Accepted: 6 September 2016

DOI: 10.1111/ef.12320

ORIGINAL ARTICLE

Climate-­induced seasonal changes in smallmouth bass growth rate potental at the southern range extent

Christopher R. Middaugh1 | Brin Kessinger2 | Daniel D. Magoulick3

1Arkansas Cooperatve Fish & Wildlife Research Unit, Department of Biological Abstract Sciences, University of Arkansas, Temperature increases due to climate change over the coming century will likely Fayeteville, AR, USA afect smallmouth bass (Micropterus dolomieu) growth in lotc systems at the south- 2Evoluton, Ecology and Organismal Biology Department, Ohio State University, ern extent of their natve range. However, the thermal response of a stream to Columbus, OH, USA warming climate conditons could be afected by the fow regime of each stream, 3U.S. Geological Survey, AR Cooperatve mitgatng the efects on smallmouth bass populatons. We developed bioenerget- Fish & Wildlife Research Unit, Department of Biological Sciences, University of ics models to compare change in smallmouth bass growth rate potental (GRP) Arkansas, Fayeteville, AR, USA from present to future projected monthly stream temperatures across two fow Correspondence regimes: runof and groundwater-dominated. Seasonal diferences in GRP between Christopher R. Middaugh, Arkansas stream types were then compared. The models were developed for fourteen Cooperatve Fish & Wildlife Research Unit, Department of Biological Sciences, streams within the Ozark–Ouachita Interior Highlands in Arkansas, Oklahoma and University of Arkansas, Fayeteville, AR, Missouri, USA, which contain smallmouth bass. In our simulatons, smallmouth bass USA. −1 −1 Email: [email protected] mean GRP during summer months decreased by 0.005 g g day in runof streams and 0.002 g g−1 day−1 in groundwater streams by the end of century. Mean GRP during winter, fall and early spring increased under future climate ­conditons within both stream types (e.g., 0.00019 g g−1 day−1 in runof and 0.0014 g g−1 day−1 in groundwater streams in spring months). We found signifcant diferences in change in GRP between runof and groundwater streams in three seasons in end-­of-­ century simulatons (spring, summer and fall). Potental diferences in stream tem- perature across fow regimes could be an important habitat component to consider when investgatng efects of climate change as fshes from various fow regimes that are relatvely close geographically could be afected diferently by warming climate conditons.

KEYWORDS bioenergetcs, climate change, growth rate potental, smallmouth bass, water temperature

1 | INTRODUCTION of organisms, for example, by limitng salmonids to high elevatons in some regions (Flebbe, 1994) and limitng the lattudinal distributon Water temperature is one of the most important abiotc conditons of smallmouth bass (Micropterus dolomieu) (Shuter, MacLean, Fry, & afectng lotc systems. Water temperature infuences metabolic rates, Regier, 1980). However, climate change is expected to increase lotc growth rates and development of many diferent organisms (Naiman & temperatures over the coming century and many lotc systems in the Turner, 2000), and it can help determine growth and survival of fshes United States are already warming (Kaushal et al., 2010). This change (Christe & Regier, 1988; Magnuson, Crowder, & Medvick, 1979). in thermal habitat could alter current distributons of organisms by Water temperatures are also critcal to structuring the distributon opening new habitats to colonisaton (e.g., smallmouth bass range

Ecol Freshw Fish. 2018;27:19–29. wileyonlinelibrary.com/journal/ef © 2016 John Wiley & Sons A/S. | 19 Published by John Wiley & Sons Ltd 20 | Middaugh et al. expanding farther north; Vander Zanden, Olden, Thorne, & Mandrak, This is likely due in part to higher thermal optma of largemouth bass 2004) and restrictng the range of species adapted to cooler water leading to a compettve advantage in warmer streams (26°C; Zweifel (e.g., Bull trout (Salvelinus confuentus); Isaak et al., 2010). et al., 1999). However, certain aspects of the fow regime, for exam- In additon to thermal paterns, hydrologic regime is also a crit- ple, the amount of groundwater infuence, could infuence compettve cal structuring component of stream ecosystems (Pof et al., 2010). interactons and beneft smallmouth bass populatons. Streams with Accordingly, many regions across the United States have had streams large amounts of groundwater input could provide thermal habitat classifed by fow regimes, such as groundwater and runof (e.g., that favours growth of smallmouth bass (Westhof & Paukert, 2014). In Leasure, Magoulick, & Longing, 2016). These classifcaton regimes additon, streams with high levels of groundwater input could increase provide clear, easily interpreted and ecologically relevant methods of recruitment of smallmouth bass relatve to non-­spring-­fed streams relatng stream hydrologic regime to ecological data (Pof et al., 2010). (Brewer, 2013). No previous work has compared smallmouth bass Groundwater-domina­ ted streams are expected to exhibit very difer- growth in groundwater-­dominated versus runof-­dominated streams. ent thermal paterns than runof-­dominated streams, with cooler water We examined how climate change could alter stream water tem- temperatures during summer months, and warmer temperatures during peratures in the Ozark–Ouachita Interior Highlands and how this winter months (Whitledge, Rabeni, Annis, & Sowa, 2006). In additon, could afect the growth rate potental (GRP) of smallmouth bass. streams with more groundwater input are likely to be less sensitve to Flow regimes have recently been predicted for every stream in the changes in warming air temperatures which could afect how fshes re- Ozark–Ouachita Interior Highlands region of Arkansas, Oklahoma and spond to climate change (Carlson, Taylor, Schlee, Zorn, & Infante, 2015). Missouri (Leasure et al., 2016). These fow regime classifcatons were Smallmouth bass is a warm-­water riverine species broadly dis- developed from ten hydrologic variables and provide a classifcaton tributed throughout North America. Because of their ecological im- of streams into three broad categories: groundwater, runof and in- portance as apex predators in many lotc systems (Rabeni, 1992), termitent with subcategories within each (Leasure et al., 2016). All smallmouth bass is a commonly studied species with many of their of our study streams were classifed as either groundwater or runof. life history traits characterised (Brewer & Orth, 2015) including bio- Groundwater streams are characterised by more fow stability and energetcs parameters (Whitledge, Hayward, Zweifel, & Rabeni, higher base fows than runof streams (Leasure et al., 2016). Many 2003). Although smallmouth bass are locally abundant in streams in runof streams have multple days of no fow each year in this region the Ozark–Ouachita Interior Highlands, this region is at the southern (Leasure et al., 2016). We did not include any intermitent streams as extent of smallmouth bass natve range and seasonal water tempera- many of these do not support year-­round smallmouth bass popula- tures exceed optmal growth levels (22°C; Zweifel, Hayward, & Rabeni, tons (Magoulick, 2000). The fow classifcatons that we use are not 1999) in many systems during a typical year. Air temperatures are ex- based on the actual groundwater contributon to the system, but in- pected to increase in the Ozark–Ouachita Interior Highlands due to stead are based on a wide range of hydrologic variables (Leasure et al., climate change over the coming century (Alder & Hostetler, 2013), po- 2016). We compared changes in seasonal GRP among streams from tentally raising some stream temperatures past habitable conditons both fow classifcatons to determine how smallmouth bass growth for smallmouth bass. could be afected diferentally between fow regimes. Although previous work has examined a potental northward ex- pansion of smallmouth bass range due to climate change (Dunlop & 2 | MATERIALS AND METHODS Shuter, 2006; Vander Zanden et al., 2004), litle work has examined how smallmouth bass could be afected by climate change at the 2.1 | Air and water temperatures southern porton of their range. Some fshes are expected to have truncated ranges due to climate change (e.g., brook trout (Salvelinus We selected fourteen streams in the Ozark–Ouachita Interior fontnalis); Meisner, 1990), but some previous work has not indicated Highlands of Arkansas, Missouri and Oklahoma that contained small- this for smallmouth bass (Mohseni, Stefan, & Eaton, 2003). In a com- mouth bass and had United States Geological Survey (USGS) river parison of smallmouth bass growth rate in four streams across a lattu- gages that collected water temperature data (Figure 1; Appendix 1). All dinal gradient, Pease and Paukert (2014) indicated a potental increase temperature data for each site were downloaded (range 1–11 years; in smallmouth bass prey consumpton and growth due to climate Appendix 1), and a monthly mean temperature for each month of change across their range. However, they did not take into account data was calculated at each site. If a month had fewer than 20 days of variaton among stream types or factors other than temperature which temperature data collected, it was excluded from analyses. Next, we could afect growth. collected monthly mean air temperatures (from the Natonal Oceanic Other research indicates that high water temperatures at the and Atmospheric Agency Natonal Center for Climate Data) for the southern extent of smallmouth bass range can be detrimental to their county where each site was located for a corresponding tme period growth and abundance. In the Ozark border region of Missouri, small- to the water temperature data. We then calculated a least-­squares mouth bass have shown declines with replacement by their compet- linear regression for each site to predict stream temperature from air tor, largemouth bass (Micropterus salmoides) (Sowa & Rabeni, 1995). In temperature data (Table 1). partcular, streams with higher maximum summer temperatures and Afer calculatng a predictve relatonship between air and water more pool area had a stronger displacement (Sowa & Rabeni, 1995). temperatures, we used the USGS Natonal Climate Change Viewer Middaugh et al. | 21

FIGURE 1 Location of sites across the Ozark–Ouachita Interior Highlands of Arkansas, Oklahoma and Missouri (shaded in grey). Each site corresponds to a USGS river gage which collects stream temperature data. Sites from runoff streams are designated with a triangle, and sites from groundwater streams are designated with a circle

TABLE 1 Linear regression results River Intercept Coefcient R2 Stream type relatng air and water temperatures from each site. Air temperature data were Bear Creek 10.04 .37 .79 Groundwater collected from the NOAA Natonal Climate Beaty Creek 6.23 .65 .92 Groundwater and Data Center, and water temperatures East Fork Black River 2.66 .84 .96 Groundwater were collected from USGS water temperature gages. All regression models Huzzah Creek 5.56 .97 .98 Groundwater were signifcant at p < 0.001 Jacks Fork 4.16 .88 .97 Groundwater Osage Creek 10.00 .49 .79 Groundwater Spavinaw Creek 8.66 .49 .88 Groundwater Sylamore Creek 4.03 .76 .97 Groundwater Big Creek 3.22 .82 .98 Runof Big River 1.80 .96 .99 Runof Bufalo River 4.67 .77 .96 Runof Illinois Bayou 1.85 .92 .96 Runof South Fork Litle Red 2.34 .88 .97 Runof Tavern Creek 2.41 .95 .99 Runof 22 | Middaugh et al.

(Alder & Hostetler, 2013) to determine modelled historical air tem- 2.2 | Bioenergetcs modelling peratures (1950–2005) and future air temperatures under climate change for two decades (2040s and 2090s). The Natonal Climate We created a bioenergetcs simulaton based on the Wisconsin bio- Change Viewer provides monthly minimum and maximum tem- energetcs model (Hansen, Johnson, Kitchell, & Schindler, 1997) but peratures downscaled to the county level for each year during the coded in program R (R Development Core Team 2008) and param- historical period, and projected monthly minimum and maximum tem- eterised for smallmouth bass (Shuter & Post, 1990; Whitledge et al., peratures for each year at the same resoluton untl the year 2099. We 2003, 2006; Wrenn, 1980; Zweifel et al., 1999). All simulatons were selected an emissions scenario of RCP 8.5 (highest CO2 emissions) run for a 300-­g fsh (275 mm (Kolander, Willis, & Murphy, 1993); ap- and used the ensemble average temperature predictons of 30 mod- proximately age 4 in the Bufalo River, AR (Whisenant & Maughan, els. We chose to only model the highest climate scenario because 1989)). We determined simulated diet based on diet samples collected of the exploratory nature of this analysis and because we focus on using gastric lavage (Kamler & Pope, 2001) from smallmouth bass cap- a comparison of two diferent future tme periods which provides a tured in the Bufalo River, AR, in May–September 2014 (n = 34; 73% contrast between a more and less severe change in air temperature. crayfsh, 24% fsh and 3% macroinvertebrate). These proportons are We calculated the mean minimum and maximum temperature of each similar to what other researchers have found for smallmouth bass month for the county corresponding to each of our streams during diets (Dauwalter & Fisher, 2008; Rabeni, 1992). Energy density of prey present and both future tme periods. Monthly mean maximum and was set at 3,063 J/g for crayfsh (Procambarus crayfsh; Eggleton & minimum temperatures were averaged to produce a monthly mean Schramm, 2002), 3,853 J/g for fsh prey (fathead minnow (Pimephales temperature for each tme period. These monthly mean temperatures promelas); Whitledge et al., 2003) and 3,421 J/g for aquatc insect lar- were then used to predict monthly mean water temperatures for each vae (average value of Chironomidae, Odonata and Ephemeroptera; site at each tme period using the previously calculated linear rela- Cummins & Wuycheck, 1971). tonships. This approach assumes no changes in river conditons (e.g., The bioenergetics model was used to calculate GRP for small- hydrologic paterns) over the coming century other than temperature mouth bass at each month for each time period. Calculating GRP al- changes. lows assessment of the effects of water temperature on smallmouth

Bear Creek Beaty Creek East Fork Black River Huzzah Creek 15 25 15 25 15 25 05 0 5 15 25 05 05 –5 0510 15 20 25 30 –5 0510 15 20 25 30 –5 0510 15 20 25 30 –5 0510 15 20 25 30 Jacks Fork Osage Creek Spavinaw Creek Sylamore Creek roundwater C) 15 25 15 25 15 25 15 25 o 05 05 05 05 –5 0510 15 20 25 30 –5 0510 15 20 25 30 –5 0510 15 20 25 30 –5 0510 15 20 25 30 Big Creek Big River Buffalo River Illinois Bayou Water temperature ( 5 25 15 25 15 25 51 05 0 0 5 15 25 05 –5 0510 15 20 25 30 –5 0510 15 20 25 30 –5 0510 15 20 25 30 –5 0510 15 20 25 30 Kiamichi River South Fork Little Red Tavern Creek Runof fG 5 5 25 52 15 25 51 51 0 0 05 –5 0510 15 20 25 30 –5 0510 15 20 25 30 –5 0510 15 20 25 30 Air temperature ( oC)

FIGURE 2 Linear regression models for each site relating mean monthly air temperature (from NOAA National Climate Data Center) and mean monthly water temperature (from USGS river gage specific to each site). Model results for each site are shown in Table 1 Middaugh et al. | 23 bass growth with all other factors being held constant (e.g., Coulter, 3 | RESULTS Sepúlveda, Troy, & Höök, 2014). We conducted ten GRP models for each month, altering the proportion of maximum consumption (p Air and water temperatures were signifcantly linearly related at all value) from 0.1 to 1.0 at 0.1 increments and outputting growth in sites with all R2 values >0.78 (Table 1; Figure 2). Groundwater streams −1 −1 grams of growth per gram of fish per day (g g day ). This range typically showed less variaton in temperature over the course of the of consumption values is much wider than a fish likely experiences year than runof streams (Figure 3). Climate simulatons indicated that naturally, but we chose to model this wide range to examine ef- across our sites, future yearly mean air temperatures will increase by fects of climate change even at unrealistically high consumption an average of 2.5°C by mid-­century and increase by an average of levels. This approach models all potential scenarios, from very low 5.8°C by the end of century. Future predicted water temperatures prey availability to assuming that prey availability will increase to corresponded with predicted future air temperatures (Figure 3). compensate for increasing smallmouth bass consumption. This was Growth rate potental models indicated a general increase in repeated for each stream and for each time period (present, mid-­ GRP for smallmouth bass during most winter, early spring and fall century and end of century). Next, mean growth rate potential for months and a general decrease in GRP during late-­spring and sum- each month in each stream was calculated and a seasonal change in mer months relatve to present water temperatures (Figure 4). GRP from present to mid-­century or end of century for each stream Runof streams typically showed an increase in GRP during winter, calculated (winter: January–March; spring: April–June; summer: spring and fall months (e.g., 0.001 g g−1 day−1 increase in March July–September; fall: October–December). Changes in growth GRP by mid-­century; 0.003 g g−1 day−1 increase in March GRP by rate potential were compared between stream types using paired the end of century), but a decrease in GRP during summer months t-­tests for each season at both time periods. All data analyses were (e.g., 0.001 g g−1 day−1 decrease in August GRP by mid-­century, conducted in program R. 0.006 g g−1 day−1 decrease in August GRP by the end of century).

Bear Creek Beaty Creek East Fork Black River Huzzah Creek 0 0 0 30 0203 02 0203 01 01 0 10 20 30 01 246810 12 24681012 246810 12 246810 12 Jacks Fork Osage Creek Spavinaw Creek Sylamore Creek roundwater 0 0 03 C) o 02 0203 0 10 20 30 0 10 20 30 01 01 246810 12 24681012 246810 12 246810 12 Big Creek Big River Buffalo River Illinois Bayou 0 0 03 Water temperature ( 02 0203 01 0 10 20 30 01 0 10 20 30 246810 12 24681012 246810 12 246810 12 South Fork Little Red Tavern Creek Runof fG

0 Time period Present Mid−century

0203 End of century 01 0 10 20 30 246810 12 24681012

Month

FIGURE 3 Predicted historical (1950–2005; solid line), mid-­century (2040–2049; long dash) and end of century (2090–2099; short dash) mean monthly stream temperatures calculated using linear regression predictive models for each stream and air temperature data from the USGS Climate Change Viewer ensemble average of all models. The top two rows are groundwater streams, and the bottom two rows are runoff streams 24 | Middaugh et al.

Bear Creek Beaty Creek East Fork Black River Huzzah Creek 0.025 0.010 0.025 0.010 0.010 – 0.005 0.025 – 0.005 0.010 0.025 – 0.005 1357911 1357911 13579111– 0.005 357911 )

– 1 Jacks Fork Osage Creek Spavinaw Creek Sylamore Creek ×day 0.025 Groundwater 0.025 – 1 0.010 0.010 0.010 0.010 – 0.005 0.025 – 0.005 – 0.005 0.025 135791113 57911 1357911 – 0.005 1357911 Big Creek Big River Buffalo River Illinois Bayou .025 0.025 0.025 0.010 0.010 0.010 Growth rate potential (g×g – 0.00 50 – 0.005 – 0.005 – 0.005 0.0101357 0.025 91113 57911 1357911 1357911 South Fork Little Red Tavern Creek Runoff Time period .025 .025 Mid-century End of century 0.010 – 0.005 0.01 00 – 0.00 50 135791113 57911 Month

FIGURE 4 Results of all growth rate potential models for adult smallmouth bass (p value range = .1–1) for present climate conditions (box plots), and mean of all growth rate potential models at mid-­century (solid line) and end of century (dashed line). The top two rows are groundwater streams, and the bottom two rows are runoff streams

Groundwater streams showed similar trends, but at a lower magni- 4 | DISCUSSION tude (e.g., 0.001 g g−1 day−1 and 0.003 g g−1 day−1 increase in March GRP for mid- ­ and end of century respectvely; 0.001 g g−1 day−1 and Our models predict that an increase in stream warming across the 0.003 g g−1 day−1 decrease in August GRP for mid- ­ and end of cen- Ozark–Ouachita Interior Highlands due to climate change will lead to an tury respectvely). increase in GRP for smallmouth bass during winter, early spring and fall In general, runof streams demonstrated more extreme changes months. However, late-­spring and summer months warmed past opt- in temperature leading to larger changes in seasonal growth rate mal temperature for smallmouth bass growth, resultng in a reducton in potental (Figures 2 and 5). Runof streams had signifcantly GRP for smallmouth bass in most of the modelled streams. Although the more positve changes in growth rate potental than groundwa- populaton-­level efects of reduced summer growth on smallmouth bass ter streams in both mid-­century and end-­of-­century simulatons are unknown, it is possible that reduced GRP during these very warm during fall months (mid-­century t = −3.54, p = .007; end of century months could lead to largemouth bass and spoted bass (Micropterus t = −3.18, p = .014; Figure 6). Groundwater streams had a more punctulatus) having a compettve advantage over smallmouth bass due positve change in GRP than runof streams in end-­of-­century sim- to higher thermal optma (Zweifel et al., 1999), potentally leading to ulatons during spring months (t = 2.80, p = .016; Figure 6). During local declines or extrpaton (Sowa & Rabeni, 1995). late spring months, runof streams were experiencing a decline in Smallmouth bass GRP was altered signifcantly diferently between GRP leading to a lower change in GRP from present conditons groundwater and runof streams in multple seasons. Previous work than seen in groundwater streams (Figure 5). Runof streams had has indicated that spring-­fed streams could have higher smallmouth a signifcantly more negatve change in growth rate potental than bass growth relatve to warmer, non-­spring-­fed streams (Whitledge groundwater streams in end-o­ f-c­ entury simulatons during summer et al., 2006). In additon, empirical work found that spring-­fed streams months (t = 2.99, p = .012; Figure 6). Middaugh et al. | 25

Bear Creek Beaty Creek East Fork Black River Huzzah Creek 0.005 0.005 0.005 – 0.005 – 0.005 –0.005 0.005 – 0.005 ) 1 – 246810 12 246810 12 246810 12 246810 12 Jacks Fork Osage Creek Spavinaw Creek Sylamore Creek ×day 1 – 0.005 0.005 0.005 0.005 Groundwater – 0.005 – 0.005 – 0.005 – 0.005

246810 12 246810 12 246810 12 246810 12 Big Creek Big River Buffalo River Illinois Bayou 0.005 0.005 0.005 0.005 – 0.005 – 0.005 – 0.005 – 0.005 Change in growth rate potential (g×g 246810 12 246810 12 246810 12 246810 12

Runoff South Fork Little Red Tavern Creek Time period 0.005 0.005 Mid-century End of century – 0.005 – 0.005

246810 12 246810 12 Month

FIGURE 5 Change in growth rate potential from present for each stream at both time periods. The solid horizontal line indicates no change from present. The top two rows are groundwater streams, and the bottom two rows are runoff streams

have higher abundances of juvenile smallmouth bass, that is, beter to four groundwater streams, and all seven runof streams exceed this reproducton (Brewer, 2013) and lotc smallmouth bass abundance threshold for at least one month out of the year. Similarly, Bouska, in Missouri are positvely associated with higher spring-­fow volumes Whitledge, and Lant (2015) identfed a threshold upper mean annual (Brewer, Rabeni, Sowa, & Annis, 2007). Groundwater input may also maximum air temperature of 31°C where smallmouth bass occurrence provide thermal resistance to streams in Michigan, promotng the con- was not associated with streams across the central United States that servaton of salmonids susceptble to the efects of climate change experienced air temperatures exceeding this level. By the end of cen- (Carlson et al., 2015). Our models predict that smallmouth bass in tury, air temperatures associated with all of our modelled streams are groundwater streams will not decline in growth rate potental during predicted to exceed this level, but water temperatures are not likely to summer months to the extent of runof streams, indicatng the impor- reach lethal limits for smallmouth bass (37°C; Wrenn, 1980). tance of these habitats for smallmouth bass populatons during future It is possible that adult smallmouth bass could behaviourally adapt temperature regimes. to changing water temperatures. Adult smallmouth bass can migrate Smallmouth bass have an upper mean weekly temperature thresh- to regulate body temperature during winter months in Ozark streams old of 27°C for sustained positve growth and viable populatons with groundwater inputs (Peterson & Rabeni, 1996; Westhof, Paukert, (Whitledge et al., 2006). Above this temperature, one would expect re- Etnger-­Dietzel, Dodd, & Siepker, 2016). Therefore, it is plausible duced annual growth, conditon and fecundity (Bagenal, 1967). Under that smallmouth bass could regulate body temperature during sum- present climate conditons, no streams that we modelled have a mean mer months by utlising spatal variatons in thermal habitat, similar to monthly temperature >27°C at any point. By mid-­century, this thresh- salmonids (e.g., rainbow trout; Ebersole, Liss, & Frissell, 2001). Many old is expected to be surpassed by one groundwater stream (Jack’s streams in this region are thermally heterogeneous with cooler water Fork) for at least one month of the year, and in six runof streams for patches due to stream stratfcaton (Mathews, 1998) and ground- at least one month of the year. By the end of century, this increases water inputs (Westhof & Paukert, 2014). Groundwater inputs into 26 | Middaugh et al.

Mid-century End of century

* 00.0 )

–1 *

200. * x day 04 –1 (gxg

laitnetopetarhtworgniegnahC * 00.0 00.0 002 0.000 0.002 0.004 –0. –0

FIGURE 6 Mean seasonal change in 00.0

004 growth rate potential from present for –0. –2 Flow classification both flow classifications at both time periods. Asterisks indicate a significant Groundwater difference (p < .05) in the mean value of the 600.0 Runoff

006 change in growth rate potential between –0. –4 Winter Spring SummerFall Winter Spring Summer Fall groundwater and runoff streams based on a paired t-­test. Error bars show standard Season error

streams can range from large springs (e.g., Alley Spring on the Jack’s ofen strongly related (Erickson & Stefan, 2000). However, these rela- Fork river; discharge = 2.7 m3/s; Mugel, Richards, & Schumacher, tonships are not as strong at temperature extremes and correlatons 2009) with long downstream infuence (Westhof & Paukert, 2014), tend to plateau at high temperatures (Erickson & Stefan, 2000). We did to small seepages associated with channel curvature with localised not observe any levelling of stream temperatures at high air tempera- infuence (Dugdale, Bergeron, & St-­Hilaire, 2015). Smallmouth bass tures (Figure 2), and our regressions were all signifcant and well ft. could migrate (Westhof et al., 2016) or use other behavioural adap- Although there may be a risk of over-­predictng stream temperatures tatons to utlise these temperature refuges. However, these refuges at the very warmest air temperatures, we believe our water tempera- may be lacking in many runof-­dominated streams and typical summer ture predictons were adequate for our modelling eforts because of drought conditons lead to isolated pools in many runof streams in our coarse tmescale and our strong regressions that we created spe- this region (Homan, Girondo, & Gagen, 2005; Magoulick, 2000). This cifcally for each stream. Further, none of our streams were headwater drying would prevent migratons of smallmouth bass in many runof streams (all either third or fourth order) which reduces variability asso- streams during summer months (Hafs, Gagen, & Whalen, 2010). More ciated with small streams. Linear relatonships between air and water data are needed on groundwater inputs into many of these systems temperatures can strongly predict both groundwater-­dominated and to beter understand available thermal refuges. The streams that we runof-­dominated streams, although groundwater streams typically modelled encompass the range of several subspecies of smallmouth have lower regression slopes (Erickson & Stefan, 2000). Similar to our bass (Interior Highland intergrade and Neosho smallmouth bass results, Whitledge et al. (2006) found strong predictve relatonships (Micropterus dolomieu velox); Brewer & Orth, 2015). We do not atempt between air and water temperature in both spring-­fed and non-spring-­ ­ to account for diferences in bioenergetcs parameters for these sub- fed streams in the Ozark Highlands of Missouri. species as no data exist; however, many of the parameters that we Another limitaton to our study is that our temperature data are use were developed in the Ozarks of Missouri (Whitledge et al., 2003). based solely on a single gage locaton for a river. In some cases, this Diferent subspecies could have diferent thermal tolerances afectng gage could be located very close to a spring and could be heavily infu- how fsh adapt and respond to climate change. enced by groundwater (Westhof & Paukert, 2014). All of our streams There are several limitatons to our study. For example, our water have groundwater input to some extent, and the locaton of these temperature predictons are relatvely simplistc. Linear air and water inputs could have afected our temperature predictons. However, temperature relatonships are commonly used to examine the pre- as we include at least six streams from each stream type, we assume dicted efects of climate change on future stream water temperatures that not all are afected in such a way. It is also possible that there (e.g., Almodóvar, Nicola, Ayllón, & Elvira, 2012; Carlson et al., 2015; may be thermal refuges present in both stream types that our mod- Pease & Paukert, 2014; Peterson & Kwak, 1999). Predictve relaton- els do not simulate. We model average stream temperatures and do ships between air and water temperatures generally improve at longer not atempt to simulate the heterogeneous thermal environment that tmescales (Morrill, Bales, & Conklin, 2005) and at the monthly scale are these streams likely contain. Another limitaton is the limited amount Middaugh et al. | 27 of water temperature data that we have for each site. For example, REFERENCES temperature predictons from Tavern Creek are based on only one year Alder,. J R., & Hostetler, S. W. (2013). USGS Natonal Climate Change Viewer. of water temperature data. However, all air and water temperature lin- US Geological Survey Retrieved from htp://www.usgs.gov/climate_ ear regressions were highly signifcant and a good ft, so we chose to landuse/clu_rd/nccv.asp. include all sites to bolster our sample size. Allan, J. D. (2004). Landscapes and riverscapes: the infuence of land use on stream ecosystems. Annual Review of Evoluton, Ecology, and Systematcs, Previous work has indicated some ways to mitgate the efects of 35, 257–284. increasing temperatures due to climate change. Stressors for small- Almodóvar, A., Nicola, G. C., Ayllón, D., & Elvira, B. (2012). Global warming mouth bass include destructon of riparian habitat leading to decreased threatens the persistence of Mediterranean brown trout. Global Change shade and increased water temperatures (Whitledge et al., 2006), Biology, 18, 1549–1560. Bagenal, T. B. (1967). A short review of fsh fecundity. In S. D. Gerking (Ed.), gravel mining (Brown, Lytle, & Brown, 1998) and other anthropogenic The biological basis of freshwater fsh producton (pp. 89–111). New York, land-­use alteratons (Allan, 2004). Increasing riparian shading through NY: Wiley and Sons Inc. restoraton eforts can decrease the daily temperature fuctuatons in Bouska, K. L., Whitledge, G. W., & Lant, C. (2015). Development and evalu- streams; however, this is most pronounced in low-­order (narrow-­width) aton of species distributon models for fourteen natve central U.S. fsh species. Hydrobiologia, 747, 159–176. streams (Whitledge et al., 2006). Mitgaton eforts can also focus on Brewer, S. K. (2013). Groundwater infuences on the distributon and abun- minimising habitat degradaton leading to the formaton of pools and dance of riverine smallmouth bass, Micropterus dolomieu, in pasture siltaton of substrates (Waters, 1995). Siltaton and increased pool area landscapes of the Midwestern USA. River Research and Applicatons, 29, can cause a reducton in preferred smallmouth bass prey and poten- 269–278. Brewer, S. K., & Orth, D. J. (2015). Smallmouth bass Micropterus dolo- tally increase risk of being outcompeted by largemouth bass (Sowa & mieu Lacepéde, 1802. In M. D. Tringali, J. M. Long, T. W. Birdsong Rabeni, 1995). Protectng thermal habitat is of even higher importance & M. S. Allen (Eds.), Black bass diversity: Multdisciplinary science for as several species of natve Ozark crayfsh have similar optmal growth conservaton (pp. 9–26). Bethesda, MD: American Fisheries Society, temperatures to smallmouth bass (22°C; Whitledge & Rabeni, 2002). Symposium 82. In general, groundwater streams are predicted to experience a much Brewer, S. K., Rabeni, C. F., Sowa, S. P., & Annis, G. (2007). Natural land- scape and stream segment atributes infuencing the distributon and smaller reducton in GRP during summer months and could provide relatve abundance of riverine smallmouth bass in Missouri. North a thermal refuge for smallmouth bass and other species not adapted American Journal of Fisheries Management, 27, 326–341. to warm temperatures. However, seasonal drying and drought condi- Brown, A. B., Lytle, M. M., & Brown, K. B. (1998). Impacts of gravel mining tons could prevent access to these refuges (Magoulick & Kobza, 2003). on gravel bed streams. Transactons of the American Fisheries Society, 127, 979–994. Regulatng fow levels and water withdrawal to preserve a minimum Carlson, A. K., Taylor, W. W., Schlee, K. M., Zorn, T. G., & Infante, D. M. fow when possible could also facilitate access of smallmouth bass to (2015). Projected impacts of climate change on stream salmonids with potental thermal habitat from groundwater inputs. Excessive water implicatons for resilience-­based management. Ecology of Freshwater withdrawals could lead to higher summer temperatures in groundwater Fish. doi:10.1111/ef.12267 Christe, G. C., & Regier, H. A. (1988). Measures of optmal thermal habitat streams, diminishing their refuge qualites for smallmouth bass. and their relatonship to yields for 4 commercial fsh species. Canadian In conclusion, we found that smallmouth bass will likely experience Journal of Fisheries and Aquatc Sciences, 45, 301–314. an increase in growth rate potental during cool months, but a decrease Coulter, D. P., Sepúlveda, M. S., Troy, C. D., & Höök, T. O. (2014). Thermal in growth rate potental during late-­spring and summer months over the habitat quality of aquatc organisms near power plant discharges: po- tental exacerbatng efects of climate warming. Fisheries Management course of the 21st century. We found that fow regime was a useful pre- and Ecology, 21, 196–210. dictor of which streams may have the strongest temperature change, Cummins, K. W., & Wuycheck, J. C. (1971). Caloric equivalents for inves- and in partcular, we suggest that monitoring and conservaton eforts tgatons in ecological energetcs. Miteilung Internatonale Vereinigung should be focused on runof fow regimes as smallmouth bass in these fuer Theorestsche und Angewandte Limnologie, 18, 1–151. Dauwalter, D. C., & Fisher, W. L. (2008). Ontogenetc and seasonal diet streams will be subject to a greater change in temperature over the com- shifs of smallmouth bass in an Ozark stream. Journal of Freshwater ing century. More work is needed to explore how these growth changes Ecology, 23, 113–121. could afect smallmouth bass at the populaton level, but it is possible Dugdale, S. J., Bergeron, N. E., & St-Hilaire, A. (2015). Spatal distributon that elevated summer temperatures will leave them more at risk of being of thermal refuges analysed in relaton to riverscape hydromorphology outcompeted by other Micropterus species, especially in runof streams. using airborne thermal infrared imagery. Remote Sensing of Environment, 160, 43–55. Dunlop, E. S., & Shuter, B. J. (2006). Natve and introduced populatons of smallmouth bass difer in concordance between climate and so- ACKNOWLEDGEMENTS matc growth. Transactons of the American Fisheries Society, 135, We thank the Natonal Science Foundaton for supportng B.K. 1175–1190. Ebersole, J. L., Liss, W. J., & Frissell, C. A. (2001). Relatonship between through a Research Experience for Undergraduates fellowship. We stream temperature, thermal refugia and rainbow trout Oncorhynchus also thank C. Paukert, J. Westhof and R. Fournier for providing com- mykiss abundance in arid-­land streams in the northwestern United ments that improved this manuscript. Two anonymous reviewers also States. Ecology of Freshwater Fish, 10, 1–10. provided helpful suggestons. Any use of trade, frm or product names Eggleton, M. A., & Schramm, H. L. (2002). Caloric densites of selected fsh prey organisms from the lower Mississippi River. Journal of Freshwater is for descriptve purposes only and does not imply endorsement by Ecology, 17, 409–414. the U.S. Government. 28 | Middaugh et al.

Erickson,. T R., & Stefan, H. G. (2000). Linear air/water temperature cor- Pof, N. L., Richter, B., Arthington, A. H., Bunn, S. E., Naiman, R. J., Kendy, relatons for streams during open water periods. Journal of Hydrologic E., … Warner, A. (2010). The Ecological Limits of Hydrologic Alteraton Engineering, 5, 317–321. (ELOHA): A new framework for developing regional environmental Flebbe, P. A. (1994). A regional view of the margin: salmonid abundance fow standards. Freshwater Biology, 55, 147–170. and distributon in the southern Appalachian Mountains of North R Development Core Team (2008). R: A language and environment for statst- Carolina and Virginia. Transactons of the American Fisheries Society, 123, cal computng. Vienna, Austria: R Foundaton for Statstcal Computng. 657–667. ISBN 3-900051-07-0. Retrieved from htp://www.R-project.org. Hafs, A. W., Gagen, C. J., & Whalen, J. K. (2010). Smallmouth bass sum- Rabeni, C. F. (1992). Trophic linkage between stream centrarchids and their mer habitat use, movement, and survival in response to low fow crayfsh prey. Canadian Journal of Fisheries and Aquatc Sciences, 49, in the Illinois Bayou, Arkansas. North American Journal of Fisheries 1714–1721. Management, 30, 604–612. Shuter, B. J., MacLean, J. A., Fry, F. E. J., & Regier, H. A. (1980). Stochastc Hansen, P., Johnson, T., Kitchell, J., & Schindler, D. E. (1997). Fish bioen- simulaton of temperature efects on frst year survival of smallmouth ergetcs [version] 3.0. Madison, WI: University of Wisconsin Sea Grant bass. Transactons of the American Fisheries Society, 109, 1–34. Insttute. Shuter, B. J., & Post, J. R. (1990). Climate, populaton viability, and the zoo- Homan, J. M., Girondo, N. M., & Gagen, C. J. (2005). Quantfcaton and geography of temperate fshes. Transactons of the American Fisheries predicton of stream dryness in the interior highlands. Journal of the Society, 119, 314–336. Arkansas Academy of Sciences, 59, 95–100. Sowa, S., & Rabeni, C. F. (1995). Regional evaluaton of the relaton of hab- Isaak, D. J., Luce, C. H., Rieman, B. E., Nagel, D. E., Peterson, E. E., Horan, D. itat to distributon and abundance of smallmouth bass and largemouth L., … Chandler, G. L. (2010). Efects of climate change and wildfre on bass in Missouri streams. Transactons of the American Fisheries Society, stream temperatures and salmonid thermal habitat in a mountain river 124, 240–251. network. Ecological Applicatons, 20, 1350–1371. Vander Zanden, M. J., Olden, J. D., Thorne, J. H., & Mandrak, N. E. Kamler, J. F., & Pope, K. L. (2001). Nonlethal methods of examining fsh (2004). Predicting occurrences and impacts of smallmouth bass stomach contents. Reviews in Fisheries Science, 9, 1–11. introductions in north temperate lakes. Ecological Applications, 14, Kaushal, S. S., Likens, G. E., Jaworski, N. A., Pace, M. L., Sides, A. M., Seekell, 132–148. D., … Wingate, R. L. (2010). Rising stream and river temperatures in Waters, T. F. (1995). Sediment in streams: Sources, biological efects and the United States. Fronters in Ecology and the Environment, 8, 461–466. control. American Fisheries Society Monograph 7. Bethesda, MD: Kolander, T. D., Willis, D. W., & Murphy, B. R. (1993). Proposed revision of American Fisheries Society. the standard weight (Ws) equaton for smallmouth bass. North American Westhof, J. T., & Paukert, C. P. (2014). Climate change simulatons predict Journal of Fisheries Management, 13, 398–400. altered biotc response in a thermally heterogeneous stream system. Leasure, D. R., Magoulick, D. D., & Longing, S. D. (2016). Natural fow re- PLoS ONE, 9(10), e111438. gimes of the Ozark-­Ouachita Interior Highlands region. River Research Westhof, J. T., Paukert, C., Etnger-Dietzel, S., Dodd, H., & Siepker, M. and Applicatons, 32, 18–35. (2016). Behavioural thermoregulaton and bioenergetcs of riverine Magnuson, J. J., Crowder, L. B., & Medvick, P. A. (1979). Temperature as an smallmouth bass associated with ambient cold-­period thermal refuge. ecological resource. American Zoologist, 19, 331–343. Ecology of Freshwater Fish, 25, 72–85. Magoulick, D. D. (2000). Spatal and temporal variaton in fsh assemblages Whisenant, K. A., & Maughan, E. (1989). Smallmouth bass and Ozark bass of drying stream pools: The role of abiotc and biotc factors. Aquatc in Bufalo Natonal River. Cooperatve Natonal Park Resources Studies Ecology, 34, 29–41. Unit Technical Report No. 28. Magoulick, D. D., & Kobza, R. M. (2003). The role of refugia for fshes during Whitledge, G. W., Hayward, R. S., Zweifel, R. D., & Rabeni, C. F. (2003). drought: a review and synthesis. Freshwater Biology, 48, 1186–1198. Development and laboratory evaluaton of a bioenergetcs model Mathews, W. J. (1998). Paterns in freshwater fsh ecology. New York, NY: for subadult and adult smallmouth bass. Transactons of the American Chapman and Hall. Fisheries Society, 132, 316–325. Meisner, J. D. (1990). Efect of climatc warming on the southern margins Whitledge, G. W., & Rabeni, C. F. (1997). Energy sources and ecological of the natve range of brook trout, Salvelinus fontnalis. Canadian Journal role of crayfshes in an Ozark stream: insights from stable isotopes of Fisheries and Aquatc Sciences, 47, 1065–1070. and gut analysis. Canadian Journal of Fisheries and Aquatc Sciences, 54, Mohseni, O., Stefan, H. G., & Eaton, J. G. (2003). Global warming and 2555–2563. potental changes in fsh habitat in U.S. streams. Climate Change, 59, Whitledge, G. W., & Rabeni, C. F. (2002). Maximum daily consumpton 389–409. and respiraton rates at four temperatures for fve species of crayfsh Morrill, J. C., Bales, R. C., & Conklin, M. H. (2005). Estmatng stream tem- from Missouri, U.S.A. (Decapoda, Orconectes spp.). Crustaceana, 75, perature from air temperature: Implicatons for future water quality. 1119–1132. Journal of Environmental Engineering, 131, 139–146. Whitledge, G. W., Rabeni, C. F., Annis, G., & Sowa, S. P. (2006). Riparian Mugel, D. M., Richards, J. M., & Schumacher, J. G. (2009). Geohydrologic inves- shading and groundwater enhance growth potental for smallmouth tgatons and landscape characteristcs of areas contributng water to springs, bass in Ozark streams. Ecological Applicatons, 16, 1461–1473. the Current River, and Jacks Fork, Ozark Natonal Scenic Riverways, Missouri. Wrenn, W. B. (1980). Efects of elevated temperature on growth and sur- U.S. Geological Survey Scientfc Investgatons Report 2009-5138. vival of smallmouth bass. Transactons of the American Fisheries Society, Naiman, R. J., & Turner, M. G. (2000). A future perspectve on North 109, 617–625. America’s freshwater ecosystems. Ecological Applicatons, 10, 958–970. Zweifel, R. D., Hayward, R. S., & Rabeni, C. F. (1999). Bioenergetcs insight Pease, A. A., & Paukert, C. P. (2014). Potental impacts of climate change into black bass distributon shifs in Ozark border region streams. North on growth and prey consumpton of stream-­dwelling smallmouth bass American Journal of Fisheries Management, 19, 192–197. in the central United States. Ecology of Freshwater Fish, 23, 336–346. Peterson, J. T., & Kwak, T. J. (1999). Modeling the efects of land use and climate change on riverine smallmouth bass. Ecological Applicatons, 9, How to cite this artcle: Middaugh CR, Kessinger B, Magoulick DD. 1391–1404. Climate-­induced seasonal changes in smallmouth bass growth rate Peterson, J. T., & Rabeni, C. F. (1996). Natural thermal refugia for tem- potental at the southern range extent. Ecol Freshw Fish. perate warmwater stream fshes. North American Journal of Fisheries 2018;27:19–29. htps://doi.org/10.1111/ef.12320 Management, 16, 738–746. Middaugh et al. | 29

APPENDIX 1 Locaton data, fow classifcaton and period of water temperature record for each USGS gage site used

River State Lattude Longitude Flow classifcaton Period of record

Bear Creek Arkansas 35.94 −92.71333 Groundwater 7/2012–5/2015 Beaty Creek Oklahoma 36.35528 −94.776111 Groundwater 3/2005–10/2014 East Fork Black River Missouri 37.55256 −90.842444 Groundwater 10/2007–5/2015 Huzzah Creek Missouri 37.97472 −91.204444 Groundwater 12/2010–5/2015 Jacks Fork Missouri 37.05611 −91.668056 Groundwater 1/2003–8/2005; 5/2010–5/2015 Osage Creek Arkansas 36.28139 −94.227778 Groundwater 11/2011–5/2015 Spavinaw Creek Oklahoma 36.3225 −94.685 Groundwater 12/2004–4/2015 Sylamore Creek Arkansas 35.99167 −92.213889 Groundwater 7/2012–5/2015 Big Creek Arkansas 35.51 −91.817472 Runof 12/2012–5/2015 Big River Missouri 37.96553 −90.574417 Runof 11/2011–5/2015 Bufalo River Arkansas 36.0225 −93.354722 Runof 5/2008–12/2014 Illinois Bayou Arkansas 35.46639 −93.041111 Runof 6/2013–5/2015 South Fork Litle Red Arkansas 35.56972 −92.621944 Runof 7/2012–5/2015 Tavern Creek Missouri 38.27778 −92.236111 Runof 6/2014–5/2015