Predicted climate change effects on fisheries habitat and...

A. Miehls

Collaborative team:

David “Bo” Bunnell Chuck Madenjian

David Warner

Paris Collingsworth

Yu-Chun Kao

Randy Claramunt

Brent Lofgren

Marjorie Perroud

Michael Murray Carlo DeMarchi

IPCC WG1 2013

A warming climate: especially at higher latitudes and over land:

Laurentian Great Lakes:

Home to more than 30 million US and Canadian citizens within its watershed.

Laurentian Great Lakes:

Contain about 20% of the world’s surface freshwater (drinking water to tens of millions).

Laurentian Great Lakes:

17,000 km of shoreline (half way around the equator). Supports numerous recreational activities.

Michigan Sea Grant, Todd Marsee

Laurentian Great Lakes:

Fisheries are a key economic driver. Anglers directly spent $2.5 billion in 2006. Multiplicative impact = $7 billion (Thayer & Loftus 2013)

Michigan Sea Grant, Todd Marsee

“Fishtown”, USA Leland, Michigan

Number that survive to the fishery

Less ice cover Less protection for eggs incubating overwinter

Climate-driven Effects on lakes

Effects on fish Effects on fish management

Ice

Michigan Sea Grant

Freeberg et al. 1990; Brown et al. 1993

Our approach: understand how climate affects fish… Our mechanistic approach:

Fish growth

Falling water levels

Less ice cover

Earlier & longer thermal stratification

Changes in fish food

Less protection

Climate-driven Effects on lakes

Effects on fish Effects on fish management

Warming temperatures Metabolism

Reduced nearshore spawning habitat

Lake productivity

Follows Jones et al. 2006

Our mechanistic approach:

Number that survive to the fishery

Overall conceptual framework:

Mechanistic approach for understanding climate-driven impacts on fisheries.

Looking backwards- Can we use historical data to detect climate signal

on biotic variables?

Looking forward- Develop localized climate predictions.

Where i) we find climate signals in historical biotic

data or ii) have previous understanding of climate impacts on biota: Forecast fish responses (i.e., growth)

Outline for today’s talk: 1. Intro Climate-driven effects on large lake ecosystems Our conceptual framework

2. Looking backward: climate-driven effects on: Algal production Prey-fish recruitment

3. Looking forward: climate-driven effects on: Water temperature, precipitation, water level, ice

cover

4. Looking forward: climate-driven effects on: Growth rates of recreationally, commercially

important fisheries.

Outline for today’s talk: 1. Intro Climate-driven effects on large lake ecosystems Our conceptual framework

2. Looking backward: climate-driven effects on: Algal production Prey-fish recruitment

3. Looking forward: climate-driven effects on: Water temperature, precipitation, water level, ice

cover

4. Looking forward: climate-driven effects on: Growth rates of recreationally, commercially

important fisheries.

David Warner and Barry Lesht

Why Worry About Algal Production?

• Lake productivity affects ecosystem services: – Availability of clean, safe drinking water – Desirability of recreational opportunities – Production of harvestable fish

Why timing matters….

Ideally, overlap in the timing….

Less ideal: algae respond differentially… and zooplankton (and fish larvae) starve

Time

Why the amount matters…

Algae

Zooplankton

Prey fish

Fish-eating fish (piscivores)

Declining nutrients in Lakes Michigan & Huron

1975 1980 1985 1990 1995 2000 2005 2010

Tota

l pho

spho

rus

load

ings

(ton

nes)

2000

3000

4000

5000

6000

7000

sprin

g to

tal p

hosp

hurs

(ug/

L)

2

3

4

5

6

7

8

EPA-GLNPO data

Lake Michigan

Research Questions 1. Has the timing and amount of algal production

changed in lakes Michigan and Huron since 1998?

2. If so, what role does climate play, relative to other variables?

Data Response – Satellite derived estimates of algae- measured as

chlorophyll-a. Note primary production was also estimated. Method validated in Lesht et al. 2013 March – May: when lake is not thermally layered. March – November: entire year without ice. May 20, 1998

Data Response – Satellite derived estimates of algae- measured as

chlorophyll-a. Note primary production was also estimated. Method validated in Lesht et al. 2013 March – May: when lake is not thermally layered. March – November: entire year without ice.

Explanatory variables: Climate ○ Air and water temperature ○ Precipitation

Nutrients ○ Spring TP concentrations in the offshore (EPA monitoring) ○ Annual total phosphorus (TP loading)

Dreissenid mussels ○ Annual density

1. Timing–does algae bloom earlier in warmer years?

Analysis

Timing: Lake Michigan 1998-2008

o Chl a (solid line)- o Spring bloom evident

from 1998~2001. o 2002 to 2008 “flatter”

time series across the year; no clear spring bloom.

o No evidence of the spring

bloom being related to water temperature between 1998-2008.

Stratification date line

1. Timing–does algae bloom earlier in warmer years?

2. Amount of algae produced- Developed regression statistical models Identified most reasonable model(s) using AICc.

Analysis

Spring: r2 =0.72 Nutrients (Spring TP, +0.77) Dreissenid mussel density (-0.31) No climate variable included in best model(s)

Results- Amount of algae produced

Annual: r2 =0.66 Dreissenid mussel density (-0.55) Water temperature (-0.47) Air temperature (+0.40)

Temperature

Alga

e

Algal growth increases with temperature

Colder years, later stratification; increases mussel access to algae.

Temperature

Alga

e

Zooplankton feeding increases with temperature.

Results- Amount of algae produced

Timing of algal “blooms”: Spring bloom is not necessarily earlier in warmer years.

How much algae is produced: Influenced by nutrients, mussels and temperature. Future research needs to tease out these interacting effects on

algal production.

Conclusions

Outline for today’s talk: 1. Intro Climate-driven effects on large lake ecosystems Our conceptual framework

2. Looking backwards: climate-driven effects on: Algal production Prey-fish recruitment

3. Looking forward: climate-driven effects on: Water temperature, precipitation, water level, ice

cover

4. Looking forward: climate-driven effects on: Growth rates of recreationally, commercially

important fisheries.

Paris D. Collingsworth, David B. Bunnell, Charles P. Madenjian, Stephen C. Riley

Transactions of the American Fisheries Society, 2014 doi:10.1080/00028487.2013.833986

Why fish recruitment (number that are produced in a given year) is important:

Leads to sustainable fish populations (& fisheries).

Recruitment is unpredictable.

The number of mature adults rarely, if ever, singularly predicts recruitment.

Other environmental predictors (including climate) are key to predicting recruitment.

Recruitment of prey fish in the Great Lakes

Prey fish support recreational fisheries (and limited commercial fisheries) with highest economic value.

Prey fish recruitment (and biomass) has been declining in several Great Lakes.

Box of Lake Michigan prey fish sampled during USGS annual

bottom trawl survey

Lake Michigan

Bloater and alewife are key prey fishes and near record-lows.

A tale of two lakes: Michigan and Huron

USGS-GLSC

A tale of two lakes: Michigan and Huron

Despite prey fish decline, salmonine fishery holds steady.

Year1998 2000 2002 2004 2006 2008 2010

Pisc

ivor

e bi

omas

s (to

nnes

)

0

2000

4000

6000

8000

10000

12000

14000

16000Chinook salmonLake trout

Lake Michigan

Tsehaye et al., Michigan State University

0

50

100

150

200

250

300

350

400

1976 1981 1986 1991 1996 2001 2006 2011

Mai

n ba

sin

biom

ass

estim

ate

(kilo

tonn

es)

Year

ROUND GOBY

TROUT PERCH

NINESPINE STICKLEBACK

DEEPWATER SCULPIN

SLIMY SCULPIN

YAO BLOATER

YOY BLOATER

YAO SMELT

YOY SMELT

YAO ALEWIFE

YOY ALEWIFE

A tale of two lakes: Michigan and Huron

Lake Huron

USGS-GLSC

Year1998 2000 2002 2004 2006 2008 2010

Pisc

ivor

e bi

omas

s (to

nnes

)

0

1000

2000

3000

4000Chinook salmonLake trout

A tale of two lakes: Michigan and Huron

Chinook salmon collapsed with alewife collapse.

Lake Huron

Brenden et al., Michigan State University

Develop statistical models to understand drivers in alewife and bloater recruitment. Traditional Ricker stock-recruit models Dynamic Linear Models (Bayesian)

Analysis

Explanatory variable Hypothesis Source

Spring-summer water temperature

(+) Facilitates faster larval growth Madenjian et al. 2005

Length of winter (-) First year overwinter mortality O’Gorman et al. 2004

Water level (+) Greater access to spawning habitat

This study

Alewife

Explanatory variable Hypothesis Source

Spring-summer water temperature

(+) Facilitates faster larval growth Madenjian et al. 2005

Length of winter (-) First year overwinter mortality O’Gorman et al. 2004

Water level (+) Greater access to spawning habitat

This study

Salmonine biomass (-) Predation on young recruits Madenjian et al. 2005

Lake productivity (+) More food for larvae O’Gorman et al. 2004

Alewife

Salmon Summer Winter Level

Sum

wi

0.0

0.2

0.4

0.6

0.8

1.0 Lake MichiganLake Huron

Alewife

Explanatory variable Hypothesis Source

Winter-spring water temperature

(+) Accelerates egg incubation and reduces predation by egg predators

Rice et al. 1987

Bloater

Explanatory variable Hypothesis Source

Winter-spring water temperature

(+) Accelerates egg incubation and reduces predation by egg predators

Rice et al. 1987

Alewife biomass (-) Predation on larval recruits Eck and Wells 1987

% female (sex ratio) (-) Balanced sex ratio increases egg fertilization rate

Brown et al. 1987; Bunnell et al. 2006

Adult condition (+) Produce eggs with more lipids to enhance larval survival

Bunnell et al. 2009

Bloater

Sex raio Condition Temp Alewife

Sum

wi

0.0

0.2

0.4

0.6

0.8

1.0Lake MichiganLake Huron

Bloater

Sex ratio

Conclusions Using historical time series, climate signals were difficult to

detect for as drivers of Lake Michigan and Huron prey fish.

Alewife- Recruitment in Lake Michigan explained by salmon predation. Lake Huron- weak explanatory power.

Bloater- Recruitment in Lakes Michigan and Huron were explained by common factors (sex ratio, alewife). No evidence of a climate signal.

Limits our ability to forecast recruitment of alewife or bloater based on future climate scenarios.

Outline for today’s talk: 1. Intro Climate-driven effects on large lake ecosystems Our conceptual framework

2. Looking backward: climate-driven effects on: Algal production Prey-fish recruitment

3. Looking forward: climate-driven effects on: Water temperature, precipitation, water level, ice

cover

4. Looking forward: climate-driven effects on: Growth rates of recreationally, commercially

important fisheries.

Brent Lofgren

Simplistic chain of causality in many studies that evaluate climate impacts

Greenhouse gases

Ecological impact (e.g., lake

temperature)

Air temperature, perhaps precipitation

Temperature Pressure

Clouds

Humidity

Wind

Sfc temperature Sfc roughness Sfc moisture

Sensible heat

Upward longwave Friction

Downward longwave

Sfc albedo

Latent heat/evap

Upward solar

Downward solar Upward longwave

Precip

More accurate chain of causality

Based on Regional Atmospheric Modeling System (RAMS) 40 km grid, with domain reaching into the Great Plains,

Hudson Bay, Atlantic Ocean, and near the Gulf of Mexico 24 vertical levels up to 17 km Array (on the same horizontal grid) of 1-D lake column

models based on Hostetler and Bartlein (1990) LEAF3 formulation of land surface Mellor-Yamada level 2.5 atmospheric boundary layer Driven by:

Canadian CRCM3 GCM SRES A2 scenario GHG concentrations

Coupled Hydrosphere-Atmosphere Research Model (CHARM)

Changes in air temperature (°C): 2057 (2043-2070) minus 1985 (1978-1993)

Winter (Dec-Feb) Summer (Jun-Aug) 4.0

3.0

2.0

1.0

4.0

3.0

2.0

1.0

Changes in Lake Michigan water temperature: 1964-1977 vs. 2058-2070

-200

-150

-100

-50

00 5 10 15 20

Feb

-200

-150

-100

-50

00 5 10 15 20

Apr

-200

-150

-100

-50

00 5 10 15 20

Jun

-200

-150

-100

-50

00 5 10 15 20

Aug

-200

-150

-100

-50

00 5 10 15 20

Oct

-200

-150

-100

-50

00 5 10 15 20

Dec

1971 2062

Dep

th (m

)

Changes in precipitation (mm/day): 2057 (2043-2070) minus 1985 (1978-1993)

Winter (Dec-Feb) Summer (Jun-Aug)

0.5

0

0.25

-0.25

-0.5

0.5

0

0.25

-0.25

-0.5

Net water balance lake water levels: 2057 (2043-2070) minus 1985 (1978-1993)

0.03

0

-0.02

0.02

0.01

-0.01

Inches/day

Changes in Lake Michigan ice thickness (m):

February (1979-1993)

1.0

0

0.5

0.25

0.75

1.0

0

0.5

0.25

0.75

February (2043-2070)

Preliminary forecasts from CHARM

Water temperatures (warmer): Summer-Fall: 2-3 °C increase throughout water column. February: deepest waters can be warmer than 4 °C.

Precipitation (wetter):

Winter: even snowier within lake effect regions. Summer: patchy, but generally more rain.

Ice cover (less):

North to south gradient.

Outline for today’s talk: 1. Intro Climate-driven effects on large lake ecosystems Our conceptual framework

2. Looking backward: climate-driven effects on: Algal production Prey-fish recruitment

3. Looking forward: climate-driven effects on: Water temperature, precipitation, water level, ice

cover

4. Looking forward: climate-driven effects on: Growth rates of recreationally, commercially

important fisheries.

Yu-Chun Kao, Chuck Madenjian

How does a warming climate affect fish growth?

Temperature

Consumption

Metabolism Waste loss

Fish growth

Prey quantity and quality

Food web

Fish physiology

Physiological optimal temperature for lake whitefish = 12 °C

Fish are ectothermic (cold-blooded)

Where they live balances competing factors: physiological optima, habitat quality, food availability, avoiding predators.

How can fish respond to warmer temperatures?

Gorskyet al. 2012

Clear Lake, Maine

They find their ideal temperatures within the lake…

13 °C

Today’s results

Forecast growth of yellow perch (cool-water fish) and lake trout (cold-water fish) in 2043 to 2070.

Simulate growth and consumption within lakes Michigan and Huron.

Assume fish occupy physiological optimum temperature, if available.

Assume different prey densities (high, baseline, low).

Yellow perch Lake trout

Effectively, fish have more days when optimal temperature is available in future….

How did we predict future fish growth?

Used the Wisconsin bioenergetics models, a mass-balance approach to account for energy

Realized consumption (PCmax) =

Metabolic cost + Waste loss + Growth

Cmax PCmax = M + W + G

Ye Yellow perch Lake trout

High prey availability

Limited prey availability

Baseline

Lake trout forecasted to grow better (relative to baseline) than yellow perch in both prey scenarios.

o Despite the number of days with “optimal temperature” increasing for yellow perch more than lake trout.

o In the future, yellow perch predicted to have higher energetic costs (metabolism, waste).

o Lake trout prey (alewife, rainbow smelt) have more calories in the fall.

Unexpected results:

Have similar predictions of future growth for:

Lake whitefish Rainbow trout Chinook salmon

Conclusion

Unless food resources increase with temperature, yellow perch and lake trout growth will likely decrease in a warming climate.

Overall take-home messages 1. Climate change will affect key fish habitat

variables (water temperature, ice cover, water level).

2. We recommend a mechanistic approach for translating those effects to fish responses.

3. Those mechanisms should include factors beyond climate. Our results revealed non-climate factors had an even greater impact: Mussels, nutrients affected algae Salmon affected alewife recruitment Prey densities affected fish growth

Overall take-home messages 4. This does not mean that forecasted climate

change effects should be ignored. Rather, these changes should be considered within a broader context. For example, use models that include the entire food-web to best understand how climate directly and indirectly influences fisheries.

Acknowledgements

USGS Patty Armenio, Roger Bergstedt , Jeff Schaeffer US FWS Greg Jacobs University of Michigan Michael Wiley, James Breck, James Diana Michigan State University Amber Peters, Iyob Tsehaye Michigan Department of Natural Resources Dave Caroffino, Ji He Ontario Ministry of Natural Resources Lloyd Mohr

Funding:

National Climate Change Wildlife Science Center

Questions? David “Bo” Bunnell dbunnell@usgs.gov

David Warner dmwarner@usgs.gov

Paris Collingsworth Collingsworth.Paris@epa.gov

Brent Lofgren brent.lofgren@noaa.gov

Yu-Chun Kao kyuchun@umich.edu

Chuck Madenjian cmadenjian@usgs.gov

Lars Jensen

Climate Warming Influence on Large Lakes

Consistent physical changes have been observed among large, deep lakes Increased water temperature, especially in surface

waters. Decrease in ice cover, lengthening of stratified period.

Fewer consistent chemical/biological observations

Lake Baikal, Siberia. Over last century:

Moore et al. 2009 BioScience

Ice-free season ~ 18 days longer. Annual surface waters warmed ~2 °C.

Lake Superior, United States & Canada

o Largest, o deepest, o northernmost Laurentian Great Lake

Star

t of

stra

tific

atio

n M

ean

sum

mer

w

ater

tem

pera

ture

% ice cover, 1979-2006

Lake Superior; Austin and Colman (2007)

Less ice = Earlier stratification & warmer summer temperatures.

% ice cover, 1979-2006

Cline et al. 2013

Changes in Lake Superior fish habitat: 1979-2006

Walleye, Chinook salmon, lean lake trout each have enjoyed more water of their preferred temperature.

Siscowet lake trout has found less preferred water.

Pothoven and Fahnenstiel 2013

J. Allen, USGS

Early 2000s: proliferation of quagga mussels in Lake Michigan

Declining nutrients in Lakes Michigan & Huron