Population Characteristics of Ozark Bass ( Ambloplites constellatus ) in the Upper White River Basin of Northern Arkansas

Ozark Bass (Ambloplites constellatus) is an understudied, endemic fish species in the Upper White River Basin of northern Arkansas. This study was part of an effort by fisheries managers to gather baseline data about the Ozark Bass to aid in understanding population dynamics and contribute to the limited data available for use in determining the efficacy of harvest regulations. Select population characteristics of Ozark Bass in two northern Arkansas streams were determined, population characteristics of Ozark Bass were compared to Shadow Bass (Ambloplites ariommus) and Rock Bass (Amblop-lites rupestris) data collected from previous studies in southern Missouri, and relative condition, length-at-age, and annual survival of Ozark Bass were compared between sample streams. Sampling occurred in Crooked Creek and the Buffalo River during summer 2013 via boat e lectroshocking. Length and weight data were recorded for all Ozark Bass collected, and fish ages were determined through selective otolith retrieval and age-length keys. Ozark Bass in Crooked Creek had greater relative condition than Ozark Bass in Buffalo River (P < 0.001). Neither Ozark bass lengths nor log-transformed weights differed (P > 0.05) between sexes for fish collected from only the Buffalo River. Ozark Bass mean annual survival was similar between Crooked pled streams, and variability in growth rates and length-at-age were observed among Ambloplites species. Results of this study contribute to the understanding of the population dynamics of the Ozark Bass that will lead to im-proved fisheries management.


Introduction
The Ozark bass (Ambloplites constellatus) is a sport fish endemic to the Upper White River Basin of southern Missouri and northern Arkansas [1]. Until 1977, Ozark bass, shadow bass (Ambloplites ariommus), and rock bass (Ambloplites rupestris) were collectively recognized as Ambloplites rupestris [2]. Although closely related to Ozark bass, shadow bass and rock bass are more widely distributed throughout the mid-western United States, in addition to being present in the White River Basin [1]. Even though angler harvest is common, Ozark bass growth and other population characteristics have been minimally studied. However, even small changes in Ozark bass population characteristics could have potentially major implications for the species as a whole due to their narrow geographic range. Consequently, knowledge of Ozark bass' population characteristics and life cycle is necessary for fisheries biologists and managers to make informed management decisions, such as maximum harvest regulations, and for predicting possible changes to population dynamics throughout the known habitat range of Ozark bass.
The Upper White River Basin in northern Arkansas and southern Missouri, which encompasses most of the known range of Ozark bass, is a unique geographic setting due to the prevalence of underlying karst geology. However, the karst geology poses potential surface water and groundwater quality concerns due to ease of contamination because of the inter-connectedness of surface water and groundwater in regions of karst geology. Consequently, altered water quality, as a result of land-use change within the habitat range and/or changes to watershed hydrology, may be a threat to Ozark bass [3] [4], especially considering that little is known about the Ozark bass' growth habits and population dynamics.
In addition to potentially shifting water quality, human activities and potential climate change may affect the sustainability of the Ozark bass population. Human activities threaten Crooked Creek and Buffalo River, two important tributaries of the White River where Ozark bass are known to inhabit. Gravel mining caused Crooked Creek to be included on the 2016 Arkansas Department of Environmental Quality's 303(d) list of impaired waterbodies due to elevated total dissolved solids [5] [6]. Annual recreational visitation to the Buffalo River increased by over 800% since the early 1970s, from around 200,000 visitors in 1974  [7], where increased visitation has also increased angling pressure. Consequently, the Arkansas Game and Fish Commission (AGFC) recently enforced stricter harvest limitations on smallmouth bass (Micropterus dolomieu) [8]. However, it is unknown if more Ozark bass will be harvested in response to the new smallmouth bass regulations.
Climate change models for the southeastern United States predict an increase in the number and duration of droughts, greater mean annual air temperature, as well as an increased frequency of heavy rain events with shorter durations in the future [9]. The impacts of a potentially changing climate on Ozark bass populations are unknown at this time, but studies have suggested other sport fish may expand their habitat range and exhibit changes in growth rate potential [10].
Population data aid fisheries biologists and managers in evaluating current regulations to ensure that overfishing does not occur and also guide the modification of population size structure. Due to the lack of historical information, describing and evaluating the present population characteristics of the Ozark bass are critical for assessment of current and potential future impacts, such as altered water quality, climate change, and/or angling pressure, to the Ozark bass population. Any information generated for an under-studied aquatic species, such as the Ozark bass, will improve aquatic ecosystem management and protect certain species from over-exploitation. Due to the relatively recent taxonomic split of the Ozark bass from other similar species, further research is warranted.
Therefore, the objectives of this study were to: 1) determine population characteristics (i.e., relative condition, length-at-age, and annual survival) for Ozark bass in northern Arkansas; 2) compare relative condition, length-at-age, and annual survival of Ozark bass between sampled streams; and 3) compare Ozark bass population characteristics to those of other more well-studied bass species (i.e., shadow bass and rock bass) within the region. It was hypothesized that: 1) Ozark bass from the Buffalo River would have similar lengths-at-age and relative condition and lower annual survival than those from Crooked Creek because of greater fishing pressure in the Buffalo River [7] [11]; and 2) there would be similarities in population characteristics among sampled Ozark bass and other Ambloplites species due to similar phylogenetic histories.

Site Description
The

Data Collection
An electrofishing boat with pulsed direct current was used to collect Ozark bass   Figure 1). In Crooked Creek, five timed runs were conducted at both the AGFC Fred Berry Conservation and Education Center (FBCEC) located near Yellville and at the AGFC Lower Pyatt Access (Table 1; Figure 1).
Ozark bass were sampled in pools and other deep habitats reachable by boat.
Each run lasted approximately 10 minutes. The weight, to the nearest ±2 grams, and the total length, to the nearest millimeter, of each fish collected were recorded. The first 10 fish consecutively collected at each site in each 10-mm size class, if available, were retained and stored on ice in a cooler for sagittal otolith retrieval. When the maximum of 10 fish in any given size class had been collected, all additional fish collected in the same size class thereafter were weighed, measured, and released. Sex was only recorded from a sub-sample of Ozark bass from the Buffalo River to compare length and weight because differences were not observed between sexes.
Sagittal otoliths were sectioned with a low-speed saw, glued to glass slides, and then sanded and polished. Annuli were counted under a microscope by two observers to determine the age of the sampled fish. Both observers independently recounted the annuli and compared results to achieve a consensus when records indicated observational differences.
Since little historic population characteristic data exist for Ozark bass, similar data were gathered for the better-studied shadow and rock bass in the region for comparison. Shadow bass data from the Spring River and Eleven Point River in Arkansas were gathered from previously published literature and age data were

Calculations and Data Analyses
Relative condition was calculated for Ozark bass from both the Buffalo River and Crooked Creek according to Le Cren [22]: where Kn is relative condition, W is weight of the individual fish, and W' is a length-specific weight derived from the regression equation of log10 transformed weight-length data for historic Ozark bass data (n = 7188) from the Upper White River Basin ( Table 2). The population age distributions were estimated for both streams using population length frequencies and age-length keys generated from fish with known lengths and ages determined from sagittal otolith observations [23]. Annual survival rates were calculated for 2-to 9-year-old Ozark bass using the population age distributions from Crooked Creek and the Buffalo River [24]. Based on prior experience and visual observation during sampling, fish younger than age 2 were not fully recruited to the sampling gear, thus these data were not included in the calculations. Von Bertalanffy equations were developed separately for the Buffalo River and Crooked Creek and used to develop length-at-age models for Ozark bass from all available age groups to compare to observed length-at-age data [25], where ages were based on age length keys. Catch per effort (CPE) was calculated for each sampling site and averaged by stream.
Data were assessed graphically using histograms, box plots, bar plots, and scatter plots to visually examine data distributions, variances, and potential outliers to ensure the assumptions of parametric tests used were met. Data were transformed, if non-normally distributed, to better meet the normality assumption of parametric tests. Independence assumptions were checked using the Durbin-Watson statistic for all analysis of variance (ANOVA) tests. The Brown Table 2. Summary of locations of Ozark bass weight-length data (n = 7188 total), collected by the Arkansas Game and Fish Commission between 1986 and 2013, that were used to develop the regression equation to determine the length-specific weight variable used in Equation (1) Forsythe test was used to assess homogeneity of variance for residuals when analyzing data using one-way ANOVA. A two-factor ANOVA was conducted separately by stream to evaluate the effects of age class, sampling sites, and their interaction on Ozark bass mean length. One-factor ANOVAs were used to examine the effects of sex on Buffalo River Ozark bass length and weight. Data from both sexes were combined for calculating and comparing other population characteristics when significant differences were not observed (P > 0.05). Relative condition was compared between streams using a Welch two-sample t-test.
A two-factor ANOVA was used to examine the effects of stream and fish species (Ozark bass, shadow bass, and rock bass) on fish length-at-age. Following Robson and Chapman [24], 95% confidence intervals (CI) were used to compare Ozark bass annual survival between streams. Significance was judged at P < 0.05 for all statistical tests and all statistical analyses were conducted using R Statistical Software [26]. Calculated length-at-age for each stream and Von Bertalanffy estimated length-at-age data were compared visually through graphical analysis. Creek (Figure 2), with the majority of the collected fish from both water bodies being 5 years old or younger.

Ozark Bass Length and Weight Comparisons between Sexes
Neither Ozark bass lengths nor log-transformed weights differed (P > 0.05) between sexes for fish collected from only the Buffalo River (female mean length = 165 mm, mean weight = 104 g, and n = 46; male mean length = 169 mm, mean weight = 110 g, and n = 50). Consequently, fish-size data were combined across

Relative Condition and Annual Survival Comparisons between Streams
Mean relative condition of Ozark bass was greater in Crooked Creek than in the Buffalo River (P < 0.001; Figure 3). Relative condition of Ozark bass from the Buffalo River ranged from 75.6% to 153% and averaged 95.9%, while relative condition of Ozark bass from Crooked Creek ranged from 83.1% to 123% and averaged 104% (Figure 3).
Neither sampling site nor the interaction of age and sampling site influenced mean length of Ozark bass in Crooked Creek or the Buffalo River (P > 0.05; Table 3), thus annual survival was not calculated for individual sampling sites.
However, mean annual survival rates of Ozark bass, age 2 to 9, were similar across streams and averaged 50% (±7% for the 95% CI) and 55% (±5% for the 95% CI) for the Buffalo River and Crooked Creek, respectively.

Length-at-Age Comparison between Streams and Von Bertalanffy Estimates
The length-at-age data for both sampled streams depicted similar trends, where length generally increased as the fish aged ( Figure 4). However, the relationship between age and length differed somewhat across streams (P < 0.001; Table 4) Ozark bass from Buffalo river were shorter compared to those from Crooked Creek at age 1, but length-at-age was greater in the Buffalo River compared to in Crooked Creek from age 2 to age 6 ( Figure 4). Length-at-age between streams after age 6 was unable to be compared because data were either unavailable (e.g., no age 8 fish were collected from the Buffalo River) or there was only one fish collected (e.g., n = 1 for both age 7 and age 9 fish collected in the Buffalo River).
Growth was greater in the Buffalo River up to age 3 and became more similar   Creek (asterisks). Error bars are standard error. Lengths-at-age between streams after age 6 were unable to be compared because data were either unavailable (e.g., no age 8 fish were collected from the Buffalo River) or there was only one fish collected (e.g., n = 1 for both age 7 and age 9 fish collected in the Buffalo River).  streams after age 3, but overall growth was greatest in the Buffalo River ( Figure   4). The greatest growth rates for both streams were observed up to age 3, after which growth rates declined in both streams (Figure 4). Von Bertalanffy terms for the Buffalo River were 258 mm for maximum attainable length (L∞), 0.346 for the growth constant (k), and −0.087 for the length-at-age 0 (t o ). The terms for Crooked Creek were 260 mm for maximum attainable length (L∞), 0.299 for the growth constant (k), and −0.328 for the length-at-age 0 (t o ). Length-at-age for both Crooked Creek and Buffalo River followed similar trends to the estimated Von Bertalanffy growth curves; however, Buffalo River Ozark bass length-at-age deviated from the modeled growth curve at age 7 likely due to the few 7-and 8-year-old fish that were collected ( Figure 5).

Length-at-Age Comparison among Ambloplites Species
Fish lengths from this study ranged from 61 (age 1) to 260 mm (age 9) for Ozark bass, while lengths used in the analysis among species ranged from 96 (age 2) to 275 mm (age 6) for shadow bass and 84 (age 2) to 259 mm (age 7) for rock bass. Although length generally increased with age across all species, an interaction was observed between age and Ambloplites species (P < 0.001; Figure 6). Shadow bass length-at-age was generally greater than or equal to that for Ozark bass, regardless of age, and greater than that for rock bass from ages 2 through 4 ( Figure 6). After age 4, rock bass length-at-age surpassed that of the other the other two species. Growth rates of Ozark bass were initially greater up to age 3 compared to that for shadow bass ( Figure 6). Rock bass growth rates were more similar to those for Ozark bass at ages 2 and 3, but were greater after age 3 ( Figure 6). After age 3, the length-at-age relationship became more similar between Ozark bass and shadow bass ( Figure 6).

Comparison of Ozark Bass Population Characteristics between Streams
Several Ozark bass population characteristics differed somewhat between Crooked Creek and the Buffalo River. Catch-per-effort was substantially greater in Crooked Creek (101.5 fish/hour) than in the Buffalo River (32.9 fish/hour).
Since CPE is sometimes related to fish density [27], there may have been a greater density (i.e., fish per linear river length) of Ozark bass in Crooked Creek   Crooked Creek is also nested within the Bull Shoals Lake Watershed [hydrologic unit code (HUC) 11010003], which has 26% less forest and 16% more agricultural land use than the Buffalo River Watershed (HUC 11010005) [28].
Increased nutrient loading in streams has been observed as percent agricultural land use increased [29]. Potentially greater concentrations of available nutrients in Crooked Creek, due to differences in land use within the watersheds, may have increased food availability to support a greater quantity of Ozark bass.  [27] observed that largemouth bass (Micropterus salmoides) were more likely to avoid boat electroshocking equipment during the daytime compared to nighttime, leading to greater CPE during nighttime sampling. Furthermore, sample timing may also have affected the size structure of fish collected between samples (e.g., with smallmouth bass) [30].
However, if time of sampling (i.e., daytime vs. nighttime) had been the primary control on observed CPE differences between streams in this study, Buffalo River CPE would likely have been greater than that in Crooked Creek, since the Buffalo River was sampled at night. Differential sample times of the day between the two streams may have been a limitation of this study, but the data collected and results generated for an understudied fish species in the Ozark Highlands provides a reasonable starting point for future, more extensive research on the Ozark bass.
Similar to CPE, it was hypothesized that relative condition of Ozark bass would be similar between streams. However, relative condition was also greater in Crooked Creek (P < 0.001; Figure 3). Greater relative condition has been linked to greater prey availability, better reproductive success, and increased growth [31]. This would suggest that food availability was greater in Crooked Creek than in the Buffalo River, which may also partially explain the greater CPE in Crooked Creek compared to the Buffalo River. Although exact fishing pressure on the streams remains unknown, differential fishing pressure is another possible explanation for observed differences in relative condition between streams, as anglers tend to keep the better-conditioned fish [32]. Even though both streams are popular fishing destinations, the Buffalo River receives more visitors and is better known because of its larger size and designation as a National Park Service unit [13] [14] and is more easily accessible by motorized boat from the White River, which is another popular fishing destination in the region.
Similar to CPE and relative condition, longer fish, in general, were collected from Crooked Creek than from the Buffalo River, with a 10 mm difference in mean total length between streams (Crooked Creek mean total length = 161.7 mm and Buffalo River mean total length = 151.2 mm). However, differences in fish length in each stream tended to increase as age increased, with greater mean length-at-age observed in Buffalo River fish after age 2 and up to age 6 ( Figure   4). A relatively large number of 5-year-old fish were collected in Crooked Creek, indicating that was a strong year class, which may have contributed to the larger average size in Crooked Creek, compared to the Buffalo River. Observed length-at-age data were similar to the Von Bertalanffy growth model for Crooked Creek (Figure 5(a)). Buffalo River fish up to age 6 and at age 9 were comparable to the Von Bertalanffy growth models, but Buffalo River fish between age 6 and 9 were either not collected or shorter in length than predicted in the model, suggesting low abundance, potentially due to limited sample size and/or growth limitations ( Figure 5(b)).
Whisenant and Maughan [17]  In contrast to CPE, relative condition, and length, the hypothesis that annual survival would be lower in the Buffalo River was rejected after observing similarities in mean annual survival between the Buffalo River and Crooked Creek (50% ± 7% and 55% ± 5% for the 95% CI, respectively), where annual survival rates across all age classes did not differ between streams. Similar annual survival rates across all age classes may have been linked to the proportionality of fishing pressure and capture success relative to the size of each stream, since annual survival is directly related to total annual mortality, where total mortality is the sum of fishing mortality (e.g., anglers harvesting fish and fish death due to hooking) and natural mortality (e.g., predators, disease, and water quality issues) [33]. In contrast to some properties, it appears that there were also several Ozark bass population characteristics that were similar between the Buffalo River and Crooked Creek.

Comparison of Population Characteristics between Ozark Bass and Other Ambloplites Species
Variability in growth rates and length-at-age were observed among Ambloplites species. With increasing age, length differences between Ozark bass and shadow bass became more similar, while differences between rock bass and Ozark bass length became more pronounced with age ( Figure 6). Shadow bass mean annual survival in the Spring River and Eleven Point River in Arkansas was reported at 64% across all age groups [21], which was greater than the mean annual survival of Ozark bass across both streams in this study (52%  [17] reported rock bass grew faster than Ozark bass after age 1. However, some of the variability among Ozark, rock, and shadow bass length-age characteristics may be due to using otoliths (Ozark and shadow bass) compared to scales (rock bass) to determine fish ages.

Potential Study Limitations
This study represented observations of Ozark bass during a single year, from a single season, and collected by a single method. The resulting size structure of Ozark bass (Figure 2) determined in this study may not have been generated in an ideal manner and may be partially biased by collection method and fish behavior at that time of the year sampling was conducted [34]. A single year of data also limits assessment of inter-annual variations in size classes among the streams sampled due to environmental conditions (i.e., differential precipitation patterns), which may partially explain the apparent strong age 5 class in Crooked Creek. In addition, length-at-age data collected for this study compared similarly to the Von Bertalanffy estimated length-at-age model, except in the Buffalo River for age 7. The deviation from the model's prediction can be attributed to low sample size in the age 7 category, as only one age 7 Ozark bass was reported from Buffalo River during this study. However, results of this study still represent a useful starting point in the characterization this understudied fish species in streams of the Ozark Highlands-Boston Mountains ecoregions that can assist fisheries biologists and managers in understanding and protecting Ozark bass from developing water quality issues and other environmental changes.
Subsequent sampling events over multiple years would yield better understanding of temporal population shifts, reveal patterns in annual variation, and provide a better representation of the actual endemic Ozark bass population.
With this information, future impacts to the Ozark bass, such as through land use or climate change, could be more easily recognized and management strategies assessed and implemented.

Conclusion
Establishing