the role of fish as sources and vectors of bacteria and...

The Role of Fish as Sources and Vectors of Bacteria and Influence of Bat Colonies on

Indicator Bacteria Levels

George GuillenPresented at Texas Watershed

Coordinator Roundtable – Bacteria Dynamics , Assessment Methods,

and BMPsJuly 27, 2011

Funded by: Harris County Public Infrastructure Department (HCPID)

Acknowledgments

Dr. Robin Brinkmeyer – Bat colony studies Jenny Wrast – Fish and Bat study – field

coordinator

Sources of E. coli within urban streams & bayous

Contaminated runoff & stormwater Malfunctioning wastewater collection

systems Improperly functioning wastewater plants Livestock & domesticated animalsWildlife and Waterfowl Reservoirs of E. Coli include:

– Algae & periphyton mats– Soils & sediments

But what about fish????

But what about bats????

Outline

Fish Study Bat Study

Outline – Fish Study

Background Field Studies Fish Aquarium Studies Conclusions Future work

a. Rainbow trout (carnivore);b. Catfish (omnivore emphasizing animal sources food);c. Carp (omnivore, emphasizing plant sources of food);d. Milkfish (microphagous planktovore).

E. coli isolate frequency

Omnivores 87.2%

Herbivores 70.0%

Carnivores 57.3%

From Gopee et al. 2000

Mammals 56%Birds 23%Crocodile 33%Turtles 4%Snakes 2%Lizards 10%Frogs 12%Fish 10%

From Gordon & Cowling, 2003

Australian species

Based on Terrestrial Vertebrates Only

POTENTIAL MECHANISMS OF E. COLI PRODUCT & TRANSPORT

Fish apparently harbor large microbial communities in their gut 107 – 109 cell/g digesta (Review: Stevens and Hume 1995).

Davis et al. 1995 documented fecal coliforms and isolated E. coli from catfish pond effluents

Fish species can harbor E. coli (Guzman et al. 2004)

Clark et al 2007, documented the presence of E. coli in some pelagic & demersal fishes

Origin of bacteria was Canadian geese & human sewage

POTENTIAL MECHANISMS OF E. COLI PRODUCT & TRANSPORT

Hansen et al. (2008) Pelagic & Benthic Fish:– Percentage of benthic fish harboring fecal coliforms

(FC) not sig. diff. than Pelagic fish– Of those that did harbor FC they found E. coli in

benthic fish at a rate 10x that of pelagic fish (e.g. 42% vs. 4%)

– Sources were identified for 65% of E. coli isolated Fecal coliforms found in every species examined,

but not every fish Not dominated by a single strain

– Concluded fish may acquire microorganisms while feeding

Objectives 1) Determine whether wild caught fish from

Harris County waterways, representing various species and trophic groups produce feces with detectable levels of indicator bacteria, E. coli

2) Determining whether farmed raised fish retained in aquaria transmit E. coli bacteria to ambient water and/or contain feces with E. coli.

3) Using literature and field data estimate potential “loading” ranges of E. coli from fish

Horsepen Bayou

White Oak Bayou

GreensBayou

Sampling Sites

Target SpeciesTrophic Group Candidate SpeciesHerbivore Armored Catfish

Grass Carp

Striped mullet

Benthic omnivore Channel catfish

Carp

Smallmouth buffalo

Insectivore Bluegill

Rio Grande Cichlid

Redear Sunfish

Longear Sunfish

Omnivore Gizzard or Threadfin shad

Gulf Menhaden

Tilapia

Piscivore Largemouth Bass

Spotted Bass

Green SunfishSpotted Gar

Benthic predator Blue catfish

Field study

Fish Collection– Tote barge Electroshocker– Euthanized in MS-222

Field study

Fecal Material Extraction– Fish measured & weighed– Large intestine removed– Fecal material extracted into pre-weighed

Bacti-bottles with 100 ml of sterile water

Fecal Matter Extraction

Laboratory Processing:

IDEXX Method

Field Study Data Analysis

E. coli – water (#/100ml), sediment (#/g), fish feces (#/g feces)

Compare sites, trophic groups (species), seasons

95% CI graphical comparison Kruskal-Wallis Multiple Comparisons with a

MINITAB Macro -Dunn’s Test - test median levels

KW AOV & Dunns Multiple Comparisons Chart E.coli in water #/100 ml

Greens

@ 4

5

Whit

e Oak

@ T

idwell

Horsep

en @

Pine

loch

2500

2000

1500

1000

500

0

Group

E.

coli

Ex

ten

de

d C

ou

nt

White Oak @ Tidwell

Horsepen @ Pineloch

Greens @ 45

Greens @ 45

White Oak @ Tidwell

Z0-ZNormal (0,1) Distribution

Boxplots with Sign Confidence IntervalsDesired C onfidence: 80.529

Family A lpha: 0.2Bonferroni Indiv idual A lpha: 0.067

Pairwise ComparisonsC omparisons: 3

|Bonferroni Z-v alue|: 1.834

NO Significant group difference in E. coli in water, but generally higher in urban

Significant group differences

KW AOV & Dunns Multiple Comparisons Chart E.coli in sediment #/g

Whit

e Oak

@ T

idwell

Horsep

en @

Pine

loch

Greens

@ 4

5

2000

1500

1000

500

0

Group

E.

coli

Ex

t C

ou

nt

pe

r g

of

sed

.

Horsepen @ Pineloch

Greens @ 45

White Oak @ Tidwell

White Oak @ Tidwell

Horsepen @ Pineloch






Significant group differences between sites in fish bacteria levels

KW AOV and Dunns Multiple Comparisons Chart

White Oak @ TidwellGreens @ 45Horsepen @ Pineloch

1400000

1200000

1000000

800000

600000

400000

200000

0

Group

E.

coli

pe

r g

ram

fis

h f

ece

s

Greens @ 45

Horsepen @ Pineloch

White Oak @ Tidwell

White Oak @ Tidwell

Greens @ 45






30.027.525.022.520.017.515.0

2500

2000

1500

1000

500

0

Water Temp (C)

E. c

oli (

#/1

00 m

l)

Greens @ 45Horsepen @ PinelochWhite Oak @ Tidwell

Sample Location

E. coli Count vs Water Temp (C)

Generally higher levels of bacteria at higher temperatures

Generally higher levels of bacteria at higher temperatures

30.027.525.022.520.017.515.0

2000

1500

1000

500

0

Water Temp (C)

E. c

oli E

xt C

ount

per

g o

f se

d.

Greens @ 45Horsepen @ PinelochWhite Oak @ Tidwell

Sample Location

E. coli per g of sed. vs Water Temp (C) (not data for Fall)

Significant differences seasonally fish fecal bacteria levels


SummerFallSpring

1400000

1200000

1000000

800000

600000

400000

200000

0

Group

E.

coli

pe

r g

ram

fis

h f

ece

s

Fall

Spring

Summer

Summer

Fall

Z0-ZNormal (0 ,1) Distr ibution





Stron

gylur

a mari

na

Pteryg

oplich

thys s

pp.

Paral

ichthy

s leth

ostig

ma

Oreoc

hrom

is sp

p.

Mugil c

epha

lus

Microp

terus

salm

oides

Lepo

mis micr

oloph

us

Lepo

mis meg

alotis

Lepo

mis mac

rochir

us

Lepo

mis gu

losus

Lepo

mis cya

nellu

s

Lepis

osteu

s ocu

latus

Ictalu

rus p

uncta

tus

Doro

soma c

eped

ianum

Cypr

inus c

arpio

Cten

opha

rynyo

don i

del la

Cichla

soma c

yano

gutta

tum

Atrac

tosteu

s spa

tula

150000

100000

50000

0

-50000

-100000

Species (Scientific name)

E. c

oli p

er g

fis

h fe

ces

95% Confidence Interval for Mean

Highest levels generally seen in L. cyanellus and L. gulosus - predators

PiscivoreOmnivoreInvertebrate FeederHerbivoreBenthic Omnivore

80000

70000

60000

50000

40000

30000

20000

10000

0

-10000

Trophic

E. c

oli p

er g

ram

fis

h fe

ces

95% Confidence Interval for Mean

No statistically significant trophic group differences


Omniv

ore

Benthic

Om

nivore

Herbiv

ore

Inve

rtebra

te F

eede

r

Pisci

v ore

1400000

1200000

1000000

800000

600000

400000

200000

0

Group

E.

coli

pe

r g

ram

fis

h f

ece

s

Benthic O mniv ore

Herbiv ore

Inv ertebrate Feeder

P isciv ore

O mniv ore

O mniv ore

Benthic O mniv ore

O mniv ore

Benthic O mniv ore

Herbiv ore

O mniv ore

Benthic O mniv ore

Herbiv ore

Inv ertebrate Feeder

Z0-ZNormal (0 ,1) Distr ibution





Estimated Loading

Few fecal production studies Varies with diet, temperature, trophic level Few true “density” studies of fish in

natural streams (requires mark recapture, multi-pass surveys or kill out – rotenone to estimate true density and biomass)

Estimated E. coli Loading from Fish

Used limited wild marine fish data in lab –fecal production rates, based on weight

Literature values of density of fish in rivers Our estimates of E. coli / gram of feces Biomass or # X fecal prod/fish X

Bacteria/feces = Daily rate of E. coliproduction

Estimated Loading of E. coli from fish

Based on our analyses, median loading estimates ranged between 935,113 and 29,703,591 E. coli MPN per hectare.

Arithmetic mean values ranged between 9,494,191 and 301,580,204 E. coli MPN per hectare.

Absolute highest estimate = 29.8 Billion Most methods yielded median and mean

E. coli levels ranging in the millions.

Estimated Loads

ValueProduction of Fish E.coli/day

Volume (cubic ml) # E.coli/ml # E. coli/100 ml

High Mean 301,580,204 10,000,000,000 0.030158 3.01580204Maximum 29,832,819,372 10,000,000,000 2.9832819 298.3281937

Aquarium Study

1st round used Bluegill (Lepomis macrochirus) 2nd round used channel catfish (Ictalurus

punctatus) 15 15 gal. aquaria

– 5 replicates Control Low density (1 fish/tank) High Density (3 fish/ tank)

Basic WQ parameters monitored daily– Additional parameters measured on bacteria sampling

days

Aquarium Study

Bacteria sampled– Pre-stocking– 1d post-stocking– 3d post-stocking– 7d post-stocking– 14d post-stocking

Blue Gill Aquarium Water E. coli (MPN) Values

(C. Catfish: 0 - E.coli)

Day# Fish

147310310310310310310

7

6

5

4

3

2

1

0

E. c

oli (

MPN

/ 1

00 m

l )

Sacrificed Fish E. coli from extracted feces

CatifshBluegill

10000000

1000000

100000

10000

1000

100

10

1

Fish

E. c

oli (

MPN

/ g

fec

es )

Conclusions Fish in the stream study seem to at least be

transporters of E.coli Supported by aquarium study using farmed fish

– Showed no increase in E.coli levels due to stocking density or over time

– However, fish fecal matter tested after 14d aquarium study showed low levels of E. coli

Diet does not seem to be very important in amount of E. coli in fish fecal matter

Wild stocks pick up bacteria from sediment, food web and transfer to other areas as they migrate

Conclusions/Future Implications

Bacteria in fish maybe indicators of bacteriological pollution in the waterbodies they inhabit

Fish seem to absorb the bacteria from their food as well as their environment

Therefore fish maybe significant transport vectors not captured by current modeling efforts

Conclusions

Seasonal and site differences contributed to probability of fish containing high indicator bacteria levels in feces

Sites with higher sediment and water bacteria generally contained fish with higher indicator bacteria

Fish in summer months generally yielded higher indicator bacteria levels.

Future Research Needs Studies (genetic and other) needed for

identifying and confirming original sources (pass through or specific to fish)

Investigate influence environmental gradient influence on bacteria levels: (pristine watershed urban) needed to evaluate trends in fish fecal E. coli vs. sediment and water levels

More lab studies needed to evaluate gut residence time vs. levels of feces bacteria levels.

(Also supplemental commercial aquarium studies). Fish Density Estimate = e.g. side scan high

definition acoustic studies

Mexican or Brazilian Free Tailed Bat (Tardarida brasiliensis)

Bat Field Study

Outline

Literature Review Historical Monitoring Data Review Field Study – upstream and downstream

sampling of colony Potential Loading Genetic sampling of fecal material and

water

Background Mexican Free Tailed bats are also considered

“guano” bats due to the large quantities of feces produced

Typically emerge from their roost areas shortly before dark to begin their nightly foraging. They can forage up to 50 miles and normally return back to their roost before sunrise

The abundance of the bats follows an annual pattern where highest numbers occurring during spring through fall. Most of these bats migrate to Mexico (early December through February) to overwinter

Background Recent estimates of the bats at the Commerce

Street Bridge Austin range up to 1.5 million per year (Tuttle 2003).

Waugh Street colony Houston size is estimated to range between 250,000 – 300,000 bats

Mexican Free Tailed bats are also considered “guano” bats due to the large quantities of feces produced

Water Quality Impacts

Unpublished data collected by the City of Austin during 1999 failed to document any impacts on Town Lake and the Colorado River

The State of California found that elevated fecal coliform levels may be caused in part by high densities of bats and other organisms that inhabit an enclosed tunnel section of San Luis Obispo creek

Site Location Seasons

Site 1. Upstream (200 ft) * Summer (Aug and Sept 08), Fall (Nov) 08, Winter (Feb) 09, Spring (April) 09

5 events total (ambient monitoring)

Site 2. Immediately below bridge (bat colony at .

Same (*) + fecal pellet sampling (April & Nov 09)

Site 3A Downstream 100 ft Source ID (water): (Aug 08, April and Nov 09)

Site 3 Downstream 200 ft. Same (*) + Source ID (water): (Aug 08, April and Nov 09)

Site 4. Downstream 1000 ft Same (*) + Source ID (water): (Aug 08, April and Nov 09)

New Bat Colonies Discovered!!

Estimates of E. coli loading into Buffalo Bayou from Waugh Street bat colony based on E. coli counts taken from fecal material suspensions and literature fecal production rates (Sgro and Wilkins 2003).

ID numbermL of dilution

waterNumber of

pelletsfeces

weight (g) <=>Extended

count E. coli1/2 extended count E. coli <=>

E. coli per gram

Using 1/2 E. coli for < DL

1 80 2 0.0174 < 1.25 0.63 < 71.8 35.92 80 2 0.0132 = 3.75 3.75 = 284.1 284.13 80 2 0.0151 < 1.25 0.63 < 82.8 41.44 80 2 0.0127 < 1.25 0.63 < 98.4 49.25 50 10 0.0384 = 275.20 275.20 = 7,166.7 7,166.76 50 10 0.0475 > 4,839.20 4,839.20 > 101,877.9 101,877.97 50 10 0.0540 = 251.80 251.80 = 4,663.0 4,663.08 50 10 0.0402 < 2.00 1.00 < 49.8 24.99 50 18 0.0773 = 2.00 2.00 = 25.9 25.9

10 40 18 0.0863 > 6,049.00 6,049.00 > 70,092.7 70,092.711 40 18 0.0539 = 5.00 5.00 = 92.8 92.812 40 18 0.0851 > 6,049.00 6,049.00 > 71,081.1 71,081.113 100 Blank 0.0027 < 1.00 0.50 < 370.4 185.2

Average 21,298.9 21,286.3

Daily loading rate in grams (g) of feces based on 300,000 bats 29,500 29,500Daily loading rate in grams (g) of feces based on 250,000 bats 24,600 24,600

Daily loading # E. coli MPN based on 300,000 bats 628,317,626 627,945,438Daily loading # E. coli MPN based on 250,000 bats 523,953,003 523,642,636

Conclusions

Based on our data and literature and our observations of the additional large bat ‘nestings’ in bridges upstream of Waugh Drive (in particular South Mason Road) and downtown at the intersections of Smith and Franklin streets that create a large bridge structure and the presence of highly visible layers of guano, bats are definitely contributing to the E. coli loads in Buffalo Bayou.

Conclusions

Since 2009, we have observed bats at the Waugh Drive Bridge as well as downtown at the Smith-Franklin Street Bridge complex year round. The absolute densities and activity of these overwintering bats would determine their overall contribution to E. coli loads in Buffalo Bayou.

Recommendations

Represent a significant but poorly understood source of nutrients and E. colibacteria in our urban watersheds.

Additional research is critically needed to characterize the interaction of bat population dynamics, seasonal and dielmovement, fecal and associated nutrient and bacteria loading, and resulting instream indicator bacteria concentrations.

THANKS to everyone who helped: Brian Muery, Julie Sandefur, Danielle Crossen, Mike Franks, Dianna Ramirez + TAMUG students

the role of fish as sources and vectors of bacteria and...

Documents