Movement of the Eurasian perch (Perca fluviatilis)
– individual responses to abiotic factors
Linda Sandberg
Degree Thesis in Biology 15 ECTS Bachelor’s level
Supervisor: Jonatan Klaminder Report passed: 10-06-2020
Abstract
Movement increases the probability for an individual to find food resources, but also increase the metabolic costs and exposure to predators. Hence, swimming behavior of fish is strongly coupled to fitness. Even though swimming activity has been studied in numerous laboratory settings, less is known about in situ activity and its dependence on abiotic factors (temperature, light conditions and barometric pressure). In this study I hypothesized that the activity increases with 1) increasing temperature and decrease with 2) barometric pressure variability and 3) average light conditions (h/day). In order to test the universality of the three hypotheses I also searched for size dependent effects. Fish activity (km/day) was measured in three lakes on individual fish (N= 14-21 per lake) using acoustic telemetry providing tracking of fish at a time resolution from seconds to hours. A positive correlation between temperature and swimming activity in line with my first hypothesis was only observed in one of the lakes. The activity decreased with increased variability in barometric pressure in two of the lakes, a finding
supporting my second hypothesis. Meanwhile increased light conditions (h/day) decreased fish activity in one of the lakes, as predicted by my third hypothesis. Nevertheless, none of my
hypotheses were valid in all three of the lakes and perch reacted differently to the abiotic factors.
One of the possible explanations for this is the importance of size differences as I noticed that the swimming activity differed between bigger and smaller individuals. My findings suggest that not only the temperature, barometric pressure and light conditions alone predict the activity in perch, but also the fish individual size, predation and the metabolic costs linked to
thermoregulation.
Key words: Eurasian perch, boreal lakes, acoustic telemetry, activity, abiotic and biotic factors.
Table of content
1 Introduction ... 1
1.1 Background ... 1
1.2 Aim ... 2
2 Method ... 2
2.1 Study species ... 2
2.2 Abiotic factors data and study locations ... 2
2.3 Tagging and retrieving the activity data ... 3
2.4 Quality control of retrieved positioning data ... 4
2.5 Statistical analyses ... 5
3 Results ... 5
3.1 Hypotheses testing in Abborrtjärn ... 5
3.2 Hypotheses testing in Stortjärn ... 6
3.3 Hypotheses testing in Tosktjärn ... 6
4 Discussion ...7
4.1 Temperature as a driver of swimming speed ...7
4.2 Air pressure as a driver of swimming speed ... 9
4.3 Solar light as a driver of swimming speed ... 9
4.4 The non-universality of abiotic factors as driver of fish behavior...10
5 Conclusions ... 11
6 Acknowledgements ... 12
7 References ... 133
Supporting figures ... 166
Appendix ... 177
1
1 Introduction
1.1 Background
According to Intergovernmental Panel on Climate Change (IPCC) the ongoing alteration of greenhouse gases in the atmosphere is resulting in changes in Earth’s climate system (IPCC 2019). The current and forecasted temperature increases will have consequences on ecosystems through shifting phenologies, distributions and interactions between species (Walther et al.
2002; Deutsch et al. 2015; Stuart-Smith et al. 2015). Knowing how climate change affects physiology functions is important in order to establish appropriate management and policy responses (McKenzie et al., 2016; Patterson et al., 2016). Nevertheless, predictions of how climate changes affect individual organisms is a challenge because of possible cascading effects from an individual level to whole ecosystem scales. In other words, model predictions based on for example species-level effects can differ from models targeting ecosystem-level effects (Brander et al., 2013; Lefevre et al., 2017). However, knowledge regarding how fish individuals are affected by abiotic factors that are expected to change in the future (water temperature, barometric pressure and light conditions) is scarce and improved knowledge regarding this theme is highly warranted by the research community.
So, how might fish be affected by abiotic factors such as temperature? Looking at the physiology of fish, they are ectotherms (Thorpe 1977), and therefore unable to regulate their body
temperature internally. Instead the body temperature of fish is regulated by the temperature of the surrounding water (Reece and Campbell 2011). In turn, the metabolism and organ function of fish is determined by the water temperature (Hochachka and Somero 2002). Subsequently the metabolic costs for the thermoregulation are linked to the resources available for reproduction, growth and development (Kooijman 2001; Angilletta 2009). Previous studies showed that rising temperature increased the basal metabolism and the food demand in perch (Brown et al. 2004;
Ohlberger et al. 2012). Other studies also showed that the activity in perch increased with increasing temperature (Neuman et al 1996; Nakayama et al. 2018) and that perch stayed motionless for longer in colder water (Saaristo et al. 2019).
Along with water temperature, the barometric pressure also has a strong influence on the activity in fish (Feyrer et al. 2015). The pressure on the surface of water is affected by the barometric pressure in the atmosphere. In turn, and according to Boyle’s law, the changes of pressure on the surface increase or decrease the pressure on the swim bladder (Calo 2015). The gas bladder seen in many fish species allows them to keep hydrostatic equilibrium. A drastic decline or rise in the barometric pressure (hPa) is expected to lower the activity in fish because their swim-bladders need to be adjusted to the new pressure condition (Chetkov 1982).
The majority of predating pelagic fish are visual feeders, and therefore strongly affected by differences in light conditions (h/day) over the course of a day (Batty et al. 1990; Ryder 1990).
Thus, the light intensity and the diel light cycle impact the activity in fish. A study showed that the light intensity especially affects the activity in fish in the evening. In the same study, also the size of the fish affected the activity, were bigger fishes usually were more active than smaller (Didrikas and Hansson 2009). Also, the activity in fish seems to differ between seasons and lakes, which probably is due to seasonal changes and differences in lake characteristics (Nakayama et al. 2018; Jacobsen et al. 2015).
Indeed, there are rationale for expecting that fish swimming activity is affected by a range of abiotic factors. In addition to these abiotic factors there are a strong coupling between fish movements and predator activities. Fish live in a “landscape of fear” where their spatial use is restricted by their perceived perception of risks from predators (Laundre et al. 2010). In other
2
words, interaction with predators may complicate the linkage between activity and abiotic factors such as temperature, barometric pressure and solar light.
1.2 Aim
With this background, it seems evident that abiotic factors can affect the swimming activity of fish. Yet, earlier studies on this subject has generally been small scaled, only including one lake or used a small number of fish individuals. Therefore the aim of this study was to assess fish behavior in three lakes and analyze how the activity of a large number (N =54) of the Eurasian perch (Perca fluviatilis) was affected by the air temperature, barometric pressure and light condition. In this assessment, I also scrutinized if the effects of the abiotic factors were size dependent. Differences in activity due to lake characteristics and season were also addressed, but the main focus was on the abiotic factors and the individual size. The study was conducted using tracking of individual fish behavior using acoustic telemetry. Fish positioning derived from acoustic telemetry providing a time resolution ranging from of 3 sec to 1 hour was used to calculate fish activity (km/day), as a function of abiotic factors from three lakes and individual size data from two lakes. A study showed that perch differed in activity and response to abiotic factors in the course of a day (Didrikas and Hansson 2009). Still in order to scale down the project, this study only analyzed differences in activity related to changed mean of each abiotic factor per day.
Based on existing literature I hypothesized that fish activity increase with 1) Increasing temperature and decrease with 2) Barometric pressure variability and 3) Average light
conditions during photoperiod. The universality of the three hypotheses was tested by searching for eventual size dependent effects.
2 Method
2.1 Study species
The Eurasian perch (Perca fluviatilis) is a eurythermal species and is a common fish native to Asia and Europe (Thorpe 1977). Important for the choice of this study-species was that it is a well-studied species. The species belongs to the order Perciformes which are ray-finned fish that feeds on other fish, for example congener species, as well as macroinvertebrates and zooplankton (Jamet 1994; Lorenzoni et al. 2007). Other studies have also unraveled the anti-predator
behavior (Christensen and Persson 1993) and the behavior in a multi-stressor environment (Saaristo et al. 2019) for the species.
2.2 Abiotic factors data and study locations
Perch swimming activity was studied in three boreal lakes. All lakes included perch and pike and were situated in the Västerbotten county. Earlier studies and netting without any catch indicated that the lakes Tosktjärn and Abborrtjärn were fishless prior to the experiment, i.e. due to earlier rotenone treatment. Meanwhile in Stortjärn the stand was natural. The perches introduced to Stortjärn were also bigger than in the two other lakes. Abborrtjärn (64°29'2.5"N 19°25'8.2"E) is a clear water lake and has an area of about 4000 m2, a maximum depth of approximately 6 m and with quagmire aligning the borders of the lake. Stortjärn (64°15'41.6"N 19°45'44.3"E) is a brown water lake and has an area of approximately 33 000 m2. Meanwhile Tosktjärn (64°30'14.8"N 19°02'36.9"E) is a brown water lake with an area of approximately 1-ha and a mean depth of 5 m.
See the location of the lakes in figure 1a below. Sampling showed that the lakes inhabited resources for feeding fish, for example zooplankton and invertebrates prior to the experiment.
3
The abiotic data (air temperature, air pressure and light conditions during photoperiod) was retrieved from SMHI from respective lake’s nearest weather station. Stortjärn is situated in Svartberget’s experimental park in the municipality of Vindeln, so the air temperature data for this lake was retrieved from the weather station in Petisträsk (SMHI 2020). Tosktjärn is situated adjacent to Ekorrträsk about 20 km from Lycksele airport, so the air temperature data was collected from Lycksele. While the last lake, Abborrtjärn is situated in Åmsele so the air temperature data was retrieved from the abiotic factors station Petisträsk A (SMHI 2020).
Atmospheric data for all the lakes was acquired from the weather station in Lycksele. Meanwhile all the hourly light condition data (h/day) was retrieved from the closest weather station in Umeå (SMHI 2020). The positon of the weather stations are marked out in figure 1a below. The experiment was between the 24/10-17/11-2014 in Tosktjärn, in Abborrtjärn between the 5-30/9- 2016 and in Stortjärn between the 15/9-10/10-2016. Tosktjärn was covered in transparent ice during the study.
Figure 1a. Map showing the location of the lakes: Abborrtjärn, Tosktjärn and Stortjärn along the road 363 and the weather stations: Lycksele, Petisträsk and Umeå. The projection is in WGS 84 and the map was created in the program Google Earth. The area is located in the Västerbotten county in Sweden.
2.3 Tagging and retrieving the activity data
The swimming activity (km/day) of fish was measured in the three lakes: Tosktjärn, Stortjärn and Abborrtjärn (Figure 1a). Tracking of fish behavior was done using acoustic telemetry. In short, this method is dependent on tagging of fish and positioning of fish position using
triangulation of transmitters based on time of their detection in a network of receivers placed in each lake. For a detailed description of the technique see Leander et al. (2019). The fish
individuals for the experiment were captured in different lakes. For Abborrtjärn, individuals of perch within the fork length range of 166±27 mm (mean ± standard deviation), were caught with a beach seine net in Lake Stöcksjön (63°45'45.9"N 20°11'54.1"E) in July 2016. Fork length (FL) is a measurement from the end of the median caudal fin rays to the anterior-most part of the fish (Anderson and Gutreuter 1983). For Stortjärn individuals of 360±3.5 mm (mean ± standard
4
deviation) were caught in the Ume River. Meanwhile, for Tosktjärn fish individuals with the fork length size 113±13 mm (mean ± standard deviation), were caught with umbrella traps in Lake Stöcksjön (N 63° 76.3976′, E 19° 76.78′), Lake Brunnsjön (N 63° 74.6706′, E 20° 04.8761′) and Tavlefjärden Bay (N 63° 83.1283′, E 20° 49.5682′).
After capture, all the fish individuals were brought to Umeå Marine Research Facility (UMF) in oxygenated tanks. In the facility they were transferred and held in a 1 m3 flow-thru reservoir at 15±1.5˚C. Then, individuals for the experiment were randomly chosen and surgically equipped with acoustic transmitters (Vemco V4 180 kHz), see Liedtke and Rub (2012) for further details on the tagging procedure. Post-operational, the individuals were left to recover for a minimum period of 10-14 days. Over the course of the recovering time the fish were feed with chironomid larvae every two days. In the end of the recovery time the individuals were classified as healthy on their appearance. Further, in order to mimic a more natural habitat in the fishless lakes, four Northern pikes were released in these respectively. The Northern pikes were caught and
transferred from the lake Tavelsjön (64°0'2.4"N 20°3'5.1"E) for Abborrjärn (sizes of 600±50 mm) and from the Öre River for Tosktjärn (sizes of 2.2-4.4 kg). The created predator density was within the natural range in boreal lakes but quite low (Pearson et al. 1996). Hence, the purpose of introducing the pikes was more to create a landscape that would be perceived with predation- risk, because this is a natural driver of perch behavior.
Tracking of the individuals was carried out for 26 days in Abborrtjärn and Stortjärn and for 22 days in Tosktjärn using the VEMCO HR2 system (Leander et al. 2019). The number of receivers in each lake was enough for good coverage, and these were fixed with buoy lines in the lakes.
Additionally satellite receivers were installed for coverage of an island in the lake Abborrtjärn.
The interval of transmission was 2 seconds and the median accuracy was 0.5 m. Then the following number of individuals: 19 for Stortjärn, 14 for Tosktjärn and 21 for Abborrtjärn were randomly chosen and transferred in oxygenated tanks to each lake. After this, the systems were left undisturbed except for compensating for observed mortality in Abborrtjärn, were pike were re-introduced on the 23rd of September. The retrieved positioning data was interpolated at 60 sec level. The 2D dimensional positioning data was interpolated on mean activity/day and individual, with approximately 500 data-points/lake and 14-21 fish individuals per lake. While the swimming speed (km/day) in all lakes was calculated by measuring the distance travelled from one point to another (minute scale). Approximately 500 interpolated data points per lake were extracted and used in calculating the mean activity for all individuals per day in each one of the lakes.
2.4 Quality control of retrieved positioning data
Firstly, to prevent biases due to initial behavioral noise during the acclimation to the lake, the first day from all lakes’ study period was excluded from analyses. For Tosktjärn the first four days were excluded due to error in the data. Note that the acclimation time is well above the 4 hours limited suggested for acclimatization in ecotoxicological behavior assays (Melvin et al.
2017). Prior to analyses, the data was quality controlled through checking for a number of factors that possibly would affect the outcome of the analyses: 1. Losses of individuals due to transmitter failure or mortalities of tagged individuals, 2. Abnormal activity (highs and lows) – possible outliers, 3. Normal distribution, 4. Standard deviations and 5. Variances in the data. The quality control of the data gave some findings. In Stortjärn three perch individuals disappeared, ID: T83 (gone from 29/9), T89 (gone from 2/10), T92 (gone from 27/9). Moreover, in the same lake four other individuals showed exceptionally low activity from a specific date and until the end of the studied period (10/10-2016). To prevent bias, activities below 1.0E-05 km/day in a continuous period of minimum three days and until the end of the study period (10/10-2016) was classified as abnormal and excluded. Thus, in Stortjärn analyses were only conducted on the data without
5
the abnormal data. Abborrtjärn and Tosktjärn did not have any individuals with abnormal activity in the lower range as in Stortjärn.
Standard deviations and variances were also calculated on the data in order to look for abnormalities. The data was checked for outliers (mean ± 3 standard deviation). Outliers in Abborrtjärn were detected but kept due to that the higher values occurred in different
individuals and under the continuous periods: 9-11/9 and 20-23/9- 2016. The same was the case for Tosktjärn that showed slighter higher activities 4-7/11-2014, so these were also kept. In Stortjärn five outliers from three different individuals were detected. Yet, the outliers in Stortjärn were kept as well, because of that the activity values around these were high as well.
Further, for the size analysis in Abborrtjärn, two equally sized groups were formed: Size class 1.
Including 11 individuals with the FL between 136-170 mm and the mean 151 mm and Size class 2.
Including 10 individuals with the FL of 171-225 mm and the mean 191 mm. Meanwhile for the size analysis in Tosktjärn two groups were formed with seven individuals in each. The fork length sizes in the groups were: Size class 1. Between 92-113 mm with the mean of 105 mm and Size class 2. Between FL = 115-151 mm with the mean of 121 mm. For Stortjärn the individual activities were not tested against the sizes due to that the data was first received in the end of the study.
2.5 Statistical analyses
Further, the data was checked for normal distribution. All the activity data in the size classes were normally distributed. For the lakes with all the individuals, all the activity data for Abborrtjärn, Tosktjärn and for Stortjärn was normally distributed making them possible to analyze with parametric tests. In order to prevent bias due to possible multicollinearity, correlation analyses were conducted between all the independent variables in the three lakes respectively (Table 1a-c, Appendix). If the correlation coefficient came out less than -0.5 or more than 0.5 (a variance inflation factor of <0.4) the variables were classified as strongly correlated.
In this data, none of the parameters came out as strongly correlated with another (Table 1a-c, Appendix). Further, the following three null hypotheses were formulated: the activity in perch does not increase with 1) increasing temperature and does not decrease with 2) Barometric pressure variability and 3) average light conditions during photoperiod.
The level for significance was set to p= <0.05, thus the null hypotheses were rejected if the p- value came out below this value, hence confirmed or neglected the null hypotheses and the initial hypotheses for the study. In order to see if there were any significant relations between the activity data and the abiotic factors or the individual sizes, statistical analyses on the data were conducted. A stepwise multiple regression was conducted in all lakes respectively, this between all the independent variables: air temperature (°C), barometric pressure (hPa, relative standard deviation per day) and average light conditions during photoperiod (h/day) and the dependent variable activity (km/day). Meanwhile, paired t-tests were conducted in order to analyze if the activity differed due to the individual sizes in Abborrtjärn or Tosktjärn. For the size classes in each lake, the mean activity of the groups was also plotted against the date in the same graph.
3 Results
3.1 Hypotheses testing in Abborrtjärn
The swimming speed in Abborrtjärn was 2.3 ± 0.73 km/day (mean ± standard deviation) during the studied period. In contrast to my first hypothesis, activity did not significantly (P>0.05)
6
increase with increasing temperature. However, the multiple regression in this lake gave a R2- value of 0.29 (adjusted R-square = 0.22) (Table 2b, Appendix) for the predictors light condition (h/day) and barometric pressure on the dependent variable activity (km/day). In the regression the relation between the independent and the dependent variables was negative (Table 2a, Appendix). The light condition and barometric pressure together predicted approximately 22-29
% of the variation in activity. That the relationship was negative, suggests that the swimming speed declined with both increasing average light conditions during photoperiod and change in barometric pressure in Abborrtjärn. In other words, both hypothesis two and three was
determined valid with the p-value of <0.001.
A deeper assessment of the activity data suggested that the measured fish activity was dependent on size. The mean activity between the size classes were significantly different with the p-value of 0.002 and the t-value -3.5. Here the mean activity in size class 1 (FL =136-170 mm) was 2.1 km/day and the variance 0.51. Meanwhile in size class 2 (FL
= 171-225 mm) the mean activity was 2.6 km/day and the variance 0.83. Between the groups the activity was generally higher in the larger individuals (Figure 2a).
3.2 Hypotheses testing in Stortjärn The swimming speed in Stortjärn was 1.1 ± 0.33 km/day (mean ± standard deviation) during the studied period: 16/9-10/10-2016. In contrast with my first hypothesis, swimming activity decreased with increased temperature (Table 3a, Appendix).
Here, the multiple regression for Stortjärn gave a R2-value of 0.46 (adjusted R-square = 0.41) (Table 3b, Appendix) and a p-value of <0.001 for the
predictors barometric pressure (hPa) and temperature on the dependent variable activity (km/day). Barometric pressure and temperature together predicted approximately 41-46 % of the variation in activity. In the regression the relation between the independent and the
dependent variables was negative (Table 3a, Appendix): a result in line with hypothesis 2 but in contrast to hypothesis 3.
3.3 Hypotheses testing in Tosktjärn
The swimming speed in Tosktjärn was 1.6 ± 0.25 km/day (mean ± standard deviation) during the studied period: 27/10-17/11-2014. In line with my first hypothesis the activity increased with increased temperature (Table 4a, Appendix). In this lake the independent variables barometric pressure (hPa) and temperature (°C) were significant predictors of the activity (km/day). The regression value for both these predictors was R2= 0.30 (adjusted R-square = 0.23) (Table 4a, Appendix) and the p-value <0.001. Barometric pressure and temperature together predicted approximately 23-30 % of the variation in activity. In the regression the relationship between the independent and the dependent variables was positive (Table 4b, Appendix). As a result,
hypothesis 2 and 3 were determined invalid. A deeper assessment of the activity data suggested that the measured fish activity was size dependent. Between the size classes in Tosktjärn the activity was generally higher in the smaller individuals (Figure 3a).
Figure 2a. The mean activity (km/day) per day in the two size-classes in Abborrtjärn between the 6- 30/9-2016. The size classes in the graph are: 1.
Individuals with the FL = 136-170 mm and 2.
individuals with the FL = 171-225 mm. Size class 1 includes 11 individuals and size class 2 includes 10 individuals.
0 1 2 3 4 5
6/9 12/9 18/9 24/9 30/9
Activity (km/day)
Date
Size class 136-170 mm Size class 171-225 mm
7
Figure 3a. The mean activity (km/day) per day in the two size classes in Tosktjärn between the 27/10-17/11-2014.
The classes includes 7 individuals each.
The paired t-test between the size classes gave the p-value <0.001 and the t-value 4.6. Here the mean activity in size class 1 (FL = 92-113 mm) was 1.8 km/day and the variance 0.19. Meanwhile the mean activity in size class 2 (FL = 115-151 mm) was 1.3 km/day and the variance 0.046 (Table 5a).
Tosktjärn Abbortjärn
Size class (92-113 mm)
Size class (115-151 mm)
Size class (136-170 mm)
Size class (171 – 225 mm)
Mean 1.8*** 1.3*** 2.1** 2.6**
Variance 0.19 0.046 0.51 0.83
N 22 22 25 25
4 Discussion
4.1 Temperature as a driver of swimming speed
Temperature came out as a significant predictor of the activity in perch in two of the lakes. In Tosktjärn my first hypothesis, along with other studies showing an increase in activity with increasing temperature were confirmed (Neuman et al 1996; Nakayama et al. 2018). A previous study showed that rising temperature increased the basal metabolism and the food demand in perch (Brown et al. 2004; Ohlberger et al. 2012). Another study also showed that the pike
foraged less with increasing temperature (Kuparinen et al. 2010). The changed basal metabolism in perch and changed foraging behavior in the predator pike, makes it reasonable for perch to increase their activity with increasing temperature. Yet, in Stortjärn, the activity decreased with an increase in temperature, thus showing a trend in complete contrast to my first hypothesis.
Meanwhile in Abborrtjärn there was no significant relation between temperature and activity. In
0 1 2 3
10/27 10/31 11/04 11/08 11/12 11/16
Activity (km/day)
Date
Size class 92-113 mm Size class 115-151 mm
Table 5a. Summary of the results from the paired t-tests in the two size classes in both Abborrtjärn and Tosktjärn.
The table shows the mean activity (km/day), variance and the sample number (N) received from the test. The activity in Tosktjärn was from the period 27/10-17/11-2014 meanwhile the activity data in Abborrtjärn was from the 6-30/9-2016. ** Significant at 0.01 level. ***Significant at 0.001 level.
8
other words, my first hypothesis was valid in Tosktjärn but invalid in Stortjärn and Abborrtjärn.
All this said, it seemed that temperature was not a universal predictor of activity in perch.
So why would not temperature function as a universal predictor of swimming activity in all lakes? Looking at the measured difference in temperature response between the lakes, this might for example be due to the temperature optima in perch. A study looked at temperature induced effects on the aerobic scope. By definition the aerobic scope is a measure on how well an
organism’s cardiovascular and respiratory system can provide oxygen over life sustaining, as example for activities such as locomotion, reproduction and growth (Clark et al. 2013; Fry 1971).
The study showed that the aerobic scope in the Eurasian perch increased progressively between 5-21°C and then flatten out (Jensen et al. 2017). I notice that the swimming activity of my studied perch largely follows these trends, where swimming activity started to respond to a temperature increase after about 5-8 °C (see Figure 4a, Supporting figures). In other words, a large part of my measurements was conducted during sub-optimal temperatures for perch where activities are expected to be low, which may explain the weak temperature response. Thus, at least in Tosktjärn the temperatures generally were below the temperature optima for perch, which might have affected the temperature response. Another important aspect of how temperature affects the activity in perch is that it increases the metabolic costs (Brown et al.
2004; Ohlberger et al. 2012). In turn the increased metabolic cost likely limits the temperature induced rise in activity to keep down the food demand, thus making the final temperature response more complex.
Also, in this study I used atmospheric temperature data. A drawback with that was that the atmospheric temperature probably changed the water temperature differently in the lakes. The final water temperature in each lake depended on the atmospheric temperature in combination with a range of morphological and location parameters (Piccolroaz et al. 2013). Comparing Stortjärn and Tosktjärn, Lake Stortjärn was approximately three times the size of Tosktjärn.
Thus the lag between the water temperature and air temperature was likely different between the lakes. The season as well as the difference in size between the lakes, might have caused the unequal response to changes in air temperature. Log transforming the temperature data might have given better results through lowering the temperature differences between the two
elements.
The risk of predation could also have affected the observed response to temperature, where the risk of predation could have been size dependent. A study showed that perch changed
ontogenetic niche as it grew and became facultative piscivores (Persson and Greenberg 1990), and in another study the predation risk for fish usually decreased with the body size (Lorenzen 2000). Looking at the individual sizes, the individuals in Stortjärn were much bigger than in the other two lakes. Therefore, the perches in Stortjärn might have behaved more like predators and/or perceived less predation-risk, and thus been less active.
Further, temperature came out as significant predictor together with light condition or
barometric pressure, which makes it hard to distinguish the temperature effect alone. Another point that was not considered in this study was the differences in temperature in the water column. A study showed that the water temperature decreased with depth. Consequently, fish needed to adjust the volume in the swim-bladder when moving vertically (Alexander 1966). Also, the Eurasian perch was demonstrated to have the antipredator behavior to for example hide in crevices on the bottom when threatened (Christensen and Persson 1993), and therefore when the individuals in my study escaped from predators they were subjected to changes in conditions.
For this study, no data on depth location of the individuals was accessible, and therefore it was unfeasible to have this approach. Yet, this study used a great size of data on perch in their natural environment and therefore probably showed fair results. Also I think that the
9
temperature optima and the predation risk probably was more important drivers of the activity than depth.
4.2 Air pressure as a driver of swimming speed
The parameter change in barometric pressure came out as a significant parameter in all lakes. In line with my second hypothesis, the activity decreased with increasing variation in barometric pressure in Stortjärn and Abborrtjärn. Here the decreased activity likely was a response to the demand of adaption (Calo 2015). Opposite to my second postulation the change in barometric pressure came out as positively related with activity in Tosktjärn. Thus, my second hypothesis was determined valid in Stortjärn and Abborrtjärn but invalid in Tosktjärn. In contrast to the effect by temperature, there are no known optimum in barometric air pressure change that could explain these between lakes differences. Instead, the differences in result between the lakes could have been due to difference in season, where the sampling in Tosktjärn was during a colder season when the lake was covered with ice. Meanwhile the sampling in Stortjärn and Abbortjärn was during the autumn. A study showed that fish reacted unequal to different abiotic factors depending on the season (Nakayama et al. 2018), even despite that the activity levels and
foraging in another study was the same between winter and summer (Jacobsen et al. 2015). Also as earlier mentioned, there possible was lag between atmospheric conditions and water
conditions, which depended on the size of the lake among other parameters (Piccolroaz et al.
2013). All the lakes differed in size, and therefore the final water condition at the same atmospheric condition probably differed as well between the lakes.
Also, as earlier mentioned the depth was not considered in this study. Even despite that perches in a study was shown to have the antipredator behavior of hiding in crevices on the bottom when threatened (Christensen and Persson 1993). A study showed that the pressure increased with depth and therefore fish needed to adjust their swim-bladders when moving vertically
(Alexander 1966). Thus the fishes in my study needed to adjust their swim-bladders when escaping from predators.
4.3 Solar light as a driver of swimming speed
The light condition only came out as a significant predictor in the lake Abborrtjärn. Therefore, my third hypothesis was determined valid in Abborrtjärn but not in Stortjärn and Tosktjärn.
Thus, the average light conditions during photoperiod was not a universal predictor of activity.
In Abborrtjärn it seemed that the brighter conditions led to more alertness and/or escaping possibilities in the fish (Pitcher and Turner 1986). More activity due to longer day length was reasonable, considering that perch in studies has been shown to be a visual forager and more active during daytime (Craig 1977; Jacobsen et al. 2002; Schleuter and Eckmann 2006;
Nakayama et al. 2018). Further, in a study in a clear water lake, the light condition came out as one of the two strongest predictors of activity (Nakayama et al. 2018). Yet, why did not light condition come out as a significant predictor in Stortjärn and Tosktjärn? The reason for this could have been due to differences between the lakes, where Abborrtjärn was the only clear water lake of the three. The perches in the brown water lakes were maybe not as strongly affected by the level of light condition due to the already reduced visual conditions.
Another possible explanation was that the activity in the brown water lakes followed a non-linear relationship with light. In studies the activity in perch was highest at intermediate light condition levels, where strong light conditions was discussed to be discouraging in an environment with predation-risk or making hunting more difficult (Turesson and Brönmark 2004; Nakayama et al 2018). Thus, the activity in the two other lakes might have been related with light condition in a way that did not cope with the linear-fit-regression tested. Also the light condition time in
10
Abborrtjärn possible was even stronger related with activity in a non-linear model. For Stortjärn, also the losses of individuals could have impacted the results. Looking on the activity data in relation to light condition in a graph, polynomial functions gives the highest R^2 values for all lakes. Yet, the difference between the R^2 values for the polynomial versus the linear function for Stortjärn and Abborrtjärn were low (<0.005). Meanwhile in Tosktjärn the R^2 value was more than 10 times higher with the polynomial function compared to the linear (0.05 compared to 0.004). In other words, my theory on differences due to non-linear relationships were not very likely in Stortjärn and Abborrtjärn, but maybe likely in Tosktjärn (Figure 5a, Supporting figures).
4.4 The non-universality of abiotic factors as driver of fish behavior The mean swimming speed of the individuals differed significantly between the two different size-classes in Abborrtjärn, where the larger individuals were more active. Also in Tosktjärn the mean activity between the size-classes differed significantly, but here the smaller individuals were more active than the larger. These results supported that my hypotheses were not universal in all lakes. Moreover, my study only focused on the relation between fish physiology and the abiotic factors: barometric pressure, temperature and light condition. Yet, studies have shown that other factors also influence the activity in perch, such as the available food resources and predator-density. For example, a study showed that perch was more active at low resource levels compared to intermediate and high resource levels (Olsson et al. 2007). Further, studies showed that the abundance of food resources for perch (zooplankton and macroinvertebrates) was connected to the nutritional status in the lakes (Taipale et al. 2016) and the season (Deininger et al. 2017). Looking at differences in activity due to nutritional status, a study showed that perch in a mesotrophic lake (total phosphorus of 30 μg/L) swam 6 km/day (Nakayama et al. 2018).
Meanwhile another study showed that perch in a more eutrophic lake (total phosphorus of 50 μg/L) swam 1 km/day (Jacobsen et al. 2015). Discussed in the former study was that a more eutrophic lake possible offered more abundance in prey and therefore lowered the activity needed for sufficient foraging. Lowered activity might also been due to higher abundance of predators in the more eutrophic lake, in order to minimize the predation risk.
Further comparing the fish activity in a mesotrophic lake with a eutrophic lake, the diel activity pattern could have differed. A study in a mesotrophic lake showed clear patterns of crepuscular activity and horizontal migration between the littoral zone and the pelagic. Meanwhile, the same study did not see this activity pattern in a hypereutrophic lake during summer. Yet the fish individuals in both lakes showed the same activity pattern during the winter (Jacobsen et al.
2015). Also discussed in the same article was that the fish in the clear water lake perceived a higher predation-risk, due to the visual condition (Jacobsen et al. 2015). These findings suggest differences in activity due to nutritional status, season and perceived predation-risk.
Discussing the observed size influenced activity in Tosktjärn and Abborrtjärn, the individuals differed in activity due to size in both lakes. Thus supporting that other factors than abiotic, for example size could have affected the activity in the perches. Further, the relations between size and activity in the two lakes were opposite. Larger individuals were more active than smaller in Abborrtjärn meanwhile in Tosktjärn the reversed pattern was distinguished. These results are supported by a study that showed major impact of size on the swimming speed (Didrikas and Hansson 2009). In other studies the swimming speed increased with increasing fork length (Baemish 1970; Peake et al. 2000), thus supporting my results in Abborrtjärn. The reason for an increased swimming speed with increased size in Abborrtjärn might have been the lowered perceived predation-risk. A study showed that the predation-risk for fish usually decreased with the body size (Lorenzen 2000). Another study on perch showed that increasing temperature increased the variation in activity between and within individuals (Nakayama et al. 2016). The difference in response due to temperature could explain the observed difference in size-related
11
response in the two lakes. Meanwhile the differences between the lakes could also been due to earlier mentioned subjects, for example dissimilar patterns in activity due to the season, the nutritional status and the perceived predation-risk within the lakes. As previously mentioned, Tosktjärn is a brown water lake and the activity data for this lake was collected during the winter season. In contrast, the activity data in the clear water lake Abborrtjärn was collected during the autumn.
Observing the sizes in the different groups, three non-overlapping size-ranges existed: a) Individuals in the small size class in Tosktjärn, b) Individuals in the small size class in
Abborrtjärn and c) Individuals in the large size class in Abborrtjärn. The individuals in the small size class swam faster than the individuals that were larger in Tosktjärn. Yet, the individuals from the large size class swam faster than the smaller individuals. Thus, another reason for the differences in the size related activity between the lakes was possible that the species followed a non-linear trend. The non-linear model could for example have shown a shift in activity due to a change in the ontogenetic niche. As previously mentioned, a study showed that perch changed their ontogenetic niche and became facultative piscivores as they grew (Persson and Greenberg 1990). Another study showed that perch became piscivoros at approximately FL = 140 mm (150 TL) (Persson and Greenberg 1990; Craig 2000). Also and as earlier mentioned, a study showed that the predation risk for fish usually decreased with the body size (Lorenzen 2000). In other words, the perch could have followed a non-linear trend in size related activity, but no
conclusion regarding this could be drawn without further investigation. Further, also the area of the lakes differed. Tosktjärn was twice as big as Abborrtjärn, which cannot be excluded as a possible cause to the observed differences between the lakes.
5 Conclusions
My study demonstrates that the activity in fish is affected by the abiotic factors, such as barometric pressure, light condition and temperature and the biotic factor individual size, but that the effects are not as straightforward as would be expected from their direct impact on fish physiology. In other words, within-lake processes likely linked to food resources and predators not necessarily linked to abiotic factors also played an important role for fish activity. The complexity and co-variation of the factors predicting activity makes it hard to entangle the impact of each factor alone. Still, my findings contribute to the field of movement ecology in perch and could be used to predict climate-change driven effects. Looking at the forecasting climate change, it will shift weather regimes which in turn could cause a fast change in the physical properties of lakes (Adrian et al., 2009). An increase in the water temperature in lakes has already been recorded (Verhoeven et al. 2006), which according to my data suggest
increased swimming activity in at least some lakes that subsequently imply increased metabolic costs; hence, if this increase in activity is larger than positive temperature effects on food resources, it will mean that perch growth rates will be reduced in response to warmer waters.
However, increasing successful foraging activity in fish have been observed in response to warming making this scenario less likely (Ohlberger et al. 2012). Still, my results provide important knowledge of how perch is affected by abiotic and biotic factors in its natural environment, which is warranted by the research community.
12
6 Acknowledgements
I would like to thank Jonatan Klaminder for being my supervisor and helping me throughout the thesis. Also I would like to thank Johan Fahlman and Gustav Hellström for preparing and providing me with the activity data. For helping me with the statistics I thank Micael Jonsson.
13
7 References
Angilletta, M.J. 2009. Thermal Adaptation: A Theoretical and Empirical Synthesis. Oxford University Press, Oxford.
Adrian, R., O’Reilly, C. M., Zagarese, H., Baines, S. B., Hessen, D. O., Keller, W., Livingstone, D.
M., Sommaruga, R., Straile, D., Donk, E. V., Weyhenmeyer, G. A. and Winder, M. Lakes as sentinels of climate change. Limnology and Oceanography. 54:2283–229.
Alexander, R.M. 1966. Physical aspects of swimbladder function. Biological Reviews. 41(1):141–176.
Beamish, F.W.H. 1970. Oxygen consumption of largemouth bass, Micropterus salmoides, in relation to swimming speed and temperature. Canadian Journal of Zoology. 48(6):1221–
1228.
Batty, R., Blaxter, J. and Richard, J. 1990. Light intensity and the feeding behaviour of herring, Clupea harengus. Marine Biology. 107(3):383–388.
Brander, K., Neuheimer, A., Andersen, KH. and Hartvig, M. 2013. Overconfidence in model projections. ICES Journal of Marine Science. 70:1065–1068.
Brown, J. H., Gillooly, J. F., Allen, A. P., Savage, V. M. and West, G. B. 2004. Toward a metabolic theory of ecology. Ecology. 85(7):1771–1789.
Calo, Adrienne. 2015. Science spotlight: Fish, Swim Bladders and Boyle’s law.
https://ww2.kqed.org/quest/2015/03/12/science-spotlight-fish-swim-bladders-and- boyles-law/ (retrieved 2020-01-20).
Clark, TD., Sandblom, E. and Jutfelt, F. 2013. Aerobic scope measurements of fishes in an era of climate change: respirometry, relevance and recommendations. The Journal of
Experimental Biology. 216:2771–2782.
Craig, J. F. 1977. Seasonal changes in the day and night activity of adult perch, Perca fluviatilis L. Journal of Fish Biology. 11(2):161–166.
Craig, J. F. 2000. Percid Fishes. Systematics, Ecology and Exploitation. Oxford: Blackwell Science.
Deininger, A., Faithfull, C. and Bergström, A.K. 2017. Nitrogen effects on the pelagic food web are modified by dissolved organic carbon. Oecologia. 184(4):901–916.
Deutsch, C.O., Ferrel, A.B., Seibel, B., Pörtner, H.-O. and Huey, R.B. 2015. Climate change tightens a metabolic constraint on marine habitats. Science. 348(6239):1132–1135.
Didrikas, T. and Hansson, S. 2009. Effects of light intensity on activity and pelagic dispersion of fish: studies with a seabed-mounted echosounder. ICES Journal of Marine Science.
66(2):388–395.
Feyrer, F., Cloern, J.E., Brown, L.R., Fish, M.A., Hieb, K.A. and Baxter, R.D. 2015. Estuarine fish communities respond to climate variability over both river and ocean basins. Global Change Biology. 21(10):3608–3619.
Fry, F.E.J. 1971. The effect of environmental factors on the physiology of fish. In Fish physiology.
1–98. San Diego, CA: Academic Press.
Hochachka, P.W. and Somero, G.N. 2002. Biochemical Adaptation: Mechanism and Process in Physiological Evolution. Oxford University Press, New York.
Jamet, J.L. 1994. Feeding activity of adult roach (Rutilus rutilus (L.)), perch (Perca fluviatilis (L.)) and ruffe (Gymnocephalus cernuus (L.)) in eutrophic Lake Aydat (France). Aquatic Sciences. 56:376 387.
Jacobsen, L., Berg, S., Baktoft, H. and Skov, C. 2015. Behavioural strategy of large perch Perca fluviatilis varies between a mesotrophic and a hypereutrophic lake. Journal of Fish Biology. 86(3):1016–1029.
Jacobsen, L., Berg, S., Broberg, M., Jepsen, N. and Skov, C. 2002. Activity and food choice of piscivorous perch (Perca fluviatilis) in a eutrophic shallow lake: a radio‐telemetry study.
Freshwater Biology. 47(12):2370–2379.
14
Jensen, D.L., Overgaard, J., Wang, T., Gesser, H. and Malte, H. 2017. Temperature effects on aerobic scope and cardiac performance of Eurasian perch (Perca fluviatilis). Journal of Thermal Biology. 68(Pt B):162–169.
Kooijman, S.A.L.M. 2001. Quantitative aspects of metabolic organization: a discussion of concepts. Philos. Trans. R. Soc. B 356, 331–349.
Kuparinen, A., Klefoth, T. and Arlinghaus, R. 2010. Abiotic factors and fishing-related correlates of angling catch rates in pike (Esox lucius). Fisheries Research. 105(2):111–117.
Laundre, J.W., Hernandez, L. and Ripple, W.J. 2010. The Landscape of Fear: Ecological Implications of Being Afraid~!2009-09-09~!2009-11-16~!2010-02-02~! The Open Ecology Journal. 3(3):1–7.
Leander, J., Klaminder, J., Jonsson, M., Brodin, T., Leonardsson, K. and Hellström, G. 2019.
The old and the new: evaluating performance of acoustic telemetry systems in tracking migrating Atlantic salmon (Salmo salar) smolt and European eel (Anguilla anguilla) around hydropower facilities. Canadian Journal of Fisheries and Aquatic Sciences.
77(1):177–187.
Lefevre, S., McKenzie, D.J. and Nilsson, G.E. 2017. Models projecting the fate of fish populations under climate change need to be based on valid physiological mechanisms. Global
Change Biology. 23(9):3449–3459.
Liedtke, T. and Rub, M. 2012. “Techniques for telemetry transmitter attachment and evaluation of transmitter effects on fish performance: Chapter 4,” in Telemetry
Techniques: A User Guide for Fisheries Research, eds N. S. Adams, J. W. Beeman, and J.
H. Eiler. Bethesda, MD: American Fisheries Society.
Lorenzen, K. 2000. Allometry of natural mortality as a basis for assessing optimal release size in fish‐stocking programmes. Canadian Journal of Fisheries and Aquatic Sciences.
57(12):2374–2381.
Lorenzoni, M., Carosi, A., Pedicillo, G. and Trusso, A. 2007. A comparative study on the feeding competition of the Eurasian perch Perca fluviatilis L. and the ruffe Gymnocephalus cernuus (L.) in Lake Piediluco (Umbria, Italy). Bull Français Pêche Pisciculture. 387:35–
57.
Mckenzie, D, J., Axelsson, Chabot, M., Claireaux, D., Cooke, G., Corner, S, J.,
De Boeck, R., Domenici, G., Guerreiro, P., Hamer, P, M., et al. 2016. Conservation physiology of marine fishes: state of the art and prospects for policy. Conservation physiology. 4(1):cow046.
Melvin, S.D., Petit, M.A., Duvignacq, M.C., Sumpter, J.P. 2017. Towards improved behavioural testing in aquatic toxicology: Acclimation and observation times are important factors when designing behavioural tests with fish. Chemosphere 180, 430-436.
Nakayama, S., Doering‐Arjes, P., Linzmaier, S., Briege, J., Klefoth, T., Pieterek, T. and
Arlinghaus, R. 2018. Fine‐scale movement ecology of a freshwater top predator, Eurasian perch (Perca fluviatilis), in response to the abiotic factors environment over the course of a year. Ecology of Freshwater Fish. 27(3):798–812.
Nakayama, S., Laskowski, K.L., Klefoth, T. and Arlinghaus, R. 2016. Between- and within- individual variation in activity increases with water temperature in wild perch.
Behavioral Ecology. 27(6):1676–1683.
Neuman, E., Thoresson, G. and Sandström, O. 1996. Swimming activity of perch, Perca fluviatilis, in relation to temperature, day‐length and consumption. Annales Zoologici Fennici. 33:669–678.
Olsson, J., Svanbäck, R. and Eklöv, P. 2007. Effects of resource level and habitat type on behavioral and morphological plasticity in Eurasian perch. Oecologia. 152:48–56.
Ohlberger, J., Mehner, T., Staaks, G. and Hölker, F. 2012. Intraspecific temperature dependence of the scaling of metabolic rate with body mass in fishes and its ecological implications.
Oikos. 121(2):245–251.
15
Patterson, D.A., Cooke, S.J., Hinch, S.G., Robinson, K.A., Young, N., Farrell, A.P. and Miller, K.M. 2016. A perspective on physiological studies supporting the provision of scientific advice for the management of Fraser River sockeye salmon (Oncorhynchus nerka).
Conservation Physiology. 4(1):cow026.
Peake, S., McKinley, R.S. and Scruton, D.A. 2000. Swimming performance of walleye ( Stizostedion vitreum). Canadian Journal of Zoology. 78(9):1686–1690.
Persson, L., Andersson, J., Wahlstrom, E. and Eklov, P. 1996. Size-specific interactions in lake systems: predator gape limitation and prey growth rate and mortality. Ecology. 77:900–
911.
Persson, L. and Greenberg, L. A. 1990. Juvenile competitive bottlenecks: The perch (Perca fluviatilis)–roach (Rutilus rutilus) interaction. Ecology. 71:44–56.
Pitcher, T. J. and Turner, J. R. 1986. Danger at dawn: Experimental support for the twilight hypothesis in shoaling minnows. Journal of Fish Biology. 29(Suppl. A):59–70.
Piccolroaz, S., Toffolon, M. and Majone, B. 2013. A simple lumped model to convert air temperature into surface water temperature in lakes. Hydrology and Earth System Sciences. 17(8):3323–3338.
Reece, J.B. and Campbell, N.A. 2011. Campbell biology. 9. [expanded and enhanced] edition.
Boston: Boston : Pearson. p. 909.
Ryder, R. 1990. Biology of percid fishes. Environmental Biology of Fishes. 27(2):157–199.
Saaristo, M., Lagesson, A., Bertram, M.G., Fick, J., Klaminder, J., Johnstone, C.P., Wong, B.B.
and Brodin, T. 2019. Behavioural effects of psychoactive pharmaceutical exposure on Eurasian perch (Perca fluviatilis) in a multi-stressor environment. Science of the Total Environment. 655:1311–1320.
Schleuter, D. and Eckmann, R. 2006. Competition between perch (Perca fluviatilis) and ruffe (Gymnocephalus cernuus): The advantage of turning night into day. Freshwater Biology.
51(2):287–297.
SMHI. 2020. Ladda ner meteorologiska observationer [Download meterological observations]. SMHI. https://www.smhi.se/data/meteorologi/ladda-ner- meteorologiska-observationer#param=Air temperature (°C)
Instant,stations=all,stationid=149340 (retrieved 2020-01-22).
Stuart-Smith, R.D., Edgar, G.J., Barrett, N.S., Kininmonth, S.J. and Bates, A.E. 2015.
Thermal biases and vulnerability to warming in the world € s marine fauna. Nature.
528(7580):88–92.
Taipale, S.J., Vuorio, K., Strandberg, U., Kahilainen, K.K., Järvinen, M., Hiltunen, M., Peltomaa, E., and Kankaala, P. 2016. Lake eutrophication and brownification downgrade
availability and transfer of essential fatty acids for human consumption. Environment International. 96:156–166.
Thorpe, J.E. 1977. Morphology, physiology, behavior, and ecology of Perca fluviatilis L., and P. Perca-Flavescens-Mitchell. Journal of the Fisheries Research Board of Canada.
34:1504–1514.
Turesson, H. and Brönmark, C. 2007. Predator‐prey encounter rates in freshwater piscivores:
Effects of prey density and water transparency. Oecologia. 153(2):281–290.
Verhoeven, J. T. A., Arheimer, B., Yin, C. and Hefting, M. M. 2006. Regional and global concerns over wetlands and water quality. Trends in Ecology and Evolution. 21(2):96–
103.
16
Supporting figures
Figure 4a. The activity data (km/day) from all lakes against temperature (°C) together in the same graph. Note that the temperature is the atmospheric and not the water temperature. A polynomial trendline gave the highest R^2 value and is plotted in the graph.
Figure 5a. The activity data (km/day) in all the lakes against the light condition (h/day) together in the same graph.
Polynomial trendlines gave the highest R^2 values and are plotted in the graph.
R² = 0,0801
0 1 2 3 4
-15 -10 -5 0 5 10 15 20
Activity (km/day)
Temperature (°C)
R² = 0,0493
R² = 0,1382
R² = 0,0502
0 0,5 1 1,5 2 2,5 3 3,5 4
0 2 4 6 8 10 12 14
Activity (km/day)
Average light condition during photoperiod (h/day) Tosktjärn Abborrtjärn Stortjärn
17
Appendix
Table 1a. The table contains the r-coefficients in the correlations in Abborrtjärn between all the independent factors:
barometric pressure (hPa, relative standard deviation/day), air temperature (°C) and light condition (average light conditions during photoperiod in h/day).
Barometric pressure
(HPA, RSTDEV/day) Air
temperature (°C)
Light condition (h/day)
Barometric pressure (HPA, RSTDEV/day) 1
Air temperature (°C) -0.21 1 Light condition
(h/day)
-0.092 -0.074 1
Table 1b. The table contains the r-coefficients in the correlations in Stortjärn between all the independent factors:
barometric pressure (hPa, relative standard deviation/day), air temperature (°C) and light condition (average light conditions during photoperiod in h/day).
Barometric pressure
(HPA, RSTDEV/day) Air
temperature (°C)
Light condition (h/day) Barometric pressure
(HPA, RSTDEV/day) 1
Air temperature (°C) -0.037 1
Light condition (h/day) 0.19 -0.36 1
Table 1c. The table contains the r-coefficients in the correlations in Tosktjärn between all the independent factors:
barometric pressure (hPa, relative standard deviation/day), air temperature (°C) and light condition (average light conditions during photoperiod in h/day).
Barometric pressure (HPA,
RSTDEV/day) Air
temperature (°C)
Light condition (h/day) Barometric pressure (hPa,
RSTDEV/day)
1
Air temperature (°C) 0.071 1
Light condition (h/day) 0.47 -0.31 1
Table 2a. Summary of the results from the multiple regression in Abborrtjärn. The table shows the coefficients, standard error, t-quota, p-value and the quantile values. Activity (km/day) is the dependent variable and the independent variables are barometric pressure (hPa, relative standard deviation/day) and light condition (average light conditions during photoperiod in h/day).
Coefficients Standard
error t-
quota p-
value Lower
95% Upper
95% Lower
95,0% Upper 95,0%
Constant 2.9 0.24 12 <0.001 2.4 3.4 2.4 3.4
Barometric pressure (hPa, RSTDEV/day)
-1.3 0.62 -2.2 0.041 -2.6 -0.063 -2.6 -0.063
Light condition (h/day)
-0.060 0.027 -2.2 0.037 -0.12 -
0.0038
-0.12 -0.0038
18
Table 2b. Summary statistics from the multiple regression showing the R-values, standard error and the sample sizes for Abborrtjärn. The dependent variable is activity (km/day) and the independent are barometric pressure (hPa, relative standard deviation/day) and light condition (average light conditions during photoperiod in h/day).
Statistics from the Regression
Multiple-R 0.54
R-square 0.29
Adjusted R-square 0.22
Standard error 0.66
Observations 25
Table 3a. Summary of the results from the multiple regression in Stortjärn. The table shows the coefficients, standard error, t-quota, p-value and the quantile values. Activity (km/day) is the dependent variable and the independent variables are barometric pressure (hPa, relative standard deviation/day) and air temperature (°C).
Coeffici
ents Standa rd error
t-quota p-
value Lower
95% Upper
95% Lower
95,0% Upper 95,0%
Constant 1.6 0.13 12 <0.001 1.3 1.8 1.3 1.8
Barometric pressure (hPa, RSTDEV/day)
-0.59 0.21 -2.9 0.009 -1.0 -0.16 -1.0 -0.16
Air temperature
(°C) -0.050 0.015 -3.3 0.003 -0.080 -0.019 -0.080 -0.019
Table 3b. Summary statistics from the multiple regression showing the R-values, standard error and the sample sizes for Stortjärn. The dependent variable is activity (km/day) and the independent are barometric pressure (hPa, relative standard deviation/day) and air temperature (°C) .
Statistics from the Regression
Multiple-R 0.68
R-square 0.46
Adjusted R-square 0.41
Standard error 0.26
Observations 25
Table 4a. Summary of the results from the multiple regression in Tosktjärn. The table shows the coefficients, standard error, t-quota, p-value and the quantile values. Activity (km/day) is the dependent variable and the independent variables are barometric pressure (hPa, relative standard deviation/day) and air temperature (°C).
Coefficients Standard
error t-
quota p-
value Lower
95% Upper
95% Lower
95,0% Upper 95,0%
Constant 1.4 0.081 18 <0.001 1.3 1.6 1.3 1.6
Barometric pressure (hPa, RSTDEV/day )
0.53 0.25 2.1 0.049 0.002 1.1 0.002 1.1
Air
temperature (°C)
-0.020 0.009 -2.1 0.049 -0.039 <0.001 -0.039 <0.001
19
Table 4b. Summary statistics from the multiple regression showing the R-values, standard error and the sample sizes for Tosktjärn. The dependent variable is activity (km/day) and the independent variables are barometric pressure (hPa, relative standard deviation/day) and air temperature (°C).
Statistics from the regression
Multiple-R 0.55
R-square 0.30
Adjusted R-square 0.23
Standard error 0.22
Observations 22
