Content uploaded by Troy Zorn
Author content
All content in this area was uploaded by Troy Zorn on Feb 06, 2018
Content may be subject to copyright.
Projected impacts of climate change on stream
salmonids with implications for resilience-based
management
Andrew K. Carlson
1
, William W. Taylor
1
, Kelsey M. Schlee
1
, Troy G. Zorn
2
, Dana M. Infante
1
1
Department of Fisheries and Wildlife, Center for Systems Integration and Sustainability, Michigan State University, East Lansing, MI 48824, USA
2
Michigan Department of Natural Resources, Marquette Fisheries Research Station, Marquette, MI 49855, USA
Accepted for publication November 16, 2015
Abstract –The sustainability of freshwater fisheries is increasingly affected by climate warming, habitat alteration,
invasive species and other drivers of global change. The State of Michigan, USA, contains ecologically,
socioeconomically valuable coldwater stream salmonid fisheries that are highly susceptible to these ecological
alterations. Thus, there is a need for future management approaches that promote resilient stream ecosystems that
absorb change amidst disturbances. Fisheries professionals in Michigan are responding to this need by designing a
comprehensive management plan for stream brook charr (Salvelinus fontinalis), brown trout (Salmo trutta) and
rainbow trout (Oncorhynchus mykiss) populations. To assist in developing such a plan, we used stream-specific
regression models to forecast thermal habitat suitability in streams throughout Michigan from 2006 to 2056 under
different predicted climate change scenarios. As baseflow index (i.e., relative groundwater input) increased, stream
thermal sensitivity (i.e., relative susceptibility to temperature change) decreased. Thus, the magnitude of temperature
warming and frequency of thermal habitat degradation were lowest in streams with the highest baseflow indices.
Thermal habitats were most suitable in rainbow trout streams as this species has a wider temperature range for
growth (12.0–22.5 °C) compared to brook charr (11.0–20.5 °C) and brown trout (12.0–20.0 °C). Our study
promotes resilience-based salmonid management by providing a methodology for stream temperature and thermal
habitat suitability prediction. Fisheries professionals can use this approach to protect coldwater habitats and drivers
of stream cooling and ultimately conserve resilient salmonid populations amidst global change.
Key words: brook charr; brown trout; rainbow trout; Michigan; temperature; resilience
Introduction
As climate warming, habitat fragmentation, invasive
species and other drivers of global change alter aqua-
tic ecosystems throughout the world, managing fish-
eries for resilience has become an important
conservation framework (Hansen et al. 2015). Resili-
ence is the capacity of a system to absorb distur-
bances and retain its structure and function (Holling
1973). Managing aquatic ecosystems for resilience is
particularly important when they contain species that
are sensitive to ecological stressors. For instance,
with relatively low thermal tolerance thresholds
(Raleigh 1982a,b; Raleigh et al. 1986), salmonid
fishes may experience reductions in growth and sur-
vival due to temperature elevation caused by climate
change. Thus, managing these species and their
ecosystems for thermal resilience is an important
task. Collaboration among scientists, managers, poli-
cymakers and public stakeholders will be important
for developing management approaches commensu-
rate with the wide geographical distribution and high
socioeconomic value of salmonid populations (Isaak
et al. 2015; Snyder et al. 2015).
Temperature exerts a fundamental physiological
influence on fish metabolism, which regulates
growth, survival and reproduction of individuals
(Dodds & Whiles 2010) and ultimately the dynamics
Correspondence: A. K. Carlson, Department of Fisheries and Wildlife, Center for Systems Integration and Sustainability, Michigan State University, East Lansing,
MI, 48824 USA. E-mail: carls422@msu.edu
190 doi: 10.1111/eff.12267
Ecology of Freshwater Fish 2017: 26: 190–204 Ó2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
ECOLOGY OF
FRESHWATER FISH
of populations and communities. For example, water
temperature is an important factor determining fish
distribution and assemblage composition (Magnuson
et al. 1979). Temperatures above species-specific
thermal maxima cause mortality; temperatures below
maxima alter fish growth, reproduction, abundance
and population size structure (Magnuson et al. 1997).
In addition, thermal warming can indirectly decrease
fish growth and survival by degrading water quality
(e.g., reduced dissolved oxygen; Ficklin et al. 2013).
Climate change is predicted to increase stream tem-
peratures and thereby alter thermal habitat suitability
(Lyons et al. 2010) and fish community composition
(Isaak et al. 2012). Moreover, dams, culverts, and
other anthropogenic barriers increase water tempera-
tures and decrease population and habitat connectiv-
ity (Lessard 2000; Hayes et al. 2006). Regardless of
mechanism, stream salmonids are sensitive to thermal
warming because they are adapted to cold and cool
water environments (Raleigh 1982a,b; Raleigh et al.
1986; Wehrly et al. 2007). Climate-driven increases
in water temperature may reduce salmonid growth,
reproduction, and survival in streams currently near
thermal optima. Thus, it is imperative that fisheries
professionals develop management strategies to pro-
mote thermal resilience and thereby conserve salmo-
nid populations in a warming climate.
The State of Michigan, USA, has ecologically and
socioeconomically valuable populations of brook charr
(Salvelinus fontinalis), brown trout (Salmo trutta), and
rainbow trout (Oncorhynchus mykiss) distributed
throughout 31,000 km of streams (Godby et al. 2007;
Tyler & Rutherford 2007). Projected air temperature
warming resulting from climate change is predicted to
increase stream temperatures in the Midwestern Uni-
ted States by 0.8–4.0 °C (Pilgrim et al. 1998; Lyons
et al. 2010). If temperatures exceed thermal optima for
these species, growth, reproduction, and/or survival
will decline, particularly during the warmest month of
the year (i.e., July; Zorn et al. 2011). Thus, Michigan
is an ideal study area for investigating how thermal
warming will impact coldwater streams and salmonid
population dynamics. This information will enable
fisheries professionals to develop resilience-based
management programmes: collaborative efforts among
scientists, managers, policymakers and public stake-
holders to maintain stream ecosystem structure and
function amidst global change.
The goal of this study was to evaluate the effects of
projected air temperature warming on salmonid ther-
mal habitat suitability in coldwater streams in Michi-
gan to facilitate the development of a resilience-based
management programme. Our first objective was to
measure the accuracy of stream-specific air–water tem-
perature regression models by backcasting stream tem-
peratures in 2006 and 2012, years with pre-existing air
and stream temperature metrics. Our second objective
was to forecast stream temperatures in 2036 and 2056
and project thermal habitat suitability for brook charr,
brown trout and rainbow trout growth and survival.
We predicted that water temperatures would increase
overall from 2006 to 2056 but unevenly among
streams and time periods (e.g., 2012–2036, 2036–
2056) due to system-specific patterns of thermal warm-
ing and temporal variability in projected carbon diox-
ide (CO
2
) emissions (Arblaster et al. 2014). We
expected that thermal habitat impairment would occur
more frequently in streams with extensive surface run-
off than in groundwater-dominated, thermally buffered
systems (Sear et al. 1999; Krider et al. 2013).
Methods
Study area
Our study encompassed 30 coldwater streams
throughout the State of Michigan, USA (Fig. 1).
Streams were chosen in three Michigan Department
of Natural Resources (MDNR) management regions
[i.e., Upper Peninsula (UP), northern Lower Penin-
sula (NLP) and southern Lower Peninsula (SLP);
Table 1] to conduct the study at a statewide scale rel-
evant for salmonid management that also spanned a
latitudinal thermal gradient in which temperatures
increased from north (47.03°N) to south (42.64°N).
In addition, streams were selected to encompass a
hydrological gradient from surface runoff to ground-
water dominance by evaluating their relative base
flow, the component of streamflow attributable to
groundwater. We used a United States Geological
Survey report (Neff et al. 2005) to obtain each
stream’s base flow index (BFI), the mean rate of base
flow divided by the corresponding mean rate of total
streamflow. Baseflow index ranges from zero to one
with increasing groundwater input (Wahl & Wahl
1988); streams in our study ranged from 0.44 to 0.73
(Table 1). All BFI calculations were made using a
digital filter hydrograph separation technique (Arnold
& Allen 1999; Kelleher et al. 2012) whereby daily
streamflow records were partitioned into groundwater
and surface runoff components. Moreover, all streams
were important from a management standpoint as
they supported recreational fisheries for brook charr,
brown trout, rainbow trout or a combination of these
coldwater species. Only streams that had necessary
historical data for development of stream-specific
temperature regression models (i.e., field-measured
air and water temperatures) were selected. We devel-
oped a list of streams meeting these criteria (i.e., lati-
tudinal gradient, hydrological gradient, recreational
importance and historical data) using information
from the MDNR ‘Better Fishing Waters’webpage
191
Resilience-based salmonid management
(MDNR 2015) and through personal communication
with MDNR Fisheries Division personnel (Tracy
Kolb, Todd Wills) and employees of the Michigan
Council of Trout Unlimited (Trout Unlimited 2015).
A minimum of five streams per region were selected
for each species to ensure regional replication. One
exception was the SLP, where brook charr and rain-
bow trout are not widely distributed and only four
streams contained one or both of these species. In
total, 16, 18 and 11 streams contained brook charr,
brown trout and rainbow trout, respectively. Eleven
streams supported more than one species, and four
supported all species (Table 1).
Stream-specific regression models
Historical air and water temperatures for each stream
were used to create stream-specific regression mod-
els. Daily air temperatures collected in summer
months (i.e., June, July and August) from 1990 to
2010 were compiled using the United States Histori-
cal Climate Network online interface (http://
cdiac.ornl.gov/). Air temperatures were measured at
the gauging station closest to each stream’s headwa-
ters, where MDNR water temperature gauges
recorded daily temperatures in summer months from
1990–2010. The most upstream gauge on each stream
was selected because temperatures are typically cool-
est and most optimal for salmonids in headwater
reaches. As such, we focused on these areas because
if their thermal habitat is degraded, downstream habi-
tat and salmonid populations will also be impaired.
Hydrological unit codes (HUCs) for each stream’s
sub-basin (HUC8) and subwatershed (HUC 12) were
identified using the National Hydrography Dataset
Plus Version 1 (NHDPlusV1) and the Watershed
Boundary Dataset (USEPA 2005). The North Ameri-
can Anthropogenic Barrier Dataset was used to locate
and omit gauges directly below dams, which elevate
temperatures compared to upstream reaches (Lessard
2000). For streams without MDNR gauges, a substi-
tute gauge on the nearest stream within the same sub-
watershed was used. Stream-specific regression
models were developed by pairing mean summer air
and water temperatures from recent years (i.e., 2002–
2010) in Microsoft Excel (Table 2). Air temperature
coefficients represented indices of stream thermal
sensitivity (i.e., relative susceptibility to temperature
change; Kelleher et al. 2012) and were coupled with
air temperature projections (see below) to predict
future water temperatures. In addition, historical
warming in each stream was evaluated by comparing
air temperatures in 1976 and 2006 and multiplying
this temperature change by each stream’s air tempera-
ture coefficient to determine the magnitude of stream
warming.
Fig. 1. Map of brook charr, brown trout
and rainbow trout streams used for air–
stream temperature modelling in Michigan,
USA. Streams and corresponding
identification numbers are listed in
Table 1. Triangles denote locations of
MDNR stream temperature gauges.
192
Carlson et al.
Air temperature projections
Three coupled climate models (CCMs) were used to
backcast mean June, July and August air tempera-
tures in 2006 and 2012 and forecast mean tempera-
tures in the same months in 2036 and 2056 for each
sub-basin: the Third Generation Coupled Global Cli-
mate Model (CGCM3, Canadian Centre for Climate
Modelling and Analysis), the CM2 Global Coupled
Climate Model (CM2, Geophysical Fluid Dynamics
Laboratory at the National Oceanic and Atmospheric
Administration) and the Hadley Centre Coupled
Model version 3 (HadCM3, Met Office, United King-
dom’s National Weather Service). All CCMs were
based on the World Climate Research Programme
(WCRP) Coupled Model Intercomparison Project
phase 3 (CMIP3) multimodel data set. Spatial down-
scaling was performed using the Bias-correction spa-
tial disaggregation (BCSD) approach to adjust the
resolution of the climate model (~200 9200 km) to
a scale germane for Michigan streams (12 912 km;
Maurer et al. 2007). The United States Forest
Service’s (USFS) Eastern Forest Environmental
Threat Assessment Center (EFETAC) in North
Carolina supplied mean June, July and August air
temperatures for Michigan sub-basins containing
coldwater salmonid streams. Projections were made
based on the CGCM3, CM2, and HadCM3 models
using area-weighted means for all years assuming the
Special Report of Emission Scenarios A2 and B1 cli-
mate forcing scenarios. The A2 scenario describes a
future world with rapid economic growth and effi-
cient energy technologies and predicts atmospheric
CO
2
concentrations to be 820 ppm in 2100. In con-
trast, the B1 scenario projects a convergent world
with a service and information economy and reduced
material consumption and predicts atmospheric CO
2
concentrations 550 ppm in 2100. Combining scenar-
ios was informative as they represent upper and
lower emission thresholds for stream temperature
prediction.
Stream temperature and thermal habitat suitability
projections
Stream temperatures were backcasted in 2006 and
2012 and forecasted in 2036 and 2056 by inputting
CCM air temperature predictions into stream-specific
Table 1. Descriptive information about 30 salmonid streams used for temperature modelling in Michigan, USA. Region refers to Michigan location [i.e., Upper
Peninsula (UP), northern Lower Peninsula (NLP) and southern Lower Peninsula (SLP)]. Sub-basin denotes the National Hydrography Dataset sub-basin (i.e.,
8-digit hydrological unit code). Map number refers to stream identifiers in Fig. 1. Species refers to salmonids present in each stream [i.e., brook charr (BKC),
brown trout (BNT) and rainbow trout (RBT)]. BFI denotes baseflow index, the mean rate of base flow divided by the corresponding mean rate of total
streamflow (Neff et al. 2005).
Stream Region Subbasin Map number Species BFI
Bark River UP Cedar-Ford 18 BNT 0.60
Bear Creek SLP Kalamazoo 14 BNT 0.58
Black River NLP Black 1 BKT 0.63
Boardman River NLP Boardman-Charlevoix 2 BKT, BNT 0.58
Bryan Creek UP Escanaba 19 BKT 0.62
Canada Creek NLP Black 3 BKT 0.63
Carp River UP Carp-Pine 20 BKT, BNT, RBT 0.64
Cedar Creek SLP Lower Grand 10 BKT, BNT 0.50
Cedar River SLP Cedar-Ford 11 BNT 0.60
Chocolay River UP Betsy-Chocolay 25 RBT 0.62
Davenport Creek UP Brevoort-Millecoquins 26 RBT 0.65
Duke Creek SLP Lower Grand 15 BKT, BNT 0.50
East Branch Fox River UP Manistique 27 BKT, BNT 0.73
Elm River UP Keweenaw Peninsula 21 RBT 0.45
Escanaba River UP Escanaba 22 BKT, BNT 0.44
Iron River UP Brule 28 BKT 0.59
Little Indian River UP Manistique 29 BKT 0.73
Little Muskegon River NLP Muskegon 4 RBT 0.62
Manistee River NLP Manistee 5 BKT, BNT, RBT 0.65
Martin Creek SLP Pere Marquette-White 16 BKT 0.61
Pere Marquette River NLP Pere Marquette-White 6 BNT 0.61
Pigeon River NLP Cheboygan 7 BNT, RBT 0.65
Pine River SLP Pine 12 BNT 0.65
Pine River NLP Manistee 8 BKT, BNT, RBT 0.49
Prairie Creek SLP Lower Grand 13 BNT 0.50
Rogue River SLP Lower Grand 17 BNT, RBT 0.50
Salmon Trout River UP Keweenaw Peninsula 23 BKT 0.45
Tahquamenon River UP Tahquamenon 24 BNT 0.55
West Branch Sturgeon River NLP Cheboygan 9 BKT, BNT, RBT 0.65
Yellow Dog River UP Dead-Kelsey 30 RBT 0.52
193
Resilience-based salmonid management
Table 2. Stream-specific temperature regression models, with summary statistics, historical warming and projected future warming for each system. F(i.e., F
1,7
) and Pvalues refer to the air temperature (A) parameter
used to predict stream temperature (S). DAand DSdenote changes in air and stream temperatures over corresponding time periods. For 2006–2012, 2012–2036 and 2036–2056, predicted changes in Aand Sare
provided for the A2 and B1 climate forcing scenarios (B1 in parentheses).
Stream name Regression SE FPR
2
ΔA1976–
2006
ΔS1976–
2006 ΔA2006–2012 ΔS2006–2012 ΔA2012–2036 ΔS2012–2036 ΔA2036–2056 ΔS2036–2056
Bark River S=0.32A +12.98 0.05 49.41 <0.01 0.86 +1.72 +0.55 +1.09 (+1.23) +0.35 (+0.39) +2.74 (+3.16) +0.88 (+1.01) 0.69 (0.65) 0.22 (0.21)
Bear Creek S=0.23A +11.61 0.03 46.74 <0.01 0.85 +0.94 +0.22 +0.58 (+0.82) +0.13 (+0.19) +2.40 (+3.10) +0.55 (+0.71) +0.72 (0.29) +0.17 (0.07)
Black River S=0.29A +8.74 0.04 57.04 <0.01 0.88 +3.17 +0.92 +0.60 (+0.59) +0.17 (+0.17) +2.58 (+3.79) +0.75 (+1.10) 0.77 (1.15) 0.22 (0.33)
Boardman River S=0.14A +11.99 0.02 72.85 <0.01 0.90 +3.17 +0.44 0.56 (0.77) 0.08 (0.11) +2.49 (+3.77) +0.35 (+0.53) 0.87 (1.04) 0.12 (0.15)
Bryan Creek S=0.42A +7.69 0.08 28.86 <0.01 0.78 +2.33 +0.98 +1.01 (+1.48) +0.43 (+0.62) +2.83 (+3.11) +1.19 (+1.31) 0.62 (0.83) 0.26 (0.35)
Canada Creek S=0.60A +6.91 0.08 61.72 <0.01 0.88 +3.17 +1.90 0.79 (0.59) 0.47 (0.36) +2.58 (+3.79) +1.55 (+2.27) 0.77 (1.15) 0.46 (0.69)
Carp River S=0.28A +12.06 0.04 41.63 <0.01 0.84 +2.00 +0.56 +0.65 (+1.14) +0.18 (+0.32) +2.78 (+3.62) +0.78 (+1.01) 0.49 (1.29) 0.14 (0.36)
Cedar Creek S=0.56A +6.32 0.09 39.10 <0.01 0.83 +0.72 +0.40 +0.58 (+0.65) +0.33 (+0.37) +2.48 (+3.14) +1.39 (+1.76) +0.48 (0.49) +0.27 (0.28)
Cedar River (SLP) S=0.25A +11.42 0.03 97.82 <0.01 0.92 +2.61 +0.65 0.68 (0.23) 0.17 (0.06) +2.74 (+3.16) +0.69 (+0.79) 0.69 (0.65) 0.17 (0.16)
Chocolay River S=0.23A +10.29 0.03 77.99 <0.01 0.91 +2.72 +0.63 0.64 (+0.43) 0.15 (+0.10) +2.71 (+3.36) +0.62 (+0.77) 0.42 (1.31) 0.10 (0.30)
Davenport Creek S=0.13A +8.97 0.01 99.55 <0.01 0.92 +2.00 +0.26 +0.55 (+0.79) +0.07 (+0.10) +2.63 (+3.53) +0.34 (+0.46) 0.54 (1.26) 0.07 (0.16)
Duke Creek S=0.48A +4.45 0.08 37.27 <0.01 0.82 +0.72 +0.35 +0.65 (+0.74) +0.31 (+0.36) +2.48 (+3.14) +1.19 (+1.51) +0.48 (0.49) +0.23 (0.24)
East Branch Fox River S=0.33A +7.73 0.04 86.45 <0.01 0.91 +2.00 +0.66 +0.56 (+0.87) +0.19 (+0.29) +2.61 (+3.44) +0.86 (+1.14) 0.50 (1.27) 0.17 (0.42)
Elm River S=0.82A +2.11 0.06 193.20 <0.01 0.96 +2.72 +2.23 +0.99 (+1.79) +0.82 (+1.47) +3.19 (+2.65) +2.62 (+2.17) 0.75 (0.39) 0.62 (0.32)
Escanaba River S=0.88A +3.03 0.14 39.26 <0.01 0.83 +2.72 +2.40 +1.01 (+1.48) +0.89 (+1.30) +2.83 (+3.11) +2.49 (+2.73) 0.62 (0.83) 0.54 (0.73)
Iron River S=0.30A +12.76 0.04 52.58 <0.01 0.87 +1.89 +0.57 0.77 (0.03) 0.23 (0.01) +3.13 (+2.82) +0.94 (+0.85) 0.74 (0.35) 0.22 (0.11)
Little Indian River S=0.06A +14.86 0.01 83.65 <0.01 0.91 +2.72 +0.16 0.61 (+0.21) 0.04 (+0.01) +2.61 (+3.44) +0.16 (+0.21) 0.50 (1.27) 0.03 (0.08)
Little
Muskegon River
S=0.34A +12.09 0.04 69.13 <0.01 0.89 +2.56 +0.87 0.79 (0.80) 0.27 (0.27) +2.48 (+3.41) +0.84 (+1.16) 0.12 (0.72) 0.04 (0.24
Manistee River S=0.13A +10.67 0.02 57.70 <0.01 0.88 +3.17 +0.41 +0.76 (+0.49) +0.10 (+0.06) +2.47 (+3.56) +0.32 (+0.46) 0.54 (0.81) 0.07 (0.11)
Martin Creek S=0.38A +9.61 0.05 62.93 <0.01 0.89 +2.11 +0.80 0.82 (0.83) 0.31 (0.31) +2.48 (+3.32) +0.94 (+1.26) +0.05 (0.55) +0.02 (0.21)
Pere
Marquette River
S=0.18A +12.50 0.02 51.89 <0.01 0.86 +2.11 +0.38 0.82 (0.83) 0.15 (0.15) +2.48 (+3.32) +0.45 (+0.60) +0.05 (0.55) +0.01 (0.10)
Pigeon River S=0.11A +11.93 0.01 60.49 <0.01 0.88 +3.17 +0.35 +0.67 (+0.50) +0.07 (+0.06) +2.52 (+3.78) +0.28 (+0.42) 0.83 (1.11) 0.09 (0.12)
Pine River (NLP) S=0.22A +10.89 0.03 73.41 <0.01 0.90 +2.11 +0.46 +0.76 (+0.49) +0.17 (+0.11) +2.47 (+3.56) +0.54 (+0.78) 0.54 (0.81) 0.12 (0.18)
Pine River (SLP) S=0.40A +9.23 0.04 119.15 <0.01 0.94 +2.56 +1.02 +0.68 (+0.72) +0.27 (+0.29) +2.49 (+3.25) +1.00 (+1.30) +0.23 (0.69) +0.09 (0.27)
Prairie Creek S=0.74A +0.95 0.07 106.03 <0.01 0.93 +1.94 +1.44 +0.58 (+0.65) +0.43 (+0.48) +2.48 (+3.14) +1.83 (+2.32) +0.48 (0.49) +0.35 (0.37)
Rogue River S=0.23A +13.17 0.04 27.46 <0.01 0.77 +0.72 +0.17 +0.65 (+0.74) +0.15 (+0.17) +2.48 (+3.14) +0.57 (+0.72) +0.48 (0.49) +0.11 (0.11)
Salmon Trout River S=0.29A +8.23 0.05 33.60 <0.01 0.80 +2.33 +0.68 +0.99 (+1.79) +0.29 (+0.52) +3.19 (+2.65) +0.92 (+0.77) 0.75 (0.39) 0.22 (0.11)
Tahquamenon River S=0.50A +12.29 0.05 115.88 <0.01 0.93 +2.00 +1.00 +0.75 (+1.32) +0.38 (+0.66) +2.72 (+3.51) +1.36 (+1.75) 0.41 (1.36) 0.21 (0.68)
West Branch
Sturgeon River
S=0.06A +11.93 0.01 67.49 <0.01 0.89 +3.17 +0.19 +0.67 (+0.50) +0.04 (+0.03) +2.52 (+3.78) +0.15 (+0.23) 0.83 (1.11) 0.05 (0.07)
Yellow Dog River S=0.38A +9.90 0.06 45.58 <0.01 0.85 +2.33 +0.89 0.63 (+0.39) 0.24 (+0.15) +2.98 (+2.96) +1.13 (+1.13) 0.63 (0.81) 0.24 (0.31)
194
Carlson et al.
regression models. Mean July stream temperatures
were projected because this month is typically the
warmest and most thermally stressful for salmonids
in Michigan (Zorn et al. 2011). To incorporate the
range of air temperatures projected by each CCM and
the intrinsic uncertainty and unique characteristics
(i.e., atmospheric pressure, sea ice rheology, forest
canopy density and soil layering) of each model, pre-
dictions were averaged across the three CCM’s. Spe-
cies-specific thermal habitat suitability status was
assigned for each stream based on conditions for
growth and survival associated with projected July
temperatures. Status 1, 2, 3, and 4 corresponded with
optimal growth, suboptimal (i.e., reduced) growth, no
growth and extirpation, respectively (Table 3). Uni-
ted States Fish and Wildlife Service (USFWS) Bio-
logical Reports (e.g., Raleigh 1982a,b; Raleigh et al.
1986) contained thermal habitat status temperature
ranges for juveniles and adults of each species. Other
sources (i.e., Fry et al. 1946; Baldwin 1957; Wurts-
baugh & Davis 1977; Elliott & Hurley 2000; Hay
et al. 2006) contained temperature ranges for juve-
niles and/or adults and were used to confirm tempera-
tures reported in the USFWS reports. When threshold
temperatures (e.g., thermal minima, maxima) differed
between juveniles and adults, we reported juvenile
temperatures under the premise that resilient salmo-
nid fisheries can only be conserved if young fish
survive to adulthood.
Analyses
The accuracy of stream-specific regression models
was evaluated by comparing each stream’s projected
temperature and thermal habitat suitability status in
2006 to its actual (i.e., field measured) temperature
and associated habitat status obtained from the
MDNR database. The association between BFI and
air temperature regression coefficients, which are
indices of stream thermal sensitivity, was assessed
using simple linear regression (Kelleher et al. 2012).
Results
Model accuracy
Stream-specific models accurately projected tempera-
ture and thermal habitat suitability status in brook
charr, brown trout, and rainbow trout streams. Under
the A2 scenario, the mean deviation between pre-
dicted and actual temperatures was 0.46 °C
(SD =0.56; Table 4). Under the B1 scenario, the
mean deviation between predicted and actual temper-
atures was 0.58 °C (SD =0.59; Table 4). Under
the A2 and B1 scenarios, stream-specific models pre-
dicted thermal habitat status with 93.0 per cent over-
all accuracy in streams with brook charr (94.0 per
cent accuracy, N=15), brown trout (89.0 per cent
accuracy, N=16) and rainbow trout (100.0 per cent
accuracy, N=11; Table 4).
Thermal habitat suitability: brook charr
Stream-specific regression models projected that cli-
mate-induced air temperature elevation will have sub-
stantial effects on stream temperature and thermal
habitat suitability, with impacts varying by BFI (i.e.,
streams with lower BFI were more thermally sensitive;
Fig. 2) as well as species, region, time period and cli-
mate forcing scenario. From 1976 to 2006, the mean
temperature of brook charr streams increased by
0.74 °C as mean air temperature increased by 2.34 °C
(Table 2). In the UP under the A2 and B1 scenarios,
thermal habitat suitability was predicted to be optimal
in the East Branch Fox and Little Indian rivers and
suboptimal in the Iron River from 2006–2056 (Fig. 3a,
b). From 2006 to 2012, thermal habitat was projected
to be optimal in Bryan Creek and the Salmon Trout
River and suboptimal in the Carp and Escanaba rivers
under both scenarios. From 2012 to 2036, thermal
habitat was projected to become suboptimal in Bryan
Creek with predicted warming by 1.19 °C (A2 scenar-
io) and 1.31 °C (B1 scenario; Table 2). Similarly, ther-
mal habitat was forecasted to become unsuitable for
growth in the Escanaba River with projected warming
by 2.49 °C (A2 scenario) and 2.73 °C (B1 scenario;
Table 2). Habitat suitability was predicted to remain
the same from 2036 to 2056 under both scenarios in all
brook charr streams evaluated.
In the NLP from 2006 to 2056, thermal habitat
was projected to be optimal in most streams (Black,
Boardman, Manistee, Pine, West Branch Sturgeon
Table 3. Thermal habitat suitability status (habitat status) designations and
corresponding temperature ranges (temperature) and growth conditions
(growth) for juvenile and adult brook charr (BKC; Fry et al. 1946; Baldwin
1957; Raleigh 1982a), brown trout (BNT; Raleigh et al. 1986; Elliott &
Hurley 2000; Hay et al. 2006) and rainbow trout (RBT; Wurtsbaugh &
Davis 1977; Raleigh 1982b).
Species Habitat status Temperature Growth
BKC 1 11.0 ≤°C<16.5 Optimal
2 16.5 ≤°C<20.5 Suboptimal
3 20.5 ≤°C<25.3 None
4°C≥25.3 Extirpation
BNT 1 12.0 ≤°C<17.0 Optimal
2 17.0 ≤°C<20.0 Suboptimal
3 20.0 ≤°C<26.2 None
4°C≥26.2 Extirpation
RBT 1 12.0 ≤°C<16.4 Optimal
2 16.4 ≤°C<22.5 Suboptimal
3 22.5 ≤°C<25.0 None
4°C≥25.0 Extirpation
195
Resilience-based salmonid management
rivers) under the A2 and B1 scenarios (Fig. 3c,d).
From 2012 to 2036, thermal habitat was forecasted to
become unsuitable for growth in Canada Creek under
the B1 scenario with predicted warming by 2.27 °C
(Fig. 3d; Table 2). In the SLP under the A2 and B1
scenarios, thermal habitat was predicted to be subop-
timal from 2006 to 2056 in Martin Creek and Cedar
Creek (Fig. 3e,f). From 2012 to 2036 under the B1
scenario, thermal habitat was forecasted to become
suboptimal in Duke Creek with predicted warming
by 1.51 °C (Fig. 3f; Table 2).
Thermal habitat suitability: brown trout
From 1976 to 2006, the mean temperature of brown
trout streams increased by 0.65 °C as mean air tem-
perature increased by 2.09 °C (Table 2). In the UP
under the A2 and B1 scenarios, thermal habitat suit-
ability was predicted to be optimal in the East Branch
Fox River from 2006 to 2056 (Fig. 4a,b). From 2006
to 2012, thermal habitat suitability was projected to
be optimal in the Bark River, suboptimal in the
Escanaba River and unsuitable in the Tahquamenon
Table 4. Actual versus projected stream temperatures and thermal habitat suitability (THS) status for brook charr (BKC), brown trout (BNT) and rainbow trout
(RBT) as predicted by stream-specific air–water temperature regression models. Temperatures are predicted in 2006 under the A2 (820 ppm atmospheric CO
2
by 2100) and B1 (550 ppm atmospheric CO
2
by 2100) scenarios. Dsymbols denote differences between projected and actual stream temperatures obtained
from the Northeast Climate Science Center. The first and second THS numbers represent statuses associated with actual and projected temperatures,
respectively.
Stream name Actual A2 ΔTHS B1 ΔTHS
Bark River 19.91 18.95 0.96 BNT: 2,2 18.88 1.03 BNT: 2,2
Bear Creek 15.75 16.54 0.79 BNT: 1,1 16.54 0.79 BNT: 1,1
Black River 15.05 14.46 0.59 BKC: 1,1 14.36 0.69 BKC: 1,1
Boardman River 15.02 14.88 0.14 BKC: 1,1 14.85 0.17 BKC: 1,1
BNT: 1,1 BNT: 1,1
Bryan Creek 16.66 15.27 1.39 BKC: 2,1 15.06 1.60 BKC: 2,1
Canada Creek 19.95 19.41 0.54 BKC: 2,2 19.09 0.86 BKC: 2,2
Carp River 17.61 17.05 0.56 BKC: 2,2 17.00 0.72 BKC: 2,2
BNT: 2,2 BNT: 2,2
RBT: 2,2 RBT: 2,2
Cedar Creek 18.30 18.27 0.04 BKC: 2,2 18.32 0.02 BKC: 2,2
BNT: 2,2 BNT: 2,2
Cedar River (SLP) 17.15 16.59 0.56 BNT: 2,1 16.46 0.69 BNT: 2,1
Chocolay River 15.23 14.77 0.46 RBT: 1,1 14.52 0.71 RBT: 1,1
Davenport Creek 11.50 11.39 0.11 RBT: 1,1 11.34 0.16 RBT: 1,1
Duke Creek 15.32 14.69 0.63 BKC: 1,1 14.73 0.59 BKC: 1,1
BNT: 1,1 BNT: 1,1
East Branch Fox River 14.23 13.76 0.47 BKC: 1,1 13.62 0.61 BKC: 1,1
BNT: 1,1 BNT: 1,1
Elm River 17.48 16.96 0.52 RBT: 2,2 16.40 1.10 RBT: 2,2
Escanaba River 19.95 18.88 1.07 BKC: 2,2 18.45 1.50 BKC: 2,2
BNT: 2,2 BNT: 2,2
Iron River 18.61 18.53 0.08 BKC: 2,2 18.34 0.27 BKC: 2,2
Little Indian River 16.16 16.09 0.07 BKC: 1,1 16.03 0.13 BKC: 1,1
Little Muskegon River 19.89 19.24 0.65 RBT: 2,2 19.22 0.67 RBT: 2,2
Manistee River 13.55 13.29 0.26 BKC: 1,1 13.29 0.26 BKC: 1,1
BNT: 1,1 BNT: 1,1
RBT: 1,1 RBT: 1,1
Martin Creek 18.33 17.86 0.47 BKC: 2,2 17.87 0.46 BKC: 2,2
Pere Marquette River 16.62 16.41 0.21 BNT: 2,2 16.41 0.21 BNT: 2,2
Pigeon River 14.45 14.12 0.33 BNT: 1,1 14.10 0.35 BNT: 1,1
RBT: 1,1 RBT: 1,1
Pine River (NLP) 15.97 15.26 0.71 BKC: 1,1 15.27 0.70 BKC: 1,1
BNT: 1,1 BNT: 1,1
RBT: 1,1 RBT: 1,1
Pine River (SLP) 18.42 17.55 0.87 BNT: 2,2 17.56 0.86 BNT: 2,2
Prairie Creek 18.35 16.57 1.78 BNT: 2,1 16.64 1.71 BNT: 2,1
Rogue River 18.36 17.95 0.41 BNT: 2,2 17.97 0.39 BNT: 2,2
RBT: 2,2 RBT: 2,2
Salmon Trout River 14.62 13.51 1.11 BKC: 1,1 13.31 1.31 BKC: 1,1
Tahquamenon River 20.02 21.16 1.14 BNT: 3,3 20.86 0.84 BNT: 3,3
West Branch Sturgeon River 13.30 13.12 0.18 BKC: 1,1 13.11 0.19 BKC: 1,1
BNT: 1,1 BNT: 1,1
RBT: 1,1 RBT: 1,1
Yellow Dog River 18.00 17.33 0.67 RBT: 2,2 16.96 1.04 RBT: 2,2
196
Carlson et al.
River under both scenarios (Fig. 4a,b). During the
same time period, thermal habitat was forecasted to
become suboptimal in the Carp River under the B1
scenario with predicted warming by 0.32 °C
(Table 2). From 2012 to 2036, thermal habitat was
projected to become unsuitable for growth in the
Escanaba River with predicted warming by 2.49 °C
(A2 scenario) and 2.73 °C (B1 scenario; Table 2).
From 2036 to 2056, habitat suitability was predicted
to remain the same under both scenarios in all brown
trout streams evaluated.
In the NLP from 2006 to 2056, thermal habitat was
forecasted to be optimal under the A2 and B1 scenar-
ios in all brown trout streams evaluated, including the
Boardman, Manistee, Pere Marquette, Pigeon, Pine
and West Branch Sturgeon rivers (Fig. 4c,d). In the
SLP under the A2 and B1 scenarios, thermal habitat
suitability was predicted to be optimal in Duke Creek
and suboptimal in the Rogue River from 2006 to 2056
(Fig. 4e,f). From 2006 to 2012, thermal habitat was
projected to be optimal in the Cedar River and subop-
timal in Cedar Creek and the Pine River under both
scenarios. During the same time period, thermal habi-
tat was predicted to become suboptimal in Prairie
Creek under the B1 scenario with projected warming
by 0.48 °C (Table 2). From 2012 to 2036, thermal
habitat was forecasted to become suboptimal in the
Cedar River and Bear Creek and unsuitable for growth
in Cedar Creek under both scenarios. From 2036–
2056 under the A2 scenario, thermal habitat was pro-
jected to become optimal in the Cedar River and
unsuitable for growth in Cedar Creek with predicted
cooling and warming by 0.17 °C and 0.27 °C respec-
tively (Fig. 4e; Table 2). Thermal habitat suitability
was forecasted to remain the same from 2036 to 2056
under the B1 scenario in all brown trout streams eval-
uated (Fig. 4f).
Thermal habitat suitability: rainbow trout
From 1976 to 2006, the mean temperature of brown
trout streams increased by 0.64 °C as mean air tem-
perature increased by 2.42 °C (Table 2). In the UP
under the A2 and B1 scenarios, thermal habitat suit-
ability was predicted to be optimal in the Chocolay
River and Davenport Creek and suboptimal in the
Carp, Elm and Yellow Dog Rivers from 2006–2056
(Fig. 5a,b). Similarly, thermal habitat was projected
to be optimal (Manistee, Pigeon, Pine and West
Branch Sturgeon rivers) and suboptimal (Little
Muskegon River) in the NLP from 2006 to 2056
under the A2 and B1 scenarios (Fig. 5c,d). Thermal
habitat in the Rogue River, the only rainbow trout
stream in the SLP, was predicted to be suboptimal
under both emissions scenarios from 2006–2056
(Fig. 5e,f).
Discussion
Our study is the first broad-scale investigation of cli-
mate-driven stream temperature warming in Michigan
with implications for salmonid management. Water
temperatures were projected to increase by 0.19–
5.94 °C in 30 coldwater salmonid streams over the
next 40 years due to predicted air temperature warm-
ing. This finding supports a previous study in the
State of Minnesota, USA (Pilgrim et al. 1998), in
which temperatures in 39 streams increased by 0.30–
6.90 °C under a doubling of atmospheric CO
2
con-
centrations comparable to the A2 climate forcing sce-
nario. In addition, our study supports a previous
investigation in the State of Wisconsin, USA (Lyons
et al. 2010), in which thermal habitat degradation
was projected for brook charr and brown trout in 282
streams forecasted to warm by 0.80–4.00 °C over the
next 45 years. In our research, the magnitude of
warming and thermal habitat impairment varied
among streams and time periods, with the greatest
warming and habitat degradation occurring from
1976 to 2006 (2.72 °C) and projected from 2012 to
2036 (2.49–2.62 °C) in the Elm and Escanaba rivers,
systems with the lowest BFI (<0.45). In contrast, the
magnitude of warming from 1976 to 2006 (0.16–
0.66 °C) and projected warming from 2012 to 2036
(0.16–0.86 °C) was considerably lower in the East
Branch Fox and Little Indian rivers, systems with the
highest BFI (0.73). This finding supported our
hypothesis that temperature and thermal habitat alter-
ations would vary spatially and temporally due to
system-specific patterns of warming. It also rein-
forces previous research indicating fish assemblages
in Michigan streams with stable discharge from
groundwater inputs are more resilient to thermal
warming than those in streams with less stable dis-
Fig. 2. Relative influence of baseflow index on thermal sensitiv-
ity (i.e., air temperature regression coefficients) of streams in the
State of Michigan, USA. The regression equation is: Thermal sen-
sitivity =1.78 9Baseflow index +1.38 (P<0.01, R
2
=0.44).
197
Resilience-based salmonid management
charge (Zorn et al. 2012). Thermal habitat degrada-
tion was projected to occur least frequently in rain-
bow trout streams as this species has a wider
temperature range (12.0–22.5 °C) for growth and sur-
vival compared to brook charr (11.0–20.5 °C) and
brown trout (12.0–20.0 °C) (Wurtsbaugh & Davis
1977).
Warmer air temperatures are projected to increase
stream temperatures both directly (i.e., advection) and
indirectly through effects on precipitation, evapora-
Fig. 3. Projected temperatures in individual Michigan brook charr streams in 2006, 2012, 2036 and 2056 under A2 (820 ppm atmospheric
CO
2
by 2100) and B1 (550 ppm by 2100) scenarios. Plots are organised by region and emissions scenario: (a) Upper Peninsula, A2 scenar-
io; (b) Upper Peninsula, B1 scenario; (c) northern Lower Peninsula, A2 scenario; (d) northern Lower Peninsula, B1 scenario; (e) southern
Lower Peninsula, A2 scenario; (f) southern Lower Peninsula, B1 scenario. Dotted lines represent transitions between thermal habitat suit-
ability statuses. Stream abbreviations are as follows: BRC =Bryan Creek, BLR =Black River, BOR =Boardman River, CAC =Canada
Creek, CAR =Carp River, CEC =Cedar Creek, DUC =Duke Creek, EFR =East Branch Fox River, ESR =Escanaba River, IRR =Iron
River, LIR =Little Indian River, MAR =Manistee River, MAC =Martin Creek, PIR =Pine River (northern Lower Peninsula),
STR =Salmon Trout River, WSR =West Branch Sturgeon River.
198
Carlson et al.
tion and transpiration. Predicted effects of climate
change include more variable (and overall lower) pre-
cipitation (Parry et al. 2007; Stoner et al. 2013),
increased evaporation (Compagnucci et al. 2001) and
decreased discharge. These impacts would increase
stream temperature by reducing the volume of water
exposed to solar radiation. However, in Michigan
streams near the Great Lakes, precipitation will likely
Fig. 4. Projected temperatures in individual Michigan brown trout streams in 2006, 2012, 2036 and 2056 under A2 (820 ppm atmospheric
CO
2
by 2100) and B1 (550 ppm by 2100) scenarios. Plots are organised by region and emissions scenario: (a) Upper Peninsula, A2 scenar-
io; (b) Upper Peninsula, B1 scenario; (c) northern Lower Peninsula, A2 scenario; (d) northern Lower Peninsula, B1 scenario; (e) southern
Lower Peninsula, A2 scenario; (f) southern Lower Peninsula, B1 scenario. Dotted lines represent transitions between thermal habitat suit-
ability statuses. Stream abbreviations are as follows: BEC =Bear Creek, BOR =Boardman River, BRC =Bryan Creek, CAR =Carp
River, CEC =Cedar Creek, CER =Cedar River, DUC =Duke Creek, EFR =East Branch Fox River, ESR =Escanaba River,
MAR =Manistee River, PMR =Pere Marquette River, PGR =Pigeon River, PNR =Pine River (southern Lower Peninsula), PIR =Pine
River (northern Lower Peninsula), PRC =Prairie Creek, ROR =Rogue River, TAR =Tahquamenon River, WSR =West Branch Stur-
geon River.
199
Resilience-based salmonid management
intensify in fall and winter due to predicted increases
in snowfall (Norton & Bolsenga 1993; Primack
2000), increasing surface runoff and water volume
and decreasing temperature compared to systems
without this source of precipitation (Kurylyk et al.
2013). Increased canopy and soil surface evaporation
is projected to decrease groundwater recharge and
increase water temperature (Ge et al. 2013), although
streams high BFI are expected to warm at a slower
rate than those with low BFI due to groundwater-dri-
ven thermal buffering (Menberg et al. 2014). Tran-
spiration of water from plants has been reported to
contribute to stream warming by reducing streamflow
and groundwater recharge (Federer & Lash 1978;
Bond et al. 2002). Effects of climate warming on
transpiration and stream temperature will depend lar-
Fig. 5. Projected temperatures in individual Michigan rainbow trout streams in 2006, 2012, 2036 and 2056 under A2 (820 ppm atmo-
spheric CO
2
by 2100) and B1 (550 ppm by 2100) scenarios. Plots are organised by region and emissions scenario: (a) Upper Peninsula,
A2 scenario; (b) Upper Peninsula, B1 scenario; (c) northern Lower Peninsula, A2 scenario; (d) northern Lower Peninsula, B1 scenario; (e)
southern Lower Peninsula, A2 scenario; (f) southern Lower Peninsula, B1 scenario. Dotted lines represent transitions between thermal habi-
tat suitability statuses. Stream abbreviations are as follows: CAR =Carp River, CHR =Chocolay River, DAC =Davenport Creek,
ELR =Elm River, LMR =Little Muskegon River, MAR =Manistee River, PGR =Pigeon River, PIR =Pine River, ROR =Rogue
River, WSR =West Branch Sturgeon River, YDR =Yellow Dog River.
200
Carlson et al.
gely on changes in the species composition and leaf
characteristics (e.g., stomatal conductance, water-use
efficiency) of riparian plants and trees (Compagnucci
et al. 2001; Kirschbaum 2004). We encourage
researchers to incorporate the thermal effects of tran-
spiration, riparian shading and watershed land cover
(Wehrly et al. 1997; Wiley et al. 2010) into future
stream temperature modelling to improve predictive
accuracy.
Projected increases in Michigan stream tempera-
tures will result in thermal habitat conditions less
conducive for salmonid growth and survival during
the warmest period of the year (Raleigh 1982a,b;
Raleigh et al. 1986), particularly in thermally sensi-
tive streams with low BFI values. This will likely be
manifested by declines in availability and changes in
spatial arrangement of optimal thermal habitat, with
potential effects on salmonid distribution and sur-
vival. For instance, if stream warming restricts cold-
water habitats to specific areas (e.g., headwater
reaches), salmonid populations will become isolated,
which could increase interspecific competition and
decrease survival (Tsuboi et al. 2013; Dugdale et al.
2015). Previous research in Michigan’s Muskegon
River suggested warmer water temperatures resulting
from climate change will restrict the distribution and
decrease the survival of brook charr, brown trout,
and rainbow trout (Steen et al. 2010). Researchers in
the State of Wisconsin, USA, predicted that by 2060
the total length of streams suitable for brown trout
would decline by 8, 33, and 88 per cent under limited
(summer air temperatures increase 1.0 °C and water
0.8 °C), moderate (air 3.0 °C and water 2.4 °C), and
major (air 5.0 °C and water 4.0 °C) climatic warming
(Lyons et al. 2010). Results were more extreme for
brook charr, with distributional declines of 44 and 94
per cent under limited and moderate warming, and
complete extirpation under major warming. Although
our models did not project extirpation of any species
in the Michigan streams evaluated, summer growth
limitation may be coupled with decreased reproduc-
tion as lower oxygen levels resulting from stream
temperature warming and/or groundwater withdrawal
may reduce egg survival or prevent spawning
(Raleigh 1982a,b; Raleigh et al. 1986). Although
annual salmonid growth may remain stable or
increase due to longer growing season length and
increased prey availability, projected reductions in
growth and survival during the warmest period of the
year have important management implications.
Management implications
We encourage scientists, biologists, policymakers and
public stakeholders to collaboratively develop resili-
ence-based management programmes to conserve sal-
monid populations amidst global change. Our
research represents a step in this direction as it pro-
jects consequences of stream thermal warming for
salmonid growth and survival in Michigan. Our
results indicate it is important for fisheries profession-
als to manage streams for thermal resilience by form-
ing public–private partnerships to protect watershed
land cover types that facilitate high groundwater
recharge (e.g., grasslands; Waco & Taylor 2010; Sii-
tari et al. 2011), preserve riparian vegetation and
associated shading (Blann et al. 2002), and maintain
longitudinal connectivity to promote salmonid move-
ment to cold headwater reaches (Drake & Taylor
1996; Hayes et al. 1998). We recommend these
strategies play prominent roles in resilience-based
management programmes for coldwater streams and
their important salmonid fisheries. As documented in
our study, streams with low thermal sensitivity (i.e.,
high BFI) will likely maintain thermal conditions
(e.g., cool summer temperatures, seasonal flow stabil-
ity; Wiley et al. 1997; Baker et al. 2003) that are
more conducive for salmonid growth and survival
than systems with high thermal sensitivity in the next
40 years. Thus, we encourage managers to allocate
resources (i.e., time, money and personnel) to priori-
tise protection of streams with low thermal sensitivity
and inform public stakeholders about realistic expec-
tations for stream fish communities amidst global
change (e.g., salmonid decline, centrarchid expan-
sion; Pease & Paukert 2014). Moreover, because
streams with high BFI were less susceptible to tem-
perature change, we suggest that managers use spatial
BFI maps as tools for understanding stream thermal
sensitivity across large geographical areas. Moreover,
we encourage managers to increase the spatial and
temporal coverage of air and stream temperature
monitoring networks and thereby expand their utility
for salmonid management.
Fisheries professionals and public stakeholders can
also collaboratively implement additional strategies
for resilience-based salmonid management. They can
promote thermally resilient salmonid populations by
removing dams and installing fish ladders at roadside
crossings and culverts to restore stream habitat con-
nectivity, particularly in cold headwater reaches,
which function as thermal refugia during warm sum-
mer months. Fisheries professionals can also foster
salmonid population resilience by implementing regu-
lations (e.g., protected slot limits, reduced creel lim-
its) that reduce harvest and increase survival during
thermally stressful periods. In addition, we recom-
mend fisheries professionals protect a diversity of sal-
monid size classes, genetic stocks and prey species
that tolerate a wide range of temperatures predicted
from climate change models (Hansen et al. 2015).
We encourage fisheries managers and policymakers
201
Resilience-based salmonid management
to provide incentives (e.g., financial assistance,
open space tax deduction, fast-track permitting;
Knight 2009) for land developers and property
owners to protect coldwater habitat and thermal
buffering mechanisms on their lands. Moreover,
because it is infeasible to protect salmonid popula-
tions and thermal habitat in all streams, it will be
necessary for fisheries professionals to implement a
triage approach with specific criteria for stream pro-
tection (e.g., species composition, habitat quality
and recreational importance). In summary, resili-
ence-based salmonid management programmes will
require effective collaboration among scientists,
biologists, policymakers and public stakeholders.
Our research promotes resilience-based management
by providing a methodology to project stream tem-
perature and thermal habitat suitability. Fisheries
professionals can use this approach to protect cold-
water habitats and drivers of stream cooling and
ultimately conserve resilient salmonid populations
amidst global change.
Acknowledgements
We thank Jennifer Moore Myers (USFS EFETAC) for assis-
tance with air temperature projection models and data acqui-
sition. We thank Kyle Herreman and Wesley Daniel
(Michigan State University), Tracy Kolb and Todd Wills
(Michigan Department of Natural Resources) and Henry
Quinlan (United States Fish and Wildlife Service) for provid-
ing initial environmental and trout population data critical for
development of this study. We acknowledge the Programme
for Climate Model Diagnosis and Intercomparison (PCMDI)
and the WCRP’s Working Group on Coupled Modelling for
assisting in procurement of the WCRP CMIP3 multimodel
data set.
References
Arblaster, J., Dufresne, J.L., Fichefet, T., Fiedlingstein, P.,
Gao, X., Gutowski, W.J. Jr, Johns, T., Krinner, G.,
Shongwe, M., Tebaldi, C., Weaver, A.J. & Wehner, M.
2014. Long-term climate change: Projections, commitments
and irreversibility. In: Stocker, T.F., Qin, D., Plattner, G.K.,
Tignor, M., Allen, S.K., Boschung, J., Nauels, A., Xia, Y.,
Bex, V. & Midgley, P.M., eds. Climate change 2013: the
physical science basis. Contribution of working group I to
the fifth assessment report of the Intergovernmental
Panel on Climate Change. Cambridge: Cambridge Univer-
sity Press, pp. 1029–1136.
Arnold, J.G. & Allen, P.M. 1999. Validation of automated
methods for estimating base flow and ground water recharge
from streamflow records. Journal of the American Water
Resources Association 35: 411–424.
Baker, M.E., Wiley, M.J., Carlson, M.L. & Seelbach, P.W.
2003. A GIS model of subsurface water potential for aquatic
resource inventory, assessment, and environmental manage-
ment. Environmental Management 32: 706–719.
Baldwin, N.S. 1957. Food consumption and growth of brook
trout at different temperatures. Transactions of the American
Fisheries Society 86: 323–328.
Blann, K., Nerbonne, J.F. & Vondracek, B. 2002. Relation-
ship of riparian buffer type to water temperature in the drift-
less area ecoregion of Minnesota. North American Journal
of Fisheries Management 22: 441–451.
Bond, B.J., Jones, J.A., Moore, G., Phillips, N., Post, D. &
McDonnell, J.J. 2002. The zone of vegetation influence on
baseflow revealed by diet patterns of streamflow and vegeta-
tion water use in a headwater basin. Hydrological Processes
16: 1671–1677.
Compagnucci, R., da Cunha, L., Hanaki, K., Howe, C., Mailu,
G., Shiklomanov, I. & Stakhiv, E. 2001. Hydrology and
water resources. In: McCarthy, J.J., Canziani, O.F., Leary,
N.A., Dokken, D.J. & White, K.S., eds. Climate change
2001: impacts, adaptation, and vulnerability. International
Panel on Climate Change Third Assessment Report. Cam-
bridge: Cambridge University Press, pp. 191–233.
Dodds, W.K. & Whiles, M.R. 2010. Freshwater ecology: con-
cepts and environmental applications of limnology, 2nd edn.
Burlington: Academic Press. xv + 829 pp.
Drake, M.T. & Taylor, W.W. 1996. Influence of spring and
summer water temperature on brook charr, Salvelinus fonti-
nalis, growth and age structure in the Ford River, Michigan.
Environmental Biology of Fishes 45: 41–51.
Dugdale, S.J., Bergeron, N.E. & St-Hilaire, A. 2015. Spatial
distribution of thermal refuges analysed in relation to river-
scape hydromorphology using airborne thermal infrared ima-
gery. Remote Sensing of Environment 160: 43–55.
Elliott, J.M. & Hurley, M.A. 2000. Daily energy intake and
growth of piscivorous brown trout, Salmo trutta. Freshwater
Biology 44: 237–245.
Federer, C.A. & Lash, D. 1978. Simulated streamflow response
to possible differences in transpiration among species of
hardwood trees. Water Resources Research 14: 1089–1097.
Ficklin, D.L., Stewart, I.T. & Maurer, E.P. 2013. Effects of
climate change on stream temperature, dissolved oxygen,
and sediment concentration in the Sierra Nevada in Califor-
nia. Water Resources Research 49: 2765–2782.
Fry, F.E.J., Hart, J.S. & Walker, K.F. 1946. Lethal temperature
relations for a sample of young speckled trout, Salvelinus
fontinalis, Vol. 54. Toronto: The University of Toronto Press.
Ge, Z.M., Kellom€
aki, S., Zhou, X., Wang, K.Y., Peltola, H.,
V€
ais€
anen, H. & Strandman, H. 2013. Effects of climate
change on evapotranspiration and soil water availability in
Norway spruce forests in southern Finland: an ecosystem
model based approach. Ecohydrology 6: 51–63.
Godby, N.A. Jr, Rutherford, E.S. & Mason, D.M. 2007. Diet,
feeding rate, growth, mortality, and production of juvenile
steelhead in a Lake Michigan tributary. North American
Journal of Fisheries Management 27: 578–592.
Hansen, G.J.A., Gaeta, J.W., Hansen, J.F. & Carpenter, S.R.
2015. Learning to manage and managing to learn: sustaining
freshwater recreational fisheries in a changing environment.
Fisheries 40: 56–64.
Hay, J., Hayes, J. & Young, R. 2006. Water quality guidelines
to protect trout fishery values. Prepared for Horizons Regio-
nal Council. Cawthorn Report No. 1025. 17 pp.
Hayes, D.B., Taylor, W.W., Drake, M., Marod, S. & Whelan,
G. 1998. The value of headwaters to brook trout (Salvelinus
202
Carlson et al.
fontinalis) in the Ford River, Michigan, USA. In: Haigh,
M.J., Krecek, J., Rajwar, G.S. & Kilmartin, M.P. (eds).
Headwaters: water resources and soil conservation. New
Delhi: Oxford and IBH Publishing Co., 495 pp.
Hayes, D.B., Dodd, H. & Lessard, J. 2006. Effects of small
dams on cold water stream fish communities. In: Nelson, J.,
Dodson, J.J., Friedland, K., Hamon, T.R., Musick, J. & Ver-
spoor, E., eds. Reconciling fisheries with conservation: Pro-
ceedings of the Fourth World Fisheries Congress. Bethesda:
American Fisheries Society Symposium 43, pp. 587–602.
Holling, C.S. 1973. Resilience and stability of ecological sys-
tems. Annual Review of Ecology and Systematics 4: 1–23.
Isaak, D.J., Muhlfeld, C.C., Todd, A.S., Al-Chokhachy, R.,
Roberts, J., Kershner, J.L., Fausch, K.D. & Hostetler, S.W.
2012. The past as prelude to the future for understanding
21st-century climate effects on Rocky Mountain trout. Fish-
eries 37: 542–556.
Isaak, D.J., Young, M.K., Nagel, D.E., Horan, D.L. & Groce,
M.C. 2015. The cold-water climate shield: delineating refu-
gia for preserving salmonid fishes through the 21st century.
Global Change Biology 21: 2540–2553.
Kelleher, C., Wagener, T., Gooseff, M., McGlynn, B.,
McGuire, K. & Marshall, L. 2012. Investigating controls on
the thermal sensitivity of Pennsylvania streams. Hydrologi-
cal Processes 26: 771–785.
Kirschbaum, M.U.F. 2004. Direct and indirect climate change
effects on photosynthesis and transpiration. Plant Biology 6:
242–253.
Knight, K. 2009. Land use planning for salmon, steelhead and
trout. Washington Department of Fish and Wildlife. Olym-
pia, Washington. Available at: http://wdfw.wa.gov/publica-
tions/00033/wdfw00033.pdf (accessed 1 October 2015).
Krider, L.A., Magner, J.A., Perry, J., Vondracek, B. & Fer-
rington, L.C. Jr 2013. Air-water temperature relationships in
the trout streams of southeastern Minnesota’s carbonate-
sandstone landscape. Journal of the American Water
Resources Association 49: 896–907.
Kurylyk, B.L., Bourque, C.P.A. & MacQuarrie, K.T.B. 2013.
Potential surface temperature and shallow groundwater tem-
perature response to climate change: an example from a small
forested catchment in east-central New Brunswick (Canada).
Hydrology and Earth System Sciences 17: 2701–2716.
Lessard, J.L. 2000. Temperature effect of dams on coldwater
fish and macroinvertebrate communities in Michigan, Master
of Science. East Lansing: Michigan State University.
Lyons, J., Stewart, J.S. & Mitro, M. 2010. Predicted effects of
climate warming on the distribution of 50 stream fishes in
Wisconsin, U.S.A. Journal of Fish Biology 77: 1867–1898.
Magnuson, J.J., Crowder, L.B. & Medvick, P.A. 1979. Tem-
perature as an ecological resource. American Zoologist 19:
331–343.
Magnuson, J.J., Webster, K.E., Assel, R.A., Bowser, C.J., Dil-
lon, P.J., Eaton, J.G., Evans, H.E., Fee, E.J., Hall, R.I.,
Mortsch, L.R., Schindler, D.W. & Quinn, F.H. 1997. Poten-
tial effects of climate change on aquatic systems: Laurentian
Great Lakes and Precambrian Shield Region. Hydrological
Processes 11: 825–871.
Maurer, E.P., Brekke, L., Pruitt, T. & Duffy, P.B. 2007. Fine-
resolution climate projections enhance regional climate
change impact studies. Eos Transactions, American Geo-
physical Union 88: 504.
Menberg, K., Blum, P., Kurylyk, B.L. & Bayer, P. 2014.
Observed groundwater temperature response to recent cli-
mate change. Hydrology and Earth System Sciences 18:
4453–4466.
Michigan Department of Natural Resources (MDNR). 2015.
Fishing –Michigan.gov/fishing. Michigan Department of
Natural Resources. Available at: http://www.michigan.gov/
dnr/0,4570,7-153-10364—,00.html (accessed 19 June 2015).
Neff, B.D., Day, S.M., Piggott, A.R. & Fuller, L.M. 2005.
Base flow in the Great Lakes basin: U.S. Geological Survey
Scientific Investigations Report 2005–5217, 23 p.
Norton, D.C. & Bolsenga, S.J. 1993. Spatiotemporal trends in
lake effect and continental snowfall in the Laurentian Great
Lakes, 1951–1980. Journal of Climate 6: 1943–1956.
Parry, M., Canziani, O., Palutikof, J., dervan Linden, P. &
Hanson, C. 2007. Climate change 2007: impacts, adaptation
and vulnerability. International Panel on Climate Change
Fourth Assessment Report.
Pease, A.A. & Paukert, C.P. 2014. Potential impacts of
climate change on growth and prey consumption of stream-
dwelling smallmouth bass in the central United States.
Ecology of Freshwater Fish 23: 336–346.
Pilgrim, J.M., Fang, X. & Stefan, H.G. 1998. Stream tempera-
ture correlations with air temperatures in Minnesota: impli-
cations for climate warming. Journal of the American Water
Resources Association 34: 1109–1121.
Primack, A.G.B. 2000. Simulation of climate-change effects
on riparian vegetation in the Pere Marquette River, Michi-
gan. Wetlands 20: 538–547.
Raleigh, R.F. 1982a. Habitat suitability index models: brook
trout: U.S. Fish and Wildlife Service Biological Report 82
(10.24). 42 pp.
Raleigh, R.F. 1982b. Habitat suitability index models: rainbow
trout: U.S. Fish and Wildlife Service Biological Report 82
(10.60). 76 pp.
Raleigh, R.F., Zuckerman, L.D. & Nelson, P.C. 1986. Habitat
suitability index models and instream flow suitability curves:
brown trout: U.S. Fish and Wildlife Service Biological
Report 82(10.124). 65 pp.
Sear, D.A., Armitage, P.D. & Dawson, F.H. 1999.
Groundwater dominated rivers. Hydrological Processes 13:
255–276.
Siitari, K.J., Taylor, W.W., Nelson, S.A.C. & Weaver, K.E.
2011. The influence of land cover composition and ground-
water on thermal habitat availability for brook charr (Salveli-
nus fontinalis) populations in the United States of America.
Ecology of Freshwater Fish 20: 431–437.
Snyder, C.D., Hitt, N.P. & Young, J.A. 2015. Accounting for
groundwater in stream fish thermal habitat responses to cli-
mate change. Ecological Applications 5: 1397–1419.
Steen, P.J., Wiley, M.J. & Schaeffer, J.S. 2010. Predicting
future changes in Muskegon River watershed game fish dis-
tributions under future land cover alteration and climate
change scenarios. Transactions of the American Fisheries
Society 139: 396–412.
Stoner, A.M.K., Hayhoe, K., Yang, X.H. & Wuebbles, D.J.
2013. An asynchronous regional regression model for statis-
tical downscaling of daily climate variables. International
Journal of Climatology 33: 2473–2494.
Trout Unlimited. 2015. Michigan TU roster. Available at:
http://www.michigantu.org/images/pdffiles/documents/MITU
203
Resilience-based salmonid management
%20Council%20Roster092315pfm.pdf (accessed 19 June
2015).
Tsuboi, J., Iwata, T., Morita, K., Endou, S., Oohama, H. &
Kouichi, K. 2013. Strategies for the conservation and man-
agement of isolated salmonid populations: lessons from
Japanese streams. Freshwater Biology 58: 908–917.
Tyler, J.A. & Rutherford, E.S. 2007. River restoration effects
on steelhead populations in the Manistee River, Michigan:
Analysis using an individual-based model. Transactions of
the American Fisheries Society 136: 1654–1673.
United States Environmental Protection Agency (USEPA).
2005. National Hydrography Dataset Plus - NHDPlus.
Waco, K.E. & Taylor, W.W. 2010. The influence of
groundwater withdrawal and land use changes on brook
charr (Salvelinus fontinalis) thermal habitat in two cold-
water tributaries in Michigan, USA. Hydrobiologia 650:
101–116.
Wahl, K.L. & Wahl, T.L. 1988. Effects of regional ground
water declines on streamflows in the Oklahoma Panhandle.
In: Waterstone, M. & Burt, R.J., eds. Proceedings of the
symposium on water-use data for water resources manage-
ment. Tucson: American Water Resources Association, pp.
239–249.
Wehrly, K.E., Wiley, M.J. & Seelbach, P.W. 1997. Land-
scape-based models that predict July thermal characteristics
of lower Michigan rivers. Michigan Department of Natural
Resources, Fisheries Research Report 2037, Ann Arbor.
Wehrly, K.E., Wang, L. & Mitro, M. 2007. Field-based esti-
mates of thermal tolerance limits for trout: Incorporating
exposure time and temperature fluctuation. Transactions of
the American Fisheries Society 136: 365–374.
Wiley, M.J., Kohler, S.L. & Seelbach, P.W. 1997. Reconciling
landscape and local views of aquatic communities: lessons
from Michigan trout streams. Freshwater Biology 37: 133–
148.
Wiley, M.J., Hyndman, D.W., Pijanowski, B.C., Kendall,
A.D., Riseng, C., Rutherford, E.S., Cheng, S.T., Carlson,
M.L., Tyler, J.A., Stevenson, R.J., Steen, P.J., Richards,
P.L., Seelbach, P.W., Koches, J.M. & Rediske, R.R. 2010.
A multi-modeling approach to evaluating climate and land
use change impacts in a Great Lakes River Basin. Hydrobi-
ologia 657: 243–262.
Wurtsbaugh, W.A. & Davis, G.E. 1977. Effects of fish size
and ration level on the growth and food conversion effi-
ciency of rainbow trout, Salmo gairdneri Richardson. Jour-
nal of Fish Biology 11: 99–104.
Zorn, T.G., Seelbach, P.W. & Wiley, M.J. 2011. Developing
user-friendly habitat suitability tools from regional stream
fish survey data. North American Journal of Fisheries Man-
agement 31: 41–55.
Zorn, T.G., Seelbach, P.W. & Rutherford, E.S. 2012. A regio-
nal-scale habitat suitability model to assess the effects of flow
reduction on fish assemblages in Michigan streams. Journal
of the American Water Resources Association 48: 871–895.
204
Carlson et al.
