Nonylphenol

Concentration [µg/l]

0 100 200 300 400 500 600

Ext

inct

ion

risk

[%]

0

20

40

60

80

100

IRDAMTWA

Institute for Environmental Research

Thomas G. Preuss1, Tido Strauss2, Hans Toni Ratte1

1Institute for Environmental Research, RWTH Aachen University, Worringerweg 1, D-52056 Aachen, Germany2Research Institute gaiac, Mies-van-der-Rohe Str. 19, D-52074 Aachen, Germany

How detailed do we have to model populationsto predict extinction probabilities and recovery time?

IntroductionThe interest for the application of population modelling within the pesticide registration is

increasing nowadays. In 2007 the LEMTox workshop was conducted to discuss the

possibilities and obstacles of population modelling for this purpose. One major conclusion

from this workshops was that modelling for environmental risk assessment should not focus

on realism but more on protectiveness. For this poster we investigated the questions which

environmental factors have to be included to do protective population modelling.

newborn

Feeding

Ageing

growth

juvenile development

born juveniles

brood size

embryo development

�yes

yes yes

no

no

no

maximal age ?

Adult? Birthing?

Food concentration

Competition

single pond treatment compared to an isolated control pond

mesocosm facility treatment compared to an isolated untreated pond

mesocosm facility treatment compared to the mesocosm facility control ponds

Effect models

no

Calculation of current development per day

Begin of dormancy?

Molt?

Chronic or naturalmortality? Adult?

yesDeath

yesDormant

no

no

yes

Deathyes no

noyes

Newly hatched larva

Additional allochtonous seeding

Emigration

Autochtonous seeding

New eggs

Next larval orthe pupal stage

Update of the developmental stage

Calculation of the actual test item concentration in the water

Calculation of the actual food level depending on population density

yes

Female?

yes

Acute mortality?

Mesocosm Mesocosm Mesocosm Mesocosm scenarioscenarioscenarioscenario Single Single Single Single pondpondpondpond scenarioscenarioscenarioscenario

uniform distribution of emerged adults

mesocosm facility with10 treatments and 3 controls

isolated untreated pond

isolated single treatment

We used two individual-based population models, which were tested previously on measured data, to investigate

the influence of environmental factors on the extinction probability and recovery time.

Extinction probability of Daphnia magna populations (IDamP) were calculated at laboratory scale. Recovery time

within a mesocosm facility was calculated for Chaoborus crystallinus. Since chaoborids are flying insects a

metapopulation approach was used to calculate autochthonous and allochthonous recovery for a typical mesocosm

facility and a single treated pond in comparison to mesocosm controls and an isolated control pond.

∫−=t

dttCWTWAftS0

)(*_1)(

Time weighted average (TWA):Ashauer et al. 2006

Immediate response (IR):The concentration-response curve

Damage assessment model (DAM):Lee et al. 2002, Ashauer et al. 2006

BoutWinB CkCk

dt

dC ×−×=

DlkCkdt

dDlrBk ×−×=

)0,max()( tresDlth −=

Fig. 3: Extinction probability for different effect model.Three different effect models were used to describe the toxicity for individual daphnids. The IR uses the concentration response curve, the

TWA is a pseudokinetic model and the semi-mechanistic DAM describes the effect by an toxicokinetic/toxicodynamic approach.

���� Sensitivity 2 to 5 times lower compared to Immediate response

���� Sensitivity 2 to 3 times lower compared to standard food level

���� Sensitivity 2 to 3 times lower with competition

IDamP-Model

Fig. 1: Extinction probability at different food concentrationExtinction probabilities were calculated at different food concentrations. Food dependent toxicity was not taken into account. Predictions

of population dynamics at different food concentrations are shown in the right figure. At lower food concentrations extinction probabilities

increase.

Fig. 2: Extinction probability at different competition scenariosCompetition was calculated for three scenarios, the competitor was insensitive against the compound. Equal competitor means the same

population dublicated, stronger competitor had higher feeding rate, weaker competitor lower. The right figure shows the population dynamic

for the three examples.

� an appropriate effect model

� food dependency

� important interactions, like competition � This can easily be done using different scenarios. For it only a few additional data are needed.

� Metapopulation approaches for dispersed migrating species (e.g. Chaoborus) � Here immigration to the stressed population as well as emmigration from

untreated populations have to be considered.

Pesticide

Concentration [µg/l]

0 10 20 30 40 50

Ext

inct

ion

risk

[%]

0

20

40

60

80

100

IRDAMTWA

The IDamP model (Preuss et al. submitted):

• calibrated on individual level

• tested on individual and population level for

• different food concentrations & scenarios

• constant exposure (3,4-Dichloroanline (3,4-DCA),

Nonylphenol (p-NP))

• variable exposure (pesticide)

Extinction probabilities were calculated for 100 days at

constant exposure under semistatic conditions.

Substances were selected due to different mode of action,

p-NP (narcotic), 3,4-DCA (embryogenesis), pesticide (high

toxicity)

ConclusionsWith adequate models sensitivity of populations at different environmental scenarios can be investigated. This will help to estimate safety factors which are

protective but not over protective. For this purpose population models should include:

3,4-DCA

Concentration [µg/l]

0 20 40 60 80 100 120 140

Ext

inct

ion

risk

[%]

0

20

40

60

80

100

2x feedingStandard 0,1x feeding0,03x feeding

Nonylphenol

Concentration [µg/l]

0 100 200 300 400 500

Ext

inct

ion

risk

[%]

0

20

40

60

80

100

Time [d]

0 10 20 30 40

Pop

ulat

ion

size

0

100

200

300

400

500

600

Time [d]

0 10 20 30 40

0.5 mgC pop-1 d-1 1.3 mgC pop-1 d-1

3,4-DCA

Concentration [µg/l]

0 20 40 60 80 100 120 140

Ext

inct

ion

risk

[%]

0

20

40

60

80

100

without competitionweak competitorstrong competitorequal competitor

Pesticide

Concentration [µg/l]

0 10 20 30 40

Ext

inct

ion

risk

[%]

0

20

40

60

80

100

IBM-Chaoborus

The individual-based model IBM-Chaoborus:

• calibrated on individual level and one mesocosm

• tested on mesocosm data for

• untreated populations

• variable exposure scenarios (insecticide)

Fig. 6: Recovery time for different scenarios

For different scenarios recovery time was calculated dependent on the concentration of a two peak scenario. Whereas recovery

was observed within the mesocosm treatment compared to the mesocosm control up to 0.030 µg/l within 30 days, no recovery was

observed within the same scenario compared to an external isolated control above 0.012 µg/l and recovery lasted longer. This

phenomena was even stronger for single treatment scenarios.

Fig. 4: Population dynamics for a three peak application of alpha-Cypermethrin over three years

Population dynamics were calculated over three years with a three peak application every year. Two scenarios were choosen a

typical mesocosm facility with control ponds very close to the treated ponds, and an isolated treated pond without

immigration from undisturbed populations.

Fig. 5: Extinction probability for a two peak application over three yearsCalculated extinction probablity for the two scenarios shown in figure 4.

���� Populations are at higher risk without allochtone recovery and over time

���� Extinction probability increases over time at yearly applications

Larvae 1

Larvae 3

Larvae 4

Pupae

Adult

Larvae 2

without allochtonous recovery

alpha-Cypermethrin [µg/L]

1e-4 1e-3 1e-2 1e-1 1e+0

Ext

inct

ion

risk

[%]

0

20

40

60

80

100 first yearsecond yearthird year

Isolatedpopulation

Jan Jan Jan Jan

Indi

v.

0

50

100

150

200

Jan Jan Jan Jan

Indi

v.

0

50

100

150

200

Jan Jan Jan Jan

Indi

v.0

50

100

150

200

Jan Jan Jan Jan

Indi

v.

0

50

100

150

200

Jan Jan Jan Jan

Indi

v.

0

50

100

150

200

0,0015 µg/L

0,015 µg/L

0,15 µg/L

with allochtonous recovery without allochtonous recovery

Jan Jan Jan Jan

Indi

v.

0

50

100

150

200

0,0015 µg/L

0,015 µg/L

0,15 µg/L

alpha-Cypermethrin [µg/L]0.000 0.005 0.010 0.015 0.020 0.025 0.030

Rec

over

y tim

e [d

ays]

0

20

40

60

80

100

Recovery time after the second application

*

*no recovery at higher concentrations*:

�Recovery time of treated populations increases for isolated populations

�Recovery time in mesocsm facilities includes a reduced control density

0 200100 300 0 200100 300 0 200100 300

Time [d]

Abu

ndan

ce

Control 5 µg/l 3,4-DCA 30 µg/l 3,4-DCA

equa

lw

eak

stro

ng

Population of interest Competitor

0 200100 300 0 200100 300 0 200100 300

Time [d]

Abu

ndan

ce

0 200100 3000 200100 300 0 200100 3000 200100 300 0 200100 3000 200100 300

Time [d]

Abu

ndan

ce

Control 5 µg/l 3,4-DCA 30 µg/l 3,4-DCA

equa

lw

eak

stro

ng

Population of interest Competitor

with allochtonous recovery

alpha-Cypermethrin [µg/L]

1e-4 1e-3 1e-2 1e-1 1e+0

Ext

inct

ion

risk

[%]

0

20

40

60

80

100 first yearsecond yearthird year

Mesocosm population(Metapopulation)