How do warm winters affect energy allocation tradeoffs in cool-water fish?

Olivia Balkcum, William Sims

Abstract

Winter is a critical period for many temperate fishes as dramatic changes in winter thermal regimes may result in tradeoffs between reproductive and somatic energy allocation patterns. Yellow Perch are found throughout North America, and reproductive development typically occurs during October-March, but not much is known about how southern populations allocate energy during overwinter periods. Using controlled laboratory studies of fish captured from wild populations, we sought to determine if Yellow Perch populations in South Carolina and Ohio’s Lake Erie have different overwinter energy allocation strategies. Additionally, we tested if fish from each population responded differently to typical warm and cold winters for their respective regions.

We hypothesized that northern populations would allocate more energy to reproductive growth than southern populations and that cold winters would result in more energy allocated towards reproduction. We examined the effect of location, winter temperature treatment, and sex on overwinter somatic and reproductive growth using a linear mixed-effects model. Results suggest that northern populations allocated more energy towards reproduction than southern populations and that colder winters promoted greater energetic investment in reproduction than warmer winters. Our findings may have important implications for setting expectations for this cool water species in response to climate change.

Introduction

The Yellow Perch, Perca flavescens, is a species that inhabits a wide latitudinal range throughout North America. This species prefers cool, well oxygenated water, and females develop eggs during the winter and spawn during early spring. This species is well studied in northern latitudes, however information on basic life history traits is lacking in the southern portions of this species range. Climate change could profoundly affect coldwater fish species (Jeppesen et.al 2010), especially those that require long, cold winters to develop viable eggs prior to spawning during spring. In this study we sought to quantify the overwinter allocation of energy to reproduction and somatic growth (that is growth in non-reproductive tissues).

We quantified reproductive and somatic growth in two populations, one from Ohio and one from South Carolina, to test if fish from each population responded differently to typical warm and cold winters for their respective regions. We hypothesized that cold winters would lead fish to allocate more energy to reproduction while warm winters would allow for greater somatic growth. Further, we hypothesized that northern populations would allocate more energy to reproductive growth than southern populations and that cold winters would result in more energy allocated towards reproduction.

Materials and Methods

Two separate experiments were conducted with Ohio and South Carolina fish. The Ohio experiment was conducted during the winter of 2011-2012. The South Carolina experiment was conducted during the winter of 2018-2019. In Ohio, Perch were collected from the central basin of Lake Erie , and in South Carolina from the savannah river in Augusta, below Clarks Hill Dam. Perch in South Carolina were collected via hook and line sampling. Ohio fish were collected via bottom trawl. Previous experiments both had a long and short winter duration, although the thermal regimes used in these treatments differed among experiments that were based on local environmental conditions. In Ohio, the long winter consisted of 107 days at 4 degrees Celsius, while the short winter consisted of 52 days at 4 degrees Celsius. In South Carolina, the long winter consisted of 42 days at 8 degrees Celsius, while the long winter consisted of 21 days at 8 degrees Celsius (Table 1). 

Fish in all treatments were fed at level sufficient to meet only their basal metabolic demands, as determined from an existing bioenergetics model. Fish were measured at the start of each experiment and a subset of fish were euthanized to obtain overall weight and gonad masses. From these euthanized fish, we used linear regression to develop relationships between fish mass and gonad mass for both males and females in Ohio and South Carolina. These relationships were then used to predict the gonad mass in live fish at the beginning of each experiment. When the experiment concluded, we measured the total weight and gonad mass of both males and females in each treatment and quantified the percent change in gonad and somatic tissues in each winter duration.

We used a one-way analysis of variance (ANOVA) in R to analyze data collected from South Carolina and Ohio. We calculated tank averages for percent change in reproductive and somatic tissues, as tank was our experimental unit. We confirmed the assumptions of normality and constant variance by examining a histogram of the residuals in a plot of residuals vs fitted values.

Table 1. Treatment Duration and Temperatures for Ohio Long and Short and South Carolina Long and Short Winters. Table describes details of each of the four winter treatments of Yellow Perch.

Table 1. Treatment Duration and Temperatures for Ohio Long and Short and South Carolina Long and Short Winters. Table describes details of each of the four winter treatments of Yellow Perch.

Results

A total of 59 fish were used for soma and 55 for gonads. The analysis of variance (ANOVA) of Ohio female gonads showed that there was a significant difference (P=0.019) in percent gonad difference between short and long treatments, with long treatment gonads having a higher difference than the shorter treatment (Figure 1). This differs from South Carolina female gonads, which were insignificant (P=0.39) and showed little difference between treatments (Figure 1). Anova of origin shows there was no significant difference between Ohio and South Carolina Female gonads(P=0.46). Both Ohio and South Carolina female percent somatic tissue (soma) change for short and long treatments were insignificant (P=0.34,0.99), although Ohio female soma was significantly lower than South Carolina (P=<0.001Figure 3). Male gonads followed a similar trend to one another, with both Ohio and South Carolina short and long treatments being insignificant (P=0.31,0.39),and having marginally significant difference between origin (Figures 2,Figure 3, P=0.003). Both Ohio and South Carolina male Soma treatments were insignificant ( (P=0.97, 0.99), however South Carolina Female and male somas were much significantly higher than Ohio somas (P=<0.001, Figure 3).

Figure 1.

Figure 1. Percent gonad difference of long and short winter treatments in females from Ohio and South Carolina (A,B) and males (C,D), with P value.

Figure 2.

Figure 2. Percent soma difference of long and short winter treatments in females from Ohio and South Carolina (A,B) and males (C,D), with P value.

Figure 3.

Figure 3. Percent gonad and soma difference of females from Ohio and South Carolina (A,C) and males (B,D), with P value.

Discussion

The ANOVA model results from this experiment reveal that only female gonad weight varied significantly between winter treatments in Ohio with the short treatment having a lower mass than the long treatment. This could be due to short winters having a higher metabolic cost; resulting in less energy going towards gonad or soma growth and more energy allocated towards base metabolic processes. Somatic growth in both northern and southern populations was not affected by winter treatments, which suggest non-reproductive growth during winter is minimal and not traded off against reproductive development.  Both South Carolina male and female fish appeared to show higher soma development during the winter than Ohio fish, showing that South Carolina fish were still actively growing during the winter. Likely taking advantage of warmer winters in southern latitudes. All male gonads showed negative percent difference values, this is likely because fish were measured after the spawning season and were spent of milt. We saw no difference between the boxplots of Ohio and South Carolina female gonads.

Ohio females concurred with our hypothesis that Ohio fish would allocate more energy to reproduction in the long winter, however all other models countered our hypothesis that cold winter would lead to more energy to reproduction, as they showed no significant difference in the allocation of energy to reproduction between long and short winter treatments. Previous studies show Ohio GSI tends to be higher than South Carolina GSI. However we did see that South Carolina fish put on more soma in the winter in comparison to Ohio fish, which concurs with our hypothesis that warm winters would allow for greater somatic growth.

Our findings are significant in that they provide insight into how shorter, warming winters can affect fish, such as Yellow Perch. When looking at current global climate change trends, we can expect for these northern populations, such as Lake Erie Yellow Perch, to undergo warmer winters in the future that are more aligned with temperatures mimicking South Carolina. Due to these warmer temperatures, Ohio fish may take on life history traits similar to that of South Carolina’s current life history traits. Changes to Ohio fish life history traits may necessitate changes in management and regulation of fish in that area.It’s possible that these effects could also be seen in other fish species such as salmonids, esox, and other members of Percidae. Fish collected from both Ohio and South Carolina experienced the same kind of water system, as well as feeding simulations, making climate the focus of the study. A common-garden experiment in this study would have possibly yielded more accurate results in regards to the winter temperature treatments, but in the case of this study would not have been possible, due to the time difference in studies, as well as the extreme temperature change that the fish may have undergone. To more accurately predict the trends of overwinter energy allocation in this study, we would have liked to have a larger sample size of both male and female fish from Ohio and South Carolina populations, as well as possibly add another origin reference to better understand the latitudinal trend of warming winters affecting growth. In studying overwinter energy allocation of fish species, we suggest future studies further look into mechanisms driving metabolic cost in overwinter allocation, as well as look more into current and well validated  bioenergetics models for Yellow Perch and other cool water species as bioenergetics models used for this study were outdated and no current models existed for this or similar species.

References