2019). The legislation of BRDs is often too rigid and
follows a “one-gear-fits-all” approach, where technical solutions are applied at the
15
fishery or regional level. In contrast, since the amount of unwanted catches is
mostly determined by the combination of gear, fishing practice and quota
availability, and since these may differ among vessels, the economic
consequences of the landing obligation can be expected at a vessel level.
Therefore, each vessel should be able to choose from a “toolbox” of BRDs to
better match the gear selectivity with specific catch goals. Moreover, in mixed
fisheries one BRD is rarely enough to cope with the spatial and temporal
variability in catch composition, as well as inter-annual variation in quota limits.
More gear options need to be identified to support alternative harvest strategies
(Eliasen et al., 2019). A toolbox of flexible gear modifications, which can be
temporarily applied to the gear without requiring major structural changes, could
enable a more dynamic adjustment of the gear selectivity at the haul-by-haul
level.
The present PhD thesis aimed at: i) developing a gear solution to prevent
potentially unwanted catches from entering the trawl, ii) determining to which
extent target and bycatch species can be separated inside the trawl, and iii)
investigating the potential of combinations of such gear solutions to achieve
optimal selectivity profiles. We used as a case study the Danish Nephrops
(Nephrops norvegicus) directed mixed trawl fishery, one of the most economically
important fisheries in Europe and the Northeast Atlantic, and the most
challenging in terms of bycatch reduction (Kelleher, 2005). In the present review,
an overview of BRDs available for the case-study fishery is presented and used
to discuss the definition of a “flexible gear modification”. According to this
definition, Paper I presents the development and testing of a flexible anterior
modification. Paper II and III advance the knowledge on a flexible posterior
modification. Finally, Paper IV addressed the question of predicting optimal
BRDs combinations, using the results obtained in Paper I-III, as well as previous
gear modifications described for the case-study fishery.
16
17
2. The outcomes of a discard ban: a global overview
Landing obligations, better known as discard bans, have been used as a
management tool all over the world in order to achieve full accountability of all
catches. Positive long-term outcomes have been described for countries outside
the EU that have been managing discard bans for decades (Diamond and
Beukers-Stewart, 2011; Condie at al., 2014; Karp et al., 2019). For example, a
discard ban for Pacific cod (Gadus macrocephalus) and pollock (Theragra
chalcogramma) has been enforced in the US Alaskan groundfish fishery since
1998 (Graham et al., 2007). In response to the ban, the fishery adopted gear
modifications to improve the selectivity of the fishing gears used and cooperation
among the vessels was observed in signalling bycatch hotspots to be avoided.
This led to a reduction of the discard rates of both pacific cod and pollock to just
0.4% and less than 1%, respectively (Graham et al., 2007). Another example is
represented by the British Columbia groundfish trawl fishery, in which the
discarding of Sebastes spp. is prohibited. Here, the management system has
introduced both incentives for the fishermen to match catches to quotas, in the
form of individual transferable quotas, and deterrents for illegal discarding, such
as a 100% coverage monitoring programme (Branch and Hilborn, 2008).
Although in these examples the discard ban is limited to few commercial species,
there are successful cases of discard bans where all commercial species are
included. This is the case of the Norwegian discard ban, which started for cod
and haddock in 1987 and was then extended gradually to include all living marine
resources in the following 30 years (Gullestad et al., 2015). Besides enforcing the
ban, the core of the Norwegian discard ban involves a set of “pragmatic
exemptions” (Gullestad et al., 2015) to increase flexibility and sustain the viability
of those fishermen that can demonstrate responsible behaviour and conduct (e.g.
allowed discard of live individuals and damaged catch in “small quantities”;
Gullestad et al., 2015).
Successful examples of discard ban have in common specific features, such as a
high level of surveillance, with serious consequences for those that violate the
rules (Hall et al., 2000; Branch and Hilborn, 2008), but also incentives for the
industry to comply to the ban (Hall and Mainprize, 2005; Stockhausen et al.,
2012) and an overall set of measures aimed at reducing the amount of unwanted
18
catches in the first place, thus simplifying their management on land (Condie et
al., 2014; Salomon et al., 2014). In particular, flexibility in gear-based technical
measures and legislation has been identified as key to obtain a reduction of
bycatch rates without causing the fishery to collapse (Condie et al., 2014; EU,
2016; Eliasen et al., 2019, Karp et al., 2019).
Consequently, for an effective implementation of the landing obligation, together
with its enforcement, it is necessary to address the question of to what extent the
selectivity of a specific fishery can be adjusted. As described in Section 1, this
question is of particular importance in mixed species fisheries, where choke
species can strongly affect the capitalization of quotas.
19
3. Case study fishery: why the Nephrops trawl fishery?
The Nephrops-directed mixed trawl fishery is one of the most profitable fisheries
in Denmark, with approximately 184 vessels targeting Nephrops for at least part
of the year of (2017 data; Danish Fisheries Agency). The main fishing areas are
in the North Sea (ICES Division IVa and IVb), Skagerrak and Kattegat (ICES
Division IIIa; Fig. 1). Total landings of Nephrops in 2017 were above 4,000
tonnes, for a value of approximately 250 million DKK (33 million Euro;
http://www.statistikbanken.dk). In addition, these vessels landed approximately
2,000 tonnes of fish, including cod (Gadus morhua), saithe (Pollachius virens),
hake (Merluccius merluccius), haddock (Melanogrammus aeglefinus), plaice
(Pleuronectes platessa), witch flounder (Glyptocephalus cynoglossus) and
monkfish (Lophius piscatorius), among others. Therefore, because of its highly
morphologically diverse catch and the recovering status of some gadoid stocks in
the area, this fishery has been classified as “very high risk” in terms of non-
compliance to the landing obligation (Anon, 2015).
Figure 1. Geographic position and ICES classification of the main fishing areas for the case-study fishery. Illustration by Dr Thomas Noack.
20
3.1 Fishing dynamics
Most vessels operating in the Danish Nephrops-directed mixed trawl fishery have
quota to land fish species which can contribute up to 2/3 of the profit of the
fishery (Danish Fisheries Agency). Consequently, the Danish fishery adopts the
so-called Combi trawls, which were designed to maximize the retention of both
Nephrops and fish species. Contrary to strictly Nephrops-directed trawls, Combi
trawls include longer sweeps, a higher headline height and an extension between
the trawl body (i.e. tapered section) and codend (Fig. 2a). All these elements
contribute to increasing fish catchability (Winger et al., 2010).
(a)
(b)
Figure 2. Schematic illustration of the trawl design and twin-rig configuration used by the Danish Nephrops-directed mixed trawl fishery. (a) Nephrops trawl scheme, modified with permission by SEAFISH Industry Authority. (b) Twin-rig scheme; artwork by Marco Nalon.
21
The trawlers, varying in length mostly between 15 and 30 m
(http://www.statistikbanken.dk), tow in general two identical trawls in a twin-rig
configuration (Fig. 2b). This system has been proven to increase catches of
benthic and demersal species, in particular Nephrops (Sangster and Breen,
1998).
3.1.1 Nephrops catchability
Nephrops norvegicus, also known as Norway Lobster, Langoustine or Scampi, is
a widely distributed arthropod crustacean of the order Decapoda. It lives on the
continental shelf of the North East Atlantic and in the Mediterranean Sea, on
muddy substrates, at depths ranging from 15 to 800 m (Chapman, 1980; Ungfors
et al., 2013). A specific type of seabed, i.e. fine cohesive mud, is essential for this
species, as Nephrops construct deep and complex burrows which are used as
refuges and for reproduction (Chapman and Rice, 1971; Rice and Chapman,
1971). When inside the burrows, Nephrops are unlikely to be caught by trawl
nets. Therefore, emergence due to foraging and mating activities, determines the
catchability of the species (Bell et al., 2006). This varies depending on biological
and environmental factors, such as molting cycle, female reproductive stage,
ambient light level, season, area, and tides (Chapman, 1980).
When outside the burrows, Nephrops react to disturbance, in particular physical
contact, by rapidly flipping the tail which propels them backwards (Newland and
Chapman, 1989). Reaction distance has been observed to vary between 0 and
55 cm, depending on the orientation of the individual with respect to the towing
direction of the gear (Newland and Chapman, 1985; Newland and Chapman,
1989). Average swimming speed (tail-flips) after tactile stimulation is 1−1.5 knots,
and the distance covered is limited to approximately 1-2 m (Newland and
Chapman, 1989). Thus, with fishing gears towed generally at 2–3 knots,
Nephrops in the path of the trawl are quickly overtaken by the footrope (Main and
Sangster, 1985; Newland and Chapman, 1989). After being stimulated, Nephrops
can, on average, rise vertically about 20–50 cm above the seabed (Newland and
Chapman, 1985). Therefore, considering standard headline heights of
approximately 1–2 m, and even though bigger individuals have been observed to
22
rise to up to 85 cm from the seabed, very few Nephrops escape above the
headline of the Combi trawl (Main and Sangster, 1985).
3.1.2 Fish catchability
Demersal fish species, sympatric in Nephrops fishing grounds, include both
roundfish and flatfish species. To be available to capture by the trawl, these
species have to be in the trawl path (i.e. area swept by the footrope). Therefore,
to effectively target fish, a trawl has to first concentrate them into the trawl path, a
result that can be achieved by exploiting fish behavioural responses to the
forward, spreading components of the trawl (Winger et al., 2010). This
mechanism is generally referred to as “herding”. Fish in the herding area (i.e.
between the doors) are stimulated by the doors and sweeps, which interact with
the seafloor producing vibrations sediment resuspensions (Glass and Wardle,
1989; Engås and Ona, 1990; Winger et al., 2010). Most species react to these
stimuli as they would in case of an approaching predator (Fernӧ and Huse,
2003). Roundfish species, in general, tend to swim away from the approaching
threat (i.e. doors and sweeps) while keeping it at the edge of their visual field.
This produce a movement described as “fountain manoeuvre” (Fig. 3a; Winger et
al., 2010). This results in individuals swimming directly into the trawl path,
exposing them to capture. Among the factors influencing the efficiency of
roundfish herding, two are known to play a fundamental role: the length of the
sweeps, with herding efficiency increasing at increasing lengths, and the angle of
the sweeps with respect to the towing direction (Winger et al., 2010). For cod and
haddock, sweeps lengths between 20 and 120 m (Engås and Godø, 1989) and
sweeps angles between 10 and 20 degrees (Strange, 1984) were found to
significantly increase catches.
In contrast, flatfish and benthic species, specialized in camouflage, are reticent to
flee and start swimming away only after direct or near contact with the doors or
sweeps (Fig. 3b; Main and Sangster, 1981). Once they flee, they move away
perpendicularly from the stimulus and either attempt to keep a constant distance
from the pursuing threat or burst to gain distance and then resettle on the
seafloor (Ryer, 2008). As a result of this slow herding process, longer sweeps (up
23
to 400 m) and small sweeps angles are necessary to leave enough time for the
individuals to reach the trawl path (Ryer et al., 2010).
Finally, the herding process was found to be size-dependent for many species,
and its efficiency to vary according to the towing speed, as a result of the species
and size-dependent differences in swimming capacity (Winger et al., 2010; He,
2011). In particular, those individuals that do not possess the energy or have the
capacity to maintain swimming speeds at least as fast as the towing speed would
be overtaken before reaching the trawl path and be exposed to capture (Winger
et al., 2010).
Figure 3. Schematic illustration of the herding process for (a) roundfish species (Winger et al., 2010) and (b) flatfish species (Main and Sangster, 1981).
Once in proximity of the trawl mouth, fish have been observed to turn around and
attempt to keep position ahead of the footrope (Main and Sangster, 1981; Glass
and Wardle, 1989; Wardle, 1993; Arimoto et al., 2010; Winger et al., 2010).
Depending on the species, size and the level of exhaustion, each individual can
either rise vertically and escape over the headline, be overtaken by the footrope,
or turn again and swim directly into the trawl (Kim and Wardle, 2003; Winger et
al., 2010).
Finally, for the most relevant bycatch species of the Nephrops-directed fishery,
fish catchability is known to vary at different light levels, as the behavioural
reactions described are mainly vision-dependent (Walsh and Hickey, 1993; Ryer
et al., 2010). Indeed, herding has been observed to cease at light levels below
fish visual thresholds (Wardle, 1993; Kim and Wardle, 1998a).
(a) (b)
24
3.2 Bycatch reduction measures before the landing obligation
Nephrops-directed trawl fisheries adopt a minimum mesh size of 70 or 90 mm
(depending on region). However, the poor selective properties of these mesh
sizes in relation to the Minimum Conservation Reference Size (MCRS) of the fish
species caught by the Danish fleet lead to high catches of undersized individuals
(Kelleher, 2005; Krag et al., 2008). To appropriately select out these undersized
individuals, a minimum mesh size of 120 mm is necessary (Graham and Ferro,
2004). Following concerns about the state of gadoids stocks, and in particular
cod (EC, 2008), since 2013 Danish trawlers targeting Nephrops are required to
use either of the following options:
a species-selective grid with 35 mm-spaced vertical bars inserted in a 70
mm square mesh codend, at least 8 m from the codline, to exclude the fish
bycatch, both undersized and commercial sized;
a size selective trawl (termed SELTRA trawl) consisting of a 90 mm
diamond mesh codend with a 3 m long escape panel inserted in the upper
netting of codend, starting at least 7 m before the codline. Depending on
the fishing area, the panel can be of either square meshes (140 mm, 3
opening angle ratio in Skagerrak; 180 mm, 4 opening angle ratio in
Kattegat) or diamond meshes (270 mm, both areas; Madsen et al., 2012;
ICES, 2014).
The escape panel is effective in reducing the catch of undersized individuals
while retaining commercial sized individuals (Frandsen et al., 2009; Briggs et al.,
2010). Therefore, it is the BRD adopted by most of the Danish Nephrops-directed
mixed trawl fishery. However, the release efficiency of an escape panel can be
more variable than that of a grid, because the escape panel relies on the
individuals actively contacting it. This varies according to size and species-
specific behaviour as well as position of the SMP in the gear (Krag et al., 2014;
Herrmann et al., 2015a; Nikolic et al., 2015).
3.3 Bycatch reduction under the landing obligation
The introduction of the landing obligation implies, in theory, that fishing activities
have to stop when the first quota is exhausted. Among the mix of species that
25
compose the bycatch in the Danish Nephrops-directed mixed trawl fishery, there
are two that represent a high risk of choking the fishery: cod and plaice (North
Sea Advisory Council, 2018). Cod capture below the MCRS (30 and 35 cm in
ICES Division IIIa and IVa-b, respectively) has been significantly reduced with the
adoption of the SMP (North Sea Advisory Council, 2018). However, unwanted
catch rates remain variable and have been observed to increase at higher cod
abundances (North Sea Advisory Council, 2018). Moreover, cod above the
MCRS used to be frequently discarded when of low-value (Category 4 and 5;
http://www.hanstholmfiskeauktion.dk/prices?lang=en). Therefore, with the full
implementation of the landing obligation in 2019, a quota deficit for cod for all the
State Members fishing in these areas is predicted to impact most mixed demersal
fisheries, including the Nephrops-directed one (North Sea Advisory Council,
2018). In contrast, there is currently a surplus in plaice quota but this species has
been classified as a potential economic choke species, i.e. a species where the
high abundance of undersized catches that has now to be sorted and stored on-
board can make the trip uneconomic (North Sea Advisory Council, 2018). Finally,
additional species can be troublesome depending on the specific fishing area and
gear design; Denmark has currently a quota deficit for hake, saithe and whiting,
all species whose catches can be abundant in ICES Division IVa and IVb (North
Sea Advisory Council, 2018).
Therefore, the current BRDs are not sufficient to prevent the impact of choke
species on the Danish Nephrops-directed mixed trawl fishery. In particular, the
SELTRA codends were not designed to prevent the catch of commercial-sized
cod and fishermen would have to adopt a grid to avoid their capture once out of
quota. This would obviously eliminate any other valuable fish bycatch as well as
potentially cause a loss of target Nephrops (Frandsen et al., 2009).
Consequently, fishermen could now be willing to voluntarily adopt BRDs which
provide alternative harvest patterns. In particular, simple BRDs, that do not
require major changes in fishing dynamics, are more likely to be adopted by
fishermen than more complex ones (Broadhurst, 2000). Furthermore, flexible
solutions that can be used to change the selectivity at the haul-by-haul level,
when required by the catch composition, would represent a valuable tool to cope
with the EU landing obligation.
26
27
4. Gear modifications for the Nephrops-directed mixed trawl
fishery
Prior to discussing which gear modifications can lead to a more flexible
selectivity, it is beneficial to explore the different types of modifications that can
be introduced in a Combi trawl. In particular, this section will focus on BRDs
developed for the Nephrops-directed fisheries, or applicable to them, with the aim
of reducing fish bycatch while maintaining Nephrops catches. For decades gear
technologists have developed BRDs which exploit species differences in terms of
morphology, size and behaviour to improve gears selectivity (Catchpole and
Gray, 2010; Graham, 2010). Dozens of BRDs and gear modifications are
documented in literature (Broadhurst, 2000; Catchpole and Revill, 2008; Graham,
2010) and private and public institutions are now collecting and organizing in
databases the sea trial results to aid the industry in identifying gear designs and
BRDs options (e.g. http://www.discardless.eu/selectivity_manual;
http://www.seafish.org/geardb/; https://tool.gearingup.eu/). The BRDs involve
modifications to different components of the trawl and are hereafter organized in
two main groups depending on their aim: preventing the catch of unwanted
individuals (anterior modifications) or select these out after they entered the trawl
(posterior modifications; Fig. 4).
Figure 4. Schematic illustration of the area of interest of each group of gear modifications. Artwork by Marco Nalon.
28
4.1 Anterior modifications
The capture process begins well ahead of the trawl, when individuals initially
detect the noise produced by the vessel, warps, doors and gear components
(Winger et al., 2010). According to Fernö and Huse (2003), the timing, type and
intensity of the response is then governed by the same behavioural trade-offs
that determine prey fleeing from predators (i.e. optimal escape theory; Ydenberg
and Dill, 1986). In general, each individual that detects a threat must sequentially
decide (1) whether and when to flee, (2) in which direction to flee, (3) how fast to
flee, and (4) how far to flee. If we consider as benefit the continuation of the
previously ongoing activity and as drawback the risk of being predated, the
behavioural response can be expressed as the result of the individual attempt to
minimize costs and maximize benefits (Fernö, 1993; Godin, 1997). This decision-
making process continues while the distance between the individual and the
predator shrinks. Once the perceived risk of being predated exceeds the benefits
of keeping position, the individual starts to flee. In general, the escape begins
with a slow adjustment in swimming direction away from the approaching
stimulus (Olsen et al. 1983; Winger et al., 2010). Therefore, the direction of the
stimulus perceived as a threat is what determines the direction of the escape. In
a standard trawl, the stimuli created by the early spreading components (i.e.
doors and sweeps) are those identified as the approaching threat (Kim and
Wardle, 1998a). Thus, fish swim away from these components and are eventually
herded towards the trawl mouth. Once they reach the trawl mouth, new stimuli
are perceived, e.g. the footrope and netting, which stimulate the individuals to
turn around. Here fish have been observed to either maintain distance ahead of
the pursuing trawl or to escape in different directions, e.g. above the headline or
below the footrope (Kim and Wardle, 1998a; Winger et al., 2010).
Anterior gear modifications are those that either mitigate these initial stimuli, thus
reducing the efficiency of the herding process, add stimuli to re-direct fish escape
out of the trawl path, or facilitate escaping opportunities once in the trawl mouth.
29
4.1.1 Doors, sweeps and bridles
As described above, doors, sweeps and bridles are the elements of the trawl that
first trigger fish escape response and direct it towards the trawl path. In particular,
their interaction with the sediment, in terms of sand cloud and vibrations
produced, create multisensorial stimuli that trigger fish response (Winger et al.,
2010). Therefore, modifications that raise either of these components off the
seafloor (Fig. 5, a and b) have successfully reduced the herding of some fish
species (Rose et al., 2010; Ryer et al., 2010; He et al., 2014; Sistiaga et al.,
2015; 2016; BIM, 2018). Moreover, species-specific herding efficiency can be
modified by shortening the length of the sweeps and bridles, thus increasing their
angle with respect to the towing direction (Mathai et al., 1984). Previous studies
have demonstrated that at wider angle of attachment of the sweeps fish have
less time to reach the trawl path and, thus, herding is less efficient (Strange,
1984). For example, at angles greater than 20 degrees, catches of cod and
haddock were significantly reduced (Strange, 1984).
4.1.2 Counter-herding and anterior fish excluder devices
Although the visual and tactile stimuli produced by doors and sweeps increase
fish catchability, the same type of stimuli, simply orientated in a different direction,
can cause an early escape response and increase fish chances of avoiding
capture. For example, higher-order multi-net configurations such as quad-rig
systems (i.e. four gears towed in parallel) catch less fish due to the additional
presence of wires to connect the different gears, which lead the fish to the outer
extremities of the catching zone (Broadhurst et al., 2013a; b). Similarly, additional
elements, such as diagonal wires, ropes and plastic banners (Fig. 5, c and d),
can be added in the herding area to re-direct fish escape away from the trawl
path (Ryer, 2008; McHugh et al., 2014, 2015; Paper I; BIM, 2018). These
devices, termed counter-herding devices, will be discussed in details in Section 5.
4.1.3 Headline height
Once in the trawl mouth, fish are stimulated by the footrope to swim in the towing
direction until fatigued or until the costs of maintaining position exceed the
benefits (Kim and Wardle, 2003; Breen et al., 2004; Winger et al., 2010).
30
Depending on the species, size and individual fitness, some fish will attempt to
swim upwards and escape over the trawl headline. Intuitively, higher headline
heights increase fish catchability (Wardle, 1986; Johnson et al., 2008; Broadhurst
et al., 2016). Therefore, to intentionally reduce the trawl efficiency in catching
species that attempt to escape upwards in the trawl mouth (i.e. haddock; Wardle,
1986), one can simply lower the headline height. Previous studies demonstrated
that the optimal headline height derives from a trade-off between having sufficient
height to maximise the capture of the target crustacean, which can swim upwards
when stimulated by the footrope, and lowering it sufficiently to minimise catches
of unwanted fish species (Eayrs, 2002; Madhu et al., 2015). However, length-
dependent swimming capacities and behaviours may reduce the efficacy of this
gear modification for smaller individuals.
4.1.4 Topless or cutaway trawl
Once in the trawl mouth, the netting of the trawl wings and mouth has been
observed to stimulate avoidance behaviour in fish (Kim and Wardle, 2003).
Depending on species and light level, the colour contrast of the twine with respect
to the background can stimulate fish to keep away from the netting even if the
mesh size would allow them to swim through (Kim and Wardle, 1998b, Winger et
al., 2010). Therefore, by perceiving the upper netting panel of the trawl mouth,
fish are stimulated to stay in proximity of the seafloor until they turn into or are
overtaken by the trawl. Consequently, trawls designed with the footrope located
ahead of the headline, remove this stimulus and can enhance fish upward
escape (Fig. 5 e). These designs, termed cutaway trawls or topless trawls, have
significantly reduced catches of roundfish and rockfish species (Thomsen, 1993;
Hannah et al., 2005; He et al., 2007; Chosid et al., 2008; Campbell et al., 2010;
Krag et al., 2015; Eayrs et al., 2017). However, inconclusive results were
obtained regarding the efficacy of the cutaway trawl on cod, with some studies
effectively reducing cod catches (Thomsen, 1993; Pol et al., 2003; Chosind et al.,
2008), and others finding no significant difference (Revill et al., 2006; Krag et al.,
2015). The efficiency of the cutaway trawl was found to depend on the headline
height, with no effect on cod detected with a headline height of 4.5 m (Krag et al.,
2015).
31
4.1.5 Netting tapering
The entry of fish into the trawl net is highly variable within and between species
and depends on multiple factors. One of these is how the trawl funnel appears to
the fish. Large trawls with a steeper-angle tapering (i.e. inclination of the netting)
that slowly reduces the funnel section and leads into an extension instead of
directly into the codend can create the illusion of an open path and incentivize
fish to turn around and swim into the trawl (Winger et al., 2010). In contrast,
smaller trawls with a wider-angle tapering leading directly into the codend, could
result in the netting being more evident to the fish. Moreover, altering the tapering
in the netting can have two subtle consequences on the trawl selectivity: i) a
wider-angle tapering may increase the probability for an individual to contact the
netting and, thus, to be selected out; and ii) the angle at which individuals contact
the netting with a wider-angle tapering may favour escapement (Broadhurst et al.,
2012). Although relatively few studies were conducted to determine the effect of
trawl length and netting tapering angle on fish catchability (Broadhurst et al.,
2012; 2015), the different features of commercial gears targeting fish and those
not targeting them are, intuitively, the result of fishermen’s experience of these
effects.
32
Figure 5. Examples of anterior gear modifications. (a) Semi-pelagic doors from Sistiaga et al. (2015); (b) Floating sweeps from He at al. (2014); (c) Anterior fish excluder from McHugh et al. (2017); (d) Hypothetical floating counter-herding device from Ryer (2008); and (e) Picture of a model topless trawl tested in a flume tank from He et al. (2007).
33
4.2 Posterior modifications
Once fish have entered the trawl, their vertical position in the trawl section is
species-dependent and can range from a few centimetres to several meters from
the lower netting panel (Winger et al., 2010). Benthic species such as monkfish
and flatfish typically stay in close proximity to the lower netting (e.g. Rose, 1995;
Bublitz, 1996), whereas demersal species, e.g. gadoids, tend to distribute
vertically across the trawl section (e.g. Main and Sangster, 1981; Thomsen,
1993). Moreover, the vertical distribution of fish inside the trawl has been proved
to vary, for some species, throughout the journey towards the codend (Holst et
al., 2009; Fryer et al., 2017). Upon reaching the narrow section immediately
ahead of the codend (i.e. trawl extension), some species, including haddock and
cod, can begin to swim erratically, in random directions (Grimaldo et al., 2008; He
et al., 2008). Indeed, as exhaustion sets in and crowding increases orderly
behaviours may be disrupted and substituted by randomly orientated burst-
swimming (Winger et al., 2010). This behaviour is likely to cause collision with
netting or other individuals. Finally, once in the codend, most fish are considered
to be exhausted and highly stressed, having endured prolonged continuous
swimming and attempted to avoid contact with the netting and other individuals.
Here, depending on their residual energy, individuals may attempt to keep
position ahead of the accumulated catch, try to escape through the meshes or
become part of the accumulated catch (Watson, 1989; Wardle, 1992; O’Neill et
al., 2003).
Posterior gear modifications, which include some of the most studied and
implemented BRDs, exploit behavioural differences in species vertical distribution
and swimming capacity, as well as morphological differences between and within
species, to separate and/or select out unwanted catches (see reviews by Glass,
2000; Graham, 2006).
4.2.1 Trawl body mesh size
To reduce the catch of undersized roundfish, while retaining the most valuable
larger sizes, an effective strategy can be to substitute the standard upper netting
mesh size with a larger one. This provides an escape opportunity for those fish
34
that rise while falling back into the trawl, but still within a size-selection process
that would retain larger individuals. Several studies tested this modification in
different fisheries (Thomsen, 1993; Briggs, 2010; Campbell et al., 2010; Kynoch
et al., 2011; Krag et al., 2014; Bayse et al., 2016). In particular, considering the
species of interest of this study, the following examples obtained significant
reductions of bycatch species: Thomsen (1993) reduced cod catches by 38% in a
whitefish fishery by replacing the 135 mm mesh size at the rear end of the
tapered section with 540 mm mesh size netting; Campbell et al. (2010)
significantly reduced catches of cod below 78 cm by replacing the 160 mm mesh
size netting with a 300 mm one, and Krag et al. (2014) obtained a significant
reduction on several important bycatch species of the Nephrops-directed fishery
by using a 800 mm mesh size section in the tapered area before the trawl
extension.
4.2.2 Horizontal separator panel
Differences in vertical escaping patterns at the mouth of the trawl can be
exploited to segregate species into two or more different compartments, by
inserting horizontal netting panels (Fig. 6 a). This design, termed separator trawl,
was originally tested by Strzysewski (1972) with a separator starting 1.5 m above
the footrope in a demersal herring trawl. Subsequently, separator trawls were
developed and tested for a variety of fisheries including the Nephrops-directed
mixed trawl fishery. Fryer et al. (2017) reviewed such studies, and analysed the
main factors affecting species-specific separation efficiency. Depending on the
species, the position of the separator panel, both vertically and horizontally in the
trawl, were found to affect the proportion of species entering each compartment
(Fryer et al., 2017). Among the species analysed, all the fish, with the exception
of monkfish, were found to be affected by the height of the separator, whereas
only cod was additionally affected by the horizontal distance from the footrope. A
higher proportion of cod entered the upper compartment when the separator
started at the extension level. Moreover, plaice was found to be affected by the
time of the day, with a higher proportion of individuals entering the upper
compartment during the night (Fryer et al., 2017).
35
4.2.3 Grids
In crustacean fisheries, grids are often applied to mechanically sort the small-
sized target crustacean, which passes through the grid into the codend, from
larger animals, which are diverted out of the trawl through an escape window.
Both rigid (e.g. Isaksen et al., 1992; Polet, 2002; Graham, 2003; Fonseca et al.,
2005) and flexible grids (e.g. Loaec et al., 2006) have been developed for
crustacean fisheries. A Nephrops trawl with a Swedish grid (35 mm bars space)
catches 80–100% less weight of commercial sized fish (Catchpole et al., 2006;
Valentinsson and Ulmestrand, 2008; Frandsen et al., 2009; Drewery et al., 2010).
Intuitively, this use of grids is not applicable if the marketable fish represent a
desired catch. Nonetheless, grids can be applied to the Nephrops-directed mixed
trawl fishery to separate fish into a different codend instead of releasing them
(Fig. 6 b; Anon, 2001; Graham and Fryer, 2006; Grimaldo et al., 2008). Moreover,
simple frames with only few bars can be used to stimulate fish, both behaviourally
and mechanically, to enter the upper codend (Krag et al., 2009a).
4.2.4 Square mesh panels (SMP)
The idea of SMPs came from maintaining mesh opening to assist fish escape
(Fig. 6 c). Indeed, due to their structure, square meshes stay open irrespective of
the longitudinal tension, unlike diamond mesh which tends to close as
longitudinal strain is applied (Robertson, 1986). Due to their relative simplicity,
they have been applied in a multitude of configurations and mesh sizes. From the
trawl body (e.g. Briggs, 2010), to the trawl extension (e.g. Krag et al., 2008), to
the codend (e.g. Herrmann et al., 2015a), SMPs have been inserted in various
positions in the trawl. However, SMPs have proved more effective when placed in
a position where the likelihood of fish contacting the meshes is higher, e.g. before
the catch accumulation zone in the codend or at the passage between tapered
section and extension/codend (Fig. 6 d; Graham and Kynoch, 2001; Graham et
al., 2003). Additional stimulation can also be added to accentuate fish contact
with the SMP or square mesh section. For example, Grimaldo et al. (2018) tested
floats and Light Emitting Diodes (LED) lights on free moving rope and increased
the escape rate of haddock; Kim and Wang (2010) tested a fluttering net panel
and a set of free ropes, successfully stimulating the escapement of juvenile red
36
sea bream (Pagrus major) in laboratory conditions; and Krag et al., 2016 used a
system of floats and ropes to discourage fish from moving through and prolong
their stay in the SMP section, thus increasing their probability of contacting the
meshes and escaping.
4.2.5 Sieve panels
A sieve panel consists of a netting panel attached inside the trawl body and/or
extension with a continuous inclination, generally an upward inclination in
Nephrops-directed fishery. Assuming that Nephrops will be concentrated in
proximity of the lower netting (Fryer et al., 2017; Karlsen et al., 2019) and that a
large enough mesh size is used in the sieve panel, most of the target species will
pass through the sieve and be retained in the codend. In contrast, bycatch
species will attempt to avoid contact with the meshes and follow the rising netting
towards an escape widow or a separate compartment. Therefore, similarly to
grids, sieve panels with the fore edge of the panel attached to the trawl lower
netting (Fig. 6 e) can be used in the Nephrops-directed fishery to select out
unwanted species too big to pass through the panel (Santos et al., 2018a;
Cosgrove et al., 2019). In contrast, sieve panels with the fore edge partly raised
would allow fish species generally found in close proximity of the lower netting,
such as monkfish and flatfish, to be retained together with Nephrops
(Valentinsson and Ulmestrand, 2008).
4.2.6 Horizontally divided codends
Among the horizontally divided trawls, the ones with the separator starting in the
trawl extension effectively separate most important bycatch species for
Nephrops-directed mixed trawl fisheries (Fryer et al., 2017). The relative height of
the compartments can be adjusted to maximize the probability of fish entering the
upper compartment while minimizing that of Nephrops. Multiple studies have
quantified the vertical separation efficiency of horizontally divided trawl codends
for both target and bycatch species, and how the addition of behavioural
stimulators can alter it (Holst et al., 2009; Krag et al., 2009b; Karlsen et al., 2019;
Paper II; Paper III). From simple frames (Krag et al., 2009b; Karlsen et al., 2019),
to LED lights (Paper II), floats and chains (Paper III), these studies modified the
37
species-specific separation into the compartments. In particular, these stimulators
were significantly effective on undersized individuals, despite them being often
considered too exhausted to respond to stimulation (Paper II; Paper III).
4.2.7 Codend configuration
Among all the components of the trawl that can be modified to reduce bycatch,
none is as studied and broadly implemented as the codend. It is well-known that
the mesh size, shape, twine material and twine thickness influence, both
mechanically and behaviourally, species possibility to escape through them
(O’Neill, 2003; Herrmann et al., 2015b). Moreover, the circumference (i.e. No. of
meshes) in the codend can affect the intensity of the water flow inside the
codend, and provide fish additional opportunities to escape through the codend
meshes by reducing the speed required to maintain position within the codend
(Rose, 1995; Broadhurst et al., 1999; O’Neill et al., 2003; Jones et al., 2008).
Therefore, codends can be modified in multiple ways to better select out
unwanted catches. In particular, the mesh size can be increased (e.g. Beutel et
al., 2008; Frandsen et al., 2011); the codend can be entirely or partly constructed
from square meshes (e.g. Frandsen et al., 2011; Wienbeck et al., 2014); the
number of diamond meshes in the codend circumference can be reduced to
enhance their opening (e.g. Broadhurst and Kennelly, 1996); and hanging ropes
(i.e. ropes shortened with respect to the stretched length of the codend) can be
inserted to prevent the codend from stretching while the catch accumulates and
thus maintaining mesh opening (e.g. Robertson and Shanks, 1989).
38
Figure 6. Examples of posterior gear modifications. (a) Separator trawl from Ferro et al. (2007); (b) Grid in combination with a horizontally divided trawl codend from Graham and Fryer (2006); (c) Square mesh codends model from SEAFISH Industry Authority; (d) From left to right: SMP before catch accumulation point, SMP at the end of the tapered section, and grid with escape window, from Drewery et al. (2010); and (e) Sieve panel in combination with a horizontally divided trawl codend from Santos et al. (2018a).
39
5. Flexible gear modifications
The previous section highlighted the wide array of options to modify the
selectivity of a Nephrops-directed mixed demersal trawl and reduce the bycatch
of fish. However, the few that have been introduced into legislation are applied to
entire fleets, without considering vessel-specific catch goals. Some gear
modifications (e.g. topless trawls, reduced headline height, and sieve panels
combined with escape windows) will always be more effective at sorting out
larger individuals due to within-species physical and behavioural limits in
swimming ability and escape responses (Revill et al., 2006; Drewery et al., 2010;
Krag et al., 2015). However, this loss of valuable catch would not be desirable
when fish quota is available. Therefore, these modifications are unlikely to be a
popular tool for all fishermen and at all times, as the cost–benefit balance of
preserving quota by eliminating the bigger individuals is likely to vary according to
quota levels and species market value fluctuations. In contrast, BRDs that
minimize the catch of undersized individuals, such as SMPs, can be perceived as
more “environmentally friendly options” but are unable to prevent the catch of
commercial sized bycatch species from choking the fishery.
Moreover, some BRDs have not been implemented or adopted voluntarily by the
fishermen due to unacceptable losses of target species (Frandsen et al., 2009;
Ingólfsson, 2011; Santos et al., 2018a). However, with the landing obligation, a
temporary loss of target species can become acceptable to the fishermen, as the
alternative solution (i.e. buying extra quota or stopping fishing) would ultimately
be more costly.
Fishermen face a much more complex reality than the one represented and
addressed by the legislation or by individual gear modifications. Even within
fishery, the fleet consists of different sized vessels using different sized gears of
varying designs. The specific location, time of day or season, weather conditions,
vessel configuration, skipper ability, quota combination, etc., all affect the catch
composition and amount of unwanted catches (Feekings et al., 2012).
Consequently, fishermen need to be able to address selectivity issues on a day-
to-day or even haul-to-haul basis. This could be achieved via gear modifications
40
designed to be simple in their application to the trawl and cost-effective in terms
of time required for their implementation.
According to the above, this thesis focused on “flexible” gear modifications, which
could provide the means to alter the selectivity of the trawl at the haul level:
5.1 Flexible anterior modifications
Among the anterior modifications previously described, counter-herding and
anterior fish excluder devices are some of the most flexible and simple options.
Because they consist of additional components added to the trawl, they do not
normally require any change to either the geometry or fishing dynamics of the
gear (McHugh et al., 2014, 2015; Paper I). Anterior fish excluder devices were
developed and tested by McHugh et al. (2014; 2015) in the Australian school
prawn (Metapenaeus macleayi) fishery. Here, otter trawls have only short bridles
and no sweeps, thus the additional components placed between the doors have
the function of making the approaching trawl more evident (McHugh et al., 2015).
In contrast, the use of counter-herding stimuli to lead fish out of the trawl path has
firstly been investigated in this study (Paper I). The potential of this type of
devices had been previously discussed but not tested due to concerns about
constraining the door spread (Ryer, 2008). In particular, engineering challenges
were expected in handling the variable tensions on the components of the
counter-herding device, for example at variations in spread due to bottom
topography and sediment characteristics. However, a careful consideration of the
materials and geometry of the counter-herding designs has proven sufficient in
preventing such problems (Paper I). Some adjustments of the spreading
mechanism (e.g. weight of the doors) can be required to prevent a reduction in
door-spread (Paper I). Nonetheless, the results imply only a minor reduction in
spread, which does not compromise the trawl geometry and can result in
improved bottom contact of the footrope (Broadhurst et al., 2014) thus increasing
Nephrops catches.
The potential of these devices has just started to be explored. Although positive
results were achieved in terms of bycatch reduction with both anterior fish
41
excluder (McHugh et al., 2014, 2015) and counter-herding devices (Paper I), only
few geometries and materials have been investigated. The mechanism that
determines the efficacy of these devices has still to be clearly understood. Due to
the forward position in the trawl and to the sediment resuspension, typical of
crustacean-directed fisheries, video observations of fish responses to counter-
herding devices are difficult to obtain. Therefore, it is unsure if fish are re-directed
out of the trawl path, as hypothesized (Fig. 7a), or if they rise vertically and
escape above the headline (Fig. 7b; Paper I). Most likely, the response varies
across species and sizes, as well as according to environmental parameters
affecting fish perception of the counter-herding device.
Figure 7. Schematic illustration of the possible swimming direction of fish in response to the counter-herding device. (a) horizontal response, and (b) vertical response. Artwork by Marco Nalon.
42
The results available show species-specific (McHugh et al., 2014, 2015; Paper I)
and size-dependent (Paper I) responses to counter-herding and anterior fish
excluder devices. Therefore, these BRDs could be used at the haul-by-haul level
to adjust both species and size composition of the catch. Different materials
should be tested to optimize the performance of the device (e.g. bouncing or
sweeping on the seafloor) and/or to make it more visible. For example, Ryer
(2008) postulated that a floating counter-herding device could be used to
selectively reduce roundfish catches and retain flatfish. Moreover, different
geometries of the counter-herding device (i.e. distance from the trawl mouth
and/or angle of attachment of the lines) should be investigated to determine if its
species and size-dependent effectiveness can be adjusted. Indeed, since
modifications to sweeps lengths and angles can significantly affect the herding
efficiency, similar effects can be expected for the counter-herding process (Ryer
et al., 2010; Winger et al., 2010).
5.2 Flexible posterior modifications
Among the array of posterior modifications, those that lead to an effective
separation of the main bycatch species from Nephrops are particularly suitable
for the necessities and goals of the Nephrops-directed mixed trawl fishery.
Indeed, the separation would not only benefit the quality of the catch (Karlsen et
al., 2015) but also allow fishermen to control their consumption of fish quota by
modifying the configuration of the upper codend. For example, large diamond or
square meshes could select out the undersized individuals of most commercial
northeast Atlantic fish species and retain the most valuable sizes (Graham and
Ferro, 2004; Frandsen et al., 2011; Wienbeck et al., 2014). After quota
exhaustion or in case of low-valued bycatch, the upper codend could even be left
open.
The main limit of these designs is the species and size-dependent separation
efficiency. Sieve panels and grids can effectively separate the catch on the basis
of size but their efficiency is limited for undersized individuals, which can pass
through together with Nephrops (Santos et al., 2018a; Cosgrove et al., 2019).
Moreover, because of the complex morphological components that determine
43
Nephrops selectivity (Frandsen et al., 2010) a rather large mesh size is required
for the sieve panel to maintain Nephrops catches in the lower codend (Santos et
al., 2018a). In contrast, horizontally divided trawl codends are very efficient in
separating Nephrops into the lower compartment (Holst et al., 2009; Karlsen et
al., 2019; Paper II; Paper III). Nonetheless, because they depend strongly on
species behaviour to achieve the separation, they can be enhanced by behaviour
stimulators. Limited research has been conducted on the application of behaviour
stimulators in combination with a horizontal separator panel (Paper II; Paper III),
while most studies have focused on increasing contact probability with escape
panels or windows. Fluttering lines, net panels, and floats have been tested as a
mean to trigger erratic swimming or avoidance reactions (Kim and Wang, 2010;
Herrmann et al., 2015a; Krag et al., 2016; Grimaldo et al., 2018; Paper III). Visual
stimuli such as black canvas or dark twine have been used to slow down fish
passage through the trawl and increase their probability of encountering a BRD
(Glass and Wardle 1995; Glass et al. 1995; He et al. 2008). LED lights have also
been used to highlight escape routes. For example, Lomeli and Wakefield (2014)
demonstrated that Chinook salmon (Oncorhynchus tshawytscha) escape rate
through escape windows increased when these were illuminated by blue LED
lights. However, although owning a great potential for species-specific
behavioural responses, the application of LED lights in trawls require further
studies. Most often the results of their application was negative (increased or
similar bycatch rate), for example when LED lights were attached in
correspondence of a grid (Hannah et al., 2015; Larsen et al., 2017), a SMP
(Grimaldo et al., 2018) or a horizontally divided trawl codend (Paper II).
Future research should develop the knowledge of species behavioural reactions
and of the main drivers of such responses. A more systematic understanding
would aid the identification of the appropriate type of stimulator, depending on the
relative position in the trawl, the average environmental conditions and the main
species of interest. Nonetheless, the results available are promising and the
continuous collection of video observations during the capture process will
improve the understanding and aid the exploitation of fish behaviour inside the
trawl.
44
45
6. Towards haul-by-haul control over selectivity
Once flexible gear modifications have been identified for a specific fishery, they
can be combined with the existing mandatory BRDs or with some of the more
complex and structural modifications described in Section 4.
Because each BRD is bound to have some limitations in efficiency, more and
more research has been conducted on the advantage of combining sequential
BRDs (Brinkhof et al., 2018; Hermann et al., 2018; Larsen et al., 2018; Paper IV).
Especially in mixed trawl fisheries, where the bycatch composition is often too
diverse to be sorted or selected out through one single selective process,
sequential processes can achieve multiple objectives, such as the reduction of
both undersized and commercial sized bycatch (Larsen et al., 2018; Paper IV).
Moreover, recent studies have highlighted that the performance of gear designs
under different, realistic catch scenarios can affect the viability of that design in a
specific fishery (Sala et al., 2015; Santos et al., 2018b; Paper IV). Therefore, by
combining sequential selective processes, the vulnerability of one BRD to specific
catch scenarios (e.g. high density of individuals around MCRS) could be
compensated by the previous or following processes (Paper IV).
In principle, as long as the BRDs do not compromise the structure of the trawl,
multiple BRDs can be combined. However, because each BRD is likely to cause
a small loss of target species, the sum of these losses can lead to unacceptable
impacts on the viability of the fishery (Paper IV). Moreover, to maximize the
advantage of adopting multiple sequential BRDs and offset the additional
complexity in gear design, the choice of BRDs should be limited to highly efficient
designs, targeting different species and size-groups.
To identify the most promising BRD combinations in well-studied fisheries such
as the Nephrops-directed mixed trawl fishery, where high numbers of BRDs have
been developed, a theoretical approach is recommended (Paper IV).
Nonetheless, future research should verify experimentally the predicted
combined selectivity of these promising combinations and determine if and to
what extent they could represent viable options for the fishery.
46
47
7. Conclusions and future work
The thesis was able to define, develop and test simple and flexible gear
modifications to reduce unwanted fish catches in the Danish Nephrops-directed
mixed trawl fishery. The study contributed with additional knowledge to the
development of both anterior (Paper I) and posterior BRDs (Paper II; III),
delivering new effective devices for the Nephrops-directed mixed trawl fishery
and advancing the understanding of species behaviour outside and inside the
trawl.
In particular, a new efficient, unique and flexible device (FLEXSELECT) was
developed and is been currently tested and adapted to other fisheries. This
counter-herding device has not only a broad applicability to several mixed trawl
fisheries around the world, but is also an ideal example of how trawl selectivity
can be substantially modified using simple and cost-effective solutions. The
results obtained in this study answered the question of feasibility and value of this
type of devices for mixed trawl fisheries. However, it raised even more questions
regarding their functioning and possible further developments. Therefore,
experimental research on counter-herding devices has just begun and it will
require both technological developments (i.e. test of different materials and
geometries) and behavioural studies to understand species-specific and length-
dependent response mechanisms.
In contrast, species separation into horizontally-divided trawl codends has
perhaps reached its maximum efficiency for the Nephrops-directed mixed trawl
fishery; at least until the understanding of species-specific behavioural responses
to stimulation is further developed. In this study, I used both tactile and visual
stimulation, as well as investigating one of the most expected responses to
artificial illumination: phototaxis. Overall, the results indicate that in the narrow
section of a Nephrops-directed trawl, behavioural responses to stimulation are
limited; when they occur, they end up often increasing the proportion of
individuals entering the lower compartment. Therefore, although the results are of
interests for future applications of both LED lights and active stimulators in other
fisheries and positions in the trawl, their development would benefit from a more
48
systematic understanding of species behaviour. This is the case, in particular, for
LED lights applications, because multiple technical parameters of the lights (e.g.
intensity, colour, pattern, and orientation) influence species responses (Nguyen
and Winger, 2019). The effect of each of these parameters needs to be
investigated individually before further interpretations of the behavioural
responses can be attempted.
Finally, the combined performance of the gear modifications tested in this study
and of other pertinent BRDs for the Nephrops-directed fishery was predicted to
support alternative harvest strategies. In particular the inclusion of flexible gear
modifications, such as the counter-herding device and the horizontally divided
trawl, would create a multi-purpose trawl, where selectivity could be adjusted to
match catch goals. This multi-purpose trawl could function as a strictly Nephrops
trawl or a mixed demersal trawl, with different species- and size-selectivity,
depending on the combination of BRDs introduced. Even though, before the
landing obligation, the cost of multiple BRDs in terms of target loss was
unacceptable, it is now the best option to maximize quota capitalization and
reduce the risk of being impacted by choke species.
With the new EU landing obligation, the role of selectivity as a management tool
and the potential uptake of BRDs by the industry are strengthening, and the
concept of one-gear-fits-all used in the legislation of fishing gears may soon be
retired. A more flexible legislation will lead to a new concept of gear design,
where the trawl can integrate multiple BRDs and fishermen can choose from a
toolbox of gear modifications to achieve the desired catch profile.
49
Reference List
Anon, 2001. Nephrops trawl discard reduction using activating selection grids
(NETRASEL). EU project FAIR-CT-98-4164, Final Report.
Anon, 2015. Recommended Measures to Achieve Compliance with the Landing
Obligation in Demersal Fisheries in North Western EU Waters. Submission
by the North Western Waters Fisheries Control Experts Group to the High
Level Group. pp. 30.
Arimoto, T., Glass, C., Zhang, X., 2010. Fish vision and its role in fish capture. In:
Behaviour of marine fishes: Capture process and conservation challenges.
Edited by P. He. Blackwell Publishing Ltd. Iowa, USA. pp 25–40.
Baudron, A.R., and Fernandes, P.G., 2015. Adverse consequences of stock
recovery: European hake, a new “choke” species under a discard ban? Fish
and Fisheries, 16: 563–575.
Bayse, S.M., Rillahan, C.B., Jones, N.F., Balzano, V., He, P., 2016. Evaluating a
large-mesh belly window to reduce bycatch in silver hake (Merluccius
bilinearis) trawls. Fisheries research, 174: 1–9.
Bell, M.C., Redant, F., Tuck, I., 2006. Nephrops species. In: Phillips, B.P. (Ed.),
Lobsters, Biology, Management, Aquaculture and Fisheries. Blackwell
Publishing Ltd., pp. 412–461.
Beutel, D., Skrobe, L., Castro, K., Ruhle, P., O’Grady, J., Knight, J., 2008.
Bycatch reduction in the Northeast USA directed haddock bottom trawl
fishery. Fisheries Research, 94: 190–198.
BIM, 2018. Assessment of Dyneema floating sweeps and fish scaring ropes in
the Irish Sea Nephrops fishery, 10 pp.
Branch, T.A., and Hilborn, R., 2008. Matching catches to quotas in a multispecies
trawl fishery: targeting and avoidance behavior under individual transferable
quotas. Canadian Journal of Fisheries and Aquatic Sciences, 65: 1435–
1446.
Breen, M., Dyson, J., O'Neill, F.G., Jones, E., Haigh, M., 2004. Swimming
endurance of haddock (Melanogrammus aeglefinus L.) at prolonged and
sustained swimming speeds, and its role in their capture by towed fishing
gears. ICES Journal of Marine Science, 61: 1071–1079.
50
Briggs, R.P., 2010. A novel escape panel for trawl nets used in the Irish Sea
Nephrops fishery. Fisheries Research, 105: 118–124.
Brinkhof, J., Olsen, S.H., Ingólfsson, Ó.A., Herrmann, B., Larsen, R.B., 2018.
Sequential codend improves quality of trawl-caught cod. PloS one, 13:
e0204328.
Broadhurst, M.K., 2000. Modifications to reduce bycatch in prawn trawls: a review
and framework for development. Reviews in Fish Biology and Fisheries, 10:
27–60.
Broadhurst, M.K., and Kennelly, S.J., 1996. Effects of the circumference of
codends and a new design of square-mesh panel in reducing unwanted by-
catch in the New South Wales oceanic prawn-trawl fishery,
Australia. Fisheries Research, 27: 203–214.
Broadhurst, M.K., Kennelly, S.J., Eayrs, S., 1999. Flow-related effects in prawn
trawl codends: potential for increasing the escape of unwanted fish through
square-meshpanels. Fishery Bulletin, 97: 1–8.
Broadhurst, M.K., Sterling, D.J., Millar, R.B., 2012. Short vs long penaeid trawls:
Effects of side taper on engineering and catching performances. Fisheries
Research, 134: 73–81.
Broadhurst, M.K., Sterling, D.J., Millar, R.B., 2013a. Progressing more
environmentally benign penaeid-trawling systems by comparing Australian
single- and multi-configurations. Fisheries Research, 146: 7–17.
Broadhurst, M.K., Sterling, D.J., Millar, R.B., 2013b. Relative engineering and
catching performances of paired penaeid-trawling systems. Fisheries
research, 143: 143–152.
Broadhurst, M.K., Sterling, D.J., and Millar, R.B., 2014. Engineering and catch
implications of variable wing-end spread on a penaeid trawl. Fisheries
Research, 153: 24–30.
Broadhurst, M.K., Sterling, D.J., Millar, R.B., 2015. Increasing lateral mesh
openings in penaeid trawls to improve selection and reduce drag. Fisheries
Research, 170: 68–75.
Broadhurst, M.K., Sterling, D.J., Millar, R.B., 2016. Confounding effects of knot
orientation in penaeid trawls. Fisheries Research, 179: 124–130.
Bublitz, C.G., 1996. Quantitative evaluation of flatfish behavior during capture by
trawl gear. Fisheries Research, 25: 293–304.
51
Campbell, R., Harcus, T., Weirman, D., Fryer, R.J., Kynoch, R.J., O’Neill, F.G.,
2010. The reduction of cod discards by inserting 300 mm diamond mesh
netting in the forward sections of a trawl gear. Fisheries Research, 102:
221–226.
Catchpole, T.L., and Gray, T.S., 2010. Reducing discards of fish at sea: a review
of European pilot projects. Journal of Environmental Management, 91: 717–
723.
Catchpole, T.L., and Revill, A.S., 2008. Gear technology in Nephrops trawl
fisheries. Reviews in Fish Biology and Fisheries, 18: 17–31.
Catchpole, T.L., Frid, C.L.J., Gray, T.S., 2006. Importance of discards from the
English Nephrops norvegicus fishery in the North Sea to marine
scavengers. Marine Ecology Progress Series, 313: 215–226.
Chapman, C.J., 1980. Ecology of juvenile and adult Nephrops. In: Cobb, J. S.
Phillips, B. F. (Eds.), The biology and management of lobsters. Vol. II
Ecology and management. Academic press, pp. 143–160.
Chapman, C.J., and Rice, A.L., 1971. Some direct observations on the ecology
and behaviour of the Norway lobster Nephrops norvegicus. Marine Biology,
10: 321–329.
Chosid, D., Pol, M., Szymanski, M., Ribas, L., Moth-Poulsen, T., 2008. Diel
variation within the species selective "Topless" trawl net. Journal of Ocean
Technology, 3: 31–58.
Condie, H.M., Grant, A., Catchpole, T.L., 2014. Incentivising selective fishing
under a policy to ban discards; lessons from European and global
fisheries. Marine Policy, 45: 287–292.
Cosgrove, R., Browne, D., Minto, C., Tyndall, P., Oliver, M., Montgomerie, M.,
McHugh, M., 2019. A game of two halves: Bycatch reduction in Nephrops
mixed fisheries. Fisheries Research, 210: 31–40.
Crowder, L.B., and Murawski, S.A., 1998. Fisheries bycatch: Implications for
management. Fisheries, 23: 8–17.
Diamond, B., and Beukers-Stewart, B.D., 2011. Fisheries Discards in the North
Sea - Waste of Resources or a Necessary Evil? Reviews in Fisheries
Science, 19: 231–245.
52
Drewery, J., Bova, D., Kynoch, R.J., Edridge, A., Fryer, R.J., O’Neill, F.G., 2010.
The selectivity of the Swedish grid and 120 mm square mesh panels in the
Scottish Nephrops trawl fishery. Fisheries Research, 106: 454–459.
Eayrs, S., 2002. Understanding fish and prawn behaviour: potential to reduce
bycatch in a tropical prawn trawl fishery. Fisheries Science, 68: 367–370.
Eayrs, S., Pol, M., Caporossi, S. T., Bouchard, C., 2017. Avoidance of Atlantic
cod (Gadus morhua) with a topless trawl in the New England groundfish
fishery. Fisheries research, 185: 145–152.
Eliasen, S.Q., Feekings, J., Krag, L., Veiga-Malta, T., Mortensen, L.O., Ulrich, C.,
2019. The landing obligation calls for a more flexible technical gear
regulation in EU waters–Greater industry involvement could support
development of gear modifications. Marine Policy, 99: 173–180.
Engås, A., and Godø, O.R., 1989. The effect of different sweep lengths on the
length composition of bottom - sampling trawl catches. Journal du Conseil
International pour l´Exploration de la Mer, 45: 263–268.
Engås, A., and Ona, E., 1990. Day and night fish distribution pattern in the net
mouth area of the Norwegian bottom-sampling trawl. ICES, 189: 123–127.
EC, 2008. Council Regulation (EC) No 1342/2008 of 18 December 2008
Establishing a Long-term Plan for Cod Stocks and the Fisheries Exploiting
Those Stocks and Repealing Regulation (EC) No 423/200.
EU, 2013. Regulation (EU) No 1380/2013 of the European Parliament and
Council of 11 December 2013 on the Common Fisheries Policy. Official
Journal of the European Union, L 354/22.
EU, 2016 On the Conservation of Fishery Resources and the Protection of
Marine Ecosystems through Technical Measures, Amending Council
Regulations (EC) No. 1967/2006, (EC) No. 1098/2007, (EC) No. 1224/2009
and Regulations (EU) No. 1343/2011 and (EU) No. 1380/2013.
Evans, S.M., Hunter, J.E., Elizal, Wahju, R.I., 1994. Composition and fate of the
catch and bycatch in the Farne Deep (North Sea) Nephrops fishery. ICES
Journal of Marine Science, 51: 155–168.
Feekings, J., Bartolino, V., Madsen, N., Catchpole, T., 2012. Fishery discards:
factors affecting their variability within a demersal trawl fishery. PloS one, 7:
e36409.
53
Fernö, A., 1993. Advances in understanding of basic behavior: consequences for
fish capture studies. ICES journal of Marine Science Symposium, 196: 5–
11.
Fernö, A., and Huse, I., 2003. Fish avoidance of survey vessels and gear: can
predictions based on the response of fi sh to predators explain the observed
variations? Presented at the ICES Symposium on Fish Behavior in
Exploited Ecosystems, June 23–26, 2003. Bergen, Norway.
Ferro, R.S.T., Jones, E.G., Kynoch, R.J., Fryer, R.J., Buckett, B.E. 2007.
Separating species using a horizontal panel in the Scottish North Sea
whitefish trawl fishery. ICES journal of Marine Science, 64: 1543–1550.
Fonseca, P., Campos, A., Larsen, R.B., Borges, T.C., Erzini, K., 2005. Using a
modified Nordmøre grid for by-catch reduction in the Portuguese
crustacean-trawl fishery. Fisheries Research, 71: 223–239.
Frandsen, R.P., Holst, R., Madsen, N., 2009. Evaluation of three levels of
selective devices relevant to management of the Danish Kattegat-Skagerrak
Nephrops fishery. Fisheries Research, 97: 243–252.
Frandsen, R.P., Madsen, N., Krag, L.A., 2010. Selectivity and escapement
behaviour of five commercial fishery species in standard square- and
diamond-mesh codends. ICES Journal of Marine Science, 67: 1721–1731.
Frandsen, R.P., Herrmann, B., Madsen, N., Krag, L.A., 2011. Development of a
codend concept to improve size selectivity of Nephrops (Nephrops
norvegicus) in a multi-species fishery. Fisheries Research, 111: 116–126.
Fryer, R.J., Summerbell, K., O’Neill, F.G., 2017. A meta-analysis of vertical
stratification in demersal trawl gears. Canadian Journal of Fisheries and
Aquatic Science, 999: 1–8.
Glass, C.W., 2000. Conservation of fish stocks through bycatch reduction: a
review. Northeastern Naturalist, 7: 395–410.
Glass, C.W., and Wardle, C.S., 1989. Comparison of the reactions of fish to a
trawl gear, at high and low light intensities. Fisheries Research, 7: 249–266.
Glass, C.W., and Wardle, C.S., 1995. Studies on the use of visual stimuli to
control fish escape from codends. II. The effect of a black tunnel on the
reaction behaviour of fish in otter trawl codends. Fisheries Research, 23:
165–174.
54
Glass, C.W., Wardle, C.S., Gosden, S.J., Racey, D.N., 1995. Studies on the use
of visual stimuli to control fish escape from codends. I. Laboratory studies
on the effect of a black tunnel on mesh penetration. Fisheries Research, 23:
157–164.
Godin, J.-G., 1997. Evading predators. In: Godin J.-G. (ed). Behavioral Ecology
of Teleost Fishes. pp 191–236. Oxford University Press.
Graham, N., 2003. By-catch reduction in the brown shrimp, Crangon crangon,
fisheries using a rigid separation Nordmøre grid (grate). Fisheries
Research, 59: 393–407.
Graham, N., 2006. Trawling: historic development, current status and future
challenges. Marine Technology Society Journal, 40: 20–24.
Graham, N., 2010. Technical measures to reduce bycatch and discards in trawl
fisheries. In He, P. (Ed.), Behavior of Marine Fishes: Capture Processes
and Conservation Challenges, pp. 237–264. Wiley-Blackwell, Arnes, IA.
Graham, N., and Kynoch, R.J., 2001. Square mesh panels in demersal trawls:
some data on haddock selectivity in relation to mesh size and position.
Fisheries Research, 49: 207–218.
Graham, N., Kynoch, R. J., Fryer, R. J., 2003. Square mesh panels in demersal
trawls: further data on haddock and whiting selectivity in relation to panel
position. Fisheries Research, 62: 361–375.
Graham, N., and Ferro, R.S.T., 2004. The Nephrops fisheries of the Northeast
Atlantic and Mediterranean: a review and assessment of fishing gear
design. ICES Cooperative Research Report No. 270.
Graham, N., and Fryer, R.J., 2006. Separation of fish from Nephrops norvegicus
into a two-tier cod-end using a selection grid. Fisheries research, 82: 111–
118.
Graham, N., Ferro, R.S.T., Karp, W.A., MacMullen, P., 2007. Fishing practice,
gear design, and the ecosystem approach—three case studies
demonstrating the effect of management strategy on gear selectivity and
discards. ICES Journal of Marine Science, 64: 744–750.
Grimaldo, E., Sistiaga, M., Larsen, R.B., 2008. Evaluation of codends with sorting
grids, exit windows, and diamond meshes: size selection and fish
behaviour. Fisheries Research, 91: 271–280.
55
Grimaldo, E., Sistiaga, M., Herrmann, B., Larsen, R.B., Brinkhof, J., Tatone, I.,
2018. Improving release efficiency of cod (Gadus morhua) and haddock
(Melanogrammus aeglefinus) in the Barents Sea demersal trawl fishery by
stimulating escape behaviour. Canadian Journal of Fisheries and Aquatic
Sciences, 75: 402–416.
Gullestad, P., Blom, G., Bakke, G., Bogstad, B., 2015. The “Discard Ban
Package”: Experiences in efforts to improve the exploitation patterns in
Norwegian fisheries. Marine Policy, 54: 1–9.
Hall, S.J., and Mainprize, B.M., 2005. Managing by-catch and discards: how
much progress are we making and how can we do better? Fish and
Fisheries, 6: 134–155.
Hall, M.A., Alverson, D.L., Metuzals, K.I., 2000. By-Catch: Problems and
Solutions. Marine Pollution Bulletin, 41: 204–219.
Hannah, R.W., Parker, S.J., Buell, T.V., 2005. Evaluation of a selective flatfish
trawl and diel variation in rockfish catchability as bycatch reduction tools in
the deepwater complex fishery off the US west coast. North American
journal of fisheries management, 25: 581–593.
Hannah, R.W., Lomeli, M.J.M., Jones, S.A., 2015. Tests of artificial light for
bycatch reduction in an ocean shrimp (Pandalus jordani) trawl: Strong but
opposite effects at the footrope and near the bycatch reduction device.
Fisheries Research, 170: 60–67.
He, P., 2011. Behavior of marine fishes: capture processes and conservation
challenges. (Ed.) John Wiley & Sons.
He, P., Goethel, D., Smith, T., 2007. Design and test of a topless shrimp trawl to
reduce pelagic fish bycatch in the Gulf of Maine pink shrimp fishery. Journal
of Northwest Atlantic Fishery Science, 38: 13–21.
He, P., Smith, T., Bouchard, C., 2008. Fish behavior and species separation for
the Gulf of Main multispecies trawls. Journal of Ocean Technology, 3: 60–
77.
He, P., Rillahan, C., Balzano, V., 2014. Reduced herding of flounders by floating
bridles: application in Gulf of Maine Northern shrimp trawls to reduce
bycatch. ICES Journal of Marine Science, 72: 1514–1524.
56
Herrmann, B., Krag, L.A., Frandsen, R., Madsen, N., Lundgren, B., Stæhr, K.-J.,
2009. Prediction of selectivity from morphological conditions: Methodology
and a case study on cod (Gadus morhua). Fisheries Research, 97: 59–71.
Herrmann, B., Wienbeck, H., Karlsen, J.D., Stepputtis, D., Dahm, E., Moderhak,
W. 2015a. Understanding the release efficiency of Atlantic cod (Gadus
morhua) from trawls with a square mesh panel: effects of panel area, panel
position, and stimulation of escape response. ICES Journal of Marine
Science, 72: 686–696.
Herrmann, B., Wienbeck, H., Stepputtis, D., Krag, L.A., Feekings, J., Moderhak,
W., 2015b. Size selection in codends made of thin-twined Dyneema netting
compared to standard codends: A case study with cod, plaice and
flounder. Fisheries research, 167: 82–91.
Herrmann, B., Sistiaga, M., Grimaldo, E., Larsen, R.B., Olsen, L., Brinkhof, J.,
Tatone, I., 2018. Size selectivity and length-dependent escape behaviour of
haddock in a sorting device combining a grid and a square mesh
panel. Canadian Journal of Fisheries and Aquatic Sciences.
https://doi.org/10.1139/cjfas-2017-0461
Holst, R., Ferro, R.S., Krag, L.A., Kynoch, R.J., Madsen, N., 2009. Quantification
of species selectivity by using separating devices at different locations in
two whitefish demersal trawls. Canadian Journal of Fisheries and Aquatic
Sciences, 66: 2052–2061.
ICES, 2014. Report of the Working Group on Mixed Fisheries Methods
(WGMIXFISH-METH), 20–24 October 2014, Nobel House, London, UK. 75
pp.
Ingólfsson, Ó.A., 2011. The effect of forced mesh opening in the upper panel of a
Nephrops trawl on size selection of Nephrops, haddock and
whiting. Fisheries Research, 108: 218–222.
Isaksen, B., Valdemarsen, J.W., Larsen, R.B., Karlsen, L., 1992. Reduction of
fish bycatch in shrimp trawl using rigid separator grid in the aft belly.
Fisheries Research, 13: 335–352.
Johnson, D.D., Rotherham, D., Gray, C.A., 2008. Sampling estuarine fish and
invertebrates using demersal otter trawls: effects of net height, tow duration
and diel period. Fisheries Research, 93: 315–323.
57
Jones, E.G., Summerbell, K., O’Neill, F., 2008. The influence of towing speed
and fish density on the behaviour of haddock in a trawl cod-end. Fisheries
Research, 94: 166–174.
Karlsen, J.D., Krag, L.A., Albertsen, C.M., Frandsen, R.P., 2015. From fishing to
fish processing: separation of fish from crustaceans in the Norway lobster-
directed multispecies trawl fishery improves seafood quality. PloS one, 10:
e0140864.
Karlsen, J.D., Krag, L.A., Herrmann, B., Lund, H.S., 2019. Using vertical
distribution to separate fish from crustaceans in a mixed species trawl
fishery. Canadian Journal of Fisheries and Aquatic Science.
doi.org/10.1139/cjfas-2018-0313
Karp, W.A., Breen, M., Borges, L., Fitzpatrick, M., Kennelly, S.J., Kolding, J-,
Nielsen, K.N., Viðarsson, J.R., Cocas, L., Leadbitter, D., 2019. Strategies
Used Throughout the World to Manage Fisheries Discards – Lessons for
Implementation of the EU Landing Obligation. In: Uhlmann S.S., Ulrich C.,
Kennelly S.J. (eds) The European Landing Obligation. Springer, Cham.
Kelleher, K., 2005. Discards in the world’s marine fisheries: an update. FAO
Fisheries Technical Paper No. 470. Food and Agriculture Organization of
the United Nations, Rome, Italy.
Kennelly, S.J., 2007. By-catch Reduction in the World’s Fisheries. (Ed.) Springer,
290 pp.
Kim, Y.H., and Wardle, C.S., 1998a. Modelling the visual stimulus of towed
fishing gear. Fisheries research, 34: 165–177.
Kim, Y.H., and Wardle, C.S., 1998b. Measuring the brightness contrast of fishing
gear, the visual stimulus for fish capture. Fisheries research, 34: 151–164.
Kim, Y.H., and Wardle, C.S., 2003. Optomotor response and erratic response:
quantitative analysis of fish reaction to towed fishing gears. Fisheries
Research, 60: 455–470.
Kim, Y.H., and Whang, D.S., 2010. An actively stimulating net panel and rope
array inside a model cod-end to increase juvenile red seabream
escapement. Fisheries Research, 106: 71–75.
Krag, L. A., Frandsen, R. P., Madsen, N., 2008. Evaluation of a simple means to
reduce discard in the Kattegat-Skagerrak Nephrops (Nephrops norvegicus)
58
fishery: Commercial testing of different codends and square-mesh
panels. Fisheries Research, 91: 175–186.
Krag, L.A., Madsen, N., Karlsen, J.D., 2009a. A study of fish behaviour in the
extension of a demersal trawl using a multi-compartment separator frame
and SIT camera system. Fisheries Research, 98: 62–66.
Krag, L.A., Holst, R., Madsen, N., 2009b. The vertical separation of fish in the aft
end of a demersal trawl. ICES Journal of Marine Science: Journal du
Conseil, 66: 772–777.
Krag, L.A., Herrmann, B., Karlsen, J.D. 2014. Inferring fish escape behaviour in
trawls based on catch comparison data: model development and evaluation
based on data from Skagerrak, Denmark. PloS One, 9: e88819.
Krag, L.A., Herrmann, B., Karlsen, J.D., Mieske, B., 2015. Species selectivity in
different sized topless trawl designs: Does size matter? Fisheries
Research, 172: 243–249.
Krag, L.A., Herrmann, B., Feekings, J., Lund, H., Karlsen, J.D., 2016. Improving
escape panel selectivity in Nephrops directed fisheries by actively
stimulating fish behaviour. Canadian Journal of Fisheries and Aquatic
Science, 74: 486–493.
Kynoch, R.J., O’Neill, F.G., Fryer, R.J., 2011. Test of 300 and 600 mm netting in
the forward sections of a Scottish whitefish trawl. Fisheries research, 108:
277–282.
Larsen, R.B., Herrmann, B., Sistiaga, M., Brinkhof, J., Tatone, I., Langård, L.,
2017. Performance of the Nordmøre grid in shrimp trawling and potential
effects of guiding funnel length and light stimulation. Marine and Coastal
Fisheries, 9: 479–492.
Larsen, R.B., Herrmann, B., Sistiaga, M., Brinkhof, J., Grimaldo, E., 2018.
Bycatch reduction in the Norwegian Deep-water Shrimp (Pandalus borealis)
fishery with a double grid selection system. Fisheries Research, 208: 267–
273.
Loaec, H., Morandeau, F., Meillat, M., Davies, P., 2006. Engineering
development of flexible selectivity grids for Nephrops. Fisheries
Research, 79: 210–218.
Lomeli, M.J., and Wakefield, W.W., 2014. Examining the potential use of artificial
illumination to enhance Chinook salmon escapement out a bycatch
59
reduction device in a Pacific hake midwater trawl. National Marine Fisheries
Service, Northwest Fisheries Science Center Report, Seattle, WA.
MacLennan, D.N., 1992. Fishing gear selectivity: an overview. Fisheries
Research, 13: 201–204.
Madsen, N., Holst, R., Frandsen, R.P., Krag, L.A., 2012. Improving the
effectiveness of escape windows in directed Norway lobster Nephrops
norvegicus trawl fisheries. Fisheries Science, 78: 965–975.
Madhu, V.R., Remesan, M.P., Pravin, P., Boopendranath, M.R., 2015.
Performance of separator trawl off Cochin, southwest coast of India.
Fisheries Technology, 52: 145–151.
Main, J., and Sangster, G., 1981. A study of the fish capture process in a bottom
trawl by direct observation from a towed underwater vehicle. Scottish
fisheries research report No. 23, pp. 24.
Main, J., and Sangster, G., 1985. Trawling experiments with a two-level net to
minimise the undersized gadoid by-catch in a Nephrops fishery. Fisheries
Research, 3: 131–145.
Mathai, T.J., Syed Abbas, M., Mhalathkar, H.N., 1984. Towards optimisation of
bridle lengths in bottom trawls. Fishery Technology, 21: 106–108.
McHugh, M.K., Broadhurst, M.K., Sterling, D.J., Millar, R.B., 2014. Comparing
and modifying penaeid beam- and otter-trawls to improve ecological
efficiencies. Fisheries Management and Ecology, 21: 299–311.
McHugh, M.K., Broadhurst, M.K., Sterling, D.J., Millar, R.B., 2015. A ‘simple
anterior fish excluder’ (SAFE) for mitigating penaeid-trawl bycatch. PLoS
ONE 10: e0123124.
McHugh, M.J., Broadhurst, M.K., and Sterling, D.J. 2017. Choosing anterior-gear
modifications to reduce the global environmental impacts of penaeid trawls.
Reviews in Fish Biology and Fisheries, 27: 111–134.
Newland, P., and Chapman, C.J., 1985. Some observations on orientation and
swimming behaviour of Nephrops in relation to trawling. ICES, pp. 1–7.
Newland, P., and Chapman, C.J., 1989. The swimming and orientation behaviour
of the Norway Lobster, Nephrops norvegicus (L.), in relation to trawling.
Fisheries Research, 8: 63–80.
60
Nguyen, K.Q., and Winger, P.D., 2019. Artificial Light in Commercial
Industrialized Fishing Applications: A Review, Reviews in Fisheries Science
& Aquaculture, 27: 106–126.
Nikolic, N., Diméet, J., Fifas, S., Salaün, M., Ravard, D., Fauconnet, L., Rochet,
M.J., 2015. Efficacy of selective devices in reducing discards in the
Nephrops trawl fishery in the Bay of Biscay. ICES Journal of Marine
Science, 72: 1869–1881.
North Sea Advisory Council, 2018. Comments on the Implementation of the
Landing Obligation in the North Sea Demersal Fisheries - Joint
Recommendation for a Delegated Act for 2019. NSAC Advice Ref. 01–
1718.
O’Neill, F.G., 2003. A theoretical study of the factors which influence the
measurement of fishing netting mesh size. Ocean engineering, 30: 2053-
2063.
Olsen, K., Angell, J., Pettersen, F., Løvik, A., 1983. Observed fi sh reactions to a
surveying vessel with special reference to herring, cod, capelin, and polar
cod. FAO Fishery Report, 300: 131–138.
Pol, M.V., Carr, H.A., Ribas, L.R., 2003. Groundfish Trawl Nets Designed to
Reduce the Catch of Atlantic Cod Gadus morhua. Northeast Consortium.
Polet, H., 2002. Selectivity experiments with sorting grids in the North Sea brown
shrimp (Crangon crangon) fishery. Fisheries Research, 54: 217–233.
Punt, A.E., Smith, D.C., Tuck, G.N., Methot, R.D., 2006. Including discard data in
fisheries stock assessments: two case studies from south-eastern Australia.
Fisheries Research, 79: 239–250.
Revill, A., Dunlin, G., Holst, R. 2006. Selective properties of the cutaway trawl
and several other commercial trawls used in the Farne Deeps North Sea
Nephrops fishery. Fisheries Research, 81: 268–275.
Rice, A.L., and Chapman, C.J., 1971. Observations on the burrows and
burrowing behaviour of two mud-dwelling decapod crustaceans, Nephrops
norvegicus and Goneplax rhomboides. Marine Biology, 10: 330–342.
Robertson, J.H.B., 1986. Design and construction of square mesh cod-ends.
Department of Agriculture and Fisheries, Scotland.
61
Robertson, J.H.B., and Shanks, A.M., 1989. Further studies of the size selection
of Nephrops by different cod-ends (No. 1/89). Scottish Fisheries Working
Paper.
Rose, C., 1995. Behaviour of North Pacific ground fish encountering trawls:
Applications to reduce bycatch. Alaska Sea Grant College Program Rep.,
no. 96-03: 235–241.
Rose, C.S., Gauvin, J.R., Hammond, C.F., 2010. Effective herding of flatfish by
cables with minimal seafloor contact. Fishery Bulletin, 108: 136–145.
Ryer, C.H., 2008. A review of flatfish behavior relative to trawls. Fisheries
Research, 90: 138–146.
Ryer, C.H., Rose, C.S., Iseri, P.J., 2010. Flatfish herding behavior in response to
trawl sweeps: a comparison of diel responses to conventional sweeps and
elevated sweeps. Fishery Bulletin, 108: 145–154.
Sala, A., Lucchetti, A., Perdichizzi, A., Herrmann, B., Rinelli, P., 2015. Is square-
mesh better selective than larger mesh? A perspective on the management
for Mediterranean trawl fisheries. Fisheries Research, 161: 182–190.
Salomon, M., and Holm-Müller, K., 2013. Towards a sustainable fisheries policy
in Europe. Fish and Fisheries, 14: 625–638.
Salomon, M., Markus, T., Dross, M., 2014. Masterstroke or paper tiger–The
reform of the EU׳s Common Fisheries Policy. Marine Policy, 47: 76–84.
Sangster, G.I., and Breen, M., 1998. Gear performance and catch comparison
trials between a single trawl and a twin rigged gear. Fisheries Research, 36:
15–26.
Santos, J., Herrmann, B., Mieske, B., Krag, L.A., Haase, S., Stepputtis, D.,
2018a. The efficiency of sieve‐panels for bycatch separation in Nephrops
trawls. Fisheries Management and Ecology.
https://doi.org/10.1111/fme.12323
Santos, J., Herrmann, B., Stepputtis, D., Günther, C., Limmer, B., Mieske, B., ...
and Bethke, E., 2018b. Predictive framework for codend size selection of
brown shrimp (Crangon crangon) in the North Sea beam-trawl fishery. PloS
one, 13: e0200464.
Sistiaga, M., Herrmann, B., Grimaldo, E., Larsen, R.B., Tatone, I., 2015. Effect of
lifting the sweeps on bottom trawling catch efficiency: A study based on the
62
Northeast arctic cod (Gadus morhua) trawl fishery. Fisheries Research, 167:
164–173.
Sistiaga, M., Herrmann, B., Grimaldo, E., Larsen, R.B., Tatone, I., 2016. The
effect of sweep bottom contact on the catch efficiency of haddock
(Melanogrammus aeglefinus). Fisheries Research, 179: 302–307.
Stockhausen, B., Officer, R.A., and Scott, R., 2012. Discard mitigation–what we
can learn from waste minimization practices in other natural resources?
Marine Policy, 36: 90–95.
Strange, E.S., 1984. Review of the fishing trials with Granton and Saro deep sea
trawl gear 1963–1967. Scottish Fisheries Working Paper 8/84.
Strzysewski, W., 1972. Observations on fish behaviour in the trawl. ICES CM
1972/B:9.
Thomsen, B., 1993. Selective flatfish trawling. In: ICES journal of Marine Science
Symposium Vol. 196, pp. 161–164.
Ulrich, C., Reeves, S. A., Vermard, Y., Holmes, S. J., Vanhee, W., 2011.
Reconciling single-species TACs in the North Sea demersal fisheries using
the Fcube mixed-fisheries advice framework. ICES Journal of Marine
Science, 68: 1535–1547.
Ungfors, A., Bell, E., Johnson, M.L., Cowing, D., Dobson, N.C., Bublitz, R.,
Sandell, J., 2013. Chapter seven – Nephrops fisheries in European waters.
In: Magnus, L.J., Mark, P.J. (Eds.), Advances in Marine Biology. Academic
Press, pp. 247–314.
Valentinsson, D., and Ulmestrand, M., 2008. Species-selective Nephrops
trawling: Swedish grid experiments. Fisheries Research, 90: 109–117.
Walsh, S.J., and Hickey, W.M., 1993. Behavioural reactions of demersal fish to
bottom trawls at various light conditions. ICES Journal of Marine Science
Sympium, 196: 68-76.
Wardle, C.S., 1986. Fish behaviour and fishing gear. In The behaviour of teleost
fishes (pp. 463–495). Springer, Boston, MA.
Wardle, C.S., 1992. Studies of fi sh behavior leading to a reduction of bycatch. In:
Jones RP (ed). Proceedings of an International Conference on Shrimp
Bycatch. May 24–27, 1992, Lake Buena Vista, Florida. pp 163–184.
Wardle, C.S., 1993. Fish behaviour and fishing gear. In: Pitcher, T.J. (Ed.),
Behavior of Teleost Fishes, pp. 609–643. Champman & Hall, London.
63
Watson, J.W., 1989. Fish behaviour and trawl design: potential for selective trawl
development. In: Campbell, C.M. (Ed.), Proceedings of the World
Symposium on Fishing Gear and Fishing Vessels. Marine Institute, St.
Johns, Newfoundland, Canada, pp. 25–29.
Wienbeck, H., Herrmann, B., Feekings, J.P., Stepputtis, D., Moderhak, W., 2014.
A comparative analysis of legislated and modified Baltic Sea trawl codends
for simultaneously improving the size selection of cod (Gadus morhua) and
plaice (Pleuronectes platessa). Fisheries research, 150: 28–37.
Winger, P.D., Eayrs, S., Glass, C.W., 2010. Fish behaviour near bottom trawls. In
He, P. (Ed.), Behavior of marine fishes: capture processes and conservation
challenges, pp. 67–102. Wiley-Blackwell, Arnes, IA.
Ydenberg, R.C., and Dill, L.M., 1986. The economics of fleeing from predators.
In: Advances in the Study of Behavior, Vol. 16, pp. 229–249. Academic
Press.
64
Papers
Due to copyright reasons, papers I to III are not included in this version of the
thesis and can be found at the following links:
Paper I: Melli, V., Karlsen, J. D., Feekings, J. P., Herrmann, B., Krag, L.
A., 2018. FLEXSELECT: counter-herding device to reduce
bycatch in crustacean trawl fisheries. Canadian Journal of
Fisheries and Aquatic Sciences, 75: 850–860. https://doi:
10.1139/cjfas-2017-0226
Paper II: Melli, V., Krag, L.A., Herrmann, B., Karlsen, J.D., 2018.
Investigating fish behavioural responses to LED lights in trawls
and potential applications for bycatch reduction in the Nephrops-
directed fishery. ICES Journal of Marine Science, 75: 1682–1692.
https://doi.org/10.1093/icesjms/fsy048
Paper III: Melli, V., Krag L.A., Herrmann, B., Karlsen J.D., 2019. Can active
behaviour stimulators improve fish separation from Nephrops
(Nephrops norvegicus) in a horizontally divided trawl codend?
Fisheries Research. https://doi.org/10.1016/j.fishres.2018.11.027
PAPER IV
Predicting optimal combinations of bycatch reduction devices in
fishing gears: a meta-analytical approach
Melli1* V., Herrmann2,3 B., Karlsen1 J.D., Feekings1 J.P. and Krag1 L.A.
1DTU Aqua, National Institute of Aquatic Resources, North Sea Science Park, DK-9850,
Hirtshals, Denmark
2SINTEF Ocean, Willemoesvej 2, DK-9850 Hirtshals, Denmark
3University of Tromsø, Breivika, N-9037 Tromsø, Norway
Abstract
Global efforts to reduce the capture of non-target species and/or undersized individuals have led to the development of a vast array of bycatch reduction devices (BRDs), in particular for mixed trawl fisheries due to their high bycatch rates. Some of these BRDs could likely benefit from being combined. However, the number of BRDs available would generate a prohibitive number of combinations to be tested in scientific trials. Therefore, in this study we proposed a meta-analytical approach to predict the species-specific, size-selectivity of a trawl with combinations of relevant BRDs, originally tested independently. We applied the method to the well-studied Nephrops (Nephrops norvegicus) directed mixed trawl fishery in the Skagerrak and Kattegat seas and included eight different BRDs: a counter-herding device, a modification of the netting in the trawl body, a horizontal separation and multiple codend configurations. This generated a total of 100 possible combinations. We predicted the size-selectivity of each combination for the target species, Nephrops, and two bycatch species of different economic value, cod (Gadus morhua) and haddock (Melanogrammus aeglefinus). Furthermore, we illustrated how to compare and investigate the performance of the combinations obtained, from both single- and multi-species perspectives, under different catch scenarios. As a result, we identified the most pertinent BRD combinations for the case-study fishery and the alternative harvest strategies that they could support. From the original set of combinations, the meta-analytical approach facilitated the identification of the 15 most pertinent options, of which one would minimize the catch of the two bycatch species considered and another maintain commercial catches of cod. Finally, we identified which interchangeable combinations would lead to a more flexible and dynamic trawl selectivity.
Keywords Bycatch reduction devices, sequential selectivity, optimal gear design, mixed demersal fishery, cod (Gadus morhua), Nephrops norvegicus
1. Introduction
Addressing the issue of unwanted catches in mixed species trawl fisheries is one
of the major challenges of fisheries science and management (Kelleher, 2005;
Graham, 2010; Pérez Roda et al., 2019). For decades, efforts to reduce the
capture of non-target species and/or undersized individuals have involved the
development of technical modifications of the fishing gear; herein termed Bycatch
Reduction Devices (BRDs; Kennelly and Broadhurst, 2002). Some BRDs consist
of modifications to the physical components of the trawl, e.g. mesh size and
characteristics, overall gear geometry, footrope and headline configuration
(Fujimori et al., 2005; Krag et al., 2010; Broadhurst et al., 2012; Herrmann et al.,
2015; Brinkhof et al., 2017). Others use the available knowledge of species-
specific behavioural responses during the catching process (Winger et al., 2010)
to modify species catchability or provide specific escape routes (Krag et al., 2008;
2015; Sistiaga et al., 2015; Lomeli et al., 2018; Melli et al., 2018a). Some BRDs
combine both aspects to exploit differences in size, shape and behaviour among
species to select out the unwanted catches (e.g. Graham and Fryer, 2006; Kim
and Wang, 2010; Karlsen et al., 2018). Together with severe management
measures (e.g discard bans, quotas and reduced fishing effort), BRDs have
contributed to reduce global discard (Zeller et al., 2017). However, specific
fisheries are still bounded to high discard rates (Pérez Roda et al., 2019).
Among the mixed trawl fisheries, those that target crustaceans are the most
challenging in terms of bycatch reduction because they catch a mix of species
with substantial morphological differences and adopt a small mesh size to retain
the target crustacean (Kelleher, 2005; Pérez Roda et al., 2019). Thus, they have
been widely studied and many BRDs have been developed for these fisheries.
This is the case, for example, for the Nephrops (Nephrops norvegicus) directed
mixed trawl fisheries in the northeast Atlantic and penaeid fisheries around the
world (see for review Broadhurst, 2000; ICES, 2004; Catchpole and Revill, 2008).
Dozens of BRDs are documented in literature for these fisheries, with various
degrees of effectiveness in reducing the bycatch and impacts on the target
species. However, because of the morphological and behavioural differences
among the species caught, achieving the desired reduction of bycatch in these
fisheries is rarely (if ever) obtained via one single modification. Among the vast
array of BRDs available, some could benefit from being combined with others.
However, given the number of modifications available, testing all the possible
combinations at sea would be extremely expensive and time consuming.
Therefore, a cost-efficient approach is to first identify the most promising
combinations that are worth further and detailed investigation.
The aim of this study is to use a meta-analytical approach to (i) utilize all the
information available from previously described BRDs (ii) demonstrate how to
compare and identify the most promising BRD combinations by modelling the
combined selectivity of a trawl with multiple pertinent BRDs; and (iii) predict the
performance of each BRD combination under different real catch scenarios, both
single- and multi-species. We used as a case study BRDs developed for the
Nephrops-directed mixed trawl fishery in the Skagerrak and Kattegat seas. This
meta-analytical approach would not only help to re-evaluate interesting designs
that were never implemented into legislation, but would also provide the tools
necessary to determine if and when their combination would represent an optimal
choice for the case-study fishery. The optimal combination will depend on the
specific catch composition, in terms of species and sizes, and on fishermen’s
individual catch goals (Engås and Soldal, 1992; Maynou and Sardà, 2001;
Feekings et al., 2012). For example, when quotas are available, fishermen may
aim at reducing undersized individuals, while when quotas become restrictive, an
additional temporary reduction in marketable sizes might be necessary.
Therefore, we identified the most promising combinations that could match the
gear selectivity with either of these harvest strategies.
2. Case-study fishery
Nephrops, also known as Norway lobster, langoustine or scampi, is one of the
economically most important fishery resources in Europe. Due to its wide
geographical range, it is fished using a wide variety of gear types, under different
technical legislations, and in different environmental conditions (ICES, 2004).
Therefore, to maintain a certain level of similarity in fishing dynamics in our case
study, we focused on the ICES sub-division IIIa. Nephrops catches in this area
are regulated by quota, where Denmark is responsible for taking the majority of
the quota (ICES, 2014). A Minimum Conservation Reference Sizes (MCRS) of 32
mm carapace length is set for EU countries fishing in this area (ICES, 2016). The
main fishing technique used to target Nephrops is demersal otter trawls, often
towed in twin- or multi-rig configurations. It is typically conducted at depths
between 30 and 200 m, on muddy grounds. Besides Nephrops, the fishery
catches several economically important fish species, including cod (Gadus
morhua) and haddock (Melanogrammus aeglefinus). These bycatch species can
be a desirable catch, as they contribute to the economic value of the fishery.
However, they are also subjected to quota management and to the EU landing
obligation (EU, 2013). In addition, cod is regulated by the EU long-term
management plan which aims at restoring depleted cod stocks (EC, 2008). To
achieve this objective, trawlers targeting Nephrops are required since February
2013 to use either a species-selective grid in combination with a 70 mm square
mesh codend (termed Swedish grid due to its main adoption by the Swedish
fleet) or a size selective codend (termed SELTRA) which consists of a 90 mm
diamond mesh codend with either a square mesh panel (SMP; 140 mm in
Skagerrak, 180 mm in Kattegat) or a diamond mesh panel (270 mm; ICES,
2014).
3. Meta-analytical approach
The selectivity of trawl gears is often described by the species- and size-selective
properties of the codend (e.g. mesh size and shape, twine thickness, and codend
circumference). However, multiple selective processes occur during the catching
process. For an individual encountering a trawl to end up being retained in the
codend, it has to be retained during each step of the capture process. Therefore,
we considered the overall selection process in the trawl as sequential. By
assuming each step to be independent from the others, we could combine
population-independent size selectivity estimates for each individual step of the
process. This enabled us to predict the size selectivity of a trawl with a
combination of BRDs tested separately in previous independent experiments. We
selected BRDs pertinent for the case-study fishery for which the data-collection
method allowed estimating absolute species-specific size selectivity, i.e. covered-
codend experiments (Wileman et al., 1996) and paired gears experiments
including a non-selective mesh size (Krag et al., 2014). Furthermore, we selected
BRDs that could be applied independently in different sections of the trawl,
without interfering with each other.
3.1 Predicting the overall trawl selectivity
For a Nephrops, cod or haddock of length l, the likelihood of entering a specific
section of the trawl requires that it is retained by the previous sections. We
divided the trawl in four sections i (Fig. 1), each with an individual retention
probability r(l)i, and modelled the overall retention probability 𝑟𝐶𝑜𝑚𝑏𝑖𝑛𝑒𝑑(𝑙) for an
individual of length l, assuming that it was available for the gear, as the product of
the size selection processes in each section of the trawl:
𝑟𝐶𝑜𝑚𝑏𝑖𝑛𝑒𝑑(𝑙) = ∏ 𝑟(𝑙)𝑖4𝑖=1 = 𝑟𝐻𝑒𝑟𝑑𝑖𝑛𝑔(𝑙) × 𝑟𝐵𝑜𝑑𝑦(𝑙) × 𝑟𝐸𝑥𝑡𝑒𝑛𝑠𝑖𝑜𝑛(𝑙) × 𝑟𝐶𝑜𝑑𝑒𝑛𝑑(𝑙) (1)
where rHerding(l), rBody(l), rExtension(l) and rCodend(l) were the size selectivity in the
respective sections of the trawl.
Figure 1 Schematic drawing of the four independent trawl sections considered in this
study.
3.2 BRDs included in the study
Three species were considered in this study: the target species, Nephrops, and
two bycatch species, cod and haddock. For each of these species, we included in
the meta-analysis seven datasets involving a total of six independent BRDs and
the design used as baseline (i.e. C0, 90 mm diamond mesh codend). The dataset
used as baseline selectivity for haddock differed respect to Nephrops and cod
due to lack of data for this species in Krag et al., (2013; Table 1). Two of the
datasets were from studies using paired gears, while the rest were analysed
according to a covered-codend design.
The three first sections of the trawl (i.e. Herding, Body and Extension) involved
one modification each. When predicting the selectivity of a trawl with combined
BRDs, they were either present (termed H1, B1, E1) or absent (H0, B0, E0). The
last section of the trawl (i.e. Codend) involved four options, numbered from C0 to
C3, Moreover, we included the option of leaving the codend open (C4) by
considering zero retention for those individuals entering that codend.
Table 1 Summary of the datasets included in the meta-analysis.
3.2.1 Herding section
The counter-herding device developed by Melli et al. (2018a) reduced the
retention of fish bycatch by leading fish outside the trawl path (Fig. 2 a). No
significant effect was detected on the catches of Nephrops, whereas the design
was very effective on haddock and less on cod. Length-dependent effects were
detected for both roundfish species.
3.2.2 Body section
The large mesh diamond panel (800 mm) tested by Krag et al. (2014) in the
upper netting of the trawl body exploited fish behaviour inside the trawl. Some
species are known to rise towards the upper netting (Winger et al., 2010) and,
thus, have a higher probability to escape through the large meshes (Fig. 2 b).
The results showed a strong effect on some roundfish, such as haddock, but a
lower effect on cod and flatfish.
3.2.3 Extension section
The data used for this section were those collected for the horizontally divided
trawl codend used as baseline in Melli et al. (2018b) and Melli et al. (2019). The
separator panel and frame influence the vertical distribution of species in the
trawl extension to segregate them in different compartments (Fig. 2c). The results
Reference Trawl section ID Type of data Description
Melli et al., 2018a HERDING H0/H1 Paired gears Counter-herding device
Krag et al., 2014 BODY B0/B1 Paired gearsTrawl with 800 mm diamond meshes
in the upper netting of trawl body
Melli et al., 2018b
and Melli et al.,
2019
EXTENSION E0/E1 Covered-Codend Horizontally divided trawl codend
Krag et al., 2013 CODEND C0 Covered-Codend90 mm diamond mesh codend;
cod and Nephrops
Krag et al., 2016 CODEND C0 Covered-Codend90 mm diamond mesh codend;
haddock
Krag et al., 2015 CODEND C1 Covered-Codend 120 mm diamond mesh codend
Krag et al., 2013 CODEND C2 Covered-Codend90 mm diamond mesh codend with
120 mm square mesh panel
Krag et al., 2015 CODEND C3 Covered-Codend120 mm diamond mesh codend with
180 mm square mesh panel
showed that Nephrops enter the lower compartment, with the exception of a few
bigger individuals. In contrast, most fish entered the upper compartment.
However, while haddock showed a strong preference for the upper compartment,
the vertical distribution of cod was found to be length-dependent, where larger
individuals entered the upper compartment in higher proportions than smaller
ones (Melli et al., 2019).
3.2.4 Codend section
A total of four codend designs were included in the meta-analysis: 90 mm
diamond (C0) from Krag et al. (2013) for Nephrops and cod and Krag et al.
(2016) for haddock; 120 mm diamond (C1) from Krag et al. (2015); 90 mm
diamond with a 120 mm SMP (C2) from Krag et al. (2013); and 120 mm diamond
with 180 mm SMP (C3) from Krag et al. (2015; Table 1; Fig. 2d). The most
important characteristics that determine their selectivity are summarized in Table
2. Using the horizontal separation in the Extension section, we could apply each
codend in either the lower or upper position and in combination with each other.
Table 2 Summary of codend specifications. Circum. = circumference in the codend; Twine thickness = twine thickness of the netting; SMP = square mesh panel; m= metre, mm = millimetre.
CodendLength
(m)
Circum.
(No.
meshes)
Codend
mesh size
(mm)
Twine
thickness
SMP
mesh
size (mm)
SMP
Length
(m)
Cover
mesh size
(mm)
C0 7 100 95.1 4 mm, Double - - 40
C1 6 92 127.4 5 mm, Double - - 40
C2 7 100 94.8 4 mm, Double 126.1 3 40
C3 6 92 126.9 5 mm, Double 180 3 40
Figure 2 Schematic drawings of the BRDs included in the study. a) Counter-herding device from Melli et al., 2018a; b) Large meshes in the upper netting of the trawl body from Krag et al., 2014; c) Horizontally divided trawl codend from Melli et al., 2018b; d) C0: 90 mm diamond codend from Krag et al., 2013; C1: 120 mm diamond codend from Krag et al., 2014; C2: 90 mm diamond codend with 120 mm SMP from Krag et al., 2013; C3:120 mm diamond codend with 180 mm SMP from Krag et al., 2014.
3.3 Estimation of size-selectivity from original datasets
For each of the datasets included in the meta-analysis, and for each species
separately, we estimated the size-dependent retention probability r(l) with the
size represented by the length l of the species (Wileman et al. 1996). Two
different approaches were followed, depending on the type of experimental data
originally collected (i.e. covered-codend or paired gears).
3.3.1 Covered-codend
Several parametric models were tested to describe the size selection, r(l, v)
where v is a vector consisting of the parameters of the model. The values of the
parameters v were then estimated so that the experimental data (averaged over
hauls) would be most likely observed, assuming that the model was able to
describe the data sufficiently well. We considered a total of nine models. The four
models Logit, Probit, Gompertz and Richard described in Wileman et al. (1996);
the dual sequential selection models applied by Zuur et al. (2001), Sistiaga et al.
(2010) and Krag et al. (2016); and the triple logistic model described by Noack et
al. (2017). The dual sequential and triple logistic models imply that two and three
selective processes, respectively, are expected to contribute to the size selection
process. These processes are assumed sequential, meaning that the proportion
of individuals exposed to the second process is assumed to consist of those that
were not exposed to the first process and additionally those that were, but were
retained. For the dual sequential model, we assumed the first process to be
modelled by a logistic curve, while considering all the four classical models (Logit,
Probit, Gompertz and Richard) for the second process. To identify the best model
for each species and dataset, we followed the procedure of inspecting goodness
of fit as described by Wileman et al. (1996), selecting the model with the lowest
Akaike information criterion (AIC) value (Akaike, 1974).
3.3.2 Paired gears
For the BRDs in the Herding and Body sections, where the experimental data
were not expected to follow an s-shaped selectivity model, we used the flexible
polynomial model of order four often applied to catch comparison of paired gears
data (Krag et al., 2014; Melli et al., 2018a). This provided 31 additional models
that were considered as candidates for describing the experimental data. The
model with the lowest AIC was selected to either describe the size-dependent
retention rate, r(l), according to Krag et al. (2014), or the catch comparison rates,
cc(l), according to Melli et al. (2018a). In the latter, cc(l) was used to estimate the
catch ratio, cr(l), using the relationship between cr and cc (Herrmann et al. 2017).
The catch ratio expresses the relative selectivity of the design when compared to
the control trawl. This step was required because the device tested by Melli et al.
(2018a) could not only reduce the retention of individuals but also increase it.
Thus, a retention rate limited to 1.0 would not fully represent the absolute effect
of the device. Using the catch ratio, a value of 1.0 implies that there is no
difference in catch respect to the control, whereas values above and below 1.0
imply increased or reduced catches, respectively.
3.3.3 Uncertainty estimation
The model chosen for each dataset, its parameters and fit statistics are
summarised in Appendix 1. Regardless of the model selected, once the model
was chosen for each dataset and each species, a double-bootstrap method with
1000 repetitions was applied to consider both within and between hauls variation
in size selectivity (Millar, 1993). For each design, the result of the bootstrap
method was a new set of data that was analysed using the identified selection
model. This bootstrap set, besides being used to estimate Efron 95% confidence
intervals (CIs; Efron, 1982) for the specific design, was an essential step to
estimate of uncertainties for the combined selectivity (see section 3.4).
3.4 Estimation of combined selectivity
Considering the BRDs included in this study, and because the modification
introduced in the Extension section is a separation into compartments, rExtension(l)
which expresses the probability of an individual of length l to enter the lower
compartment, Eq. (1) becomes:
𝑟𝐶𝑜𝑚𝑏𝑖𝑛𝑒𝑑(𝑙) = 𝑟𝐻𝑒𝑟𝑑𝑖𝑛𝑔(𝑙) × 𝑟𝐵𝑜𝑑𝑦(𝑙) × [𝑟𝐸𝑥𝑡𝑒𝑛𝑠𝑖𝑜𝑛(𝑙) × 𝑟𝐶𝑜𝑑𝑒𝑛𝑑𝐿(𝑙) + (1.0 − 𝑟𝐸𝑥𝑡𝑒𝑛𝑠𝑖𝑜𝑛(𝑙)) ×
𝑟𝐶𝑜𝑑𝑒𝑛𝑑𝑈(𝑙)] (2)
where rCodendL(l) is the size selectivity of the lower codend and rCodendU(l) of the
upper one. When no separation is included in the trawl rExtension(l) equals one,
meaning that all individuals enter the only codend available. When no BRD is
inserted in the Herding and Body sections, rHerding(l) and rBody(l) are assumed to
equal one, meaning that individuals that enter that section are retained as they
would in a standard trawl.
According to Eq. (2), we defined as baseline of this study a trawl with no BRD in
the Herding area (rHerding(l)=1.0), no BRD in the Body section (rBody(l)=1.0), no
separation in the Extension (rExtension(l)=1) and a 90 mm diamond codend (C0). All
the possible combinations of BRDs were obtained by substituting the size
selectivity in the pertinent sections in Eq. (2).
To estimate 95% Efron CIs for each rCombined(l), we used the bootstrap sets
obtained in section 3.3.3 for each original design. Because these bootstrap sets
were obtained independently, a new bootstrap set of results for rCombined(l) was
created using:
𝑟𝐶𝑜𝑚𝑏𝑖𝑛𝑒𝑑(𝑙)𝑖 = 𝑟𝐻𝑒𝑟𝑑𝑖𝑛𝑔(𝑙)𝑖 × 𝑟𝐵𝑜𝑑𝑦(𝑙)𝑖 ×
[𝑟𝐸𝑥𝑡𝑒𝑛𝑠𝑖𝑜𝑛(𝑙)𝑖 × 𝑟𝐶𝑜𝑑𝑒𝑛𝑑𝐿(𝑙)𝑖 + (1.0 − 𝑟𝐸𝑥𝑡𝑒𝑛𝑠𝑖𝑜𝑛(𝑙)𝑖) × 𝑟𝐶𝑜𝑑𝑒𝑛𝑑𝑈(𝑙)𝑖] 𝑖 ∈ [1 … 1000] (3)
where i denotes the bootstrap repetition index (Herrmann et al. 2018). In Eq. (3)
the 1000 bootstrap sets generated from the original datasets were either
multiplied or summed to obtain the new bootstrap set for the combined
configuration. Based on this final bootstrap set, 95% Efron Percentile CIs for
rCombined(l) were estimated.
All the analyses were conducted with the software SELNET (Herrmann et al.,
2012).
4. Identification of the most promising combinations
Once the selectivity of the BRD combinations was modelled, its applicability and
relevance for the case-study fishery was investigated through different tools.
These tools allowed us to select out all the combinations that did not produce the
desired effects, in terms of target and bycatch catches, and ultimately determine
which BRD combinations have the highest potential for optimizing the selectivity
of the Nephrops-directed fishery in Skagerrak and Kattegat.
In particular, we used three tools to visualize and summarize the performance of
the combinations: delta selectivity, cumulative population caught, and
performance indicators.
4.1 Delta selectivity
The first tool to investigate the species-specific, population-independent
performance of a BRD combination entailed comparing it to the size-selectivity of
the baseline design. If rB(l) is the size selectivity of the baseline trawl, and rC(l)
the size selectivity of the combination of interest, then the difference in selectivity,
Δr(l) is:
Δ𝑟(𝑙) = 𝑟C(𝑙) − 𝑟B(𝑙) (4)
Uncertainties for Δr(l) were estimated using the approach described in (section
3.1), but by subtracting the two independently generated bootstrap sets. In
general, Δr(l) spans between -1 and 1, where values above 0.0 imply that the
combination has higher retention probability for individuals of length l than the
baseline design, while values below 0.0 imply a lower retention probability.
Δr(l) was estimated separately for each species of interest and used to identify
the species-specific length-range significantly affected by the combination of
BRDs. Ideally, for a good combination, Δr(l) would be close to or above 0.0 for
target species and below 0.0 for unwanted species.
4.2 Cumulative population caught
For the second tool, we investigated the performance of each combination under
three realistic population scenarios for each species (P1-P3). The three
population structures nPop(l) for each species were estimated using the original
datasets included in this study, by pooling data over hauls for hauls with more
than 20 individuals. The data included for each population are summarized in
Appendix 2. For covered-codend datasets, data from codend and cover were
summed to reflect the population entering the trawl. In contrast, for paired gears
datasets, only data from the control trawl were used to obtain the population
entering the trawl. Within species, the populations differed in length-range
represented, density of each length-class and mode/s (i.e. most frequent length
class represented). For each population, uncertainties (95% Efron CIs) were
obtained based on a double bootstrap method. This considered both the
between-hauls variability in the structure of the population entering the codend
and within-haul variability deriving from limited numbers of fish or Nephrops
entering the codend in that specific haul, as well as the potential effect of
subsampling. Specifically, the double bootstrap procedure accounted for
between-hauls variability by selecting hauls h with replacement from the h
number of hauls selected from the dataset. Within-haul uncertainty was
accounted for by resampling with replacement the fish or Nephrops length-
measured, followed by raising the numbers according to the subsampling ratios
within each compartment. The number resampled for each compartment in this
inner bootstrap loop equalled the total number of individuals length-measured in
the respective compartment in the selected haul. 1000 bootstrap repetitions were
conducted and used to calculate the 95% Efron CIs for the population nPop(l).
Using the size-selection curves predicted in section 3.4 for each BRD
combination, and applying them to nPop(l), we obtained simulated catches,
nCatch(l). We visualized the population caught as the cumulative distribution
function for the catch:
𝐶𝐷𝐹_𝑛𝐶𝑎𝑡𝑐ℎ(𝐿) = ∑ {𝑟𝑐𝑜𝑚𝑏𝑖𝑛𝑒𝑑(𝑙) × 𝑛𝑃𝑜𝑝(𝑙)}𝐿𝑙=0 (5)
For each 𝐶𝐷𝐹_𝑛𝐶𝑎𝑡𝑐ℎ(𝐿) we calculated 95% CIs based on the bootstrap sets for
𝑟𝑐𝑜𝑚𝑏𝑖𝑛𝑒𝑑(𝑙) and 𝑛𝑃𝑜𝑝(𝑙) using the approach previously described for 𝑟𝑐𝑜𝑚𝑏𝑖𝑛𝑒𝑑(𝑙).
Ideally, a good BRD combination would lead to catching more commercial sized
than undersized individuals regardless of the population structure. Because
𝐶𝐷𝐹_𝑛𝐶𝑎𝑡𝑐ℎ(𝐿) expresses the retention rate up to a certain length, the rate at the
species-specific MCRS denotes the proportion of catch that is undersized for a
given population scenario. Moreover, BRD combinations whose efficiency was
significantly affected by the population structure had 𝐶𝐷𝐹_𝑛𝐶𝑎𝑡𝑐ℎ(𝐿) with non-
overlapping CIs for the different scenarios.
4.3 Performance indicators
For the third tool, we converted the number of individuals per length class into
weights and used them to calculate summary indicators (Sala et al., 2015). This
is particularly useful to evaluate the usability of a BRD combination in a fishery
because quotas are typically expressed in weight, not in number of individuals.
For cod and haddock, we used the length-weight relationship available on
fishbase.org for ICES Division IIIa. For Nephrops we used the data from the Data
Collection Framework (DCF) and International Bottom Trawl Survey (IBTS)
programs in Skagerrak and Kattegat. The specific values of the factors a and b
used for the length-weight conversion are provided in the Supplementary
Material.
Using the size-selection curves predicted in section 3.1 for each BRD
combination, and applying them to the population expressed in weight, 𝑤𝑙 ×
𝑛𝑃𝑜𝑝(𝑙), we obtained simulated catches in weight, 𝑤𝑙 × 𝑟𝑐𝑜𝑚𝑏𝑖𝑛𝑒𝑑(𝑙) × 𝑛𝑃𝑜𝑝(𝑙).
𝑤𝑙 is the weight for length class l obtained by the species specific relationship
𝑤𝑙 = 𝑎 × 𝑙𝑏. These were then summarized by calculating three different indicators
(wP−, wP+, and WDiscardRatio), for each of the species-specific nPop(l)
separately. wP− and wP+ were used to express the percentage of weight retained
for individuals below and above the species-specific MRCS, respectively, for a
specific combination of BRDs. Ideally, a selective gear would have a low wP− for
both the target and bycatch species. In contrast, wP+, should be high for the
target species and either high or low for the bycatch species, depending on the
catch goals of the individual fisherman. The wDiscardRatio was calculated to
express the percentage of weight of undersized individuals respect to the total
weight of the catch. Normally, a low wDiscardRatio would imply that the gear is
well suited to the catch scenario. However, BRDs that strongly reduce the weight
of the commercial sized bycatch, thus enhancing high quota savings, will have a
relatively high value of wDiscardRatio. The indicators were calculated by:
𝑤𝑃− = 100 ∑ {𝑎×𝑙𝑏×𝑟𝑐𝑜𝑚𝑏𝑖𝑛𝑒𝑑(𝑙)×𝑛𝑃𝑜𝑝(𝑙)}𝑙<𝑀𝐶𝑅𝑆
∑ {𝑎×𝑙𝑏×𝑛𝑃𝑜𝑝(𝑙)}𝑙<𝑀𝐶𝑅𝑆
𝑤𝑃+ = 100 ∑ {𝑎×𝑙𝑏×𝑟𝑐𝑜𝑚𝑏𝑖𝑛𝑒𝑑(𝑙)×𝑛𝑃𝑜𝑝(𝑙)}𝑙>MCRS
∑ {𝑎×𝑙𝑏×𝑛𝑃𝑜𝑝(𝑙)}𝑙>𝑀𝐶𝑅𝑆 (6)
𝑤𝐷𝑖𝑠𝑐𝑎𝑟𝑑𝑅𝑎𝑡𝑖𝑜 = 100 ∑ {𝑎 × 𝑙𝑏 × 𝑟𝑐𝑜𝑚𝑏𝑖𝑛𝑒𝑑(𝑙) × 𝑛𝑃𝑜𝑝(𝑙)}𝑙<𝑀𝐶𝑅𝑆
∑ {𝑎 × 𝑙𝑏 × 𝑟𝑐𝑜𝑚𝑏𝑖𝑛𝑒𝑑(𝑙) × 𝑛𝑃𝑜𝑝(𝑙)}𝑙
We used the MRCS for the ICES division IIIa: 32 mm carapace length for
Nephrops, 30 cm and 27 cm total length for cod and haddock, respectively. All
indicators (wP−, wP+ and wDiscardRatio) were estimated with uncertainties for
each species and population scenario, using the bootstrap set for rCombined(l) and
nPop(l). Specifically, by first calculating the values for the indicators based on
each bootstrap repetition result for rCombined(l) and nPop(l) synchronous in (5) to
obtain a bootstrap set for the indicator values. Finally, based on the resulting
bootstrap set, 95% Efron CIs were obtained for each of the indicators.
Because uncertainties are typically wider at the tails of the length range
represented in the data, and since the conversion into weights accentuate the
influence of the larger and less represented length classes when estimating the
indicators, we restricted the length range for each of the species analysed
according to the strength of the data in the original datasets. In particular, we set
the minimum length class as the smallest length class including at least five
individuals in all the datasets. Similarly, we determine the maximum length of the
range as the largest length class with at least five individuals in all the datasets.
The length range was, therefore, restricted to 20.5–76.5 cm and 18.5–43.5 cm for
cod and haddock, respectively, and to 20.5–59.5 mm for Nephrops. This
approach prevented the less represented length classes from compromising the
information contained in the main bulk of data.
4.4 Multispecies comparison of the best combinations
Once a subset of the overall combination was identified, we could simulate and
compare their performance under a multispecies catch scenario. Following the
process described in section 4.2, but restricting the number of hauls selected to
those including all the three species of interest, we estimated a multispecies set
of populations, nPop(l). We then used these populations to calculate the
indicators described above. Each indicator was used to grade the overall
performance of the combinations, and identify the best options, depending on the
hypothetical catch goals (e.g. maximum quota saving or maximum economic
output).
5. Results and discussion
By using this meta-analytical approach we could identify hidden potential for
improving selectivity in the case-study fishery, i.e. Nephrops-directed trawl
fishery, deriving from the combination of previously described, pertinent BRDs.
Nonetheless, the predicted selectivity curves and performances of the BRD
combinations are theoretical, assuming that when combined the BRDs would
perform as they do when applied individually. However, for this assumption to be
true, the combined application of the BRDs in the trawl would need to be carefully
planned. For example, when inserting a codend with a SMP as the lower codend
of a horizontally divided trawl, the escape probability through the meshes of the
SMP would likely be affected by the obstruction represented by the upper
codend. Therefore, to perform experimentally as predicted, the design would
need to prevent proximity between the two codends. This additional complexity
may be justified if the improvement in selectivity of the BRD combination is
substantial. The major outcome of this meta-analysis is indeed the identification
the most promising combinations that could be worthwhile the time and cost
outlay associated with experimental investigation.
5.1 Prediction of combined BRDs selectivity
From the datasets included in this study, we obtained a total of 100 possible
combined designs for Nephrops and cod. Since data for haddock were
unavailable for C2 (i.e. 90 mm diamond mesh size codend with a 120 mm SMP),
the number of possible combinations for haddock was 64. For all the species,
four combinations had rCombined(l) equal to 0.0, relative to the theoretical option of
fishing with an open codend (C4) when no separation in the extension was
included (E0). Thus, the number of combined selectivity curves was 96 for
Nephrops and cod (Fig. 3 and 4) and 60 for haddock (Fig. 5). In each figure, the
combinations that were identified as the most promising for the case-study fishery
at the end of the elimination process described in section 5.2 are highlighted.
By combining multiple BRDs we were able to produce alternative selectivity
patterns (Fig. 3, 4 and 5), respect to the traditional S-shaped selection curve of
trawl gears (Dickson et al., 1995; Wileman et al., 1996). In the simple codends
C0 and C1 (i.e. with no SMP) the retention probability increased with the size of
the individuals until reaching the 100% retention level. The inclusion of a SMP did
already alter the shape of the selectivity curve, to the point that the curve
appeared to be split into two sections with different steepness (e.g. Fig. 3,
H0B0E0C3). This is caused by the sequential selection processes of the SMP
and codend, respectively. Consequently, the addition of selection processes (e.g.
BRDs) increased the level of complexity of the selectivity curves (e.g. Fig. 4,
H1B1E0C0), sometimes increasing the retention of the smaller length classes
and/or decreasing the retention of the larger ones (e.g. Fig. 4, H1B1E0C0).
Fig
ure
3. P
red
icte
d s
ele
ctiv
ity c
urv
es (s
olid
lines) a
nd
95
% E
fron
CIs
(da
sh
ed
lines) o
f the
96
co
mb
ina
tion
s fo
r Ne
ph
rop
s. G
rey rib
bo
ns h
igh
ligh
t the
most p
ertin
ent
co
mb
inatio
ns s
ele
cte
d a
fter th
e e
va
lua
tion
pro
cess. T
he
figure
is re
ad
as a
ma
trix w
ith e
ach
plo
t inclu
din
g c
urv
es w
ith a
co
mbin
atio
n o
f the
BR
Ds in
dic
ate
d a
s c
olu
mn
an
d ro
w title
s. C
olu
mn
s a
dd
the
ho
rizo
nta
l sep
ara
tion
(E1
) with
the
spe
cifie
d c
ode
nd
as lo
we
r co
den
d. R
ow
s a
dd
BR
Ds in
the
traw
l bo
dy (B
1), h
erd
ing
are
a (H
1), o
r bo
th
(H1B
1). B
RD
s in
unspe
cifie
d s
ectio
ns a
re a
bse
nt (i.e
. H0
; B0; E
0). T
he
first p
lot (u
ppe
r left c
orn
er) s
ho
ws th
e s
ele
ctiv
ity o
f a tra
wl w
ith th
e fo
ur s
ing
le c
od
en
ds: C
0=
Red;
C1=
Gre
en
; C2=
Blu
e; C
3=
Pu
rple
. In c
olu
mn
2 to
6, th
e c
olo
ur o
f each
cu
rve
co
rresp
onds to
the
rela
tive
co
de
nd
in th
e u
pp
er p
ositio
n. T
he
colo
ur P
ink is
use
d fo
r the
op
en
co
de
nd
(C4
).
Fig
ure
4
. P
red
icte
d se
lectivity cu
rve
s (s
olid
lin
es)
an
d 9
5%
E
fron
C
Is (d
ashe
d lin
es)
of
the
96
com
bin
atio
ns fo
r co
d.
Gre
y rib
bon
s h
igh
ligh
t th
e m
ost
pe
rtin
ent
co
mb
ination
s s
ele
cte
d a
fte
r th
e e
va
lua
tio
n p
rocess.
Th
e f
igure
is r
ead
as a
ma
trix
with
ea
ch
plo
t in
clu
din
g c
urv
es w
ith
a c
om
bin
atio
n o
f th
e B
RD
s in
dic
ate
d a
s c
olu
mn
an
d r
ow
title
s.
Colu
mn
s a
dd
the
ho
rizo
nta
l sep
ara
tio
n (
E1
) w
ith
th
e s
pe
cifie
d c
ode
nd
as lo
we
r co
den
d.
Ro
ws a
dd
BR
Ds in t
he
tra
wl b
od
y (
B1
), h
erd
ing
are
a (
H1
), o
r b
oth
(H1B
1).
BR
Ds in
unspe
cifie
d s
ectio
ns a
re a
bse
nt
(i.e
. H
0;
B0;
E0
). T
he
first
plo
t (u
ppe
r le
ft c
orn
er)
sho
ws t
he
se
lectivity o
f a
tra
wl w
ith
th
e f
ou
r sin
gle
cod
en
ds:
C0
=R
ed;
C1=
Gre
en
; C
2=
Blu
e;
C3=
Pu
rple
. In
co
lum
n 2
to
6,
the
colo
ur
of
each
cu
rve
co
rre
sp
onds t
o t
he
re
lative
co
de
nd
in
th
e u
pp
er
positio
n.
Th
e c
olo
ur
Pin
k i
s u
se
d f
or
the
op
en
co
de
nd
(C
4).
Fig
ure
5. P
redic
ted
se
lectiv
ity c
urv
es (s
olid
lines) a
nd
95%
Efro
n C
Is (d
ash
ed lin
es) o
f the
60
com
bin
atio
ns fo
r ha
dd
ock. G
rey rib
bo
ns h
ighlig
ht th
e m
ost p
ertin
en
t
co
mb
inatio
ns s
ele
cte
d a
fter th
e e
va
lua
tion
pro
cess. T
he
figure
is re
ad
as a
ma
trix w
ith e
ach
plo
t inclu
din
g c
urv
es w
ith a
co
mbin
atio
n o
f the
BR
Ds in
dic
ate
d a
s c
olu
mn
an
d ro
w title
s. C
olu
mn
s a
dd
the
ho
rizo
nta
l sep
ara
tion
(E1
) with
the
spe
cifie
d c
ode
nd
as lo
we
r co
den
d. R
ow
s a
dd
BR
Ds in
the
traw
l bo
dy (B
1), h
erd
ing
are
a (H
1), o
r bo
th
(H1B
1). B
RD
s in
unspe
cifie
d s
ectio
ns a
re a
bse
nt (i.e
. H0
; B0; E
0). T
he
first p
lot (u
ppe
r left c
orn
er) s
ho
ws th
e s
ele
ctiv
ity o
f a tra
wl w
ith th
e fo
ur s
ing
le c
od
en
ds: C
0=
Red;
C1=
Gre
en
; C3=
Pu
rple
. In c
olu
mn
2 to
6, th
e c
olo
ur o
f each
cu
rve
co
rrespo
nds to
the
rela
tive
co
de
nd
in th
e u
pp
er p
ositio
n. T
he
co
lou
r Pin
k is
use
d fo
r the
op
en
co
den
d(C
4).
Before interpreting the combined selectivity patterns, in particular of the most
complex designs, it is important to observe their uncertainties. For each predicted
selectivity curve, the 95% Efron CIs reflected the strength of the data and the
consistency (between-hauls variation) of the effect in the original datasets. Thus,
combinations of BRDs with high binomial noise in one or more of the original
datasets resulted in wide CIs. In particular, this is the case for the tails of the
length-range of each species, where the dataset with the most restricted length
range limited the inferential power for that combination. This result prevented
predictions which were not supported by the original experimental data.
Examples can be observed in Fig. 3, where the combined selectivity curves of H1
and H1B1 for Nephrops resembled a bell-shaped curve (Dickson et al., 1995;
Lövgren et al., 2016), with a high retention of the central length classes and a low
retention of the smaller and larger classes. However, as expressed by the wide
CIs, the effect on the larger classes is inconclusive and should not be interpreted.
Moreover, many of the combined selectivity curves involving the counter-herding
device (H1) exceeded retention rates of 1.0 (e.g. Fig. 3 columns 1–4 of rows 3–4,
and Fig. 4 all curves of rows 3–4). This was caused by the use of the catch ratio
to describe the effect of the counter-herding device, which in some cases
increased the number of individuals entering the trawl.
Of the variety of selectivity patterns that could be achieved by combining the
BRDs included in this study, only few combinations would be viable for the case-
study fishery. In particular, high retention rates of commercial sized Nephrops
were an essential requirement for a Nephrops-directed fishery. Indeed, the BRD
combinations highlighted tended to all have a similar selective pattern in the main
length-range of Nephrops (Fig. 3). In contrast, the desirable effects on the two
bycatch species, cod and haddock, were more complicated to evaluate, since
they can vary depending on the specific catch goals. Some of the highlighted
BRD combinations strongly reduced the retention of haddock but not that of cod;
others reduced the undersized bycatch but retained the commercial sized cod;
others reduced both the undersized and commercial sized fractions of both
species. In the following section we have illustrated the elimination process
undertaken to identify the most pertinent (highlighted) combinations for the case-
study fishery that could be worth further experimental investigations.
5.2 Identification of the most promising combinations
To determine the most relevant combinations, we inspected both the absolute
selectivity and the performance under different catch scenarios of each
combination, relative to the three species of interest. This elimination process
was iterative, as each of the tools used (section 4) provided different information
about the efficiency and applicability of the BRD combinations. However, the
indicators (third tool) proved to be the most efficient measure to determine if the
BRD combination could represent a viable option for the case-study fishery. In
particular, all the BRD combinations predicted to cause a major loss of
commercial sized Nephrops, under any of the considered population scenarios,
were excluded. In contrast, the other two tools provided a more detailed and
length-based information, and were used in few cases where the absolute
selectivity curve (section 5.1) and the indicators were not sufficient to clearly
select or eliminate the BRD combination from the pool.
Hereafter we illustrated the information provided by each tool through examples,
which are meant to clarify the desired properties of a pertinent BRD combination
rather than describing the selectivity pattern of all the combinations obtained.
5.2.1 Delta selectivity
The species-specific absolute selectivity of BRD combinations was compared to
that of the baseline gear to identify significant changes in selectivity (Fig. 6). In
the example, where the combination H0B0E1C1C0 is compared to the baseline,
the Delta selectivity highlighted significant major losses of commercial sized
Nephrops (red curve). Moreover, the combination resulted in a moderate but
significant reduction of cod between 20 and 40 cm (green curve), but no
significant change in haddock catches (blue curve).
Indeed, the combination H0B0E1C1C0 included a horizontally divided codend
with a lower codend of 120 mm diamond meshes and an upper codend of 90 mm
diamond meshes. Intuitively, this combination would not be applicable to a
Nephrops-directed fishery, as most Nephrops enter the lower compartment (Melli
et al., 2018b; Melli et al., 2019; Karlsen et al., 2018) and would not be retained
with a 120 mm diamond mesh size (Krag et al., 2008). Similarly, most haddock
enter the upper compartment and would not be appropriately selected out by the
90 mm diamond mesh size (Graham and Ferro, 2004; Melli et al., 2018b). In
contrast, the reduction in cod between 20 and 40 cm highlighted how this length-
range enters in greater proportion the lower compartment and would, therefore,
benefit from encountering a 120 mm mesh size.
Figure 6 Predicted selectivity with 95% Efron CIs (solid lines with ribbons) of the combination H0B0E1C1C0 for Nephrops (red), cod (green) and haddock (blue). Delta selectivity with 95% Efron CIs (solid line with dashed lines) of the H0B0E1C1C0 selectivity respect to the baseline (H0B0E0C0).
5.2.2 Cumulative population caught
In terms of bycatch reduction, the cumulative population caught was used to
inspect the variability in discard ratios of BRD combinations, under different
population scenarios (Fig. 7). Indeed, when implementing a BRD to reduce the
bycatch of commercial species, a first objective is generally to reduce the catch of
undersized individuals, more than that of valuable sizes. However, the efficiency
of some of the BRD combinations in selecting out undersized individuals was
found to be strongly affected by the structure of the population encountered (Fig.
7). For example, in the third population scenario (P3), where the mode in the
population is close to the MCRS for cod (30 cm), approximately 75% of the
population caught with the combination H1B0E1C2C1 consisted of undersized
individuals. This was caused by the BRDs included in this combination
(H1B0E1C2C1), which were a counter-herding device, a horizontally divided
codend, a lower codend of 90 mm diamond mesh size with a 120 mm SMP and
an upper codend with 120 mm diamond meshes. Although this combination could
have been expected to be a viable option for the case-study fishery, the results
showed that it is less performing than other combinations with a lower amount of
BRDs involved. This is because none of the BRDs included in H1B0E1C2C1
were very effective in improving the selectivity for cod around 30 cm, which enter
more frequently the lower compartment and have a lower escape rate through
SMPs (Krag et al., 2015; Melli et al., 2018a). Accordingly, combinations with
population-dependent efficiencies in reducing undersized catches were
considered less desirable than those with lower but more stable efficiency.
Figure 7 On the left, cumulative population caught with 95% Efron CIs (solid lines with ribbons) with the combination H1B0E1C2C1 under three population scenarios of cod. The vertical dashed line indicates the MCRS (30 cm). On the right, structure of the three populations used with 95% Efron CIs (solid lines with ribbons).
5.2.3 Performance indicators
Finally, the performance of the BRD combinations was investigated from a
management-fishermen perspective by calculating the weight indicators to
determine how fishermen’s quotas and incomes may be affected. This tool is
particularly useful to overview the performance of the combinations under
realistic catch scenarios and to summarize a large quantity of information,
including some of those conveyed by the other tools (e.g. proportion of
undersized caught). All the approximately 1000 indicators for all the BRD
combinations and for each of the population scenarios (P1-P3 per species and a
multispecies scenario) are presented as Supplementary Material. Here, to
illustrate the interpretation of the indicators, we present a subset of BRD
combinations (Table 3). The subset includes the indicators for cod population
scenarios for: the baseline design, two combinations that minimally affected the
retention of cod (red), two combinations with a moderate effect (yellow) and two
combinations that minimized cod catches (green). The baseline design retained
on average between 66.6 and 78.2% of the weight of undersized cod (wP-). The
addition of a horizontally divided codend with a SMP in the lower codend
(H0B0E1C2C0) or of the counter-herding device (H1B0E0C0) did not significantly
reduce the weight of cod below the MCRS (Table 3). In contrast, pairing the
counter-herding device with a large mesh panel in the trawl body (H1B1E0C0)
reduced significantly both undersized (wP- between 33.0 and 34.2%) and
commercial sized catches (wP+ between 33.3 and 38.8%) because of their
complementary efficacy (Krag et al., 2014; Melli et al., 2018a). An even lower
retention of undersized cod could be achieved by combining the counter-herding
device with a horizontally divided trawl codend having a 90 mm mesh size and
120 mm SMP in the lower compartment and 120 mm mesh size in the upper
compartment (H1B0E1C2C1). Moreover, this combination retained on average a
higher percentage of commercial cod (in weight), a result that could be desirable
within the case-study fishery (Table 3). However, in all these examples, the
wDiscardRatio (i.e. percentage of weight discarded respect to the total caught)
was significantly and substantially higher in the scenario with high density of
individuals around the MCRS (P3; Table 3).
BRD combinations highly effective on cod (e.g. H0B1E0C2 and H1B1E1C2C4),
not only reduced the percentage of weight retained below and above the MCRS,
but also had lower wDiscardRatio for the most critical scenario (P3).
Table 3 Indicators for the baseline design (H0B0E0C0) and six examples of combinations for cod, under three population scenarios (P1, P2 and P3). 95 % Efron CIs are shown within parenthesis. The examples are ordered according to their mean wP-, colours are used to highlight the efficiency of the combination in reducing catches of cod: red = low effect; yellow = medium effect; green = high effect.
5.3 Most promising BRD combinations for the case-study fishery
After inspecting the performance of the BRD combinations, we identified 15
combinations for Nephrops and cod, out of the original set of 96 combinations,
which could be applicable to the case-study fishery (Table 4). Of these 15
combinations, only 10 included predictions for haddock, due to the lack of data
for the 90 mm diamond mesh size codend with a 120 mm SMP (C2). All the
selected combinations retained on average a weight of commercial sized
Nephrops within 15% from the baseline design. This derived mainly from having
a lower codend of 90 mm diamond mesh size, whenever the horizontal
separation was introduced. Only one of the selected BRD combinations had a
wP− (%) wP+ (%) wDiscardRatio (%)
P1 66.6 (53.9 – 77.6) 98.7 (98.0 – 99.3) 2.8 (1.6 – 4.9)
P2 78.2 (70.5 – 86.1) 96.0 (94.3 – 97.2) 8.1 (5.4 – 11.1)
P3 69.0 (57.0 – 78.5) 94.0 (91.7 – 96.4) 59.9 (48.3 – 66.8)
P1 52.7 (41.4 – 63.0) 95.5 (93.7 – 97.1) 2.3 (1.3 – 4.0)
P2 63.6 (56.2 – 71.2) 90.7 (88.0 – 93.2) 7.0 (4.7 – 9.8)
P3 54.7 (44.4 – 63.2) 86.7 (83.0 – 91.7) 56.2 (43.2 – 63.7)
P1 55.8 (41.7 – 67.5) 63.4 (50.0 – 81.4) 3.6 (1.8 – 6.3)
P2 61.7 (50.1 – 73.8) 63.2 (51.3 – 79.0) 9.5 (6.0 – 13.3)
P3 57.6 (43.4 – 68.6) 62.9 (52.2 – 77.8) 65.0 (52.4 – 71.7)
P1 33.6 (23.4 – 42.7) 38.8 (28.3 – 51.9) 3.6 (1.7 – 6.8)
P2 33.0 (24.7 – 42.4) 35.9 (27.8 – 48.1) 9.0 (5.4 – 13.6)
P3 34.2 (24.0 – 43.6) 33.3 (26.4 – 44.9) 67.6 (51.8 – 76.0)
P1 12.4 (7.6 – 16.6) 55.0 (41.9 – 71.3) 1.0 (0.5 – 1.8)
P2 15.8 (10.9 – 20.5) 47.4 (36.3 – 61.3) 3.5 (2.0 – 5.4)
P3 12.6 (7.8 – 16.5) 40.9 (32.2 – 55.5) 38.6 (22.9 – 49.7)
P1 6.1 (3.2 – 10.9) 52.9 (43.4 – 64.6) 0.5 (0.2 – 1.0)
P2 9.4 (5.8 – 15.6) 43.1 (35.9 – 55.1) 2.3 (1.2 – 3.9)
P3 6.2 (3.0 – 10.6) 34.6 (25.3 – 49.6) 26.9 (11.5 – 40.9)
P1 1.2 (0.6 – 2.4) 8.3 (5.1 – 11.5) 0.6 (0.3 – 1.5)
P2 1.8 (1.0 – 3.1) 6.8 (4.4 – 9.7) 2.8 (1.3 – 5.2)
P3 1.3 (0.6 – 2.3) 5.4 (3.3 – 8.6) 32.1 (14.9 – 49.3)
H0B1E0C2
H1B1E1C2C4
H0B0E1C2C0
H1B0E0C0
H0B0E0C0
H1B1E0C0
H1B0E1C2C1
different lower codend, C2, in combination with a 90 mm diamond codend as
upper codend (Table 4). Furthermore, out of the 15 BRD combinations selected,
10 included the counter-herding device (Melli et al., 2018a) and six the large
mesh size in the upper netting of the trawl body (Krag et al., 2014). Only three of
the selected combinations included the maximum level of complexity (i.e. No. of
BRDs) possible in this study. This was mainly caused by the potential loss of
commercial sized Nephrops associated with each additional BRD introduced in
the trawl, which could eventually add up to an unacceptable level. However,
these were also the BRD combinations predicted to be most effective in reducing
the overall bycatch (Fig. 8).
Table 4 Description of the BRDs included in the 15 most pertinent combinations. H=Herding area; B=Trawl body; E=Trawl extension.
When comparing the performance of the selected BRD combinations under a
multispecies catch scenario (see Appendix 2 for the structures of the populations
considered and the Supplementary Material for the indicators values), we
identified some clear potential harvest strategy for the fishing vessels operating in
the Skagerrak-Kattegat seas (Fig. 8). In Figure 8, the #0 indicates the baseline
design; under the catch scenario considered, the undersized bycatch retained by
the baseline design consisted of 75.3% (66.2–84.0) for cod and a highly variable
H B E Lower codend Upper codend
H0B0E1C0C1 1 - - YES 90 mm diamond 120 mm diamond
H0B0E1C0C2 2 - - YES 90 mm diamond 90 mm + 120 mm SMP
H0B0E1C0C3 3 - - YES 90 mm diamond 120 mm + 180 mm SMP
H0B1E0C0 4 - YES - 90 mm diamond -
H0B1E1C0C2 5 - YES - 90 mm diamond 90 mm + 120 mm SMP
H1B0E0C0 6 YES - - 90 mm diamond -
H1B0E1C0C1 7 YES - YES 90 mm diamond 120 mm diamond
H1B0E1C0C2 8 YES - YES 90 mm diamond 90 mm + 120 mm SMP
H1B0E1C0C3 9 YES - YES 90 mm diamond 120 mm + 180 mm SMP
H1B0E1C0C4 10 YES - YES 90 mm diamond open
H1B0E1C2C0 11 YES - YES 90 mm + 120 mm SMP 90 mm diamond
H1B1E0C0 12 YES YES - 90 mm diamond -
H1B1E1C0C1 13 YES YES YES 90 mm diamond 120 mm diamond
H1B1E1C0C2 14 YES YES YES 90 mm diamond 90 mm + 120 mm SMP
H1B1E1C0C3 15 YES YES YES 90 mm diamond 120 mm + 180 mm SMP
BRDs includedCombination ID
percentage of haddock (10.7–67.7%). Moreover, catches of commercial sized
bycatch were 97.4% (96.4–98.2) and 62.0% (26.0–92.0) for cod and haddock,
respectively. Respect to the baseline design, most of the identified BRD
combinations had desirable catch profiles: they did not affect significantly the
weight of commercial sized Nephrops retained; and they caught less than 50% of
the weight of undersized bycatch, both cod and haddock (highlighted sections in
Fig. 8). One exception, the combination #6 (H1B0E0C0), was predicted to retain
on average 60.6% (48.3–73.0) of the weight of undersized cod. In terms of
commercial sized, and thus valuable, bycatch a desirable catch profile could be
to either strongly reduce catches or to maintain them as high as possible,
depending on quota availability and market values. However, all the BRD
combinations identified as most promising similarly minimized the percentage of
commercial sized haddock retained, with the exception of combination #1
(H0B0E1C0C1). Nonetheless, this effect may be desirable in a Nephrops-
directed fishery when considering handling and storage costs, as well as the
relatively low market price for haddock
(http://www.hanstholmfiskeauktion.dk/prices?lang=en).
Figure 8 Two species comparisons of the performance of the most promising BRD combinations (15 for Nephrops and cod, and 10 for haddock) under the multispecies catch scenario. On the left column, percentage (in weight) of undersized bycatch retained (wP-). On the right column, percentage (in weight) of commercial sized bycatch retained (wP+). The first two rows show the percentage (in weight) of bycatch with respect to the percentage (in weight) of target catches (i.e. commercial sized Nephrops). Highlighted sections indicate desirable performances. MCRS = Minimum Conservation Reference Size.
If fishermen were to preserve their fish quotas by reducing both undersized and
commercial sized roundfish catches and obtain relatively clean Nephrops
catches, the optimal choice of BRD combinations would be #15. Indeed, by
including a BRD in each of the four sections of the trawl considered in this study,
this combination achieved overall retention below 25% and 1% of the weight of
cod and haddock, respectively (Fig. 8). This would allow fishermen to continue
fishing for Nephrops even when approaching exhaustion of roundfish quotas. In
contrast, if fishermen were to minimize the bycatch of undersized roundfish, while
maintaining the majority of the income deriving from commercial sized cod, for
example when cod quota is available, the BRD combinations #2 (H0B0E1C0C2)
and #7 (H1B0E1C0C1) could represent the best options (Fig. 8). Although many
other BRD combinations achieved similar results, these two had the advantage of
maintaining on average the same percentage retained of undersized Nephrops
as the baseline design (see Supplementary Materials for the Indicators values).
In particular, #2 retained 83.0 % (78.3–87.6) of commercial cod catches and
although data for haddock were not available for this BRD combination, haddock
catches can be expected to be low due to its high escape rate through 120 mm
SMPs (Krag et al., 2008; Fryer et al., 2014). Moreover, other BRD combinations
could be preferred if fishermen were to shift from one harvest strategy (e.g.
maximum income) to the other (e.g. quota saving) without having to return to the
harbour. For example, the BRD combination #2 can be converted into
combination #8 by simply adding the counter-herding device and to #10 by
leaving the upper codend open. Therefore, when encountering a bycatch hotspot,
fishermen could drastically reduce roundfish catches, without having to change
fishing area or interrupt fishing activities. The combination of BRDs can allow
fishermen to address selectivity issues on a day-to-day or even haul-to-haul
basis, increasing their ability to adapt to changes in species availability and
annual quotas. Furthermore, because trawl catches are highly variable and the
proportion and composition of the unwanted fraction varies according to multiple
factors (Engås and Soldal, 1992; Feekings et al., 2012), this approach would
make trawl selectivity more flexible and dynamic.
6. Conclusions
The meta-analytical approach described in this study offers the opportunity to fully
exploit the knowledge available concerning bycatch reduction measures for well-
studied fisheries and fishing areas. It creates the means to predict and compare
the performance of combinations of multiple BRDs, from both single- and
multispecies perspectives, and to identify solutions that could support flexible
harvest strategies. We hope that this approach will initiate further discussion
about new multi-purpose trawl designs, where a flexible selectivity can be
achieved by inserting or removing BRDs, depending on the populations
encountered and the individual catch goals.
The meta-analysis allowed the identification of interesting BRD combinations for
the Nephrops-directed fishery which would be worth experimental validation. For
future references, to optimize the predictive power of the analysis, some caution
should be used when choosing which BRD types and experimental datasets to
include: 1) the choice should be limited to BRDs with substantial effects; limited
effects would result in inconclusive predictions; 2) within species, homogeneity of
length-range among the studies included is essential, as the dataset with the
most restrictive range will affect the overall uncertainties; 3) a multi-species
approach including target and bycatch species is always recommended,
especially when considering mixed fisheries.
7. References
Akaike, H., 1974. A new look at the statistical model identification. IEEE
Transactions on Automatic Control, 19: 716–723.
Brinkhof, J., Larsen, R.B., Herrmann, B., Grimaldo, E., 2017. Improving catch
efficiency by changing ground gear design: Case study of Northeast Atlantic
cod (Gadus morhua) in the Barents Sea bottom trawl fishery. Fisheries
research, 186: 269–282.
Broadhurst, M.K., 2000. Modifications to reduce bycatch in prawn trawls: a review
and framework for development. Reviews in Fish Biology and Fisheries, 10:
27–60.
Broadhurst, M.K., Sterling, D.J., Cullis, B.R., 2012. Effects of otter boards on
catches of an Australian penaeid. Fisheries Research, 131: 67–75.
Catchpole, T.L., and Revill, A.S., 2008. Gear technology in Nephrops trawl
fisheries. Reviews in Fish Biology and Fisheries, 18: 17–31.
Dickson, W., Smith, A., Walsh, S., 1995. Methodology Manual: Measurement of
Fishing Gear Selectivity. Canada Dept. of Fisheries and Oceans, Ottawa.
Efron, B., 1982. The jackknife, the bootstrap and other resampling plans. SIAM
Monograph No. 38, CBSM-NSF.
Engås, A., and Soldal, A.V., 1992. Diurnal variations in bottom trawl catch rates
of cod and haddock and their influence on abundance indices. ICES Journal
of Marine Science, 49: 89–95.
EC, 2008. Council Regulation (EC) No 1342/2008 of 18 December 2008
Establishing a Long-term Plan for Cod Stocks and the Fisheries Exploiting
Those Stocks and Repealing Regulation (EC) No 423/200.
EU, 2013. Regulation (EU) No 1380/2013 of the European Parliament and
Council of 11 December 2013 on the Common Fisheries Policy. Official
Journal of the European Union, L 354/22.
Feekings, J., Bartolino, V., Madsen, N., Catchpole, T., 2012. Fishery discards:
factors affecting their variability within a demersal trawl fishery. PloS one, 7:
e36409.
Fryer, R.J., O'Neill, F.G., Edridge, A., 2016. A meta-analysis of haddock size-
selection data. Fish and fisheries, 17: 358–374.
Fujimori, Y., China, K., Oshima, T., Miyashita, K., Honda, S., 2005. The influence
of warp length on trawl dimension and catch of walleye pollock Theragra
chalcogramma in a bottom trawl survey. Fisheries Science, 71: 738–747.
Graham, N., 2010. Technical measures to reduce bycatch and discards in trawl
fisheries. In He, P. (Ed.), Behavior of Marine Fishes: Capture Processes
and Conservation Challenges, pp. 237–264. Wiley-Blackwell, Arnes, IA.
Graham, N., and Ferro, R.S.T., 2004. The Nephrops fisheries of the Northeast
Atlantic and Mediterranean: a review and assessment of fishing gear
design. ICES Cooperative Research Report No. 270.
Graham, N., and Fryer, R.J., 2006. Separation of fish from Nephrops norvegicus
into a two-tier cod-end using a selection grid. Fisheries Research, 82: 111–
118.
Herrmann, B., Sistiaga, M.B., Nielsen, K.N., Larsen, R.B., 2012. Understanding
the size selectivity of redfish (Sebastes spp.) in North Atlantic trawl
codends. Journal of Northwest Atlantic Fishery Science, 44: 1–13.
Herrmann, B., Wienbeck, H., Karlsen, J.D., Stepputtis, D., Dahm, E., Moderhak,
W., 2015. Understanding the release efficiency of Atlantic cod (Gadus
morhua) from trawls with a square mesh panel: effects of panel area, panel
position, and stimulation of escape response. ICES Journal of Marine
Science, 72: 686–696.
Herrmann, B., Sistiaga, M., Rindahl, L., Tatone, I., 2017. Estimation of the effect
of gear design changes on catch efficiency: Methodology and a case study
for a Spanish longline fishery targeting hake (Merluccius
merluccius). Fisheries Research, 185: 153–160.
Herrmann, B., Sistiaga, M., Grimaldo, E., Larsen, R.B., Olsen, L., Brinkhof, J.,
Tatone, I., 2018. Size selectivity and length-dependent escape behaviour of
haddock in a sorting device combining a grid and a square mesh
panel. Canadian Journal of Fisheries and Aquatic Sciences, (ja).
https://doi.org/10.1139/cjfas-2017-0461
ICES, 2004. The Nephrops fisheries of the Northeast Atlantic and Mediterranean
– A review and assessment of fishing gear design. ICES Cooperative
Research Report, No. 270. 40 pp.
ICES, 2014. Report of the Working Group on Mixed Fisheries Methods
(WGMIXFISH-METH), 20–24 October 2014, Nobel House, London, UK. 75
pp.
ICES, 2016. Report of the Working Group on the Assessment of Demersal
Stocks in the North Sea and Skagerrak (WGNSSK), 26 April–5 May 2016.
ICES CM 2016/ACOM:14.
Karlsen, J.D., Krag, L.A., Herrmann, B., Lund, H.S., 2018. Using vertical
distribution to separate fish from crustaceans in a mixed species trawl
fishery. Canadian Journal of Fisheries and Aquatic Sciences.
https://doi.org/10.1139/cjfas-2018-0313
Kelleher, K., 2005. Discards in the world’s marine fisheries: an update. FAO
Fisheries Technical Paper No. 470. Food and Agriculture Organization of
the United Nations, Rome, Italy.
Kennelly, S.J., and Broadhurst, M.K., 2002. By-catch begone: changes in the
philosophy of fishing technology. Fish and Fisheries, 3: 340–355.
Kim, Y.H., and Whang, D.S., 2010. An actively stimulating net panel and rope
array inside a model cod-end to increase juvenile red seabream
escapement. Fisheries Research, 106: 71–75.
Krag, L.A., Frandsen, R.P., Madsen, N., 2008. Evaluation of a simple means to
reduce discard in the Kattegat-Skagerrak Nephrops (Nephrops norvegicus)
fishery: Commercial testing of different codends and square-mesh
panels. Fisheries Research, 91: 175–186.
Krag, L.A., Holst, R., Madsen, N., Hansen, K., Frandsen, R., 2010. Selective
haddock (Melanogrammus aeglefinus) trawling: avoiding cod (Gadus
morhua) bycatch. Fisheries Research, 101: 20–26.
Krag, L.A., Poulsen, M.S., Vinther, M., Herrmann, B., Madsen, N., Frandsen, R.,
Karlsen, J.D., 2013. Dokumentation af selektiv effekt af SELTRA, 180 pp.
Krag, L.A., Herrmann, B., Karlsen, J.D., 2014. Inferring fish escape behaviour in
trawls based on catch comparison data: model development and evaluation
based on data from Skagerrak, Denmark. PloS one, 9: e88819.
Krag, L.A., Herrmann, B., Karlsen, J.D., Mieske, B., 2015. Species selectivity in
different sized topless trawl designs: Does size matter? Fisheries Research,
172: 243–249.
Krag, L.A., Herrmann, B., Feekings, J., Karlsen, J.D., 2016. Escape panels in
trawls – a consistent management tool? Aquatic Living Resources, 29: 306.
Lomeli, M.J., Waldo Wakefield, W., Herrmann, B., 2018. Illuminating the
Headrope of a Selective Flatfish Trawl: Effect on Catches of Groundfishes,
Including Pacific Halibut. Marine and Coastal Fisheries, 10: 118–131.
Lövgren, J., Herrmann, B., Feekings, J., 2016. Bell-shaped size selection in a
bottom trawl: A case study for Nephrops directed fishery with reduced
catches of cod. Fisheries Research, 184: 26–35.
Maynou, F., and Sardà, F., 2001. Influence of environmental factors on
commercial trawl catches of Nephrops norvegicus (L.). ICES Journal of
Marine Science, 58: 1318–1325.
Melli, V., Karlsen, J.D., Feekings, J.P., Herrmann, B., Krag, L.A., 2018a.
FLEXSELECT: counter-herding device to reduce bycatch in crustacean
trawl fisheries. Canadian Journal of Fisheries and Aquatic Sciences, 75:
850–860.
Melli, V., Krag, L.A., Herrmann, B., Karlsen, J.D. 2018b. Investigating fish
behavioural responses to LED lights in trawls and potential applications for
bycatch reduction in the Nephrops-directed fishery. ICES Journal of Marine
Science, 75: 1682–1692.
Melli, V., Krag L.A., Herrmann, B., Karlsen J.D., 2019. Can active behaviour
stimulators improve fish separation from Nephrops (Nephrops norvegicus)
in a horizontally divided trawl codend? Fisheries Research.
https://doi.org/10.1016/j.fishres.2018.11.027
Millar, R.B., 1993. Incorporation of between-haul variation using bootstrapping
and nonparametric estimation of selection curves. Fishery Bulletin, 91: 564–
572.
Noack, T., Frandsen, R.P., Krag, L.A., Mieske, B., Madsen, N., 2017. Codend
selectivity in a commercial Danish anchor seine. Fisheries research, 186:
283–291.
Pérez Roda, M.A. (ed.), Gilman, E., Huntington, T., Kennelly, S.J., Suuronen, P.,
Chaloupka, M., Medley, P., 2019. A third assessment of global marine
fisheries discards. FAO Fisheries and Aquaculture Technical Paper No.
633. Rome, FAO. 78 pp.
Sala, A., Lucchetti, A., Perdichizzi, A., Herrmann, B., Rinelli, P., 2015. Is square-
mesh better selective than larger mesh? A perspective on the management
for Mediterranean trawl fisheries. Fisheries Research, 161: 182–190.
Sistiaga, M., Herrmann, B., Grimaldo, E., Larsen, R.B., 2010. Assessment of dual
selection in grid based selectivity systems. Fisheries Research, 105: 187–
199.
Sistiaga, M., Herrmann, B., Grimaldo, E., Larsen, R.B., Tatone, I., 2015. Effect of
lifting the sweeps on bottom trawling catch efficiency: A study based on the
Northeast arctic cod (Gadus morhua) trawl fishery. Fisheries Research, 167:
164–173.
Wileman, D., Ferro, R.S.T., Fonteyne, R., Millar, R.B., 1996. Manual of methods
of measuring the selectivity of towed fishing gears. ICES Cooperative
Research Report, 215.
Winger, P.D., Eayrs, S., Glass, C.W., 2010. Fish behaviour near bottom trawls. In
He, P. (Ed.), Behavior of marine fishes: capture processes and conservation
challenges, pp. 67–102. Wiley-Blackwell, Arnes, IA.
Zeller, D., Cashion, T., Palomares, M., Pauly, D., 2018. Global marine fisheries
discards: a synthesis of reconstructed data. Fish and Fisheries, 19: 30–39.
Zuur, G., Fryer, R.J., Ferro, R.S.T., Tokay, T., 2001. Modeling the size selectivity
of a trawl codend and an associated square mesh panel. ICES Journal of
Marine Science, 58: 657.
APPENDIX 1
In this appendix we describe the models used for the size-selectivity in each of
the original datasets and species included in the meta-analyses. Moreover, we
report the fit statistics for each model fit.
In case of poor fit statistics (p-value <0.05; deviance >>DoF), the model curve
plots and the residuals were examined to determine whether there were
structural problems in describing the experimental data with the model or if it
could be a case of data overdispersion (Wileman et al., 1996). When no
systematic structure was detected, we considered the low p-values to be a
consequence of overdispersion in the data. Such cases are frequent, especially
when subsampling occured, and have been reported before in all the original
studies included in this meta-analysis (Krag et al., 2013; Krag et al., 2014; Krag
et al., 2015; Krag et al., 2016; Melli et al., 2018a; Melli et al., 2018b; Melli et al.,
2019).
1. Paired gears datasets
1.1 Herding area and trawl body
Data for these two Bycatch Reduction Devices (BRDs) were collected using
paired gears, i.e. a modified test trawl towed in parallel with a control trawl. For
each species, length-dependent count data for each gear were used to estimate
the size-dependent catch comparison rate cc(l) with 95% Efron confidence
intervals (Efron, 1982). The catch comparison rate cc(l) expresses the probability
of a catching an individual of length l with the test trawl given that it was available
to either trawl.
To model cc(l) we used a highly flexible model, often applied to this type of
experiments (Krag et al., 2014; Melli et al., 2018a):
𝑐𝑐(𝑙, 𝒗) = exp (𝑓(𝑙,𝒗))
1.0+exp (𝑓(𝑙,𝒗)) (1)
where f is a polynomial of the fourth order with coefficients v0,…,v4 so v =
(v0,…,v4). We used f (l,v) in the following form:
𝑓(𝑙, 𝒗) = ∑ 𝑣𝑖 × (𝑙
100)
𝑖4𝑖=0 = 𝑣0 + 𝑣1 ×
𝑙
100+ 𝑣2 ×
𝑙2
1002 + ⋯ + 𝑣4 × 𝑙4
1004 (2)
where the length l is divided by 100 to improve the numerical stability of the
model fitting by preventing numerical overflow due to lengths being raised to
powers in the polynomials. Leaving out one or more of the parameters v0…v4 in
equation (4) provided 31 additional models that were considered as potential
models to describe cc(l,v). We then applied model averaging to describe cc(l,v),
ranking the models according to how likely they were compared to each other
(Burnham and Anderson, 2002). The individual models were ranked and
weighted according to their Akaike's Information Criterion (AIC) values (Akaike,
1974; Burnham and Anderson, 2002; Herrmann et al., 2017) and models with
AIC values within +10 the value of the model with the lowest AIC, were
considered to contribute to cc(l,v) (Katsanevakis, 2006; Herrmann et al., 2017).
Fit statistics highlighted overdispertion in the data for both cod and haddock in
the dataset used for the Herding area (Melli et al., 2018a) and for cod in the trawl
body dataset (Table 1).
Tabel 1. Fit statistics for the modelled catch comparisons.
Cod Nephrops Haddock
p-value Deviance DoF
p-value Deviance DoF
p-value Deviance DoF
Herding area
0.03* 100.75 76 0.06 53.49 39 0.01* 61.50 39
Trawl Body
0.01* 109.94 76 0.55 45.03 47 0.62 35.75 39
2. Covered-codend datasets
2.1 Trawl Extension
The BRD introduced in the trawl extension was a horizontal separation into two
compartments; all individuals that entered the trawl were assumed to be caught
in either the upper or lower compartment because of the mesh size used (40 mm
T90) that is non-selective for the species considered. We were interested in
estimating the length-dependent probability for an individual to enter the upper
compartment, cUPPER(l). According to Krag et al. (2014), we used a length-
dependent model containing four parameters (c1, c2, L50C, and SRC):
cUPPER(𝑙) = 𝑐1 + (c2 ― 𝑐1) ×exp [(
ln(9)
𝑆𝑅𝐶×(𝑙−𝐿50𝐶)]
1.0+ exp [(ln(9)
𝑆𝑅𝐶×(𝑙−𝐿50𝐶)]
(3)
In a Eq. (3) the probability for an individual to enter the upper compartment,
cUPPER(l), follows a logistic curve within two asymptoms, c1 and c2. The constants
c1 and c2 are constrained to the interval [0.0; 1.0] and represent the asymptotic
probabilty of entering the upper compartment for the largest and smallest
individuals, respectively. L50C is the length at which cUPPER(l) is the mean of c1
and c2. SRC defines how quickly cUPPER(l) shifts from a value close to c1 to a value
close to c2 with increasing length in the vicinity of L50C. Thus, if SRC is close to
0.0, the change in cUPPER(l) will appear over a small length range, whereas if SRC
has a value far from 0.0 the change in cUPPER(l) will cover a wider length span.
Model fits statistics (p-value, deviance, DoF, R2) and parameters for cUPPER(l) of
each species are summarized in Table 2.
Tabel 2. Fit statistic for the modelled cUPPER(l)
Parameters Cod Nephrops Haddock
L50C 16.29 35.97 12.92
SRC 4.91 5.11 2.38
c1 0.76 0.22 0.77
c2 0.30 0.11 0.00
p-value 0.31 0.34 0.03*
Deviance 79.59 47.21 57.96
DoF 74 44 39
2.2 Codends
For each species and each codend separately, we tested different parametric
models to estimate the retention rate at length, r(l, v), where v is a vector
consisting of the parameters of the model. We chose the model with the lowest
individual Akaike information criterion (AIC) value (Akaike, 1974).
2.2.2 Nephrops
The triple logistic model (Eq. 4) was found to describe best the size selectivity of
Nephrops in the codends C0, C1 and C3 with the retention probability described
by:
𝑟(𝑙, 𝑐1, 𝐿501, 𝑆𝑅1, 𝑐2, 𝐿502, 𝑆𝑅2, 𝐿503, 𝑆𝑅3) = 𝑐1 × 𝐿𝑜𝑔𝑖𝑡(𝑙, 𝐿501, 𝑆𝑅1) + 𝑐2 ×
𝐿𝑜𝑔𝑖𝑡(𝑙, 𝐿502, 𝑆𝑅2) + (1.0 − 𝑐1 − 𝑐2) × 𝐿𝑜𝑔𝑖𝑡(𝑙, 𝐿503, 𝑆𝑅3) (4)
The triple logistic model is constructed by assuming that there are three different
selective processes which contribute to the overall selectivity, i.e. it is the sum of
three logit models in which the weights of the contributions add up to 1.0 (Noack
et al., 2017). These processes are determined by the multiple possible contacts
modes of Nephrops with the codend meshes (Frandsen et al., 2010). In the triple
logistic model, a fraction of individuals, c1, will be subjected to one logistic size
selection process with parameters L501 and SR1; another fraction c2 will be
subjected to a second logistic size selection process with parameters L502 and
SR2; the remaining fraction (1.0–c1–c2) will be subjected to a third logistic curve
with parameters L503 and SR3. The contact ratio parameters c1 and c2 indicate
the probability for an individual to have its selectivity determined by the first and
second process, respectively (Herrmann et al., 2013). Thus, they are numbers
between 0.0 and 1.0.
In contrast, the selectivity of Nephrops in the codend C2 was found to be
described best by a Dual sequential selection curve (Eq. 5) with the first process
modelled by a logistic curve and the second by the size selection model
“Gompertz” (Wileman, 1996). This model implies that the selectivity of the codend
is the result of two sequential selective processes. The first process is described
by a logistic selection curve with parameters L501 (i.e. length of fish with a 50%
retention probability) and SR1 (i.e. difference in length between fish with 75% and
25% retention probabilities) while the second process is described by a
“Gompertz” selection curve, with parameters L502 and SR2. Because the two
processes are sequential, the proportion of individuals that are exposed to the
second process is assumed to consist of those that did not attempt to escape in
the first process and additionally those that attempted to, but were retained.
Therefore, c1 represents the assumed length-independent probability that the
size selection of the individual will be defined by both selection processes (double
escape attempt), while 1.0 – c1 represents the probability of the individual
encountering only the second process. Thus, c1 is a number between 0.0 and
1.0.
𝑟(𝑙, 𝑐1, 𝐿501, 𝑆𝑅1, 𝐿502, 𝑆𝑅2) = (1.0 − 𝑐1) × 𝐺𝑜𝑚𝑝𝑒𝑟𝑡𝑧(𝑙, 𝐿502, 𝑆𝑅2) + 𝑐1 ×
𝐿𝑜𝑔𝑖𝑡(𝑙, 𝐿501, 𝑆𝑅1) × 𝐺𝑜𝑚𝑝𝑒𝑟𝑡𝑧(𝑙, 𝐿502, 𝑆𝑅2) (5)
Model fits statistics (p-value, deviance, DoF, R2) and parameters for the size
selectivity of Nephrops are summarized in Table 3.
Tabel 3. Fit statistics of the modelled size-selectivity for Nephrops in the four codends C0, C1, C2 and C3.
Parameters C0 C1 C2 C3
L50 31.14 47.91 34.98 54.70
SR 10.20 17.13 19.21 65.97
1/δ - - - -
L501 48.72 52.37 28.80 66.10
SR1 4.03 22.53 5.11 13.60
L502 30.97 47.33 33.79 44.63
SR2 7.10 5.00 25.54 0.10
L503 0.10 33.75 - 0.58
SR3 76.31 1.12 - 0.09
c1 0.09 0.63 0.73 1.78
c2 0.75 0.27 - 0.10
Model 4 4 5 4
p-value 0.96 0.42 0.06 0.93
Deviance 30.44 43.10 64.83 29.39
DoF 46 42 49 42
2.2.1 Cod
A Dual sequential size selection curve was found to describe best the selectivity
of cod in the 90 mm diamond mesh size codend (C0) and in the 120 mm
diamond codend with a 180 mm Square Mesh Panel (SMP; C3). For both
codends the two selective processes were modelled using a logistic curve and a
“Probit” curve, respectively (Eq. 6).
𝑟(𝑙, 𝑐1, 𝐿501, 𝑆𝑅1, 𝐿502, 𝑆𝑅2) =
(1.0 − 𝑐1) × 𝑃𝑟𝑜𝑏𝑖𝑡(𝑙, 𝐿502, 𝑆𝑅2) + 𝑐1 × 𝐿𝑜𝑔𝑖𝑡(𝑙, 𝐿501, 𝑆𝑅1) × 𝑃𝑟𝑜𝑏𝑖𝑡(𝑙, 𝐿502, 𝑆𝑅2) (6)
Although a dual sequential size selection model is often expected when the
codend include a SMP (e.g. C3), a second selective process can occur also in
simple codends (e.g. C0) for example during haul-back of the gear (Madsen et
al., 2012).
Similarly, the selectivity of cod in a 90 mm diamond mesh size codend with a 120
mm SMP (C2) was found to be described best by a Dual sequential size selection
curve, but with both processes modelled by a logistic curve (Eq. 7).
𝑟(𝑙, 𝑐1, 𝐿501, 𝑆𝑅1, 𝐿502, 𝑆𝑅2) =
(1.0 − 𝑐1) × 𝐿𝑜𝑔𝑖𝑡(𝑙, 𝐿502, 𝑆𝑅2) + 𝑐1 × 𝐿𝑜𝑔𝑖𝑡(𝑙, 𝐿501, 𝑆𝑅1) × 𝐿𝑜𝑔𝑖𝑡(𝑙, 𝐿502, 𝑆𝑅2) (7)
Finally, the selectivity of cod in a 120 mm diamond mesh size codend (C1) was
described best by the classical size selection model “Richard” (Wileman, 1996).
This is described not only by the parameters L50 and SR, but also by an
additional parameter (1/δ) that describes the asymmetry of the curve.
Model fits statistics (p-value, deviance, DoF, R2) and parameters for the size
selectivity of cod are summarized in Table 3.
Tabel 3. Fit statistics of the modelled size-selectivity for cod in the four codends C0, C1, C2 and C3.
Parameters C0 C1 C2 C3
L50 22.21 37.67 39.27 66.27
SR 7.57 13.35 14.14 13.84
1/δ - 0.39 - -
L501 19.63 - 44.01 68.08
SR1 3.68 - 4.86 7.31
L502 19.29 - 29.83 36.63
SR2 14.96 - 6.67 27.01
L503 - - - -
SR3 - - - -
c1 0.98 - 0.53 0.73
c2 - - - -
Model 6 Richard 7 6
p-value 0.90 1.00 0.98 0.79
Deviance 56.90 49.59 43.04 73.40
DoF 72 88 64 84
2.2.3 Haddock
The dataset used to estimate haddock size selectivity in a 90 mm diamond mesh
size codend included a 270 mm SMP (Krag et al., 2016). Therefore, data for
haddock in this codend were considered to have a binomial distribution, because
individuals escaping from both the SMP and the codend were collected in the
same cover. Following Krag et al. (2016), we estimated the selectivity of the
codend indirectly based on the length-dependent retention data for the combined
selection of SMP and codend. Indeed, the overall selectivity of a codend with a
SMP is generally modelled as Dual selection model with two logistic curves (Eq.
7), where the first selection process is described by L50SMP and SRSMP and the
second by L50codend and SRcodend. Therefore, in the meta-analysis we considered
only the parameters estimated for the logistic curve describing the selectivity of
the 90 mm diamond mesh size codend.
Finally, the selectivity of haddock in the codend C1 and C3 was found to be
described best by the size selection model “Richard” (l, L50, SR, 1/δ) and
“Gompertz” (l, L50, SR), respectively (Wileman, 1996).
Model fits statistics (p-value, deviance, DoF, R2) and parameters for the size
selectivity of haddock are summarized in Table 5.
Tabel 4. Fit statistics of the modelled size-selectivity for haddock in codends C0, C1 and C3.
Parameters C0 C1 C3
L50 52.96 29.61 111.39
SR 21.10 8.69 78.50
1/δ - 2.94 -
L501 53.01 - -
SR1 0.10 - -
L502 28.01 - -
SR2 7.98 - -
L503 - - -
SR3 - - -
c1 0.67 - -
c2 - - -
Model 7 Richard Gompertz
p-value <0.01* 0.98 0.89
Deviance 71.87 24.49 23.41
DoF 38 41 33
References
Akaike, H., 1974. A new look at the statistical model identification. IEEE
Transactions on Automatic Control 19, 716–722.
Burnham, K.P., and Anderson, D.R., 2002. Model Selection and Multimodel
Inference: A Practical Information-theoretic Approach, 2nd ed. Springer,
New York.
Efron, B., 1982. The jackknife, the bootstrap and other resampling plans. SIAM
Monograph No. 38, CBSM-NSF.
Frandsen, R. P., Herrmann, B., Madsen, N., 2010. A simulation-based attempt to
quantify the morphological component of size selection of Nephrops
norvegicus in trawl codends. Fisheries Research, 101: 156–167.
Herrmann, B., Sistiaga, M., Nielsen, K.N., Larsen, R.B., 2012. Understanding the
size selectivity of redfish (Sebastes spp.) in North Atlantic trawl codends.
Journal of Northwest Atlantic Fishery Science, 44: 1–13.
Herrmann, B., Sistiaga, M., Rindahl, L., Tatone, I., 2017. Estimation of the effect
of gear design changes in catch efficiency: methodology and a case study
for a Spanish longline fishery targeting hake (Merluccius merluccius).
Fisheries Research, 185: 153–160.
Katsanevakis, S., 2006. Modeling fish growth: Model selection, multi-model
inference and model selection uncertainty. Fisheries Research, 81: 229–
235.
Krag, L.A., Poulsen, M.S., Vinther, M., Herrmann, B., Madsen, N., Frandsen, R.,
Karlsen, J.D., 2013. Dokumentation af selektiv effekt af SELTRA, 180 pp.
Krag, L.A., Herrmann, B., Karlsen, J.D., 2014. Inferring fish escape behaviour in
trawls based on catch comparison data: model development and evaluation
based on data from Skagerrak, Denmark. PloS one, 9: e88819.
Krag, L.A., Herrmann, B., Karlsen, J.D., Mieske, B., 2015. Species selectivity in
different sized topless trawl designs: Does size matter? Fisheries Research,
172: 243–249.
Krag, L.A., Herrmann, B., Feekings, J., Karlsen, J.D., 2016. Escape panels in
trawls – a consistent management tool? Aquatic Living Resources, 29: 306.
Madsen, N., Herrmann, B., Frandsen, R.P., Krag, L. A., 2012. Comparing
selectivity of a standard and turned mesh T90 codend during towing and
haul-back. Aquatic Living Resources, 25: 231–240.
Melli, V., Karlsen, J.D., Feekings, J.P., Herrmann, B., Krag, L.A., 2018a.
FLEXSELECT: counter-herding device to reduce bycatch in crustacean
trawl fisheries. Canadian Journal of Fisheries and Aquatic Sciences, 75:
850–860.
Melli, V., Krag, L.A., Herrmann, B., Karlsen, J.D. 2018b. Investigating fish
behavioural responses to LED lights in trawls and potential applications for
bycatch reduction in the Nephrops-directed fishery. ICES Journal of Marine
Science, 75: 1682–1692.
Melli, V., Krag L.A., Herrmann, B., Karlsen J.D., 2019. Can active behaviour
stimulators improve fish separation from Nephrops (Nephrops norvegicus)
in a horizontally divided trawl codend? Fisheries Research.
https://doi.org/10.1016/j.fishres.2018.11.027
Noack, T., Frandsen, R.P., Krag, L.A., Mieske, B., Madsen, N., 2017. Codend
selectivity in a commercial Danish anchor seine. Fisheries research, 186:
283–291.
Wileman, D.A., Ferro, R.S.T., Fonteyne, R., Millar, R.B., 1996. Manual of
Methods of Measuring the Selectivity of Towed Fishing Gears. ICES
Cooperative Research Report No. 215, ICES, Copenhagen, Denmark.
APPENDIX 2
In this appendix we describe the populations used when investigating the
performance of each combination under realistic catch scenarios for each of the
species considered. The populations were generated using the original datasets
included in this study, by pooling data over hauls for hauls with more than 20
individuals (Table 1).
Table 1. Summary of the data used to generate each population the three population
scenarios for each of the species analysed.
For the multispecies scenario, hauls from the dataset by Krag et al. (2014)
containing more than 20 individuals for all the species considered were included
(Table 2).
Table 2. Summary of the data used to generate the multispecies scenario.
Species Population Original dataset No. of hauls No. of individuals
P1 Krag et al., 2016 8 6438
P2 Krag et al., 2014 22 12172
P3 Melli et al., 2019 4 7014
P1 Krag et al., 2015 25 3018
P2 Melli et al., 2018a 12 2333
P3 Melli et al., 2019 6 3835
P1 Melli et al., 2018b; 2019 14 5753
P2 Krag et al., 2014 22 4793
P3 Krag et al., 2015 15 4550
Nephrops
Cod
Haddock
Species Original dataset No. of hauls No. of individuals
Nephrops Krag et al., 2014 22 12172
Cod Krag et al., 2014 22 4803
Haddock Krag et al., 2014 22 4793
Fig. 1 illustrates the structure of the resulting populations (P1-P3 for each species
and the Multispecies scenario), as well as the 95% Efron (Efron, 1972)
Confidence Intervals obtained by the bootstrapping procedure.
Figure 1. Frequencies of the length classes represented in each single-species and
multispecies population scenario. Lengths are Carapace Length (mm) for Nephrops and
Total Length (cm) for cod and haddock.
References
Efron, B., 1982. The jackknife, the bootstrap and other resampling plans. SIAM
Monograph No. 38, CBSM-NSF.
Krag, L.A., Herrmann, B., Karlsen, J.D., 2014. Inferring fish escape behaviour in
trawls based on catch comparison data: model development and evaluation
based on data from Skagerrak, Denmark. PloS one, 9: e88819.
Krag, L.A., Herrmann, B., Karlsen, J.D., Mieske, B., 2015. Species selectivity in
different sized topless trawl designs: Does size matter? Fisheries Research,
172: 243–249.
Krag, L.A., Herrmann, B., Feekings, J., Karlsen, J.D., 2016. Escape panels in
trawls – a consistent management tool? Aquatic Living Resources, 29: 306.
Melli, V., Karlsen, J.D., Feekings, J.P., Herrmann, B., Krag, L.A., 2018a.
FLEXSELECT: counter-herding device to reduce bycatch in crustacean
trawl fisheries. Canadian Journal of Fisheries and Aquatic Sciences, 75:
850–860.
Melli, V., Krag, L.A., Herrmann, B., Karlsen, J.D. 2018b. Investigating fish
behavioural responses to LED lights in trawls and potential applications for
bycatch reduction in the Nephrops-directed fishery. ICES Journal of Marine
Science, 75: 1682–1692.
Melli, V., Krag L.A., Herrmann, B., Karlsen J.D., 2019. Can active behaviour
stimulators improve fish separation from Nephrops (Nephrops norvegicus)
in a horizontally divided trawl codend? Fisheries Research.
https://doi.org/10.1016/j.fishres.2018.11.027
SUPPLEMENTARY MATERIAL
WP
-W
P+
Dis
ca
rd R
ati
oW
P-
WP
+D
isc
ard
Ra
tio
WP
-W
P+
Dis
ca
rd R
ati
o
H0B
0E
0C
04
7 (
33
.1 -
62
.3)
85
.2 (
77
.1 -
91
.1)
0.6
(0
.3 -
1.0
)3
9.1
(2
7.5
- 5
3.6
)8
7.8
(7
9.7
- 9
3.0
)2
.0 (
1.3
- 3
.0)
36
.8 (
25
.6 -
52
.1)
80
.5 (
71
.3 -
89
.3)
19
.1 (
7.9
- 2
4.0
)
H0B
0E
0C
16
.8 (
1.7
- 1
8.1
)3
6.5
(2
8.0
- 4
6.8
)0
.2 (
0 -
0.6
)5
.9 (
1.2
- 1
6.3
)4
5.4
(3
3.5
- 5
5.4
)0
.6 (
0.1
- 1
.7)
5.7
(0
.9 -
15
.9)
31
.4 (
24
.2 -
49
.2)
8.6
(1
.0 -
19
.4)
H0B
0E
0C
23
3 (
24
.1 -
44
.8)
66
.8 (
61
.4 -
71
.8)
0.5
(0
.3 -
0.9
)2
6.6
(1
8.9
- 3
6.5
)7
0.6
(6
3.8
- 7
6.0
)1
.7 (
1.1
- 2
.5)
24
.7 (
17
.2 -
34
.5)
62
.1 (
55
.9 -
71
.2)
17
.0 (
6.7
- 2
1.9
)
H0B
0E
0C
33
2.9
(1
7.0
- 4
1.9
)3
8.3
(2
8.9
- 4
7.1
)0
.9 (
0.5
- 1
.4)
32
.9 (
11
.2 -
41
.5)
41
.8 (
32
.4 -
49
.2)
3.4
(1
.3 -
4.6
)3
2.9
(7
.6 -
41
.5)
37
.2 (
27
.1 -
46
.9)
31
.4 (
7.6
- 3
4.9
)
H0B
0E
1C
0C
14
2.1
(2
9.4
- 5
5.8
)7
5.4
(6
8.8
- 8
1.6
)0
.6 (
0.3
- 1
.0)
35
.2 (
24
.8 -
48
.6)
79
.2 (
71
.9 -
84
.7)
1.9
(1
.3 -
2.9
)3
3.2
(2
3.0
- 4
6.8
)7
1.3
(6
3.5
- 8
0.2
)1
9.4
(7
.6 -
24
.3)
H0B
0E
1C
0C
24
5.3
(3
3.0
- 5
8.7
)8
1.5
(7
4.6
- 8
6.3
)0
.6 (
0.3
- 1
.0)
37
.6 (
27
.3 -
50
.5)
84
.3 (
77
.3 -
88
.7)
2.0
(1
.3 -
2.9
)3
5.4
(2
5.6
- 4
9.0
)7
7.0
(6
9.1
- 8
4.9
)1
9.2
(7
.7 -
23
.8)
H0B
0E
1C
0C
34
5.3
(3
2.4
- 5
8.0
)7
5.6
(6
8.3
- 8
1.6
)0
.7 (
0.3
- 1
.0)
38
.3 (
27
.3 -
50
.7)
78
.2 (
71
.3 -
83
.7)
2.1
(1
.4 -
3.1
)3
6.3
(2
5.4
- 4
9.0
)7
2.1
(6
3.8
- 8
0.7
)2
0.7
(8
.4 -
25
.1)
H0B
0E
1C
0C
44
1.3
(2
8.3
- 5
5.0
)6
7.8
(6
0.5
- 7
5.2
)0
.7 (
0.3
- 1
.1)
34
.5 (
24
.1 -
48
.0)
69
.6 (
62
.1 -
76
.3)
2.2
(1
.4 -
3.3
)3
2.5
(2
2.1
- 4
6.2
)6
5.0
(5
6.5
- 7
3.2
)2
0.5
(8
.3 -
25
.5)
H0B
0E
1C
1C
01
1.6
(6
.7 -
21
.9)
46
.3 (
39
.1 -
54
.2)
0.3
(0
.1 -
0.6
)9
.9 (
5.2
- 1
9.7
)5
4.0
(4
3.1
- 6
2.3
)0
.8 (
0.4
- 1
.7)
9.3
(4
.5 -
19
.0)
40
.6 (
34
.1 -
56
.7)
10
.6 (
2.9
- 1
8.4
)
H0B
0E
1C
1C
21
0.0
(5
.4 -
20
.0)
42
.6 (
35
.6 -
50
.8)
0.3
(0
.1 -
0.6
)8
.4 (
4.1
- 1
7.7
)5
0.5
(4
0.0
- 5
8.7
)0
.7 (
0.4
- 1
.7)
7.9
(3
.4 -
16
.7)
37
.1 (
30
.5 -
53
.2)
9.9
(2
.5 -
18
.3)
H0B
0E
1C
1C
39
.9 (
4.9
- 2
0.2
)3
6.7
(3
0.0
- 4
5.0
)0
.3 (
0.1
- 0
.7)
9.1
(3
.9 -
18
.2)
44
.5 (
34
.5 -
52
.8)
0.9
(0
.4 -
1.9
)8
.9 (
3.3
- 1
7.7
)3
2.2
(2
6.2
- 4
6.9
)1
2.4
(3
.0 -
21
.0)
H0B
0E
1C
1C
46
.0 (
1.5
- 1
6.1
)2
8.9
(2
2.4
- 3
7.6
)0
.2 (
0.1
- 0
.6)
5.2
(1
.1 -
14
.7)
35
.8 (
26
.1 -
44
.5)
0.7
(0
.1 -
1.9
)5
.1 (
0.8
- 1
4.4
)2
5.1
(1
9.0
- 3
8.3
)9
.4 (
1.1
- 2
1.2
)
H0B
0E
1C
2C
03
4.7
(2
6.7
- 4
4.9
)7
0.5
(6
5.8
- 7
4.4
)0
.5 (
0.3
- 0
.9)
28
.0 (
21
.1 -
37
.1)
74
.1 (
67
.8 -
78
.7)
1.7
(1
.2 -
2.4
)2
6.1
(1
8.8
- 3
5.0
)6
5.5
(5
9.9
- 7
4.1
)1
7.1
(6
.5 -
21
.5)
H0B
0E
1C
2C
12
9.9
(2
1.7
- 4
0.1
)6
0.7
(5
5.7
- 6
5.7
)0
.5 (
0.3
- 0
.9)
24
.1 (
17
.2 -
32
.9)
65
.6 (
58
.9 -
70
.9)
1.6
(1
.1 -
2.4
)2
2.4
(1
5.7
- 3
1.4
)5
6.4
(5
0.5
- 6
6.0
)1
7.1
(6
.2 -
21
.8)
H0B
0E
1C
2C
33
3.0
(2
4.6
- 4
2.9
)6
0.9
(5
6.0
- 6
5.5
)0
.6 (
0.3
- 0
.9)
27
.3 (
19
.7 -
35
.6)
64
.6 (
58
.4 -
69
.6)
1.9
(1
.3 -
2.7
)2
5.6
(1
7.8
- 3
3.8
)5
7.1
(5
1.1
- 6
5.4
)1
8.8
(7
.0 -
23
.3)
H0B
0E
1C
2C
42
9.1
(2
0.5
- 3
9.4
)5
3.1
(4
7.9
- 5
8.9
)0
.6 (
0.3
- 1
.0)
23
.4 (
16
.5 -
32
.2)
55
.9 (
49
.8 -
62
.1)
1.8
(1
.2 -
2.8
)2
1.8
(1
5.0
- 3
0.7
)5
0.0
(4
3.9
- 5
7.7
)1
8.4
(7
.0 -
23
.4)
H0B
0E
1C
3C
03
4.6
(2
0.3
- 4
2.9
)4
7.9
(3
9.3
- 5
5.4
)0
.8 (
0.4
- 1
.2)
33
.7 (
14
.2 -
41
.4)
51
.3 (
43
.1 -
58
.1)
2.9
(1
.3 -
3.8
)3
3.4
(1
0.4
- 4
1.1
)4
5.6
(3
6.4
- 5
4.5
)2
7.5
(8
.8 -
31
.1)
H0B
0E
1C
3C
12
9.8
(1
6.0
- 3
8.1
)3
8.1
(3
0.8
- 4
5.9
)0
.9 (
0.4
- 1
.3)
29
.8 (
10
.4 -
37
.4)
42
.7 (
34
.4 -
49
.1)
3.0
(1
.2 -
4.1
)2
9.8
(7
.1 -
37
.5)
36
.4 (
28
.2 -
45
.7)
29
.7 (
7.4
- 3
3.3
)
H0B
0E
1C
3C
23
3.0
(1
9.0
- 4
1.1
)4
4.2
(3
6.1
- 5
1.6
)0
.8 (
0.4
- 1
.2)
32
.2 (
12
.7 -
39
.9)
47
.8 (
40
.0 -
54
.4)
2.9
(1
.3 -
3.9
)3
2.0
(9
.0 -
39
.7)
42
.1 (
33
.3 -
51
.0)
28
.2 (
8.0
- 3
1.6
)
H0B
0E
1C
3C
42
9.0
(1
5.0
- 3
6.8
)3
0.5
(2
3.0
- 3
8.0
)1
.0 (
0.5
- 1
.5)
29
.1 (
10
.0 -
36
.7)
33
.1 (
25
.4 -
39
.6)
3.8
(1
.5 -
5.0
)2
9.1
(6
.8 -
36
.7)
30
.1 (
21
.7 -
38
.2)
33
.3 (
7.8
- 3
7.1
)
H0B
0E
1C
4C
05
.7 (
2.8
- 9
.5)
17
.4 (
11
.5 -
20
.6)
0.4
(0
.2 -
0.7
)4
.6 (
2.2
- 8
.2)
18
.2 (
12
.6 -
22
.2)
1.1
(0
.6 -
2.2
)4
.3 (
1.9
- 7
.8)
15
.5 (
10
.5 -
19
.4)
12
.5 (
4.7
- 2
0.1
)
H0B
0E
1C
4C
10
.8 (
0.2
- 2
.7)
7.6
(4
.9 -
10
.2)
0.1
(0
.0 -
0.4
)0
.7 (
0.1
- 2
.3)
9.6
(6
.1 -
13
.2)
0.3
(0
.1 -
1.1
)0
.7 (
0.1
- 2
.2)
6.3
(4
.1 -
10
.4)
5.1
(0
.6 -
14
.7)
H0B
0E
1C
4C
24
.0 (
2.2
- 7
.1)
13
.7 (
9.3
- 1
6.1
)0
.3 (
0.2
- 0
.7)
3.1
(1
.6 -
5.8
)1
4.7
(1
0.1
- 1
8.2
)0
.9 (
0.5
- 1
.9)
2.9
(1
.3 -
5.3
)1
2.1
(8
.2 -
15
.3)
11
.0 (
4.0
- 1
8.5
)
H0B
0E
1C
4C
34
.0 (
1.8
- 6
.8)
7.8
(4
.9 -
10
.3)
0.6
(0
.2 -
1.1
)3
.8 (
1.2
- 6
.7)
8.7
(5
.8 -
11
.5)
1.9
(0
.7 -
3.6
)3
.8 (
0.8
- 6
.7)
7.1
(4
.6 -
10
.0)
21
.6 (
4.5
- 3
1.0
)
H0B
1E
0C
03
5.0
(2
2.9
- 4
9.9
)8
0.1
(6
8.7
- 8
7.4
)0
.5 (
0.2
- 0
.8)
28
.3 (
18
.1 -
42
.6)
83
.9 (
71
.8 -
90
.3)
1.5
(0
.9 -
2.4
)2
6.3
(1
6.2
- 4
0.8
)7
3.7
(6
3.4
- 8
4.3
)1
5.6
(5
.5 -
21
.7)
H0B
1E
0C
15
.1 (
1.3
- 1
4.4
)3
4.9
(2
5.9
- 4
4.3
)0
.2 (
0.0
- 0
.5)
4.3
(0
.9 -
13
.0)
44
.2 (
31
.2 -
53
.4)
0.4
(0
.1 -
1.4
)4
.1 (
0.6
- 1
2.4
)2
9.5
(2
2.1
- 4
5.7
)6
.6 (
0.8
- 1
7.0
)
H0B
1E
0C
22
4.7
(1
7.1
- 3
7.4
)6
3.0
(5
5.1
- 6
9.2
)0
.4 (
0.2
- 0
.7)
19
.3 (
12
.7 -
29
.7)
67
.7 (
57
.1 -
73
.3)
1.3
(0
.8 -
2.1
)1
7.7
(1
0.7
- 2
7.8
)5
7.1
(4
9.5
- 6
7.6
)1
3.8
(5
.2 -
19
.8)
H0B
1E
0C
32
4.5
(1
2.5
- 3
5.0
)3
6.1
(2
6.8
- 4
4.4
)0
.7 (
0.4
- 1
.2)
23
.4 (
8.4
- 3
4.2
)4
0.0
(3
0.4
- 4
7.2
)2
.6 (
1.0
- 4
.0)
23
.2 (
5.0
- 3
4.0
)3
4.0
(2
4.4
- 4
3.6
)2
6.0
(5
.8 -
32
.2)
H0B
1E
1C
0C
13
1.4
(2
0.8
- 4
5.6
)7
1.0
(6
1.1
- 7
8.4
)0
.5 (
0.2
- 0
.8)
25
.5 (
16
.3 -
38
.2)
75
.8 (
64
.4 -
82
.0)
1.5
(0
.9 -
2.4
)2
3.7
(1
4.6
- 3
6.9
)6
5.4
(5
6.1
- 7
6.4
)1
5.8
(5
.5 -
21
.6)
H0B
1E
1C
0C
23
3.8
(2
2.9
- 4
7.8
)7
6.7
(6
6.4
- 8
3.4
)0
.5 (
0.2
- 0
.8)
27
.3 (
17
.7 -
40
.4)
80
.6 (
69
.1 -
86
.4)
1.5
(0
.9 -
2.4
)2
5.3
(1
5.9
- 3
8.5
)7
0.6
(6
1.0
- 8
0.5
)1
5.6
(5
.8 -
21
.5)
H0B
1E
1C
0C
33
3.7
(2
2.6
- 4
7.7
)7
1.1
(6
1.0
- 7
8.4
)0
.5 (
0.3
- 0
.9)
27
.7 (
17
.6 -
40
.8)
74
.8 (
63
.8 -
80
.8)
1.6
(1
.0 -
2.6
)2
5.9
(1
6.0
- 3
8.8
)6
5.9
(5
6.1
- 7
5.7
)1
6.9
(6
.2 -
22
.9)
H0B
1E
1C
0C
43
0.8
(2
0.0
- 4
4.6
)6
3.7
(5
4.0
- 7
1.7
)0
.5 (
0.3
- 0
.9)
25
.0 (
15
.6 -
37
.6)
66
.4 (
56
.0 -
73
.3)
1.7
(1
.0 -
2.7
)2
3.2
(1
4.1
- 3
6.4
)5
9.4
(4
9.7
- 6
8.9
)1
6.8
(5
.9 -
23
.1)
H0B
1E
1C
1C
08
.7 (
4.8
- 1
7.7
)4
4.0
(3
5.3
- 5
1.8
)0
.2 (
0.1
- 0
.5)
7.1
(3
.5 -
15
.3)
52
.2 (
39
.7 -
60
.3)
0.6
(0
.3 -
1.4
)6
.7 (
3.0
- 1
4.5
)3
7.9
(3
0.5
- 5
3.2
)8
.3 (
2.1
- 1
6.4
)
H0B
1E
1C
1C
27
.4 (
3.9
- 1
6.2
)4
0.6
(3
2.3
- 4
8.1
)0
.2 (
0.1
- 0
.5)
6.1
(2
.9 -
14
.1)
48
.9 (
36
.4 -
56
.8)
0.6
(0
.3 -
1.3
)5
.7 (
2.3
- 1
3.2
)3
4.7
(2
7.7
- 5
0.1
)7
.8 (
2.0
- 1
6.1
)
H0B
1E
1C
1C
37
.4 (
3.6
- 1
6.2
)3
5.0
(2
7.5
- 4
3.3
)0
.2 (
0.1
- 0
.6)
6.5
(2
.8 -
14
.2)
43
.1 (
31
.6 -
51
.1)
0.7
(0
.3 -
1.6
)6
.3 (
2.2
- 1
3.8
)3
0.1
(2
3.7
- 4
4.0
)9
.7 (
2.1
- 1
7.9
)
H0B
1E
1C
1C
44
.4 (
1.1
- 1
2.3
)2
7.6
(2
0.3
- 3
5.8
)0
.2 (
0.0
- 0
.5)
3.8
(0
.8 -
11
.4)
34
.8 (
24
.2 -
43
.0)
0.5
(0
.1 -
1.5
)3
.6 (
0.6
- 1
1.0
)2
3.5
(1
7.3
- 3
6.3
)7
.3 (
0.9
- 1
8.2
)
H0B
1E
1C
2C
02
5.9
(1
8.5
- 3
8.2
)6
6.5
(5
8.5
- 7
2.2
)0
.4 (
0.2
- 0
.7)
20
.4 (
13
.9 -
30
.7)
71
.0 (
60
.1 -
76
.3)
1.3
(0
.8 -
2.1
)1
8.7
(1
2.1
- 2
8.2
)6
0.3
(5
2.9
- 7
0.3
)1
3.8
(5
.2 -
19
.4)
Co
mb
ina
tio
nP
1P
2P
3
Ta
ble
1. Nephrops
: w
eig
ht
indic
ato
rs f
or
the 9
6 B
RD
co
mb
inatio
ns e
ach
of
the 3
re
alis
tic p
op
ula
tion
sce
na
rios.
Th
e v
alu
es o
f th
e f
acto
rs u
sed
for
the le
ng
th-w
eig
ht
co
nve
rsio
n w
ere
: a
= 0
.00
076
5 a
nd
b =
2.9
80
25
WP
-W
P+
Dis
ca
rd R
atio
WP
-W
P+
Dis
ca
rd R
atio
WP
-W
P+
Dis
ca
rd R
atio
H0B
1E
1C
2C
12
2.3
(15
.2 - 3
3.7
)5
7.4
(50
.0 - 6
3.1
)0
.4 (0
.2 - 0
.7)
17
.5 (1
1.6
- 27
.1)
63
.0 (5
2.5
- 68
.5)
1.2
(0.8
- 2.1
)1
6.1
(9.9
- 25
.5)
51
.9 (4
5.0
- 62
.6)
13
.8 (4
.7 - 1
9.7
)
H0B
1E
1C
2C
32
4.6
(17
.1 - 3
6.0
)5
7.5
(49
.9 - 6
2.9
)0
.5 (0
.3 - 0
.8)
19
.8 (1
3.0
- 28
.9)
61
.9 (5
2.3
- 67
.2)
1.4
(0.9
- 2.3
)1
8.3
(11
.1 - 2
7.5
)5
2.5
(45
.4 - 6
1.2
)1
5.3
(5.8
- 20
.9)
H0B
1E
1C
2C
42
1.7
(14
.6 - 3
3.0
)5
0.0
(43
.2 - 5
6.2
)0
.5 (0
.3 - 0
.8)
17
.0 (1
1.0
- 26
.4)
53
.6 (4
4.9
- 59
.3)
1.4
(0.9
- 2.3
)1
5.6
(9.3
- 24
.8)
45
.9 (3
9.1
- 54
.7)
15
.0 (5
.5 - 2
1.0
)
H0B
1E
1C
3C
02
5.8
(14
.7 - 3
5.8
)4
5.1
(36
.2 - 5
2.4
)0
.6 (0
.3 - 1
.0)
24
.1 (1
0.4
- 34
.1)
49
.1 (3
9.5
- 55
.8)
2.1
(1.0
- 3.3
)2
3.6
(8.1
- 33
.8)
41
.8 (3
2.7
- 51
.3)
22
.6 (6
.2 - 2
8.5
)
H0B
1E
1C
3C
12
2.1
(11
.5 - 3
1.9
)3
6.0
(28
.2 - 4
3.2
)0
.7 (0
.3 - 1
.1)
21
.2 (7
.9 - 3
0.7
)4
1.1
(32
.1 - 4
7.3
)2
.3 (0
.9 - 3
.5)
21
.0 (5
.1 - 3
0.8
)3
3.4
(25
.3 - 4
2.9
)2
4.5
(5.7
- 30
.7)
H0B
1E
1C
3C
22
4.5
(13
.6 - 3
4.3
)4
1.7
(33
.3 - 4
8.6
)0
.6 (0
.3 - 1
.0)
23
.0 (9
.5 - 3
3.2
)4
5.8
(36
.8 - 5
2.0
)2
.2 (0
.9 - 3
.4)
22
.6 (6
.9 - 3
2.7
)3
8.6
(30
.4 - 4
7.8
)2
3.2
(6.1
- 29
.2)
H0B
1E
1C
3C
42
1.5
(10
.7 - 3
0.9
)2
8.7
(21
.2 - 3
6.0
)0
.8 (0
.4 - 1
.3)
20
.7 (7
.2 - 2
9.9
)3
1.7
(23
.8 - 3
7.7
)2
.8 (1
.1 - 4
.3)
20
.5 (4
.7 - 3
0.0
)2
7.4
(19
.4 - 3
5.6
)2
7.9
(6.5
- 34
.3)
H0B
1E
1C
4C
04
.2 (2
.1 - 7
.6)
16
.4 (1
0.6
- 19
.5)
0.3
(0.1
- 0.6
)3
.4 (1
.6 - 6
.3)
17
.5 (1
2.0
- 21
.5)
0.9
(0.4
- 1.8
)3
.1 (1
.3 - 6
.0)
14
.4 (9
.6 - 1
8.0
)1
0.0
(3.5
- 17
.9)
H0B
1E
1C
4C
10
.6 (0
.2 - 2
.0)
7.3
(4.8
- 9.6
)0
.1 (0
.0 - 0
.3)
0.5
(0.1
- 1.8
)9
.4 (5
.8 - 1
2.7
)0
.2 (0
.1 - 0
.9)
0.5
(0.1
- 1.7
)6
.0 (3
.8 - 9
.8)
3.9
(0.5
- 11
.9)
H0B
1E
1C
4C
23
.0 (1
.6 - 5
.6)
13
.0 (8
.6 - 1
5.3
)0
.3 (0
.1 - 0
.6)
2.3
(1.2
- 4.4
)1
4.2
(9.6
- 17
.5)
0.7
(0.4
- 1.6
)2
.1 (0
.9 - 4
.0)
11
.2 (7
.5 - 1
4.3
)8
.7 (3
.0 - 1
6.2
)
H0B
1E
1C
4C
32
.9 (1
.3 - 5
.3)
7.4
(4.7
- 9.5
)0
.4 (0
.2 - 0
.9)
2.7
(0.9
- 5.0
)8
.3 (5
.4 - 1
0.8
)1
.4 (0
.5 - 2
.9)
2.7
(0.6
- 4.9
)6
.6 (4
.2 - 9
.3)
17
.4 (3
.7 - 2
7.6
)
H1B
0E
0C
05
3.0
(32
.7 - 7
4.7
)9
2.5
(74
.2 - 1
11
.4)
0.6
(0.3
- 1.1
)4
5.9
(25
.9 - 6
9.8
)9
9.1
(73
.1 - 1
28
.7)
2.0
(1.1
- 3.7
)4
3.9
(22
.3 - 6
8.3
)8
6.9
(70
.6 - 1
07
.7)
20
.7 (8
.8 - 2
8.6
)
H1B
0E
0C
17
.7 (1
.9 - 2
0.3
)4
0.5
(28
.7 - 5
7.2
)0
.2 (0
.0 - 0
.6)
7.0
(1.4
- 19
.7)
52
.7 (3
2.0
- 78
.0)
0.6
(0.1
- 1.9
)6
.9 (1
.0 - 1
9.5
)3
4.6
(24
.5 - 5
9.1
)9
.3 (0
.9 - 2
2.0
)
H1B
0E
0C
23
7.3
(23
.7 - 5
4.1
)7
2.8
(57
.9 - 8
7.8
)0
.6 (0
.3 - 1
.0)
31
.0 (1
7.4
- 47
.6)
80
.1 (5
8.6
- 10
6.9
)1
.7 (1
.0 - 3
.1)
29
.3 (1
4.7
- 46
.9)
67
.3 (5
4.7
- 86
.8)
18
.4 (6
.8 - 2
6.5
)
H1B
0E
0C
33
7.4
(19
.1 - 5
2.1
)4
1.8
(29
.6 - 5
4.6
)1
.0 (0
.4 - 1
.6)
39
.7 (1
2.0
- 57
.5)
47
.5 (3
2.1
- 67
.3)
3.6
(1.3
- 6.2
)4
0.0
(8.2
- 57
.8)
40
.4 (2
8.2
- 54
.4)
33
.8 (7
.4 - 4
1.7
)
H1B
0E
1C
0C
14
7.6
(29
.3 - 6
7.7
)8
2.0
(65
.6 - 9
8.9
)0
.6 (0
.3 - 1
.1)
41
.4 (2
2.9
- 62
.5)
89
.7 (6
6.1
- 11
7.6
)2
.0 (1
.1 - 3
.6)
39
.6 (1
9.8
- 61
.9)
77
.1 (6
2.9
- 98
.8)
20
.9 (8
.8 - 2
9.0
)
H1B
0E
1C
0C
25
1.1
(32
.6 - 7
1.3
)8
8.5
(71
.2 - 1
05
.4)
0.6
(0.3
- 1.1
)4
4.2
(25
.3 - 6
6.3
)9
5.2
(71
.0 - 1
24
.7)
2.0
(1.1
- 3.6
)4
2.2
(21
.4 - 6
5.5
)8
3.2
(68
.5 - 1
02
.9)
20
.7 (8
.9 - 2
8.7
)
H1B
0E
1C
0C
35
1.1
(31
.8 - 7
0.8
)8
2.1
(66
.2 - 9
9.1
)0
.7 (0
.3 - 1
.2)
45
.2 (2
5.1
- 66
.4)
88
.4 (6
5.5
- 11
4.9
)2
.2 (1
.2 - 4
.0)
43
.4 (2
1.9
- 65
.3)
77
.9 (6
4.0
- 97
.2)
22
.3 (9
.6 - 3
0.3
)
H1B
0E
1C
0C
44
6.6
(28
.4 - 6
6.7
)7
3.6
(59
.0 - 8
9.1
)0
.7 (0
.3 - 1
.2)
40
.5 (2
2.4
- 61
.8)
78
.5 (5
7.4
- 10
2.0
)2
.3 (1
.2 - 4
.1)
38
.8 (1
9.2
- 61
.2)
70
.1 (5
7.0
- 87
.7)
22
.2 (9
.6 - 3
0.6
)
H1B
0E
1C
1C
01
3.2
(7.0
- 25
.4)
51
.0 (3
8.0
- 66
.6)
0.3
(0.1
- 0.7
)1
1.6
(5.6
- 24
.2)
62
.1 (4
0.4
- 89
.5)
0.8
(0.4
- 2.0
)1
1.2
(4.7
- 23
.7)
44
.4 (3
3.6
- 67
.1)
11
.5 (3
.1 - 2
1.7
)
H1B
0E
1C
1C
21
1.3
(5.8
- 22
.9)
47
.0 (3
4.9
- 62
.0)
0.3
(0.1
- 0.6
)9
.9 (4
.3 - 2
1.7
)5
8.2
(37
.4 - 8
4.2
)0
.8 (0
.3 - 1
.9)
9.5
(3.4
- 21
.1)
40
.7 (3
0.4
- 63
.2)
10
.7 (3
.0 - 2
1.2
)
H1B
0E
1C
1C
31
1.3
(5.2
- 23
.3)
40
.6 (2
9.5
- 55
.0)
0.3
(0.1
- 0.7
)1
0.8
(4.2
- 22
.2)
51
.4 (3
2.6
- 75
.2)
0.9
(0.4
- 2.3
)1
0.7
(3.7
- 21
.7)
35
.4 (2
6.0
- 57
.3)
13
.5 (3
.0 - 2
4.3
)
H1B
0E
1C
1C
46
.8 (1
.6 - 1
8.6
)3
2.1
(22
.4 - 4
6.0
)0
.2 (0
.1 - 0
.7)
6.2
(1.2
- 17
.5)
41
.5 (2
5.7
- 61
.6)
0.7
(0.1
- 2.1
)6
.1 (0
.9 - 1
7.3
)2
7.6
(19
.5 - 4
7.4
)1
0.2
(1.1
- 23
.4)
H1B
0E
1C
2C
03
9.2
(25
.9 - 5
5.3
)7
6.8
(62
.0 - 9
2.5
)0
.6 (0
.3 - 1
.0)
32
.8 (1
8.8
- 49
.3)
84
.0 (6
2.0
- 11
1.2
)1
.7 (1
.0 - 3
.1)
31
.0 (1
6.2
- 47
.3)
71
.0 (5
8.3
- 90
.6)
18
.4 (7
.3 - 2
6.5
)
H1B
0E
1C
2C
13
3.7
(21
.6 - 4
9.0
)6
6.3
(52
.7 - 8
0.4
)0
.6 (0
.3 - 1
.0)
28
.2 (1
5.7
- 42
.9)
74
.6 (5
4.0
- 10
1.1
)1
.7 (1
.0 - 3
.0)
26
.6 (1
3.5
- 41
.8)
61
.2 (4
9.8
- 80
.4)
18
.4 (6
.7 - 2
6.4
)
H1B
0E
1C
2C
33
7.3
(23
.9 - 5
2.3
)6
6.4
(53
.2 - 8
0.5
)0
.6 (0
.3 - 1
.1)
32
.0 (1
7.8
- 46
.9)
73
.3 (5
3.8
- 97
.5)
1.9
(1.1
- 3.5
)3
0.5
(15
.3 - 4
6.3
)6
1.9
(50
.6 - 7
9.9
)2
0.3
(7.5
- 28
.1)
H1B
0E
1C
2C
43
2.8
(20
.6 - 4
7.8
)5
7.8
(46
.1 - 7
0.3
)0
.6 (0
.3 - 1
.1)
27
.4 (1
5.5
- 41
.9)
63
.4 (4
5.9
- 83
.3)
1.9
(1.1
- 3.4
)2
5.8
(13
.1 - 4
1.0
)5
4.2
(43
.7 - 6
9.4
)1
9.8
(7.8
- 28
.1)
H1B
0E
1C
3C
03
9.3
(21
.4 - 5
3.3
)5
2.2
(39
.4 - 6
5.4
)0
.8 (0
.4 - 1
.4)
40
.5 (1
4.7
- 57
.3)
58
.2 (4
1.3
- 80
.7)
3.0
(1.2
- 5.1
)4
0.5
(11
.4 - 5
7.9
)4
9.4
(37
.6 - 6
4.3
)2
9.7
(8.5
- 37
.7)
H1B
0E
1C
3C
13
3.9
(17
.6 - 4
7.5
)4
1.7
(30
.8 - 5
4.1
)0
.9 (0
.4 - 1
.5)
35
.9 (1
1.4
- 52
.7)
48
.8 (3
3.1
- 69
.4)
3.2
(1.2
- 5.6
)3
6.2
(8.1
- 53
.3)
39
.6 (2
9.1
- 54
.3)
32
.0 (7
.2 - 4
0.5
)
H1B
0E
1C
3C
23
7.4
(20
.2 - 5
1.4
)4
8.2
(36
.3 - 6
0.8
)0
.8 (0
.4 - 1
.4)
38
.8 (1
3.8
- 55
.5)
54
.3 (3
8.5
- 75
.6)
3.1
(1.2
- 5.3
)3
8.8
(10
.3 - 5
6.2
)4
5.7
(34
.6 - 5
9.6
)3
0.5
(8.3
- 38
.5)
H1B
0E
1C
3C
43
2.9
(16
.7 - 4
5.9
)3
3.3
(23
.5 - 4
3.8
)1
.1 (0
.5 - 1
.8)
35
.1 (1
0.6
- 51
.3)
37
.6 (2
5.4
- 53
.1)
4.0
(1.4
- 6.8
)3
5.4
(7.5
- 51
.9)
32
.6 (2
2.7
- 44
.0)
35
.9 (8
.1 - 4
4.3
)
H1B
0E
1C
4C
06
.4 (3
.1 - 1
1.0
)1
8.9
(12
.0 - 2
4.6
)0
.4 (0
.2 - 0
.8)
5.4
(2.4
- 10
.2)
20
.6 (1
2.4
- 30
.1)
1.2
(0.5
- 2.7
)5
.1 (2
.0 - 1
0.0
)1
6.8
(10
.8 - 2
3.0
)1
3.6
(4.5
- 24
.2)
H1B
0E
1C
4C
10
.9 (0
.2 - 3
.0)
8.5
(5.1
- 12
.4)
0.1
(0.0
- 0.4
)0
.8 (0
.2 - 2
.8)
11
.2 (6
.0 - 1
7.9
)0
.3 (0
.1 - 1
.3)
0.8
(0.1
- 2.8
)7
.0 (4
.1 - 1
2.0
)5
.5 (0
.6 - 1
6.9
)
H1B
0E
1C
4C
24
.5 (2
.3 - 8
.2)
15
.0 (9
.6 - 1
9.6
)0
.3 (0
.1 - 0
.7)
3.7
(1.7
- 7.2
)1
6.7
(10
.2 - 2
3.9
)1
.0 (0
.4 - 2
.2)
3.4
(1.4
- 6.9
)1
3.1
(8.5
- 18
.0)
11
.9 (4
.0 - 2
2.0
)
H1B
0E
1C
4C
34
.5 (1
.8 - 8
.1)
8.6
(5.1
- 12
.0)
0.6
(0.2
- 1.2
)4
.6 (1
.3 - 8
.9)
9.9
(5.7
- 15
.3)
2.0
(0.6
- 4.6
)4
.6 (0
.8 - 9
.0)
7.8
(4.7
- 11
.6)
23
.5 (4
.4 - 3
6.0
)
H1B
1E
0C
03
9.5
(23
.4 - 5
9.9
)8
7.2
(67
.0 - 1
06
.5)
0.5
(0.2
- 1.0
)3
3.2
(18
.3 - 5
3.8
)9
5.0
(67
.9 - 1
23
.7)
1.5
(0.8
- 2.9
)3
1.3
(15
.1 - 5
2.2
)7
9.8
(62
.8 - 1
01
.7)
16
.8 (6
.9 - 2
5.9
)
H1B
1E
0C
15
.7 (1
.4 - 1
6.4
)3
8.8
(26
.4 - 5
4.2
)0
.2 (0
.0 - 0
.5)
5.1
(1.0
- 15
.1)
51
.3 (3
0.0
- 74
.9)
0.4
(0.1
- 1.5
)4
.9 (0
.7 - 1
4.9
)3
2.7
(22
.5 - 5
5.0
)7
.1 (0
.8 - 1
8.5
)
H1B
1E
0C
22
7.8
(17
.2 - 4
3.5
)6
8.8
(52
.7 - 8
3)
0.4
(0.2
- 0.9
)2
2.5
(12
.4 - 3
8.4
)7
7.0
(54
.1 - 1
01
.7)
1.3
(0.7
- 2.6
)2
0.9
(10
.2 - 3
6.2
)6
2.0
(48
.8 - 7
9.6
)1
4.8
(5.9
- 23
.3)
H1B
1E
0C
32
7.8
(13
.2 - 4
1.1
)3
9.5
(27
.0 - 5
1.6
)0
.8 (0
.3 - 1
.4)
28
.1 (8
.7 - 4
5.5
)4
5.7
(30
.0 - 6
3.7
)2
.7 (0
.9 - 4
.9)
28
.1 (5
.9 - 4
5.6
)3
7.0
(25
.5 - 5
1.1
)2
8.2
(5.5
- 38
.0)
Co
mb
ina
tion
P1
P2
P3
Ta
ble
1 (c
on
tinu
ed
)
WP
-W
P+
Dis
ca
rd R
ati
oW
P-
WP
+D
isc
ard
Ra
tio
WP
-W
P+
Dis
ca
rd R
ati
o
H1B
1E
1C
0C
13
5.4
(2
0.7
- 5
3.5
)7
7.4
(5
9.7
- 9
4.4
)0
.5 (
0.2
- 1
.0)
29
.8 (
16
.3 -
49
.1)
86
.1 (
60
.6 -
11
3.2
)1
.5 (
0.8
- 2
.9)
28
.2 (
13
.6 -
47
.6)
70
.9 (
55
.9 -
90
.4)
17
.0 (
6.5
- 2
5.7
)
H1B
1E
1C
0C
23
8.1
(2
2.8
- 5
7.8
)8
3.5
(6
4.4
- 1
01
.4)
0.5
(0
.2 -
1.0
)3
1.9
(1
7.9
- 5
2.1
)9
1.3
(6
5.2
- 1
17
.9)
1.5
(0
.8 -
2.9
)3
0.1
(1
4.8
- 5
0.3
)7
6.5
(6
0.7
- 9
6.6
)1
6.9
(7
.0 -
25
.7)
H1B
1E
1C
0C
33
8.1
(2
2.7
- 5
6.3
)7
7.4
(5
9.8
- 9
4.8
)0
.5 (
0.3
- 1
.0)
32
.5 (
17
.6 -
52
.8)
84
.7 (
60
.3 -
11
0.1
)1
.7 (
0.9
- 3
.2)
30
.9 (
14
.8 -
50
.5)
71
.4 (
56
.5 -
90
.4)
18
.3 (
7.3
- 2
7.0
)
H1B
1E
1C
0C
43
4.8
(2
0.1
- 5
2.3
)6
9.3
(5
3.1
- 8
5.3
)0
.5 (
0.3
- 1
.0)
29
.3 (
16
.0 -
48
.7)
75
.2 (
52
.8 -
96
.5)
1.7
(0
.9 -
3.2
)2
7.6
(1
3.2
- 4
7.1
)6
4.2
(5
0.7
- 8
2.4
)1
8.2
(7
.3 -
27
.2)
H1B
1E
1C
1C
09
.8 (
5.2
- 2
0.7
)4
8.6
(3
5.1
- 6
3.1
)0
.2 (
0.1
- 0
.6)
8.4
(3
.6 -
19
.4)
60
.2 (
37
.7 -
84
.5)
0.6
(0
.3 -
1.6
)7
.9 (
3.1
- 1
9.2
)4
1.6
(3
0.6
- 6
2.9
)9
.0 (
2.5
- 1
8.5
)
H1B
1E
1C
1C
28
.4 (
4.1
- 1
8.7
)4
4.9
(3
2.1
- 5
8.6
)0
.2 (
0.1
- 0
.5)
7.1
(3
.0 -
16
.8)
56
.6 (
35
.2 -
79
.7)
0.6
(0
.2 -
1.6
)6
.7 (
2.4
- 1
6.6
)3
8.2
(2
8.0
- 5
8.4
)8
.4 (
2.2
- 1
8.2
)
H1B
1E
1C
1C
38
.4 (
3.8
- 1
8.4
)3
8.8
(2
7.2
- 5
1.7
)0
.2 (
0.1
- 0
.6)
7.7
(3
.0 -
17
.2)
50
.0 (
30
.4 -
71
.6)
0.7
(0
.3 -
1.8
)7
.5 (
2.4
- 1
7.1
)3
3.2
(2
3.7
- 5
3.9
)1
0.5
(2
.4 -
20
.6)
H1B
1E
1C
1C
45
.0 (
1.2
- 1
4.1
)3
0.7
(2
0.5
- 4
3.6
)0
.2 (
0.0
- 0
.6)
4.5
(0
.9 -
13
.2)
40
.4 (
23
.8 -
57
.9)
0.5
(0
.1 -
1.7
)4
.3 (
0.6
- 1
3.0
)2
6.0
(1
7.8
- 4
4.6
)7
.9 (
0.9
- 2
0.3
)
H1B
1E
1C
2C
02
9.2
(1
8.6
- 4
4.6
)7
2.6
(5
5.7
- 8
7.3
)0
.4 (
0.2
- 0
.9)
23
.8 (
13
.4 -
39
.0)
80
.7 (
57
.3 -
10
5.5
)1
.3 (
0.7
- 2
.5)
22
.1 (
11
.0 -
37
.0)
65
.4 (
51
.8 -
82
.7)
14
.9 (
6.2
- 2
3.1
)
H1B
1E
1C
2C
12
5.1
(1
5.5
- 3
9.0
)6
2.8
(4
8.0
- 7
6.2
)0
.4 (
0.2
- 0
.9)
20
.4 (
11
.2 -
34
.5)
71
.8 (
50
.0 -
95
.9)
1.3
(0
.7 -
2.5
)1
9.0
(9
.4 -
32
.4)
56
.5 (
44
.2 -
73
.8)
14
.8 (
5.8
- 2
3.3
)
H1B
1E
1C
2C
32
7.8
(1
7.3
- 4
2.9
)6
2.8
(4
8.4
- 7
5.9
)0
.5 (
0.2
- 0
.9)
23
.1 (
12
.5 -
37
.6)
70
.4 (
49
.5 -
93
.1)
1.4
(0
.8 -
2.8
)2
1.7
(1
0.3
- 3
6.1
)5
7.0
(4
5.1
- 7
3.7
)1
6.4
(6
.5 -
24
.8)
H1B
1E
1C
2C
42
4.5
(1
4.9
- 3
8.0
)5
4.6
(4
1.4
- 6
6.3
)0
.5 (
0.2
- 1
.0)
19
.8 (
10
.9 -
33
.4)
60
.9 (
42
.3 -
79
.3)
1.4
(0
.8 -
2.8
)1
8.5
(9
.0 -
31
.8)
49
.8 (
39
.1 -
64
.5)
16
.1 (
6.4
- 2
4.9
)
H1B
1E
1C
3C
02
9.2
(1
5.0
- 4
2.2
)4
9.3
(3
5.9
- 6
2.2
)0
.6 (
0.3
- 1
.2)
28
.8 (
10
.4 -
46
.0)
55
.9 (
38
.0 -
76
.9)
2.2
(0
.9 -
4.1
)2
8.5
(8
.0 -
45
.6)
45
.4 (
33
.9 -
60
.9)
24
.5 (
6.9
- 3
4.0
)
H1B
1E
1C
3C
12
5.1
(1
2.3
- 3
7.8
)3
9.5
(2
8.3
- 5
1.0
)0
.7 (
0.3
- 1
.3)
25
.4 (
8.1
- 4
1.4
)4
7.0
(3
0.7
- 6
5.8
)2
.4 (
0.8
- 4
.4)
25
.4 (
6.0
- 4
1.8
)3
6.5
(2
6.4
- 5
1.6
)2
6.5
(5
.3 -
36
.2)
H1B
1E
1C
3C
22
7.8
(1
4.3
- 4
0.4
)4
5.6
(3
3.1
- 5
7.8
)0
.7 (
0.3
- 1
.2)
27
.5 (
9.7
- 4
4.0
)5
2.3
(3
5.5
- 7
2.5
)2
.3 (
0.9
- 4
.2)
27
.3 (
7.4
- 4
3.8
)4
2.0
(3
1.0
- 5
7.1
)2
5.1
(6
.8 -
34
.7)
H1B
1E
1C
3C
42
4.4
(1
1.5
- 3
6.9
)3
1.4
(2
1.5
- 4
1.5
)0
.8 (
0.4
- 1
.5)
24
.9 (
7.2
- 4
0.1
)3
6.1
(2
3.4
- 5
0.5
)3
.0 (
1.0
- 5
.4)
24
.8 (
5.5
- 4
0.6
)2
9.8
(2
0.4
- 4
1.4
)3
0.1
(6
.0 -
40
.0)
H1B
1E
1C
4C
04
.8 (
2.3
- 9
.1)
17
.9 (
11
.2 -
23
.4)
0.3
(0
.1 -
0.7
)3
.9 (
1.7
- 7
.8)
19
.8 (
11
.7 -
28
.9)
0.9
(0
.4 -
2.2
)3
.6 (
1.4
- 7
.6)
15
.6 (
9.8
- 2
1.3
)1
0.8
(3
.5 -
20
.8)
H1B
1E
1C
4C
10
.7 (
0.2
- 2
.3)
8.1
(4
.8 -
11
.7)
0.1
(0
.0 -
0.4
)0
.6 (
0.1
- 2
.1)
10
.9 (
5.6
- 1
7.2
)0
.2 (
0.1
- 1
.0)
0.6
(0
.1 -
2.0
)6
.7 (
3.9
- 1
1.3
)4
.2 (
0.5
- 1
3.7
)
H1B
1E
1C
4C
23
.4 (
1.7
- 6
.7)
14
.2 (
8.9
- 1
8.6
)0
.3 (
0.1
- 0
.6)
2.7
(1
.2 -
5.5
)1
6.2
(9
.3 -
23
.3)
0.7
(0
.3 -
1.9
)2
.4 (
1.0
- 5
.3)
12
.2 (
7.8
- 1
7.1
)9
.4 (
2.8
- 1
9.3
)
H1B
1E
1C
4C
33
.3 (
1.4
- 6
.4)
8.1
(4
.8 -
11
.2)
0.4
(0
.2 -
1.1
)3
.3 (
0.9
- 6
.7)
9.6
(5
.4 -
14
.8)
1.5
(0
.5 -
3.6
)3
.2 (
0.6
- 6
.6)
7.2
(4
.3 -
10
.6)
18
.9 (
3.4
- 3
1.3
)
Co
mb
ina
tio
nP
1P
2P
3
Ta
ble
1 (
co
nti
nu
ed
)
WP
-W
P+
Dis
ca
rd R
atio
WP
-W
P+
Dis
ca
rd R
atio
WP
-W
P+
Dis
ca
rd R
atio
H0B
0E
0C
06
6.6
(53
.9 - 7
7.6
)9
8.7
(98
.0- 9
9.3
)2
.8 (1
.6 - 4
.9)
78
.2 (7
0.5
- 86
.1)
96
(94
.3 - 9
7.2
)8
.1 (5
.4 - 1
1.1
)6
9.0
(57
.0 - 7
8.5
)9
4.0
(91
.7 - 9
6.4
)5
9.9
(48
.3 - 6
6.8
)
H0B
0E
0C
11
6.4
(9.6
- 21
.8)
85
.9 (8
1.5
- 89
)0
.8 (0
.4 - 1
.5)
20
.9 (1
4.4
- 26
.3)
73
.4 (6
8.1
- 78
.2)
3.0
(1.7
- 4.6
)1
6.7
(10
.1 - 2
1.4
)6
2.9
(53
.6 - 7
3.8
)3
5.0
(19
.5 - 4
6.0
)
H0B
0E
0C
21
0.8
(5.7
- 18
.6)
85
.1 (7
9.8
- 90
.0)
0.5
(0.3
- 1.0
)1
7.8
(11
.0 - 2
7.7
)7
3.8
(67
.6 - 8
0.1
)2
.5 (1
.4 - 4
.1)
11
.0 (5
.6 - 1
8.0
)6
3.4
(53
.3 - 7
6.4
)2
6.0
(13
.0 - 3
7.8
)
H0B
0E
0C
37
.9 (4
.4 - 1
2.9
)3
4.7
(28
.8 - 4
1.6
)1
.0 (0
.4 - 2
.1)
9.1
(5.5
- 14
.3)
28
.5 (2
3.2
- 35
.5)
3.3
(1.6
- 5.8
)8
.0 (4
.4 - 1
3.4
)2
2.1
(15
.1 - 3
4.9
)4
2.3
(21
.5 - 6
0.8
)
H0B
0E
1C
0C
12
8.9
(22
.1 - 3
6.3
)8
9.0
(85
.0 - 9
1.4
)1
.4 (0
.7 - 2
.5)
34
.7 (2
8.1
- 41
)7
8.7
(74
.4 - 8
2.7
)4
.5 (2
.8 - 6
.7)
29
.6 (2
2.8
- 36
.2)
70
.4 (6
2.2
- 79
.2)
46
.1 (3
2.5
- 55
.4)
H0B
0E
1C
0C
22
4.8
(18
.1 - 3
3.0
)8
8.3
(84
.1 - 9
2.2
)1
.2 (0
.6 - 2
.1)
32
.4 (2
5.6
- 40
.6)
79
.1 (7
3.9
- 83
.9)
4.2
(2.6
- 6.1
)2
5.3
(18
.8 - 3
2.6
)7
0.7
(62
.2 - 8
0.5
)4
2.1
(28
.5 - 5
1.8
)
H0B
0E
1C
0C
32
2.5
(16
.9 - 2
9.5
)5
0.0
(43
.3 - 5
6.1
)1
.9 (1
.0 - 3
.5)
25
.8 (2
0.0
- 32
.5)
44
.6 (3
8.4
- 51
.3)
5.9
(3.7
- 8.7
)2
3.0
(17
.3 - 2
9.8
)3
9.3
(31
.5 - 4
9.6
)5
4.4
(38
.8 - 6
5.1
)
H0B
0E
1C
0C
41
6.6
(11
.4 - 2
3.7
)2
3.5
(16
.8 - 2
8.9
)2
.9 (1
.6 - 5
.4)
18
.9 (1
3.5
- 25
.0)
22
.9 (1
6.5
- 28
.2)
8.2
(5.6
- 12
.0)
17
.0 (1
1.7
- 24
,0)
22
.4 (1
6.1
- 27
.8)
60
.7 (4
9.6
- 69
.3)
H0B
0E
1C
1C
05
4.1
(43
.1 - 6
3.8
)9
5.7
(94
.1 - 9
6.9
)2
.4 (1
.3 - 4
.1)
64
.4 (5
7.4
- 71
.7)
90
.6 (8
8.1
- 92
.9)
7.1
(4.7
- 9.9
)5
6.1
(46
.3 - 6
4.6
)8
6.6
(82
.9 - 9
1.1
)5
6.8
(44
.1 - 6
4.3
)
H0B
0E
1C
1C
21
2.3
(8.0
- 18
.1)
85
.3 (8
0.7
- 89
.3)
0.6
(0.3
- 1.1
)1
8.5
(13
.1 - 2
5.8
)7
3.7
(68
.1 - 7
9.2
)2
.6 (1
.6 - 4
.1)
12
.4 (8
.1 - 1
7.4
)6
3.3
(53
.7 - 7
5.6
)2
8.5
(16
.5 - 3
8.9
)
H0B
0E
1C
1C
31
0.0
(6.7
- 14
.3)
46
.9 (4
0.4
- 52
.7)
0.9
(0.5
- 1.8
)1
1.9
(8.4
- 16
.2)
39
.2 (3
3.7
- 45
.7)
3.2
(1.8
- 5.2
)1
0.1
(6.7
- 14
.4)
31
.9 (2
4.0
- 43
.5)
39
.2 (2
2.9
- 53
.9)
H0B
0E
1C
1C
44
.1 (2
.2 - 6
.2)
20
.5 (1
4.5
- 25
.0)
0.8
(0.4
- 1.7
)5
.0 (3
.0 - 7
.2)
17
.5 (1
2.5
- 21
.9)
3.0
(1.7
- 5.0
)4
.1 (2
.3 - 6
.1)
15
.0 (1
0.4
- 19
.7)
35
.8 (2
1.2
- 48
.5)
H0B
0E
1C
2C
05
2.7
(41
.4 - 6
3.0
)9
5.5
(93
.7 - 9
7.1
)2
.3 (1
.3 - 4
.0)
63
.6 (5
6.2
- 71
.2)
90
.7 (8
8.0
- 93
.2)
7.0
(4.7
- 9.8
)5
4.7
(44
.4 - 6
3.2
)8
6.7
(83
.0 - 9
1.7
)5
6.2
(43
.2 - 6
3.7
)
H0B
0E
1C
2C
11
5.0
(9.8
- 19
.4)
85
.7 (8
1.5
- 88
.7)
0.7
(0.4
- 1.3
)2
0.1
(14
.9 - 2
4.9
)7
3.5
(68
.4 - 7
8.0
)2
.9 (1
.7 - 4
.4)
15
.3 (1
0.0
- 19
.0)
63
.1 (5
4.2
- 74
.2)
33
.0 (1
9.1
- 43
.3)
H0B
0E
1C
2C
38
.5 (5
.5 - 1
2.9
)4
6.7
(40
.3 - 5
2.9
)0
.8 (0
.4 - 1
.5)
11
.2 (7
.9 - 1
6.2
)3
9.3
(33
.4 - 4
5.8
)3
.0 (1
.7 - 4
.8)
8.7
(5.6
- 13
.3)
32
.0 (2
4.2
- 43
.6)
35
.6 (1
9.9
- 50
.6)
H0B
0E
1C
2C
42
.6 (1
.3 - 4
.9)
20
.3 (1
4.4
- 25
.1)
0.6
(0.3
- 1.1
)4
.3 (2
.4 - 7
.0)
17
.6 (1
2.8
- 21
.9)
2.6
(1.4
- 4.4
)2
.7 (1
.3 - 4
.7)
15
.1 (1
0.5
- 20
.0)
26
.5 (1
3.4
- 39
.6)
H0B
0E
1C
3C
05
2.0
(41
.1 - 6
2.3
)8
3.5
(79
.6 - 8
8.2
)2
.6 (1
.4 - 4
.5)
61
.5 (5
4.1
- 69
.4)
79
.9 (7
5.7
- 85
.0)
7.7
(5.0
- 10
.6)
53
.9 (4
4.0
- 62
.5)
76
.8 (7
2.2
- 83
.2)
58
.8 (4
6.0
- 66
.0)
H0B
0E
1C
3C
11
4.3
(9.1
- 18
.6)
73
.7 (6
9.1
- 78
.7)
0.8
(0.4
- 1.5
)1
8.0
(13
.0 - 2
2.4
)6
2.7
(57
.5 - 6
8.4
)3
.0 (1
.7 - 4
.7)
14
.5 (9
.6 - 1
8.5
)5
3.2
(45
.5 - 6
4.9
)3
5.7
(20
.0 - 4
6.6
)
H0B
0E
1C
3C
21
0.1
(6.4
- 16
.1)
73
.1 (6
7.9
- 78
.9)
0.6
(0.3
- 1.1
)1
5.7
(10
.3 - 2
3.6
)6
3.0
(57
.4 - 6
9.6
)2
.6 (1
.5 - 4
.1)
10
.3 (6
.1 - 1
5.7
)5
3.6
(44
.9 - 6
7.0
)2
8.0
(15
.2 - 3
9.2
)
H0B
0E
1C
3C
42
.0 (1
.0 - 3
.6)
8.3
(5.7
- 10
.7)
1.0
(0.5
- 2.2
)2
.2 (1
.2 - 3
.7)
6.8
(4.7
- 9.2
)3
.4 (1
.7 - 6
.3)
2.0
(1.0
- 3.7
)5
.3 (3
.2 - 8
.6)
43
.1 (2
2.8
- 62
.8)
H0B
0E
1C
4C
05
0.0
(38
.6 - 6
0.6
)7
5.2
(69
.9 - 8
2.0
)2
.8 (1
.5 - 4
.8)
59
.3 (5
1.7
- 67
.6)
73
.1 (6
7.9
- 79
.9)
8.1
(5.3
- 11
.0)
52
.0 (4
1.9
- 60
.9)
71
.6 (6
6.4
- 78
.4)
59
.6 (4
7.4
- 66
.4)
H0B
0E
1C
4C
11
2.3
(7.2
- 16
.8)
65
.5 (6
0.2
- 72
.2)
0.8
(0.4
- 1.4
)1
5.9
(10
.7 - 2
0.4
)5
5.9
(50
.7 - 6
2.4
)3
.0 (1
.7 - 4
.5)
12
.6 (7
.4 - 1
6.6
)4
7.9
(41
.2 - 5
8.3
)3
4.8
(18
.8 - 4
5.8
)
H0B
0E
1C
4C
28
.2 (4
.3 - 1
4.2
)6
4.8
(59
.3 - 7
2.2
)0
.5 (0
.3 - 1
.0)
13
.5 (8
.3 - 2
1.0
)5
6.2
(50
.1 - 6
4.0
)2
.5 (1
.4 - 4
.1)
8.3
(4.2
- 13
.5)
48
.3 (4
0.8
- 60
.1)
25
.9 (1
2.7
- 37
.5)
H0B
0E
1C
4C
35
.9 (3
.1 - 9
.8)
26
.5 (2
1.9
- 32
.6)
0.9
(0.4
- 2.0
)6
.9 (4
.1 - 1
0.9
)2
1.7
(17
.6 - 2
7.7
)3
.3 (1
.6 - 5
.7)
6.0
(3.3
- 9.9
)1
6.9
(11
.5 - 2
6.9
)4
2.0
(21
.0 - 6
0.0
)
H0B
1E
0C
03
9.7
(30
.2 - 4
9.0
)5
9.3
(49
.4 - 7
1.2
)2
.8 (1
.4 - 5
.1)
41
.8 (3
4.6
- 50
.3)
53
.8 (4
6.5
- 65
.0)
7.7
(4.6
- 11
.2)
40
.7 (3
1.5
- 50
.0)
49
.4 (4
1.7
- 61
.1)
62
.7 (4
6.3
- 71
.5)
H0B
1E
0C
19
.7 (5
.4 - 1
3.1
)5
3.3
(43
.6 - 6
4.9
)0
.8 (0
.3 - 1
.6)
11
.1 (7
.4 - 1
4.7
)4
2.9
(35
.7 - 5
4.7
)2
.7 (1
.4 - 4
.4)
9.8
(5.6
- 13
.1)
34
.4 (2
5.5
- 48
.8)
36
.7 (1
8.3
- 50
.5)
H0B
1E
0C
26
.1 (3
.2 - 1
0.9
)5
2.9
(43
.4 - 6
4.6
)0
.5 (0
.2 - 1
.0)
9.4
(5.8
- 15
.6)
43
.1 (3
5.9
- 55
.1)
2.3
(1.2
- 3.9
)6
.2 (3
.0 - 1
0.6
)3
4.6
(25
.3 - 4
9.6
)2
6.9
(11
.5 - 4
0.9
)
H0B
1E
0C
34
.7 (2
.5 - 8
.0)
25
.4 (1
8.8
- 32
.1)
0.8
(0.3
- 1.8
)4
.9 (2
.8 - 8
.0)
19
.6 (1
4.4
- 26
.8)
2.6
(1.2
- 4.9
)4
.7 (2
.5 - 8
.3)
13
.8 (7
.4 - 2
6.7
)4
0.9
(17
.5 - 6
4.7
)
H0B
1E
1C
0C
11
7.2
(12
.6 - 2
2.1
)5
4.8
(45
.2 - 6
6.1
)1
.3 (0
.7 - 2
.6)
18
.5 (1
4.3
- 23
.2)
45
.5 (3
8.6
- 56
.8)
4.2
(2.3
- 6.5
)1
7.5
(12
.8 - 2
2.3
)3
7.9
(29
.4 - 5
1.6
)4
8.3
(29
.9 - 5
9.9
)
H0B
1E
1C
0C
21
4.6
(10
.3 - 1
9.9
)5
4.5
(45
.1 - 6
6.0
)1
.1 (0
.6 - 2
.2)
17
.2 (1
3.0
- 22
.5)
45
.7 (3
8.3
- 57
.4)
3.9
(2.2
- 6.0
)1
4.8
(10
.4 - 1
9.8
)3
8.1
(29
.4 - 5
2.4
)4
4.1
(26
.5 - 5
6.2
)
H0B
1E
1C
0C
31
3.5
(9.5
- 18
.4)
33
.5 (2
6.1
- 41
.1)
1.7
(0.8
- 3.4
)1
3.8
(10
.1 - 1
8.3
)2
7.8
(22
.2 - 3
5.7
)5
.1 (2
.9 - 8
.1)
13
.6 (9
.6 - 1
8.7
)2
2.3
(15
.3 - 3
4.5
)5
5.4
(34
.6 - 6
9.4
)
H0B
1E
1C
0C
41
0.0
(6.5
- 14
.5)
14
.1 (9
.7 - 1
8.4
)2
.9 (1
.5 - 5
.8)
10
.1 (7
.0 - 1
3.9
)1
2.8
(9.1
- 17
.1)
7.8
(4.7
- 11
.8)
10
.1 (6
.7 - 1
4.8
)1
1.8
(8.3
- 15
.8)
63
.5 (4
8.1
- 73
.6)
H0B
1E
1C
1C
03
2.2
(24
.1 - 3
9.7
)5
7.9
(48
.3 - 6
9.8
)2
.3 (1
.2 - 4
.3)
34
.4 (2
8.1
- 41
.2)
51
.2 (4
4.0
- 62
.6)
6.8
(3.9
- 10
.0)
33
.1 (2
5.2
- 40
.4)
45
.8 (3
8.2
- 58
.3)
59
.5 (4
2.0
- 69
.1)
H0B
1E
1C
1C
27
.1 (4
.5 - 1
0.6
)5
3.0
(43
.6 - 6
4.5
)0
.6 (0
.3 - 1
.1)
9.8
(6.8
- 14
.6)
43
.1 (3
5.9
- 55
.0)
2.4
(1.3
- 3.9
)7
.1 (4
.4 - 1
0.5
)3
4.5
(25
.7 - 4
9.7
)2
9.6
(15
.1 - 4
3.1
)
H0B
1E
1C
1C
36
.0 (3
.8 - 8
.7)
32
.0 (2
4.7
- 39
.5)
0.8
(0.4
- 1.7
)6
.4 (4
.3 - 9
.1)
25
.2 (1
9.5
- 33
.3)
2.7
(1.3
- 4.6
)6
.0 (3
.8 - 8
.8)
18
.7 (1
1.6
- 31
.6)
39
.3 (1
9.3
- 58
.1)
H0B
1E
1C
1C
42
.4 (1
.3 - 3
.8)
12
.7 (8
.6 - 1
6.6
)0
.8 (0
.4 - 1
.8)
2.7
(1.6
- 4.0
)1
0.2
(7.2
- 14
.0)
2.7
(1.4
- 4.7
)2
.4 (1
.3 - 3
.7)
8.2
(5.2
- 12
.2)
37
.5 (1
9.4
- 53
.0)
H0B
1E
1C
2C
03
1.3
(23
.2 - 3
9.0
)5
7.8
(48
.1 - 6
9.5
)2
.3 (1
.2 - 4
.2)
34
.0 (2
7.5
- 40
.8)
51
.2 (4
4.2
- 62
.8)
6.7
(3.8
- 9.9
)3
2.2
(24
.2 - 3
9.4
)4
5.8
(38
.1 - 5
8.4
)5
8.8
(41
.1 - 6
8.7
)
H0B
1E
1C
2C
18
.8 (5
.6 - 1
1.5
)5
3.2
(43
.6 - 6
4.8
)0
.7 (0
.3 - 1
.4)
10
.7 (7
.7 - 1
3.6
)4
3.0
(35
.8 - 5
4.8
)2
.6 (1
.4 - 4
.1)
8.9
(5.6
- 11
.5)
34
.4 (2
5.8
- 49
.0)
34
.4 (1
7.7
- 47
.9)
Co
mb
ina
tion
P1
P2
P3
Ta
ble
2. C
od
: we
igh
t indic
ato
rs fo
r the 9
6 B
RD
co
mb
inatio
ns e
ach
of th
e 3
realis
tic p
op
ula
tion
sce
na
rios. T
he v
alu
es o
f the fa
cto
rs u
sed
for th
e
leng
th-w
eig
ht c
onve
rsio
n w
ere
: a =
0.0
05
87
an
d b
= 3
.14
0
WP
-W
P+
Dis
ca
rd R
ati
oW
P-
WP
+D
isc
ard
Ra
tio
WP
-W
P+
Dis
ca
rd R
ati
o
H0B
1E
1C
2C
35
.0 (
3.1
- 7
.9)
32
.0 (
24
.7 -
39
.4)
0.7
(0
.3 -
1.5
)5
.9 (
4.0
- 8
.7)
25
.2 (
19
.6 -
33
.1)
2.5
(1
.3 -
4.2
)5
.1 (
3.1
- 7
.9)
18
.8 (
11
.7 -
31
.8)
35
.5 (
16
.6 -
55
.0)
H0B
1E
1C
2C
41
.5 (
0.7
- 2
.9)
12
.6 (
8.6
- 1
6.4
)0
.5 (
0.2
- 1
.1)
2.3
(1
.3 -
3.8
)1
0.3
(7
.1 -
14
.1)
2.3
(1
.2 -
4.3
)1
.5 (
0.7
- 2
.8)
8.2
(5
.2 -
12
.2)
27
.4 (
12
.2 -
43
.4)
H0B
1E
1C
3C
03
1.0
(2
2.9
- 3
8.2
)5
1.2
(4
2.6
- 6
2.3
)2
.5 (
1.3
- 4
.6)
32
.9 (
26
.6 -
39
.9)
45
.6 (
39
.3 -
56
.3)
7.2
(4
.2 -
10
.6)
31
.8 (
23
.9 -
39
.1)
40
.9 (
34
.0 -
53
.3)
61
.3 (
43
.4 -
70
.9)
H0B
1E
1C
3C
18
.5 (
5.2
- 1
1.2
)4
6.7
(3
8.0
- 5
7.2
)0
.8 (
0.4
- 1
.5)
9.6
(6
.7 -
12
.6)
37
.4 (
30
.8 -
48
.2)
2.7
(1
.4 -
4.4
)8
.5 (
5.3
- 1
1.2
)2
9.5
(2
1.8
- 4
3.7
)3
7.1
(1
8.9
- 5
1.5
)
H0B
1E
1C
3C
25
.8 (
3.5
- 9
.5)
46
.4 (
38
.0 -
56
.6)
0.5
(0
.3 -
1.0
)8
.3 (
5.3
- 1
2.8
)3
7.5
(3
1.2
- 4
8.6
)2
.3 (
1.2
- 3
.9)
5.9
(3
.3 -
9.3
)2
9.6
(2
1.6
- 4
4.2
)2
8.8
(1
3.5
- 4
3.5
)
H0B
1E
1C
3C
41
.2 (
0.6
- 2
.2)
6.1
(3
.9 -
8.2
)0
.8 (
0.4
- 2
.1)
1.2
(0
.6 -
2.0
)4
.7 (
3.0
- 6
.8)
2.6
(1
.2 -
5.4
)1
.2 (
0.6
- 2
.3)
3.3
(1
.6 -
6.3
)4
1.9
(1
9.0
- 6
6.8
)
H0B
1E
1C
4C
02
9.8
(2
1.8
- 3
7.2
)4
5.2
(3
7.1
- 5
5.8
)2
.8 (
1.4
- 5
.0)
31
.7 (
25
.5 -
38
.6)
40
.9 (
35
.1 -
50
.9)
7.7
(4
.5 -
11
.1)
30
.7 (
22
.7 -
38
.0)
37
.6 (
31
.8 -
48
.7)
62
.4 (
45
.1 -
71
.4)
H0B
1E
1C
4C
17
.3 (
4.1
- 1
0.0
)4
0.6
(3
3.0
- 5
0.8
)0
.8 (
0.3
- 1
.5)
8.4
(5
.6 -
11
.2)
32
.7 (
27
.0 -
42
.5)
2.7
(1
.4 -
4.3
)7
.4 (
4.2
- 9
.8)
26
.2 (
19
.7 -
38
.3)
36
.4 (
17
.4 -
50
.1)
H0B
1E
1C
4C
24
.6 (
2.3
- 8
.1)
40
.3 (
32
.8 -
50
.8)
0.5
(0
.2 -
1.0
)7
.1 (
4.3
- 1
1.5
)3
2.8
(2
7.3
- 4
2.9
)2
.3 (
1.2
- 3
.9)
4.7
(2
.3 -
8.2
)2
6.3
(1
9.8
- 3
9.2
)2
6.7
(1
1.2
- 4
0.6
)
H0B
1E
1C
4C
33
.5 (
1.8
- 6
.1)
19
.3 (
14
.4 -
24
.6)
0.8
(0
.3 -
1.8
)3
.7 (
2.1
- 6
.1)
15
.0 (
10
.9 -
20
.9)
2.6
(1
.2 -
4.7
)3
.5 (
1.8
- 6
.1)
10
.5 (
5.7
- 2
1.0
)4
0.6
(1
6.5
- 6
4.2
)
H1B
0E
0C
05
5.8
(4
1.7
- 6
7.5
)6
3.4
(5
0.0
- 8
1.4
)3
.6 (
1.8
- 6
.3)
61
.7 (
50
.1 -
73
.8)
63
.2 (
51
.3 -
79
.0)
9.5
(6
.0 -
13
.3)
57
.6 (
43
.4 -
68
.6)
62
.9 (
52
.2 -
77
.8)
65
.0 (
52
.4 -
71
.7)
H1B
0E
0C
11
3.7
(7
.5 -
18
.4)
55
.1 (
42
.0 -
71
.4)
1.1
(0
.5 -
2.0
)1
6.4
(1
0.7
- 2
1.6
)4
7.3
(3
6.4
- 6
1.5
)3
.6 (
2.0
- 5
.7)
13
.9 (
7.8
- 1
8.3
)4
0.8
(3
1.8
- 5
5.3
)4
0.9
(2
3.7
- 5
2.1
)
H1B
0E
0C
28
.8 (
4.5
- 1
5.2
)5
4.5
(4
1.5
- 7
0.9
)0
.7 (
0.3
- 1
.3)
13
.9 (
8.3
- 2
2.2
)4
7.6
(3
6.7
- 6
1.8
)3
.1 (
1.6
- 5
.1)
9.0
(4
.3 -
15
.0)
41
.1 (
31
.7 -
55
.5)
30
.7 (
15
.2 -
44
.7)
H1B
0E
0C
36
.6 (
3.4
- 1
1.1
)2
3.1
(1
5.9
- 3
1.4
)1
.2 (
0.5
- 2
.9)
7.2
(4
.0 -
11
.5)
19
.1 (
13
.1 -
27
.2)
3.9
(1
.8 -
7.0
)6
.7 (
3.5
- 1
1.2
)1
4.8
(9
.5 -
24
.8)
47
.8 (
24
.0 -
66
.2)
H1B
0E
1C
0C
12
4.2
(1
7.3
- 3
1.0
)5
7.1
(4
3.9
- 7
3.8
)1
.8 (
0.9
- 3
.3)
27
.4 (
20
.4 -
34
.4)
51
.1 (
39
.9 -
65
.1)
5.5
(3
.3 -
8.2
)2
4.7
(1
7.5
- 3
1.4
)4
6.1
(3
6.9
- 6
0.4
)5
2.1
(3
6.3
- 6
1.8
)
H1B
0E
1C
0C
21
2.2
(8
.1 -
17
.1)
35
.6 (
25
.3 -
48
.3)
1.5
(0
.6 -
2.9
)2
5.5
(1
8.8
- 3
4.1
)5
1.3
(4
0.3
- 6
5.7
)5
.1 (
3.0
- 7
.7)
21
.0 (
14
.8 -
28
.5)
46
.3 (
37
.2 -
60
.5)
48
.0 (
32
.8 -
58
.5)
H1B
0E
1C
0C
31
8.9
(1
3.2
- 2
5.2
)3
2.7
(2
3.8
- 4
2.5
)2
.4 (
1.2
- 4
.7)
20
.3 (
14
.6 -
26
.4)
29
.6 (
22
.1 -
38
.7)
6.9
(4
.1 -
10
.7)
19
.2 (
13
.4 -
25
.5)
26
.3 (
19
.6 -
36
.2)
59
.8 (
42
.6 -
70
.5)
H1B
0E
1C
0C
41
3.9
(9
.0 -
19
.7)
15
.1 (
10
.2 -
20
.3)
3.8
(2
.0 -
7.5
)1
4.9
(1
0.2
- 2
0.3
)1
5.1
(1
0.4
- 2
0.2
)9
.6 (
6.2
- 1
3.9
)1
4.2
(9
.2 -
20
.3)
15
.0 (
10
.5 -
20
.2)
65
.8 (
53
.9 -
74
.1)
H1B
0E
1C
1C
04
5.3
(3
3.6
- 5
5.5
)6
1.4
(4
8.1
- 7
9.1
)3
.1 (
1.5
- 5
.3)
50
.8 (
40
.5 -
61
.5)
59
.4 (
47
.9 -
74
.7)
8.4
(5
.1 -
12
.1)
46
.8 (
34
.8 -
56
.7)
57
.6 (
47
.9 -
72
.2)
62
.3 (
48
.6 -
69
.6)
H1B
0E
1C
1C
21
0.1
(6
.3 -
15
.1)
54
.7 (
41
.6 -
70
.7)
0.8
(0
.4 -
1.5
)1
4.5
(9
.9 -
21
.4)
47
.5 (
36
.6 -
61
.6)
3.2
(1
.8 -
5.0
)1
0.2
(6
.3 -
15
.1)
41
.1 (
32
.1 -
55
.2)
33
.6 (
19
.0 -
46
.0)
H1B
0E
1C
1C
38
.4 (
5.3
- 1
2.1
)3
0.7
(2
2.1
- 4
0.5
)1
.2 (
0.5
- 2
.4)
9.4
(6
.2 -
13
.0)
25
.8 (
18
.8 -
34
.8)
3.8
(2
.1 -
6.2
)8
.4 (
5.3
- 1
2.3
)2
1.0
(1
4.6
- 3
1.3
)4
5.0
(2
5.8
- 5
9.9
)
H1B
0E
1C
1C
43
.4 (
1.8
- 5
.2)
13
.1 (
8.6
- 1
7.7
)1
.1 (
0.5
- 2
.3)
4.0
(2
.3 -
5.8
)1
1.3
(7
.4 -
15
.5)
3.6
(2
.0 -
6.1
)3
.4 (
1.8
- 5
.2)
9.7
(6
.4 -
14
.0)
41
.7 (
24
.6 -
55
.4)
H1B
0E
1C
2C
04
4.0
(3
2.4
- 5
4.7
)6
1.3
(4
8.1
- 7
8.9
)3
.0 (
1.5
- 5
.2)
50
.2 (
40
.0 -
61
.0)
59
.5 (
48
.1 -
74
.9)
8.3
(5
.1 -
11
.9)
45
.5 (
33
.8 -
55
.7)
57
.7 (
48
.0 -
72
.0)
61
.6 (
47
.6 -
69
.0)
H1B
0E
1C
2C
11
2.4
(7
.6 -
16
.6)
55
.0 (
41
.9 -
71
.3)
1.0
(0
.5 -
1.8
)1
5.8
(1
0.9
- 2
0.5
)4
7.4
(3
6.3
- 6
1.3
)3
.5 (
2.0
- 5
.4)
12
.6 (
7.8
- 1
6.5
)4
0.9
(3
2.2
- 5
5.5
)3
8.6
(2
2.9
- 4
9.7
)
H1B
0E
1C
2C
37
.1 (
4.3
- 1
1.2
)3
0.6
(2
2.0
- 4
0.3
)1
.0 (
0.5
- 2
.1)
8.8
(5
.8 -
12
.8)
25
.9 (
18
.9 -
34
.8)
3.5
(1
.9 -
6.1
)7
.2 (
4.3
- 1
1.2
)2
1.1
(1
4.7
- 3
1.3
)4
1.0
(2
2.8
- 5
6.9
)
H1B
0E
1C
2C
42
.2 (
1.0
- 4
.0)
13
.0 (
8.4
- 1
7.7
)0
.7 (
0.3
- 1
.4)
3.4
(1
.9 -
5.8
)1
1.4
(7
.6 -
15
.5)
3.1
(1
.7 -
5.5
)2
.2 (
1.0
- 3
.9)
9.8
(6
.5 -
14
.2)
31
.3 (
16
.0 -
46
.6)
H1B
0E
1C
3C
04
3.5
(3
2.1
- 5
3.8
)5
3.8
(4
2.2
- 6
9.5
)3
.4 (
1.7
- 5
.9)
48
.5 (
38
.6 -
59
.0)
52
.7 (
42
.5 -
66
.9)
9.0
(5
.5 -
12
.7)
45
.0 (
33
.1 -
54
.7)
51
.4 (
42
.9 -
64
.8)
64
.0 (
50
.5 -
71
.0)
H1B
0E
1C
3C
11
1.9
(7
.2 -
15
.9)
47
.5 (
36
.5 -
62
.0)
1.1
(0
.5 -
2.0
)1
4.2
(9
.5 -
18
.5)
40
.6 (
31
.0 -
53
.3)
3.6
(2
.0 -
5.7
)1
2.1
(7
.4 -
16
.0)
34
.6 (
27
.0 -
48
.3)
41
.5 (
24
.2 -
53
.1)
H1B
0E
1C
3C
28
.3 (
4.9
- 1
3.3
)4
7.0
(3
6.0
- 6
1.7
)0
.8 (
0.3
- 1
.4)
12
.3 (
7.7
- 1
9.0
)4
0.8
(3
1.3
- 5
3.7
)3
.2 (
1.7
- 5
.1)
8.4
(4
.8 -
13
.1)
34
.8 (
26
.9 -
48
.2)
33
.0 (
17
.7 -
46
.4)
H1B
0E
1C
3C
41
.7 (
0.8
- 3
.0)
5.5
(3
.3 -
7.9
)1
.3 (
0.5
- 3
.1)
1.7
(0
.9 -
3.0
)4
.5 (
2.8
- 6
.5)
3.9
(1
.9 -
7.3
)1
.6 (
0.8
- 3
.1)
3.5
(2
.0 -
6.1
)4
8.7
(2
5.2
- 6
8.7
)
H1B
0E
1C
4C
04
1.8
(3
0.2
- 5
2.3
)4
8.3
(3
7.8
- 6
3.6
)3
.6 (
1.7
- 6
.1)
46
.8 (
36
.8 -
57
.3)
48
.2 (
38
.9 -
61
.3)
9.5
(5
.8 -
13
.2)
43
.3 (
31
.6 -
53
.2)
47
.9 (
39
.5 -
60
.1)
64
.8 (
51
.6 -
71
.3)
H1B
0E
1C
4C
11
0.3
(5
.5 -
14
.0)
42
.0 (
32
.0 -
55
.3)
1.0
(0
.4 -
1.9
)1
2.5
(7
.9 -
16
.7)
36
.0 (
27
.4 -
48
.1)
3.6
(1
.9 -
5.6
)1
0.4
(5
.7 -
14
.1)
31
.1 (
24
.3 -
43
)4
0.6
(2
2.4
- 5
1.6
)
H1B
0E
1C
4C
26
.7 (
3.4
- 1
1.5
)4
1.5
(3
2.0
- 5
5.2
)0
.7 (
0.3
- 1
.3)
10
.6 (
6.2
- 1
7.1
)3
6.3
(2
7.8
- 4
7.8
)3
.1 (
1.6
- 5
.1)
6.8
(3
.1 -
11
.1)
31
.3 (
24
.2 -
43
.3)
30
.6 (
14
.8 -
44
.3)
H1B
0E
1C
4C
34
.9 (
2.5
- 8
.2)
17
.6 (
12
.2 -
24
.4)
1.2
(0
.5 -
2.8
)5
.4 (
3.1
- 8
.8)
14
.5 (
9.9
- 2
0.9
)3
.9 (
1.8
- 6
.9)
5.0
(2
.6 -
8.4
)1
1.2
(7
.2 -
19
.4)
47
.5 (
23
.8 -
65
.4)
H1B
1E
0C
03
3.6
(2
3.4
- 4
2.7
)3
8.8
(2
8.3
- 5
1.9
)3
.6 (
1.7
- 6
.8)
33
.0 (
24
.7 -
42
.4)
35
.9 (
27
.8 -
48
.1)
9.0
(5
.4 -
13
.6)
34
.2 (
24
.0 -
43
.6)
33
.3 (
26
.4 -
44
.9)
67
.6 (
51
.8 -
76
.0)
H1B
1E
0C
18
.2 (
4.3
- 1
1.1
)3
4.9
(2
4.8
- 4
7.3
)1
.0 (
0.4
- 2
.1)
8.8
(5
.5 -
11
.9)
28
.2 (
20
.8 -
39
.5)
3.2
(1
.7 -
5.7
)8
.2 (
4.3
- 1
1.0
)2
2.6
(1
5.4
- 3
5.4
)4
2.4
(2
1.9
- 5
7.3
)
H1B
1E
0C
25
.0 (
2.5
- 8
.9)
34
.6 (
24
.4 -
47
.3)
0.6
(0
.3 -
1.3
)7
.4 (
4.4
- 1
2.4
)2
8.3
(2
0.6
- 3
9.8
)2
.7 (
1.3
- 5
.0)
5.1
(2
.3 -
8.8
)2
2.7
(1
5.9
- 3
5.8
)3
1.5
(1
4.1
- 4
8.0
)
H1B
1E
0C
34
.0 (
2.0
- 6
.8)
17
.2 (
10
.8 -
24
.5)
1.0
(0
.4 -
2.6
)3
.8 (
2.1
- 6
.4)
13
.4 (
8.5
- 2
0.5
)3
.0 (
1.3
- 6
.3)
4.0
(2
.0 -
6.9
)9
.4 (
4.7
- 1
8.8
)4
6.2
(2
0.0
- 7
0.1
)
H1B
1E
1C
0C
11
4.6
(9
.6 -
19
.3)
35
.8 (
25
.6 -
48
.4)
1.7
(0
.8 -
3.4
)1
4.6
(1
0.6
- 1
9.5
)3
0.0
(2
2.6
- 4
1.4
)5
.0 (
2.7
- 8
.0)
14
.6 (
9.9
- 1
9.3
)2
5.1
(1
8.1
- 3
7.6
)5
4.2
(3
4.1
- 6
5.7
)
Ta
ble
2 (
co
nti
nu
ed
)
Co
mb
ina
tio
nP
1P
2P
3
WP
-W
P+
Dis
ca
rd R
atio
WP
-W
P+
Dis
ca
rd R
atio
WP
-W
P+
Dis
ca
rd R
atio
H1B
1E
1C
0C
21
2.2
(8.1
- 17
.1)
35
.6 (2
5.3
- 48
.3)
1.5
(0.6
- 2.9
)1
3.6
(9.7
- 19
.1)
30
.1 (2
2.3
- 41
.6)
4.6
(2.6
- 7.5
)1
2.4
(8.2
- 17
.3)
25
.2 (1
8.5
- 37
.7)
49
.9 (3
0.0
- 63
.1)
H1B
1E
1C
0C
31
1.4
(7.4
- 15
.7)
22
.4 (1
4.8
- 30
.4)
2.1
(1.0
- 4.6
)1
0.9
(7.6
- 14
.9)
18
.8 (1
2.9
- 26
.7)
5.9
(3.2
- 10
.0)
11
.5 (7
.5 - 1
5.8
)1
5.1
(9.6
- 24
.3)
60
.7 (3
8.1
- 74
.8)
H1B
1E
1C
0C
48
.4 (5
.3 - 1
2.5
)9
.2 (5
.9 - 1
2.8
)3
.8 (1
.7 - 7
.6)
8.0
(5.3
- 11
.5)
8.6
(5.8
- 11
.8)
9.1
(5.5
- 14
.3)
8.5
(5.3
- 12
.5)
7.9
(5.4
- 11
.1)
68
.5 (5
3.0
- 78
.2)
H1B
1E
1C
1C
02
7.2
(18
.7 - 3
4.5
)3
7.8
(27
.5 - 5
0.8
)3
.0 (1
.4 - 5
.7)
27
.2 (2
0.2
- 34
.7)
34
.0 (2
6.2
- 46
.0)
7.9
(4.5
- 12
.0)
27
.7 (1
9.2
- 35
.2)
30
.7 (2
3.8
- 42
.6)
64
.8 (4
6.8
- 74
.1)
H1B
1E
1C
1C
25
.8 (3
.5 - 8
.9)
34
.7 (2
4.5
- 47
.3)
0.7
(0.3
- 1.5
)7
.7 (5
.1 - 1
1.8
)2
8.3
(20
.7 - 3
9.9
)2
.9 (1
.4 - 4
.9)
5.9
(3.5
- 9.0
)2
2.7
(15
.8 - 3
5.5
)3
4.6
(17
.4 - 4
9.1
)
H1B
1E
1C
1C
35
.1 (3
.0 - 7
.4)
21
.4 (1
4.0
- 29
.3)
1.0
(0.4
- 2.3
)5
.0 (3
.2 - 7
.3)
17
.0 (1
1.2
- 24
.7)
3.1
(1.6
- 5.8
)5
.0 (3
.0 - 7
.4)
12
.5 (7
.3 - 2
2.1
)4
4.8
(22
.6 - 6
4.5
)
H1B
1E
1C
1C
42
.1 (1
.0 - 3
.2)
8.3
(5.1
- 11
.5)
1.1
(0.4
- 2.4
)2
.1 (1
.2 - 3
.2)
6.7
(4.3
- 9.7
)3
.3 (1
.7 - 6
.0)
2.0
(1.0
- 3.1
)5
.4 (3
.2 - 8
.6)
43
.4 (2
2.8
- 59
.8)
H1B
1E
1C
2C
02
6.4
(18
.0 - 3
3.7
)3
7.8
(27
.4 - 5
0.9
)2
.9 (1
.4 - 5
.6)
26
.8 (2
0.0
- 34
.5)
34
.1 (2
6.2
- 45
.8)
7.8
(4.4
- 11
.9)
27
.0 (1
8.8
- 34
.3)
30
.7 (2
4.0
- 42
.4)
64
.1 (4
6.1
- 73
.6)
H1B
1E
1C
2C
17
.4 (4
.3 - 9
.9)
34
.8 (2
4.7
- 47
.3)
0.9
(0.4
- 1.9
)8
.4 (5
.6 - 1
1.4
)2
8.2
(20
.8 - 3
9.5
)3
.1 (1
.6 - 5
.4)
7.4
(4.3
- 10
.0)
22
.6 (1
5.6
- 35
.4)
40
.0 (2
0.7
- 54
.9)
H1B
1E
1C
2C
34
.2 (2
.5 - 6
.7)
1.4
(14
.1 - 2
9.2
)0
.8 (0
.4 - 2
.0)
4.7
(2.9
- 7.0
)1
7.0
(11
.2 - 2
4.8
)2
.9 (1
.4 - 5
.5)
4.2
(2.4
- 6.7
)1
2.6
(7.3
- 22
.2)
40
.7 (1
9.1
- 60
.9)
H1B
1E
1C
2C
41
.2 (0
.6 - 2
.4)
8.3
(5.1
- 11
.5)
0.6
(0.3
- 1.5
)1
.8 (1
.0 - 3
.1)
6.8
(4.4
- 9.7
)2
.8 (1
.3 - 5
.2)
1.3
(0.6
- 2.3
)5
.4 (3
.3 - 8
.6)
32
.1 (1
4.9
- 49
.3)
H1B
1E
1C
3C
02
6.2
(17
.9 - 3
3.5
)3
3.6
(24
.4 - 4
5.6
)3
.2 (1
.5 - 6
.2)
26
.0 (1
9.2
- 33
.4)
30
.5 (2
3.3
- 41
.3)
8.4
(4.8
- 12
.7)
26
.7 (1
8.5
- 34
.1)
27
.6 (2
1.5
- 38
.4)
66
.3 (4
8.3
- 75
.4)
H1B
1E
1C
3C
17
.1 (4
.1 - 9
.8)
30
.7 (2
1.7
- 42
.0)
1.0
(0.4
- 2.0
)7
.6 (5
.0 - 1
0.3
)2
4.7
(18
.1 - 3
5.4
)3
.2 (1
.6 - 5
.5)
7.1
(4.1
- 9.7
)1
9.4
(12
.9 - 3
1.5
)4
2.8
(21
.8 - 5
8.1
)
H1B
1E
1C
3C
24
.8 (2
.7 - 8
.0)
30
.5 (2
1.4
- 41
.8)
0.7
(0.3
- 1.4
)6
.5 (4
.0 - 1
0.4
)2
4.8
(17
.9 - 3
5.7
)2
.8 (1
.3 - 5
.0)
4.9
(2.6
- 7.8
)1
9.5
(13
.3 - 3
1.6
)3
3.6
(16
.0 - 4
9.5
)
H1B
1E
1C
3C
41
.0 (0
.5 - 1
.9)
4.1
(2.3
- 5.9
)1
.0 (0
.4 - 2
.8)
0.9
(0.4
- 1.6
)3
.2 (1
.8 - 4
.9)
3.0
(1.3
- 6.6
)1
.0 (0
.4 - 1
.9)
2.2
(1.0
- 4.5
)4
7.2
(21
.5 - 7
2.7
)
H1B
1E
1C
4C
02
5.1
(16
.9 - 3
2.2
)2
9.5
(21
.7 - 4
0.7
)3
.5 (1
.6 - 6
.6)
25
.0 (1
8.5
- 32
.2)
27
.3 (2
1.1
- 37
.4)
9.0
(5.2
- 13
.2)
25
.7 (1
7.7
- 33
.2)
25
.3 (2
0.1
- 35
.1)
67
.4 (5
0.5
- 75
.8)
H1B
1E
1C
4C
16
.1 (3
.1 - 8
.4)
26
.6 (1
8.9
- 37
.2)
1.0
(0.4
- 2.0
)6
.7 (4
.1 - 9
.2)
21
.5 (1
5.8
- 30
.8)
3.2
(1.6
- 5.6
)6
.2 (3
.2 - 8
.4)
17
.2 (1
1.9
- 27
.5)
42
.1 (2
0.9
- 56
.5)
H1B
1E
1C
4C
23
.8 (1
.9 - 6
.7)
26
.4 (1
8.8
- 36
.9)
0.6
(0.3
- 1.3
)5
.6 (3
.3 - 9
.3)
21
.6 (1
5.6
- 31
.1)
2.7
(1.2
- 4.8
)3
.9 (1
.7 - 6
.6)
17
.3 (1
2.1
- 27
.6)
31
.3 (1
3.6
- 47
.4)
H1B
1E
1C
4C
33
.0 (1
.4 - 5
.1)
13
.1 (8
.3 - 1
9.2
)1
.0 (0
.4 - 2
.5)
2.9
(1.6
- 4.8
)1
0.2
(6.4
- 16
.0)
3.0
(1.3
- 6.1
)3
.0 (1
.4 - 5
.1)
7.2
(3.6
- 14
.8)
45
.8 (1
9.4
- 69
.4)
Ta
ble
2 (c
on
tinu
ed
)
Co
mb
ina
tion
P1
P2
P3
WP
-W
P+
Dis
ca
rd R
ati
oW
P-
WP
+D
isc
ard
Ra
tio
WP
-W
P+
Dis
ca
rd R
ati
o
H0B
0E
0C
03
0.7
(1
0.3
- 6
5.9
)6
3.7
(2
8.1
- 9
1.8
)3
7.0
(2
6.5
- 4
8.8
)1
4.6
(5
.6 -
31
.3)
74
.6 (
32
.4 -
96
.9)
1.9
(1
.3 -
3.4
)1
6.7
(6
.2 -
38
.8)
89
.2 (
51
.3 -
99
.2)
1.9
(0
.6 -
5.7
)
H0B
0E
0C
12
2.1
(1
0.5
- 3
2.5
)5
4.4
(4
4.1
- 6
2.6
)3
3.2
(1
9.9
- 4
3.4
)8
.8 (
2.9
- 1
6.8
)6
4.4
(5
7.2
- 7
0.9
)1
.3 (
0.5
- 2
.8)
10
.1 (
2.9
- 1
9.5
)8
1.5
(7
7.0
- 8
5.9
)1
.3 (
0.3
- 3
.5)
H0B
0E
0C
32
.0 (
0.7
- 3
.8)
3.3
(1
.2 -
7.5
)4
2.6
(2
6.2
- 5
1.1
)1
.4 (
0.4
- 2
.6)
3.5
(1
.3 -
8.0
)4
.0 (
1.4
- 6
.4)
1.6
(0
.4 -
2.9
)4
.6 (
1.7
- 1
2.4
)3
.4 (
0.7
- 7
.8)
H0B
0E
1C
0C
12
4.1
(1
3.2
- 3
6.0
)5
6.6
(4
4.6
- 6
6.3
)3
4.2
(2
3.3
- 4
4.0
)1
0.2
(4
.6 -
17
.9)
66
.8 (
55
.2 -
74
.8)
1.5
(0
.8 -
2.7
)1
1.6
(5
.1 -
20
.4)
83
.3 (
73
.6 -
88
.2)
1.4
(0
.4 -
3.8
)
H0B
0E
1C
0C
38
.7 (
3.8
- 1
7.5
)1
7.4
(8
.8 -
26
.3)
37
.9 (
28
.2 -
48
.7)
4.5
(2
.1 -
8.8
)2
0.0
(9
.7 -
28
.5)
2.2
(1
.5 -
3.8
)5
.1 (
2.4
- 1
0.4
)2
4.3
(1
4.8
- 3
2.3
)2
.2 (
0.7
- 6
.0)
H0B
0E
1C
0C
47
.1 (
2.4
- 1
6.0
)1
4.9
(6
.2 -
23
.6)
37
.0 (
26
.7 -
49
.4)
3.4
(1
.3 -
7.6
)1
7.4
(7
.1 -
25
.6)
1.9
(1
.3 -
3.5
)3
.9 (
1.4
- 9
.4)
20
.8 (
11
.1 -
27
.3)
1.9
(0
.6 -
6.1
)
H0B
0E
1C
1C
02
8.7
(1
2.6
- 5
4.8
)6
1.5
(3
3.2
- 8
3.3
)3
6.3
(2
6.4
- 4
7.2
)1
3.3
(6
.0 -
26
.4)
72
.2 (
39
.5 -
89
.3)
1.8
(1
.2 -
3.1
)1
5.2
(6
.7 -
31
.8)
87
.4 (
58
.0 -
95
.3)
1.8
(0
.6 -
5.2
)
H0B
0E
1C
1C
36
.7 (
3.5
- 9
.6)
15
.2 (
11
.1 -
19
.8)
34
.9 (
22
.8 -
44
.8)
3.2
(1
.3 -
5.3
)1
7.7
(1
3.3
- 2
2.4
)1
.8 (
0.8
- 3
.1)
3.6
(1
.4 -
6.0
)2
2.5
(1
7.4
- 2
9.3
)1
.6 (
0.4
- 3
.9)
H0B
0E
1C
1C
45
.2 (
2.3
- 7
.9)
12
.7 (
8.8
- 1
6.4
)3
3.2
(2
0.0
- 4
3.7
)2
.1 (
0.7
- 4
.1)
15
.0 (
10
.6 -
19
.2)
1.3
(0
.5 -
2.8
)2
.4 (
0.7
- 4
.7)
19
.0 (
13
.6 -
24
.1)
1.3
(0
.3 -
3.5
)
H0B
0E
1C
3C
02
4.0
(8
.2 -
51
.4)
49
.6 (
22
.1 -
72
.7)
37
.1 (
26
.8 -
48
.9)
11
.5 (
4.5
- 2
4.5
)5
8.0
(2
5.7
- 7
7.1
)1
.9 (
1.3
- 3
.4)
13
.2 (
5.0
- 2
9.8
)6
9.5
(4
0.4
- 8
0.5
)2
.0 (
0.6
- 5
.8)
H0B
0E
1C
3C
11
7.4
(8
.6 -
26
.1)
42
.5 (
34
.6 -
49
.9)
33
.4 (
20
.3 -
43
.5)
7.1
(2
.5 -
13
.2)
50
.2 (
44
.1 -
57
.0)
1.4
(0
.6 -
2.8
)8
.1 (
2.6
- 1
5.4
)6
3.6
(5
8.5
- 7
0.1
)1
.3 (
0.3
- 3
.4)
H0B
0E
1C
3C
40
.5 (
0.1
- 0
.9)
0.8
(0
.3 -
1.7
)4
2.6
(2
6.3
- 5
1.4
)0
.3 (
0.1
- 0
.7)
0.8
(0
.3 -
1.9
)4
.0 (
1.4
- 6
.8)
0.4
(0
.1 -
0.7
)1
.1 (
0.4
- 2
.9)
3.4
(0
.7 -
8.0
)
H0B
0E
1C
4C
02
3.5
(7
.9 -
50
.8)
48
.9 (
21
.3 -
71
.6)
37
.0 (
26
.4 -
48
.8)
11
.2 (
4.2
- 2
4.1
)5
7.2
(2
5.0
- 7
6.1
)1
.9 (
1.3
- 3
.4)
12
.8 (
4.7
- 2
9.3
)6
8.4
(3
9.1
- 7
9.6
)1
.9 (
0.6
- 5
.7)
H0B
0E
1C
4C
11
6.9
(8
.2 -
25
.6)
41
.7 (
34
.0 -
49
.2)
33
.2 (
19
.9 -
43
.4)
6.7
(2
.3 -
12
.9)
49
.4 (
43
.1 -
56
.4)
1.3
(0
.5 -
2.8
)7
.7 (
2.1
- 1
5.1
)6
2.5
(5
7.1
- 6
9.0
)1
.3 (
0.3
- 3
.4)
H0B
0E
1C
4C
31
.5 (
0.5
- 2
.8)
2.5
(0
.9 -
5.9
)4
2.6
(2
6.0
- 5
0.8
)1
.1 (
0.3
- 2
.0)
2.7
(1
.0 -
6.3
)4
.0 (
1.4
- 6
.4)
1.2
(0
.3 -
2.2
)3
.6 (
1.3
- 9
.6)
3.4
(0
.7 -
7.8
)
H0B
1E
0C
08
.2 (
2.7
- 1
7.9
)1
1.8
(5
.2 -
18
.3)
45
.8 (
32
.6 -
57
.9)
4.6
(1
.8 -
10
.7)
11
.2 (
5.0
- 1
7.0
)3
.9 (
2.5
- 7
.7)
5.6
(2
.2 -
13
.4)
15
.0 (
8.1
- 2
1.7
)3
.8 (
1.1
- 1
1.8
)
H0B
1E
0C
15
.9 (
2.6
- 8
.9)
10
.1 (
7.5
- 1
3.0
)4
1.4
(2
4.1
- 5
3.0
)2
.6 (
0.8
- 6
.2)
9.7
(7
.7 -
12
.6)
2.6
(0
.9 -
6.2
)3
.3 (
0.9
- 7
.1)
13
.8 (
9.7
- 1
9.4
)2
.4 (
0.4
- 7
.2)
H0B
1E
0C
30
.5 (
0.2
- 1
.0)
0.6
(0
.2 -
1.5
)5
1.8
(3
1.3
- 6
1.2
)0
.5 (
0.1
- 1
.1)
0.5
(0
.2 -
1.3
)8
.6 (
2.7
- 1
5.8
)0
.5 (
0.2
- 1
.1)
0.8
(0
.3 -
2.4
)6
.7 (
1.3
- 1
5.5
)
H0B
1E
1C
0C
16
.4 (
3.4
- 9
.7)
10
.5 (
7.7
- 1
3.8
)4
2.6
(2
8.4
- 5
4.1
)3
.0 (
1.4
- 6
.3)
10
.0 (
7.7
- 1
3.3
)2
.9 (
1.4
- 6
.2)
3.8
(1
.6 -
7.5
)1
4.0
(9
.5 -
19
.9)
2.8
(0
.7 -
7.7
)
H0B
1E
1C
0C
32
.3 (
1.0
- 4
.7)
3.2
(1
.6 -
5.5
)4
6.8
(3
4.1
- 5
8.1
)1
.4 (
0.7
- 3
.2)
3.0
(1
.5 -
4.9
)4
.6 (
2.9
- 8
.8)
1.7
(0
.8 -
3.8
)4
.1 (
2.2
- 6
.7)
4.2
(1
.4 -
11
.8)
H0B
1E
1C
0C
41
.9 (
0.6
- 4
.2)
2.8
(1
.1 -
4.7
)4
5.8
(3
3.1
- 5
8.6
)1
.1 (
0.4
- 2
.7)
2.6
(1
.1 -
4.3
)3
.9 (
2.5
- 8
.3)
1.3
(0
.5 -
3.3
)3
.5 (
1.7
- 5
.4)
3.8
(1
.2 -
11
.9)
H0B
1E
1C
1C
07
.6 (
3.3
- 1
5.1
)1
1.4
(6
.1 -
16
.6)
45
.0 (
31
.9 -
56
.7)
4.1
(2
.0 -
8.7
)1
0.8
(6
.1 -
15
.6)
3.6
(2
.3 -
7.0
)5
.1 (
2.3
- 1
1.2
)1
4.7
(8
.7 -
21
.0)
3.5
(1
.1 -
10
.2)
H0B
1E
1C
1C
31
.8 (
0.9
- 2
.6)
2.8
(1
.9 -
4.0
)4
3.4
(2
7.4
- 5
4.7
)1
.0 (
0.4
- 1
.9)
2.7
(1
.9 -
3.9
)3
.5 (
1.5
- 7
.4)
1.2
(0
.5 -
2.2
)3
.8 (
2.5
- 6
.2)
3.1
(0
.8 -
7.8
)
H0B
1E
1C
1C
41
.4 (
0.6
- 2
.2)
2.4
(1
.5 -
3.4
)4
1.4
(2
4.4
- 5
3.6
)0
.6 (
0.2
- 1
.5)
2.3
(1
.6 -
3.2
)2
.6 (
0.9
- 6
.5)
0.8
(0
.2 -
1.7
)3
.2 (
2.0
- 4
.9)
2.4
(0
.5 -
7.2
)
H0B
1E
1C
3C
06
.4 (
2.2
- 1
3.9
)9
.2 (
4.1
- 1
4.2
)4
5.9
(3
2.8
- 5
7.8
)3
.6 (
1.5
- 8
.3)
8.7
(3
.9 -
13
.3)
4.0
(2
.5 -
7.7
)4
.4 (
1.8
- 1
0.5
)1
1.7
(6
.4 -
17
.0)
3.8
(1
.2 -
11
.8)
H0B
1E
1C
3C
14
.6 (
2.2
- 7
.0)
7.9
(5
.9 -
10
.2)
41
.6 (
24
.5 -
53
.2)
2.1
(0
.7 -
5.0
)7
.5 (
6.0
- 1
0.0
)2
.7 (
1.0
- 6
.3)
2.6
(0
.8 -
5.6
)1
0.7
(7
.5 -
15
.4)
2.5
(0
.5 -
7.2
)
H0B
1E
1C
3C
40
.1 (
0.0
- 0
.2)
0.1
(0
.0 -
0.4
)5
1.8
(3
1.3
- 6
1.7
)0
.1 (
0.0
- 0
.3)
0.1
(0
.0 -
0.3
)8
.7 (
2.8
- 1
6.6
)0
.1 (
0.0
- 0
.3)
0.2
(0
.1 -
0.6
)6
.7 (
1.3
- 1
5.8
)
H0B
1E
1C
4C
06
.3 (
2.1
- 1
3.8
)9
.1 (
3.9
- 1
4.0
)4
5.8
(3
2.4
- 5
7.8
)3
.5 (
1.4
- 8
.1)
8.6
(3
.8 -
13
.2)
3.9
(2
.5 -
7.6
)4
.3 (
1.6
- 1
0.4
)1
1.5
(6
.2 -
16
.7)
3.8
(1
.1 -
11
.7)
H0B
1E
1C
4C
14
.5 (
2.0
- 6
.9)
7.8
(5
.7 -
10
.0)
41
.4 (
23
.9 -
52
.9)
2.0
(0
.6 -
4.8
)7
.4 (
5.9
- 9
.8)
2.6
(0
.9 -
6.2
)2
.5 (
0.6
- 5
.5)
10
.6 (
7.4
- 1
5.1
)2
.4 (
0.4
- 7
.1)
H0B
1E
1C
4C
30
.4 (
0.1
- 0
.8)
0.5
(0
.2 -
1.1
)5
1.8
(3
1.0
- 6
1.0
)0
.4 (
0.1
- 0
.8)
0.4
(0
.2 -
1.0
)8
.6 (
2.7
- 1
5.6
)0
.4 (
0.1
- 0
.8)
0.6
(0
.2 -
1.9
)6
.7 (
1.3
- 1
5.5
)
H1B
0E
0C
07
.8 (
2.7
- 1
9.9
)9
.2 (
3.8
- 1
7.4
)5
0.8
(3
9.9
- 6
1.4
)5
.5 (
2.2
- 1
2.9
)1
0.7
(4
.2 -
17
.4)
4.9
(3
.2 -
9.7
)6
.2 (
2.4
- 1
5.7
)1
0.0
(4
.9 -
14
.6)
6.1
(2
.1 -
18
.0)
H1B
0E
0C
15
.5 (
2.4
- 1
0.1
)7
.6 (
5.4
- 1
1.7
)4
6.9
(2
9.1
- 5
7.6
)2
.9 (
0.7
- 7
.3)
9.2
(6
.2 -
13
.2)
3.0
(1
.0 -
7.9
)3
.5 (
0.9
- 8
.3)
9.0
(6
.4 -
12
.7)
3.9
(0
.7 -
11
.7)
H1B
0E
0C
30
.5 (
0.2
- 1
.1)
0.5
(0
.2 -
1.1
)5
8.1
(3
9.8
- 6
4.9
)0
.7 (
0.2
- 1
.3)
0.5
(0
.2 -
1.2
)1
1.8
(3
.8 -
20
.1)
0.6
(0
.2 -
1.2
)0
.5 (
0.2
- 1
.4)
11
.7 (
2.3
- 2
5.7
)
H1B
0E
1C
0C
16
.0 (
3.1
- 1
1.5
)8
.0 (
5.5
- 1
2.4
)4
8.0
(3
4.0
- 5
8.0
)3
.5 (
1.4
- 7
.4)
9.6
(6
.3 -
13
.9)
3.5
(1
.7 -
7.6
)4
.2 (
1.6
- 8
.7)
9.3
(6
.4 -
12
.9)
4.5
(1
.2 -
12
.4)
H1B
0E
1C
0C
32
.2 (
0.9
- 5
.3)
2.5
(1
.2 -
4.7
)5
2.0
(4
1.3
- 6
1.7
)1
.8 (
0.9
- 3
.7)
2.9
(1
.3 -
4.6
)5
.9 (
3.7
- 1
1.4
)1
.9 (
0.9
- 4
.3)
2.7
(1
.5 -
4.1
)7
.0 (
2.5
- 1
9.4
)
H1B
0E
1C
0C
41
.8 (
0.6
- 4
.8)
2.1
(0
.8 -
4.2
)5
0.8
(4
0.1
- 6
1.8
)1
.3 (
0.5
- 3
.1)
2.5
(0
.9 -
4.2
)4
.9 (
3.3
- 1
0.3
)1
.5 (
0.5
- 3
.8)
2.3
(1
.1 -
3.5
)6
.1 (
2.1
- 1
8.3
)
H1B
0E
1C
1C
07
.2 (
3.1
- 1
7.0
)8
.8 (
4.5
- 1
5.7
)5
0.0
(3
8.8
- 6
0.2
)4
.9 (
2.2
- 1
0.7
)1
0.4
(5
.1 -
16
.4)
4.5
(2
.9 -
8.8
)5
.6 (
2.4
- 1
3.0
)9
.8 (
5.6
- 1
4.0
)5
.7 (
1.9
- 1
6.0
)
H1B
0E
1C
1C
31
.7 (
0.8
- 3
.0)
2.1
(1
.4 -
3.3
)4
9.1
(3
4.3
- 5
8.5
)1
.2 (
0.4
- 2
.3)
2.5
(1
.6 -
3.8
)4
.5 (
1.9
- 9
.3)
1.3
(0
.5 -
2.5
)2
.5 (
1.6
- 3
.8)
5.2
(1
.3 -
13
.2)
Co
mb
ina
tio
nP
1P
2P
3
Ta
ble
3.
Ha
dd
oc
k:
we
igh
t in
dic
ato
rs f
or
the 6
0 B
RD
co
mb
inatio
ns e
ach
of
the 3
re
alis
tic p
op
ula
tion
sce
na
rios.
Th
e v
alu
es o
f th
e f
acto
rs u
sed
fo
r
the le
ng
th-w
eig
ht
co
nve
rsio
n w
ere
: a
= 0
.00
65 a
nd
b =
3.1
08
3
WP
-W
P+
Dis
ca
rd R
atio
WP
-W
P+
Dis
ca
rd R
atio
WP
-W
P+
Dis
ca
rd R
atio
H1B
0E
1C
1C
41
.3 (0
.5 - 2
.4)
1.8
(1.1
- 2.7
)4
6.9
(29
.6 - 5
7.8
)0
.7 (0
.2 - 1
.7)
2.1
(1.3
- 3.1
)3
.0 (1
.0 - 8
.2)
0.8
(0.2
- 1.9
)2
.1 (1
.3 - 3
.1)
4.0
(0.7
- 11
.7)
H1B
0E
1C
3C
06
.1 (2
.2 - 1
5.7
)7
.1 (3
.0 - 1
3.7
)5
0.9
(40
.1 - 6
1.4
)4
.4 (1
.8 - 1
0.1
)8
.3 (3
.3 - 1
3.9
)5
.0 (3
.2 - 9
.9)
4.9
(1.9
- 12
.4)
7.8
(4.0
- 11
.4)
6.2
(2.1
- 18
.0)
H1B
0E
1C
3C
14
.4 (1
.9 - 8
.0)
6.0
(4.2
- 9.2
)4
7.1
(30
.1 - 5
7.7
)2
.4 (0
.7 - 5
.7)
7.2
(4.8
- 10
.4)
3.2
(1.1
- 8.0
)2
.9 (0
.8 - 6
.5)
7.0
(5.0
- 10
.1)
4.1
(0.8
- 11
.8)
H1B
0E
1C
3C
40
.1 (0
.0 - 0
.3)
0.1
(0.0
- 0.3
)5
8.1
(39
.9 - 6
5.1
)0
.2 (0
.0 - 0
.3)
0.1
(0.0
- 0.3
)1
2.0
(4.0
- 21
.0)
0.1
(0.0
- 0.3
)0
.1 (0
.0 - 0
.3)
11
.8 (2
.3 - 2
5.8
)
H1B
0E
1C
4C
05
.9 (2
.0 - 1
5.5
)7
.0 (2
.9 - 1
3.6
)5
0.8
(39
.9 - 6
1.4
)4
.2 (1
.7 - 9
.8)
8.2
(3.2
- 13
.7)
4.9
(3.1
- 9.7
)4
.8 (1
.8 - 1
2.3
)7
.7 (3
.9 - 1
1.2
)6
.1 (2
.0 - 1
8.0
)
H1B
0E
1C
4C
14
.2 (1
.9 - 7
.9)
5.9
(4.1
- 9.2
)4
6.9
(29
.1 - 5
7.6
)2
.2 (0
.6 - 5
.6)
7.1
(4.7
- 10
.2)
3.0
(1.0
- 7.9
)2
.7 (0
.7 - 6
.4)
6.9
(4.8
- 9.9
)3
.9 (0
.7 - 1
1.7
)
H1B
0E
1C
4C
30
.4 (0
.1 - 0
.9)
0.4
(0.1
- 0.9
)5
8.1
(39
.6 - 6
4.9
)0
.5 (0
.1 - 1
.0)
0.4
(0.1
- 0.9
)1
1.8
(3.8
- 19
.9)
0.5
(0.1
- 1.0
)0
.4 (0
.1 - 1
.1)
11
.7 (2
.3 - 2
5.6
)
H1B
1E
0C
02
.1 (0
.7 - 5
.5)
1.7
(0.7
- 3.4
)6
0.9
(48
.5 - 7
0.2
)1
.9 (0
.8 - 5
.1)
1.6
(0.6
- 2.9
)1
0.6
(6.5
- 22
.3)
2.2
(0.8
- 5.9
)1
.6 (0
.7 - 2
.8)
12
.7 (4
.1 - 3
5.1
)
H1B
1E
0C
11
.5 (0
.6 - 2
.8)
1.4
(0.9
- 2.2
)5
6.9
(36
.9 - 6
6.9
)0
.9 (0
.2 - 3
.0)
1.4
(0.9
- 2.2
)6
.3 (1
.6 - 1
8.6
)1
.2 (0
.3 - 3
.1)
1.4
(0.9
- 2.5
)8
.1 (1
.3 - 2
4.7
)
H1B
1E
0C
30
.1 (0
.0 - 0
.3)
0.1
(0.0
- 0.2
)6
8.0
(48
.1 - 7
4.0
)0
.2 (0
.1 - 0
.6)
0.1
(0.0
- 0.2
)2
4.8
(8.0
- 42
.9)
0.2
(0.1
- 0.5
)0
.1 (0
.0 - 0
.2)
23
.1 (4
.9 - 4
5.2
)
H1B
1E
1C
0C
11
.6 (0
.8 - 3
.1)
1.4
(1.0
- 2.4
)5
8.0
(41
.9 - 6
7.3
)1
.2 (0
.5 - 3
.1)
1.4
(0.9
- 2.3
)7
.5 (3
.3 - 1
8.3
)1
.4 (0
.5 - 3
.3)
1.5
(0.9
- 2.5
)9
.3 (2
.4 - 2
5.3
)
H1B
1E
1C
0C
30
.6 (0
.3 - 1
.5)
0.5
(0.2
- 0.9
)6
2.1
(50
.2 - 7
0.6
)0
.6 (0
.3 - 1
.5)
0.4
(0.2
- 0.8
)1
2.8
(7.6
- 26
.9)
0.7
(0.3
- 1.6
)0
.4 (0
.2 - 0
.8)
14
.4 (5
.0 - 3
5.7
)
H1B
1E
1C
0C
40
.5 (0
.2 - 1
.3)
0.4
(0.1
- 0.8
)6
0.9
(48
.6 - 7
0.5
)0
.4 (0
.2 - 1
.3)
0.4
(0.1
- 0.7
)1
0.7
(6.6
- 23
.7)
0.5
(0.2
- 1.4
)0
.4 (0
.2 - 0
.7)
12
.7 (4
.2 - 3
5.7
)
H1B
1E
1C
1C
02
.0 (0
.8 - 4
.7)
1.6
(0.8
- 3.0
)6
0.1
(47
.7 - 6
9.4
)1
.7 (0
.7 - 4
.2)
1.5
(0.8
- 2.7
)9
.7 (6
.0 - 1
9.9
)1
.9 (0
.9 - 4
.8)
1.5
(0.8
- 2.7
)1
1.7
(3.6
- 31
.4)
H1B
1E
1C
1C
30
.5 (0
.2 - 0
.8)
0.4
(0.3
- 0.6
)5
9.2
(41
.7 - 6
7.7
)0
.4 (0
.1 - 1
.0)
0.4
(0.2
- 0.6
)9
.7 (3
.8 - 2
2.6
)0
.5 (0
.2 - 1
.0)
0.4
(0.2
- 0.7
)1
0.9
(2.5
- 26
.9)
H1B
1E
1C
1C
40
.3 (0
.1 - 0
.7)
0.3
(0.2
- 0.5
)5
6.9
(36
.9 - 6
7.2
)0
.2 (0
.0 - 0
.7)
0.3
(0.2
- 0.5
)6
.4 (1
.7 - 1
9.2
)0
.3 (0
.1 - 0
.7)
0.3
(0.2
- 0.6
)8
.2 (1
.3 - 2
4.7
)
H1B
1E
1C
3C
01
.7 (0
.6 - 4
.3)
1.3
(0.5
- 2.6
)6
1.0
(48
.7 - 7
0.2
)1
.5 (0
.7 - 3
.9)
1.2
(0.5
- 2.3
)1
0.8
(6.7
- 22
.5)
1.7
(0.7
- 4.6
)1
.2 (0
.6 - 2
.2)
12
.8 (4
.2 - 3
4.2
)
H1B
1E
1C
3C
11
.2 (0
.5 - 2
.2)
1.1
(0.7
- 1.7
)5
7.2
(37
.6 - 6
7.0
)0
.8 (0
.2 - 2
.4)
1.1
(0.7
- 1.7
)6
.7 (2
.0 - 1
8.8
)1
.0 (0
.3 - 2
.4)
1.1
(0.7
- 2.0
)8
.4 (1
.5 - 2
4.9
)
H1B
1E
1C
3C
40
.0 (0
.0 - 0
.1)
0.0
(0.0
- 0.1
)6
8.0
(48
.3 - 7
4.1
)0
.1 (0
.0 - 0
.1)
0.0
(0.0
- 0.0
)2
5.0
(8.6
- 44
.3)
0.1
(0.0
- 0.1
)0
.0 (0
.0 - 0
.1)
23
.2 (4
.9 - 4
6.1
)
H1B
1E
1C
4C
01
.6 (0
.5 - 4
.2)
1.3
(0.5
- 2.6
)6
0.9
(48
.5 - 7
0.1
)1
.5 (0
.6 - 3
.8)
1.2
(0.5
- 2.3
)1
0.6
(6.5
- 22
.1)
1.7
(0.6
- 4.5
)1
.2 (0
.6 - 2
.1)
12
.6 (4
.0 - 3
4.5
)
H1B
1E
1C
4C
11
.1 (0
.5 - 2
.2)
1.1
(0.7
- 1.7
)5
6.9
(36
.8 - 6
6.9
)0
.7 (0
.2 - 2
.3)
1.1
(0.7
- 1.7
)6
.3 (1
.6 - 1
8.4
)0
.9 (0
.2 - 2
.4)
1.1
(0.7
- 1.9
)8
.1 (1
.3 - 2
4.7
)
H1B
1E
1C
4C
30
.1 (0
.0 - 0
.2)
0.1
(0.0
- 0.2
)6
8.0
(48
.0 - 7
4.0
)0
.2 (0
.0 - 0
.4)
0.1
(0.0
- 0.2
)2
4.7
(7.9
- 42
.6)
0.2
(0.0
- 0.4
)0
.1 (0
.0 - 0
.2)
23
.1 (4
.8 - 4
4.9
)
Ta
ble
3 (c
on
tinu
ed
)
Co
mb
ina
tion
P1
P2
P3
WP
-W
P+
Dis
ca
rd R
ati
oW
P-
WP
+D
isc
ard
Ra
tio
WP
-W
P+
Dis
ca
rd R
ati
o
H0B
0E
0C
04
7 (
33
.1 -
62
.3)
85
.2 (
77
.1 -
91
.1)
0.6
(0
.3 -
1.0
)7
5.3
(6
6.2
- 8
4.0
)9
7.4
(9
6.4
- 9
8.2
)1
9.6
(1
3.3
- 2
6.1
)3
0.7
(1
0.3
- 6
5.9
)6
3.7
(2
8.1
- 9
1.8
)3
7.0
(2
6.5
- 4
8.8
)
H0B
0E
0C
16
.8 (
1.7
- 1
8.1
)3
6.5
(2
8.0
- 4
6.8
)0
.2 (
0 -
0.6
)1
9.2
(1
2.5
- 2
4.3
)7
8.5
(7
4.3
- 8
2.2
)7
.2 (
4.0
- 1
0.6
)2
2.1
(1
0.5
- 3
2.5
)5
4.4
(4
4.1
- 6
2.6
)3
3.2
(1
9.9
- 4
3.4
)
H0B
0E
0C
23
3 (
24
.1 -
44
.8)
66
.8 (
61
.4 -
71
.8)
0.5
(0
.3 -
0.9
)1
5.0
(8
.9 -
23
.4)
78
.5 (
72
.5 -
84
.4)
5.7
(3
.1 -
9.3
) -
- -
H0B
0E
0C
33
2.9
(1
7.0
- 4
1.9
)3
8.3
(2
8.9
- 4
7.1
)0
.9 (
0.5
- 1
.4)
8.7
(5
.1 -
14
.0)
23
.6 (
18
.7 -
30
.3)
10
.4 (
5.4
- 1
8.0
)2
.0 (
0.7
- 3
.8)
3.3
(1
.2 -
7.5
)4
2.6
(2
6.2
- 5
1.1
)
H0B
0E
1C
0C
14
2.1
(2
9.4
- 5
5.8
)7
5.4
(6
8.8
- 8
1.6
)0
.6 (
0.3
- 1
.0)
32
.8 (
26
.2 -
39
.2)
83
.0 (
79
.4 -
86
.1)
11
.1 (
7.1
- 1
5.7
)2
4.1
(1
3.2
- 3
6.0
)5
6.6
(4
4.6
- 6
6.3
)3
4.2
(2
3.3
- 4
4.0
)
H0B
0E
1C
0C
24
5.3
(3
3.0
- 5
8.7
)8
1.5
(7
4.6
- 8
6.3
)0
.6 (
0.3
- 1
.0)
29
.6 (
23
.1 -
37
.5)
83
.0 (
78
.3 -
87
.6)
10
.1 (
6.4
- 1
4.7
) -
- -
H0B
0E
1C
0C
34
5.3
(3
2.4
- 5
8.0
)7
5.6
(6
8.3
- 8
1.6
)0
.7 (
0.3
- 1
.0)
24
.8 (
19
.2 -
31
.6)
41
.2 (
34
.9 -
47
.2)
16
.0 (
10
.6 -
22
.9)
8.7
(3
.8 -
17
.5)
17
.4 (
8.8
- 2
6.3
)3
7.9
(2
8.2
- 4
8.7
)
H0B
0E
1C
0C
44
1.3
(2
8.3
- 5
5.0
)6
7.8
(6
0.5
- 7
5.2
)0
.7 (
0.3
- 1
.1)
18
.3 (
12
.9 -
24
.8)
23
.2 (
16
.6 -
28
.5)
19
.8 (
13
.8 -
28
.1)
7.1
(2
.4 -
16
.0)
14
.9 (
6.2
- 2
3.6
)3
7.0
(2
6.7
- 4
9.4
)
H0B
0E
1C
1C
01
1.6
(6
.7 -
21
.9)
46
.3 (
39
.1 -
54
.2)
0.3
(0
.1 -
0.6
)6
1.7
(5
3.6
- 6
9.7
)9
2.9
(9
1.2
- 9
4.7
)1
7.3
(1
1.4
- 2
3.5
)2
8.7
(1
2.6
- 5
4.8
)6
1.5
(3
3.2
- 8
3.3
)3
6.3
(2
6.4
- 4
7.2
)
H0B
0E
1C
1C
21
0.0
(5
.4 -
20
.0)
42
.6 (
35
.6 -
50
.8)
0.3
(0
.1 -
0.6
)1
6.0
(1
1.2
- 2
2.5
)7
8.5
(7
3.6
- 8
3.6
)6
.1 (
3.7
- 9
.5)
- -
-
H0B
0E
1C
1C
39
.9 (
4.9
- 2
0.2
)3
6.7
(3
0.0
- 4
5.0
)0
.3 (
0.1
- 0
.7)
11
.2 (
7.8
- 1
5.4
)3
6.7
(3
1.0
- 4
2.1
)8
.8 (
5.3
- 1
3.7
)6
.7 (
3.5
- 9
.6)
15
.2 (
11
.1 -
19
.8)
34
.9 (
22
.8 -
44
.8)
H0B
0E
1C
1C
46
.0 (
1.5
- 1
6.1
)2
8.9
(2
2.4
- 3
7.6
)0
.2 (
0.1
- 0
.6)
4.7
(2
.7 -
6.7
)1
8.7
(1
3.4
- 2
3.2
)7
.3 (
4.2
- 1
1.7
)5
.2 (
2.3
- 7
.9)
12
.7 (
8.8
- 1
6.4
)3
3.2
(2
0.0
- 4
3.7
)
H0B
0E
1C
2C
03
4.7
(2
6.7
- 4
4.9
)7
0.5
(6
5.8
- 7
4.4
)0
.5 (
0.3
- 0
.9)
60
.7 (
52
.4 -
68
.5)
92
.9 (
90
.7 -
95
.0)
17
.1 (
11
.2 -
23
.3)
- -
-
H0B
0E
1C
2C
12
9.9
(2
1.7
- 4
0.1
)6
0.7
(5
5.7
- 6
5.7
)0
.5 (
0.3
- 0
.9)
18
.2 (
13
.0 -
22
.5)
78
.5 (
74
.4 -
82
.2)
6.8
(4
.0 -
9.9
) -
- -
H0B
0E
1C
2C
33
3.0
(2
4.6
- 4
2.9
)6
0.9
(5
6.0
- 6
5.5
)0
.6 (
0.3
- 0
.9)
10
.2 (
7.0
- 1
5.0
)3
6.7
(3
1.0
- 4
2.6
)8
.1 (
4.9
- 1
2.9
) -
- -
H0B
0E
1C
2C
42
9.1
(2
0.5
- 3
9.4
)5
3.1
(4
7.9
- 5
8.9
)0
.6 (
0.3
- 1
.0)
3.6
(2
.0 -
6.0
)1
8.7
(1
3.3
- 2
3.4
)5
.8 (
3.2
- 1
0.2
) -
- -
H0B
0E
1C
3C
03
4.6
(2
0.3
- 4
2.9
)4
7.9
(3
9.3
- 5
5.4
)0
.8 (
0.4
- 1
.2)
59
.1 (
50
.7 -
67
.5)
79
.8 (
75
.5 -
85
.2)
18
.9 (
12
.7 -
25
.3)
24
.0 (
8.2
- 5
1.4
)4
9.6
(2
2.1
- 7
2.7
)3
7.1
(2
6.8
- 4
8.9
)
H0B
0E
1C
3C
12
9.8
(1
6.0
- 3
8.1
)3
8.1
(3
0.8
- 4
5.9
)0
.9 (
0.4
- 1
.3)
16
.7 (
11
.6 -
21
.1)
65
.4 (
61
.2 -
70
.4)
7.4
(4
.4 -
10
.8)
17
.4 (
8.6
- 2
6.1
)4
2.5
(3
4.6
- 4
9.9
)3
3.4
(2
0.3
- 4
3.5
)
H0B
0E
1C
3C
23
3.0
(1
9.0
- 4
1.1
)4
4.2
(3
6.1
- 5
1.6
)0
.8 (
0.4
- 1
.2)
13
.5 (
8.6
- 2
0.4
)6
5.4
(6
0.0
- 7
1.7
)6
.1 (
3.5
- 9
.7)
- -
-
H0B
0E
1C
3C
42
9.0
(1
5.0
- 3
6.8
)3
0.5
(2
3.0
- 3
8.0
)1
.0 (
0.5
- 1
.5)
2.1
(1
.1 -
3.7
)5
.6 (
3.8
- 7
.6)
10
.5 (
5.7
- 1
9.6
)0
.5 (
0.1
- 0
.9)
0.8
(0
.3 -
1.7
)4
2.6
(2
6.3
- 5
1.4
)
H0B
0E
1C
4C
05
.7 (
2.8
- 9
.5)
17
.4 (
11
.5 -
20
.6)
0.4
(0
.2 -
0.7
)5
7.0
(4
8.3
- 6
5.9
)7
4.2
(6
8.7
- 8
1.1
)1
9.5
(1
2.9
- 2
5.8
)2
3.5
(7
.9 -
50
.8)
48
.9 (
21
.3 -
71
.6)
37
.0 (
26
.4 -
48
.8)
H0B
0E
1C
4C
10
.8 (
0.2
- 2
.7)
7.6
(4
.9 -
10
.2)
0.1
(0
.0 -
0.4
)1
4.6
(9
.4 -
19
.0)
59
.8 (
54
.7 -
66
.1)
7.1
(3
.9 -
10
.5)
16
.9 (
8.2
- 2
5.6
)4
1.7
(3
4.0
- 4
9.2
)3
3.2
(1
9.9
- 4
3.4
)
H0B
0E
1C
4C
24
.0 (
2.2
- 7
.1)
13
.7 (
9.3
- 1
6.1
)0
.3 (
0.2
- 0
.7)
11
.4 (
6.7
- 1
8.0
)5
9.8
(5
4.1
- 6
7.2
)5
.7 (
3.0
- 9
.3)
- -
-
H0B
0E
1C
4C
34
.0 (
1.8
- 6
.8)
7.8
(4
.9 -
10
.3)
0.6
(0
.2 -
1.1
)6
.6 (
3.7
- 1
0.6
)1
8.0
(1
4.1
- 2
3.9
)1
0.3
(5
.3 -
17
.3)
1.5
(0
.5 -
2.8
)2
.5 (
0.9
- 5
.9)
42
.6 (
26
.0 -
50
.8)
H0B
1E
0C
03
5.0
(2
2.9
- 4
9.9
)8
0.1
(6
8.7
- 8
7.4
)0
.5 (
0.2
- 0
.8)
41
.6 (
34
.0 -
50
.0)
49
.1 (
41
.9 -
58
.0)
21
.1 (
14
.0 -
29
.1)
8.2
(2
.7 -
17
.9)
11
.8 (
5.2
- 1
8.3
)4
5.8
(3
2.6
- 5
7.9
)
H0B
1E
0C
15
.1 (
1.3
- 1
4.4
)3
4.9
(2
5.9
- 4
4.3
)0
.2 (
0.0
- 0
.5)
10
.6 (
6.7
- 1
4.1
)4
0.2
(3
3.5
- 4
8.3
)7
.7 (
4.2
- 1
1.8
)5
.9 (
2.6
- 8
.9)
10
.1 (
7.5
- 1
3.0
)4
1.4
(2
4.1
- 5
3.0
)
H0B
1E
0C
22
4.7
(1
7.1
- 3
7.4
)6
3.0
(5
5.1
- 6
9.2
)0
.4 (
0.2
- 0
.7)
8.1
(4
.8 -
13
.4)
40
.1 (
33
.4 -
48
.9)
6.0
(3
.2 -
10
.3)
- -
-
H0B
1E
0C
32
4.5
(1
2.5
- 3
5.0
)3
6.1
(2
6.8
- 4
4.4
)0
.7 (
0.4
- 1
.2)
4.8
(2
.7 -
8.1
)1
3.9
(1
0.3
- 1
9.2
)9
.8 (
4.9
- 1
7.2
)0
.5 (
0.2
- 1
.0)
0.6
(0
.2 -
1.5
)5
1.8
(3
1.3
- 6
1.2
)
H0B
1E
1C
0C
13
1.4
(2
0.8
- 4
5.6
)7
1.0
(6
1.1
- 7
8.4
)0
.5 (
0.2
- 0
.8)
18
.1 (
13
.9 -
23
.1)
42
.3 (
35
.6 -
50
.6)
11
.9 (
7.2
- 1
7.2
)6
.4 (
3.4
- 9
.7)
10
.5 (
7.7
- 1
3.8
)4
2.6
(2
8.4
- 5
4.1
)
H0B
1E
1C
0C
23
3.8
(2
2.9
- 4
7.8
)7
6.7
(6
6.4
- 8
3.4
)0
.5 (
0.2
- 0
.8)
16
.3 (
12
.0 -
21
.5)
42
.3 (
35
.5 -
50
.8)
10
.8 (
6.7
- 1
6.2
) -
- -
Co
mb
ina
tio
nNephrops
Co
dH
ad
do
ck
Ta
ble
4.
Mu
ltis
pe
cie
s s
ce
na
rio
: w
eig
ht
indic
ato
rs f
or
the 9
6 B
RD
co
mb
inatio
ns e
ach
of
the m
ultis
pecie
s c
atc
h s
ce
na
rios.
WP
-W
P+
Dis
ca
rd R
atio
WP
-W
P+
Dis
ca
rd R
atio
WP
-W
P+
Dis
ca
rd R
atio
H0B
1E
1C
0C
33
3.7
(22
.6 - 4
7.7
)7
1.1
(61
.0 - 7
8.4
)0
.5 (0
.3 - 0
.9)
13
.7 (1
0.0
- 18
.4)
22
.3 (1
7.7
- 27
.9)
16
.2 (1
0.4
- 24
.0)
2.3
(1.0
- 4.7
)3
.2 (1
.6 - 5
.5)
46
.8 (3
4.1
- 58
.1)
H0B
1E
1C
0C
43
0.8
(20
.0 - 4
4.6
)6
3.7
(54
.0 - 7
1.7
)0
.5 (0
.3 - 0
.9)
10
.1 (7
.0 - 1
4.3
)1
1.7
(8.2
- 15
.0)
21
.4 (1
4.4
- 31
.1)
1.9
(0.6
- 4.2
)2
.8 (1
.1 - 4
.7)
45
.8 (3
3.1
- 58
.6)
H0B
1E
1C
1C
08
.7 (4
.8 - 1
7.7
)4
4.0
(35
.3 - 5
1.8
)0
.2 (0
.1 - 0
.5)
34
.1 (2
7.4
- 40
.9)
46
.9 (4
0.0
- 55
.9)
18
.6 (1
2.0
- 25
.9)
7.6
(3.3
- 15
.1)
11
.4 (6
.1 - 1
6.6
)4
5.0
(31
.9 - 5
6.7
)
H0B
1E
1C
1C
27
.4 (3
.9 - 1
6.2
)4
0.6
(32
.3 - 4
8.1
)0
.2 (0
.1 - 0
.5)
8.7
(5.9
- 12
.7)
40
.1 (3
3.5
- 48
.4)
6.4
(3.7
- 10
.4)
- -
-
H0B
1E
1C
1C
37
.4 (3
.6 - 1
6.2
)3
5.0
(27
.5 - 4
3.3
)0
.2 (0
.1 - 0
.6)
6.2
(4.1
- 8.9
)2
0.2
(16
.1 - 2
5.2
)8
.8 (5
.2 - 1
4.3
)1
.8 (0
.9 - 2
.6)
2.8
(1.9
- 4.0
)4
3.4
(27
.4 - 5
4.7
)
H0B
1E
1C
1C
44
.4 (1
.1 - 1
2.3
)2
7.6
(20
.3 - 3
5.8
)0
.2 (0
.0 - 0
.5)
2.6
(1.5
- 3.8
)9
.6 (6
.6 - 1
2.5
)7
.8 (4
.3 - 1
2.6
)1
.4 (0
.6 - 2
.2)
2.4
(1.5
- 3.4
)4
1.4
(24
.4 - 5
3.6
)
H0B
1E
1C
2C
02
5.9
(18
.5 - 3
8.2
)6
6.5
(58
.5 - 7
2.2
)0
.4 (0
.2 - 0
.7)
33
.5 (2
6.7
- 40
.4)
46
.9 (3
9.9
- 55
.8)
18
.3 (1
1.9
- 25
.6)
- -
-
H0B
1E
1C
2C
12
2.3
(15
.2 - 3
3.7
)5
7.4
(50
.0 - 6
3.1
)0
.4 (0
.2 - 0
.7)
10
.0 (6
.8 - 1
2.8
)4
0.2
(33
.7 - 4
8.4
)7
.3 (4
.1 - 1
1.0
) -
- -
H0B
1E
1C
2C
32
4.6
(17
.1 - 3
6.0
)5
7.5
(49
.9 - 6
2.9
)0
.5 (0
.3 - 0
.8)
5.6
(3.7
- 8.5
)2
0.2
(16
.1 - 2
5.5
)8
.0 (4
.5 - 1
3.7
) -
- -
H0B
1E
1C
2C
42
1.7
(14
.6 - 3
3.0
)5
0.0
(43
.2 - 5
6.2
)0
.5 (0
.3 - 0
.8)
2.0
(1.1
- 3.4
)9
.6 (6
.6 - 1
2.6
)6
.1 (3
.3 - 1
1.4
) -
- -
H0B
1E
1C
3C
02
5.8
(14
.7 - 3
5.8
)4
5.1
(36
.2 - 5
2.4
)0
.6 (0
.3 - 1
.0)
32
.7 (2
6.1
- 39
.6)
40
.7 (3
5.1
- 49
.2)
20
.2 (1
3.2
- 27
.9)
6.4
(2.2
- 13
.9)
9.2
(4.1
- 14
.2)
45
.9 (3
2.8
- 57
.8)
H0B
1E
1C
3C
12
2.1
(11
.5 - 3
1.9
)3
6.0
(28
.2 - 4
3.2
)0
.7 (0
.3 - 1
.1)
9.2
(6.1
- 12
.0)
33
.9 (2
8.5
- 41
.6)
7.9
(4.4
- 11
.8)
4.6
(2.2
- 7.0
)7
.9 (5
.9 - 1
0.2
)4
1.6
(24
.5 - 5
3.2
)
H0B
1E
1C
3C
22
4.5
(13
.6 - 3
4.3
)4
1.7
(33
.3 - 4
8.6
)0
.6 (0
.3 - 1
.0)
7.3
(4.5
- 11
.5)
33
.9 (2
8.3
- 42
.0)
6.4
(3.5
- 10
.7)
- -
-
H0B
1E
1C
3C
42
1.5
(10
.7 - 3
0.9
)2
8.7
(21
.2 - 3
6.0
)0
.8 (0
.4 - 1
.3)
1.2
(0.6
- 2.1
)3
.3 (2
.2 - 4
.6)
9.9
(5.3
- 18
.8)
0.1
(0.0
- 0.2
)0
.1 (0
.0 - 0
.4)
51
.8 (3
1.3
- 61
.7)
H0B
1E
1C
4C
04
.2 (2
.1 - 7
.6)
16
.4 (1
0.6
- 19
.5)
0.3
(0.1
- 0.6
)3
1.5
(25
.0 - 3
8.6
)3
7.4
(32
.0 - 4
5.7
)2
1.0
(13
.6 - 2
8.6
)6
.3 (2
.1 - 1
3.8
)9
.1 (3
.9 - 1
4.0
)4
5.8
(32
.4 - 5
7.8
)
H0B
1E
1C
4C
10
.6 (0
.2 - 2
.0)
7.3
(4.8
- 9.6
)0
.1 (0
.0 - 0
.3)
8.0
(5.0
- 10
.7)
30
.6 (2
5.6
- 38
.2)
7.6
(4.1
- 11
.5)
4.5
(2.0
- 6.9
)7
.8 (5
.7 - 1
0.0
)4
1.4
(23
.9 - 5
2.9
)
H0B
1E
1C
4C
23
.0 (1
.6 - 5
.6)
13
.0 (8
.6 - 1
5.3
)0
.3 (0
.1 - 0
.6)
6.2
(3.6
- 10
.3)
30
.6 (2
5.5
- 38
.8)
6.0
(3.1
- 10
.0)
- -
-
H0B
1E
1C
4C
32
.9 (1
.3 - 5
.3)
7.4
(4.7
- 9.5
)0
.4 (0
.2 - 0
.9)
3.6
(2.0
- 6.1
)1
0.6
(8.0
- 14
.8)
9.7
(4.8
- 16
.6)
0.4
(0.1
- 0.8
)0
.5 (0
.2 - 1
.1)
51
.8 (3
1.0
- 61
.0)
H1B
0E
0C
05
3.0
(32
.7 - 7
4.7
)9
2.5
(74
.2 - 1
11
.4)
0.6
(0.3
- 1.1
)6
0.6
(48
.3 - 7
3.0
)6
1.9
(49
.8 - 8
1.8
)2
3.6
(14
.9 - 3
2.2
)7
.8 (2
.7 - 1
9.9
)9
.2 (3
.8 - 1
7.4
)5
0.8
(39
.9 - 6
1.4
)
H1B
0E
0C
17
.7 (1
.9 - 2
0.3
)4
0.5
(28
.7 - 5
7.2
)0
.2 (0
.0 - 0
.6)
15
.4 (9
.5 - 2
0.3
)4
9.2
(39
.1 - 6
6.4
)9
.0 (4
.6 - 1
3.8
)5
.5 (2
.4 - 1
0.1
)7
.6 (5
.4 - 1
1.7
)4
6.9
(29
.1 - 5
7.6
)
H1B
0E
0C
23
7.3
(23
.7 - 5
4.1
)7
2.8
(57
.9 - 8
7.8
)0
.6 (0
.3 - 1
.0)
12
.0 (6
.7 - 1
9.0
)4
9.1
(38
.9 - 6
6.3
)7
.1 (3
.6 - 1
2.1
) -
- -
H1B
0E
0C
33
7.4
(19
.1 - 5
2.1
)4
1.8
(29
.6 - 5
4.6
)1
.0 (0
.4 - 1
.6)
7.0
(3.8
- 11
.3)
15
.2 (1
1.2
- 22
.1)
12
.6 (6
.3 - 2
1.4
)0
.5 (0
.2 - 1
.1)
0.5
(0.2
- 1.1
)5
8.1
(39
.8 - 6
4.9
)
H1B
0E
1C
0C
14
7.6
(29
.3 - 6
7.7
)8
2.0
(65
.6 - 9
8.9
)0
.6 (0
.3 - 1
.1)
26
.4 (1
9.5
- 33
.4)
52
.2 (4
1.6
- 70
.1)
13
.7 (8
.0 - 1
9.7
)6
.0 (3
.1 - 1
1.5
)8
.0 (5
.5 - 1
2.4
)4
8.0
(34
.0 - 5
8.0
)
H1B
0E
1C
0C
25
1.1
(32
.6 - 7
1.3
)8
8.5
(71
.2 - 1
05
.4)
0.6
(0.3
- 1.1
)2
3.8
(17
.5 - 3
1.7
)5
2.2
(41
.7 - 6
9.5
)1
2.5
(7.4
- 18
.6)
- -
-
H1B
0E
1C
0C
35
1.1
(31
.8 - 7
0.8
)8
2.1
(66
.2 - 9
9.1
)0
.7 (0
.3 - 1
.2)
20
.0 (1
4.2
- 26
.0)
26
.4 (2
0.3
- 35
.5)
19
.3 (1
2.0
- 27
.6)
2.2
(0.9
- 5.3
)2
.5 (1
.2 - 4
.7)
52
.0 (4
1.3
- 61
.7)
H1B
0E
1C
0C
44
6.6
(28
.4 - 6
6.7
)7
3.6
(59
.0 - 8
9.1
)0
.7 (0
.3 - 1
.2)
14
.7 (9
.9 - 2
0.1
)1
4.8
(10
.0 - 2
0.5
)2
3.9
(15
.6 - 3
3.9
)1
.8 (0
.6 - 4
.8)
2.1
(0.8
- 4.2
)5
0.8
(40
.1 - 6
1.8
)
H1B
0E
1C
1C
01
3.2
(7.0
- 25
.4)
51
.0 (3
8.0
- 66
.6)
0.3
(0.1
- 0.7
)4
9.6
(38
.9 - 6
0.0
)5
8.9
(47
.2 - 7
7.9
)2
1.0
(12
.9 - 2
9.0
)7
.2 (3
.1 - 1
7.0
)8
.8 (4
.5 - 1
5.7
)5
0.0
(38
.8 - 6
0.2
)
H1B
0E
1C
1C
21
1.3
(5.8
- 22
.9)
47
.0 (3
4.9
- 62
.0)
0.3
(0.1
- 0.6
)1
2.8
(8.4
- 18
.7)
49
.1 (3
9.1
- 66
.0)
7.6
(4.1
- 12
.3)
- -
-
H1B
0E
1C
1C
31
1.3
(5.2
- 23
.3)
40
.6 (2
9.5
- 55
.0)
0.3
(0.1
- 0.7
)9
.0 (5
.8 - 1
2.8
)2
3.3
(17
.9 - 3
1.4
)1
0.9
(6.3
- 17
.4)
1.7
(0.8
- 3.0
)2
.1 (1
.4 - 3
.3)
49
.1 (3
4.3
- 58
.5)
H1B
0E
1C
1C
46
.8 (1
.6 - 1
8.6
)3
2.1
(22
.4 - 4
6.0
)0
.2 (0
.1 - 0
.7)
3.7
(2.1
- 5.6
)1
1.7
(7.9
- 16
.5)
9.1
(4.8
- 14
.8)
1.3
(0.5
- 2.4
)1
.8 (1
.1 - 2
.7)
46
.9 (2
9.6
- 57
.8)
Ta
ble
4 (c
on
tinu
ed
)
Co
mb
ina
tion
P1
P2
P3
WP
-W
P+
Dis
ca
rd R
ati
oW
P-
WP
+D
isc
ard
Ra
tio
WP
-W
P+
Dis
ca
rd R
ati
o
H1B
0E
1C
2C
03
9.2
(2
5.9
- 5
5.3
)7
6.8
(6
2.0
- 9
2.5
)0
.6 (
0.3
- 1
.0)
48
.8 (
38
.3 -
59
.2)
58
.8 (
47
.4 -
78
.4)
20
.7 (
12
.7 -
28
.8)
- -
-
H1B
0E
1C
2C
13
3.7
(2
1.6
- 4
9.0
)6
6.3
(5
2.7
- 8
0.4
)0
.6 (
0.3
- 1
.0)
14
.6 (
9.6
- 1
8.8
)4
9.2
(3
9.1
- 6
6.9
)8
.5 (
4.5
- 1
3.0
) -
- -
H1B
0E
1C
2C
33
7.3
(2
3.9
- 5
2.3
)6
6.4
(5
3.2
- 8
0.5
)0
.6 (
0.3
- 1
.1)
8.2
(5
.2 -
12
.1)
23
.3 (
17
.9 -
31
.8)
9.9
(5
.3 -
16
.5)
- -
-
H1B
0E
1C
2C
43
2.8
(2
0.6
- 4
7.8
)5
7.8
(4
6.1
- 7
0.3
)0
.6 (
0.3
- 1
.1)
2.9
(1
.5 -
5.1
)1
1.7
(8
.0 -
16
.6)
7.2
(3
.7 -
13
.0)
- -
-
H1B
0E
1C
3C
03
9.3
(2
1.4
- 5
3.3
)5
2.2
(3
9.4
- 6
5.4
)0
.8 (
0.4
- 1
.4)
47
.6 (
37
.0 -
57
.8)
50
.8 (
40
.7 -
67
.6)
22
.8 (
14
.0 -
31
.0)
6.1
(2
.2 -
15
.7)
7.1
(3
.0 -
13
.7)
50
.9 (
40
.1 -
61
.4)
H1B
0E
1C
3C
13
3.9
(1
7.6
- 4
7.5
)4
1.7
(3
0.8
- 5
4.1
)0
.9 (
0.4
- 1
.5)
13
.4 (
8.7
- 1
7.6
)4
1.1
(3
2.5
- 5
5.8
)9
.3 (
4.8
- 1
4.0
)4
.4 (
1.9
- 8
.0)
6.0
(4
.2 -
9.2
)4
7.1
(3
0.1
- 5
7.7
)
H1B
0E
1C
3C
23
7.4
(2
0.2
- 5
1.4
)4
8.2
(3
6.3
- 6
0.8
)0
.8 (
0.4
- 1
.4)
10
.8 (
6.5
- 1
6.5
)4
1.0
(3
2.6
- 5
5.9
)7
.6 (
4.0
- 1
2.4
) -
- -
H1B
0E
1C
3C
43
2.9
(1
6.7
- 4
5.9
)3
3.3
(2
3.5
- 4
3.8
)1
.1 (
0.5
- 1
.8)
1.7
(0
.8 -
3.0
)3
.6 (
2.3
- 5
.3)
12
.8 (
6.4
- 2
2.9
)0
.1 (
0.0
- 0
.3)
0.1
(0
.0 -
0.3
)5
8.1
(3
9.9
- 6
5.1
)
H1B
0E
1C
4C
06
.4 (
3.1
- 1
1.0
)1
8.9
(1
2.0
- 2
4.6
)0
.4 (
0.2
- 0
.8)
45
.9 (
35
.5 -
56
.1)
47
.1 (
37
.6 -
63
.3)
23
.5 (
14
.5 -
31
.7)
5.9
(2
.0 -
15
.5)
7.0
(2
.9 -
13
.6)
50
.8 (
39
.9 -
61
.4)
H1B
0E
1C
4C
10
.9 (
0.2
- 3
.0)
8.5
(5
.1 -
12
.4)
0.1
(0
.0 -
0.4
)1
1.7
(7
.0 -
15
.7)
37
.4 (
29
.6 -
51
.4)
9.0
(4
.5 -
13
.6)
4.2
(1
.9 -
7.9
)5
.9 (
4.1
- 9
.2)
46
.9 (
29
.1 -
57
.6)
H1B
0E
1C
4C
24
.5 (
2.3
- 8
.2)
15
.0 (
9.6
- 1
9.6
)0
.3 (
0.1
- 0
.7)
9.1
(5
.1 -
14
.6)
37
.4 (
29
.3 -
51
.3)
7.1
(3
.6 -
11
.9)
- -
-
H1B
0E
1C
4C
34
.5 (
1.8
- 8
.1)
8.6
(5
.1 -
12
.0)
0.6
(0
.2 -
1.2
)5
.3 (
2.9
- 8
.7)
11
.6 (
8.4
- 1
7.2
)1
2.6
(6
.1 -
21
.3)
0.4
(0
.1 -
0.9
)0
.4 (
0.1
- 0
.9)
58
.1 (
39
.6 -
64
.9)
H1B
1E
0C
03
9.5
(2
3.4
- 5
9.9
)8
7.2
(6
7.0
- 1
06
.5)
0.5
(0
.2 -
1.0
)3
3.6
(2
4.8
- 4
2.6
)3
1.5
(2
4.2
- 4
3.0
)2
5.1
(1
6.0
- 3
5.4
)2
.1 (
0.7
- 5
.5)
1.7
(0
.7 -
3.4
)6
0.9
(4
8.5
- 7
0.2
)
H1B
1E
0C
15
.7 (
1.4
- 1
6.4
)3
8.8
(2
6.4
- 5
4.2
)0
.2 (
0.0
- 0
.5)
8.5
(5
.0 -
11
.6)
25
.5 (
19
.4 -
35
.5)
9.5
(4
.8 -
15
.0)
1.5
(0
.6 -
2.8
)1
.4 (
0.9
- 2
.2)
56
.9 (
36
.9 -
66
.9)
H1B
1E
0C
22
7.8
(1
7.2
- 4
3.5
)6
8.8
(5
2.7
- 8
3)
0.4
(0
.2 -
0.9
)6
.5 (
3.5
- 1
0.9
)2
5.5
(1
9.3
- 3
5.8
)7
.4 (
3.7
- 1
3.0
) -
- -
H1B
1E
0C
32
7.8
(1
3.2
- 4
1.1
)3
9.5
(2
7.0
- 5
1.6
)0
.8 (
0.3
- 1
.4)
3.9
(2
.0 -
6.5
)9
.2 (
6.3
- 1
3.6
)1
1.7
(5
.5 -
20
.6)
0.1
(0
.0 -
0.3
)0
.1 (
0.0
- 0
.2)
68
.0 (
48
.1 -
74
.0)
H1B
1E
1C
0C
13
5.4
(2
0.7
- 5
3.5
)7
7.4
(5
9.7
- 9
4.4
)0
.5 (
0.2
- 1
.0)
14
.6 (
10
.3 -
19
.2)
27
.0 (
20
.7 -
37
.2)
14
.6 (
8.3
- 2
1.8
)1
.6 (
0.8
- 3
.1)
1.4
(1
.0 -
2.4
)5
8.0
(4
1.9
- 6
7.3
)
H1B
1E
1C
0C
23
8.1
(2
2.8
- 5
7.8
)8
3.5
(6
4.4
- 1
01
.4)
0.5
(0
.2 -
1.0
)1
3.1
(9
.2 -
18
.5)
26
.9 (
20
.6 -
37
.5)
13
.3 (
7.5
- 2
0.4
) -
- -
H1B
1E
1C
0C
33
8.1
(2
2.7
- 5
6.3
)7
7.4
(5
9.8
- 9
4.8
)0
.5 (
0.3
- 1
.0)
11
.1 (
7.6
- 1
5.1
)1
4.6
(1
0.7
- 2
0.1
)1
9.4
(1
1.7
- 2
8.8
)0
.6 (
0.3
- 1
.5)
0.5
(0
.2 -
0.9
)6
2.1
(5
0.2
- 7
0.6
)
H1B
1E
1C
0C
43
4.8
(2
0.1
- 5
2.3
)6
9.3
(5
3.1
- 8
5.3
)0
.5 (
0.3
- 1
.0)
8.2
(5
.3 -
11
.9)
7.5
(5
.1 -
10
.6)
25
.5 (
16
.5 -
36
.7)
0.5
(0
.2 -
1.3
)0
.4 (
0.1
- 0
.8)
60
.9 (
48
.6 -
70
.5)
H1B
1E
1C
1C
09
.8 (
5.2
- 2
0.7
)4
8.6
(3
5.1
- 6
3.1
)0
.2 (
0.1
- 0
.6)
27
.5 (
20
.1 -
35
.0)
30
.1 (
23
.1 -
41
.3)
22
.3 (
13
.6 -
32
.1)
2.0
(0
.8 -
4.7
)1
.6 (
0.8
- 3
.0)
60
.1 (
47
.7 -
69
.4)
H1B
1E
1C
1C
28
.4 (
4.1
- 1
8.7
)4
4.9
(3
2.1
- 5
8.6
)0
.2 (
0.1
- 0
.5)
7.0
(4
.5 -
10
.7)
25
.5 (
19
.4 -
35
.5)
7.9
(4
.1 -
13
.2)
- -
-
H1B
1E
1C
1C
38
.4 (
3.8
- 1
8.4
)3
8.8
(2
7.2
- 5
1.7
)0
.2 (
0.1
- 0
.6)
5.0
(3
.1 -
7.4
)1
3.1
(9
.5 -
18
.4)
10
.7 (
5.8
- 1
7.8
)0
.5 (
0.2
- 0
.8)
0.4
(0
.3 -
0.6
)5
9.2
(4
1.7
- 6
7.7
)
H1B
1E
1C
1C
45
.0 (
1.2
- 1
4.1
)3
0.7
(2
0.5
- 4
3.6
)0
.2 (
0.0
- 0
.6)
2.1
(1
.2 -
3.2
)6
.1 (
4.0
- 8
.6)
9.7
(5
.1 -
16
.1)
0.3
(0
.1 -
0.7
)0
.3 (
0.2
- 0
.5)
56
.9 (
36
.9 -
67
.2)
H1B
1E
1C
2C
02
9.2
(1
8.6
- 4
4.6
)7
2.6
(5
5.7
- 8
7.3
)0
.4 (
0.2
- 0
.9)
27
.0 (
19
.8 -
34
.4)
30
.1 (
23
.1 -
41
.2)
22
.0 (
13
.4 -
31
.5)
- -
-
H1B
1E
1C
2C
12
5.1
(1
5.5
- 3
9.0
)6
2.8
(4
8.0
- 7
6.2
)0
.4 (
0.2
- 0
.9)
8.0
(5
.1 -
10
.7)
25
.5 (
19
.4 -
35
.7)
9.0
(4
.7 -
14
.3)
- -
-
H1B
1E
1C
2C
32
7.8
(1
7.3
- 4
2.9
)6
2.8
(4
8.4
- 7
5.9
)0
.5 (
0.2
- 0
.9)
4.5
(2
.7 -
6.8
)1
3.1
(9
.5 -
18
.4)
9.8
(5
.1 -
16
.6)
- -
-
H1B
1E
1C
2C
42
4.5
(1
4.9
- 3
8.0
)5
4.6
(4
1.4
- 6
6.3
)0
.5 (
0.2
- 1
.0)
1.6
(0
.8 -
2.8
)6
.1 (
4.0
- 8
.7)
7.5
(3
.8 -
14
.1)
- -
-
H1B
1E
1C
3C
02
9.2
(1
5.0
- 4
2.2
)4
9.3
(3
5.9
- 6
2.2
)0
.6 (
0.3
- 1
.2)
26
.4 (
19
.3 -
33
.6)
26
.2 (
20
.3 -
36
.3)
24
.1 (
14
.7 -
34
.1)
1.7
(0
.6 -
4.3
)1
.3 (
0.5
- 2
.6)
61
.0 (
48
.7 -
70
.2)
H1B
1E
1C
3C
12
5.1
(1
2.3
- 3
7.8
)3
9.5
(2
8.3
- 5
1.0
)0
.7 (
0.3
- 1
.3)
7.4
(4
.6 -
9.9
)2
1.6
(1
6.4
- 3
0.4
)9
.7 (
5.0
- 1
5.2
)1
.2 (
0.5
- 2
.2)
1.1
(0
.7 -
1.7
)5
7.2
(3
7.6
- 6
7.0
)
Ta
ble
4 (
co
nti
nu
ed
)
Co
mb
ina
tio
nP
1P
2P
3
WP
-W
P+
Dis
ca
rd R
atio
WP
-W
P+
Dis
ca
rd R
atio
WP
-W
P+
Dis
ca
rd R
atio
H1B
1E
1C
3C
22
7.8
(14
.3 - 4
0.4
)4
5.6
(33
.1 - 5
7.8
)0
.7 (0
.3 - 1
.2)
5.9
(3.5
- 9.4
)2
1.6
(16
.3 - 3
0.6
)7
.9 (4
.0 - 1
3.5
) -
- -
H1B
1E
1C
3C
42
4.4
(11
.5 - 3
6.9
)3
1.4
(21
.5 - 4
1.5
)0
.8 (0
.4 - 1
.5)
0.9
(0.4
- 1.7
)2
.2 (1
.3 - 3
.4)
11
.9 (5
.9 - 2
2.5
)0
.0 (0
.0 - 0
.1)
0.0
(0.0
- 0.1
)6
8.0
(48
.3 - 7
4.1
)
H1B
1E
1C
4C
04
.8 (2
.3 - 9
.1)
17
.9 (1
1.2
- 23
.4)
0.3
(0.1
- 0.7
)2
5.4
(18
.5 - 3
2.6
)2
4.0
(18
.6 - 3
3.8
)2
5.0
(15
.6 - 3
4.9
)1
.6 (0
.5 - 4
.2)
1.3
(0.5
- 2.6
)6
0.9
(48
.5 - 7
0.1
)
H1B
1E
1C
4C
10
.7 (0
.2 - 2
.3)
8.1
(4.8
- 11
.7)
0.1
(0.0
- 0.4
)6
.4 (3
.7 - 8
.8)
19
.4 (1
4.8
- 27
.9)
9.5
(4.7
- 14
.7)
1.1
(0.5
- 2.2
)1
.1 (0
.7 - 1
.7)
56
.9 (3
6.8
- 66
.9)
H1B
1E
1C
4C
23
.4 (1
.7 - 6
.7)
14
.2 (8
.9 - 1
8.6
)0
.3 (0
.1 - 0
.6)
4.9
(2.7
- 8.4
)1
9.4
(14
.6 - 2
7.8
)7
.4 (3
.5 - 1
2.8
) -
- -
H1B
1E
1C
4C
33
.3 (1
.4 - 6
.4)
8.1
(4.8
- 11
.2)
0.4
(0.2
- 1.1
)2
.9 (1
.5 - 5
.0)
7.0
(4.9
- 10
.7)
11
.6 (5
.5 - 2
0.2
)0
.1 (0
.0 - 0
.2)
0.1
(0.0
- 0.2
)6
8.0
(48
.0 - 7
4.0
)
Ta
ble
4 (c
on
tinu
ed
)
Co
mb
ina
tion
P1
P2
P3
Technical University of Denmark
DTU AquaKemitorvetDK-2800 Kgs. Lyngby
www.aqua.dtu.dk