Printer Friendly

Mountain Lion (Puma concolor) Population Characteristics in the Little Missouri Badlands of North Dakota.


Mountain lions (Puma concolor) historically ranged across most of the American continents, and the species ranks as having the most extensive range of any terrestrial mammal in the Western Hemisphere, excluding humans (Logan and Sweanor, 2001). This distinction is testimony to the incredible adaptability of mountain lions, as evidenced by both the variety of habitats they occupy and prey they consume (Fecske et al., 2011, Logan and Sweanor, 2001). Mountain lions have ecological impacts on the ecosystems they inhabit but also elicit a wide range of sentiments from people (Jansen, 2011). This interesting juxtaposition of both positive and negative consequences associated with having mountain lions on the landscape can leave management agencies with difficult and complex management decisions; this is particularly true when dealing with newly recolonized populations with little data available to help guide discussions (Jenks, 2018).

In North Dakota mountain lions historically ranged throughout the state, although they were considered scarce in the open prairies that typify the eastern part of the state (Bailey, 1926). Human persecution led to the species disappearance and believed extirpation by the early 1900s (Bailey, 1926). Decades later, mountain lions have naturally recolonized areas of suitable habitat within the state, and since 2005, have been primarily managed by the North Dakota Game and Fish Department (NDGFD) via an annual, limited-take harvest season (NDGFD, 2006).

The North Dakota mountain lion population warrants investigation for several reasons. First, the population is considered relatively small yet experiences considerable levels of mortality, including hunting (Wilckens, 2014; NDGFD, 2017). Second, it is considered vulnerable to over-exploitation because it is semi-isolated from other mountain lion populations by large expanses of agricultural and grassland habitat (NDGFD, 2006). Further, because this population occurs on the eastern fringe of the species' current range (excluding the remnant Florida population), this population represents a potential source of dispersing individuals that could contribute to continued eastward range expansion (LaRue and Nielsen, 2008; Juarez et al., 2016; LaRue and Nielsen, 2016). Finally, this population exists in habitats that are atypical of most other studied mountain lion populations (i.e. juniper-forested draws and buttes, not mountains or deserts) and can serve as a model population in future decisions as mountain lions continue to expand into other non-traditional habitats.

Despite these considerations, robust data regarding many aspects of this population have remained scant. Therefore, we initiated a study to improve our understanding of this population and its demographic status. Our objectives were to: (1) estimate annual survival rates and document sources of mortality of mountain lions in North Dakota, (2) use statistical population reconstruction (SPR) to build a population model, and (3) use the population model to produce estimates of mountain lion population abundance and investigate population trend. With these results managers will be better equipped to make responsible and well-informed management decisions in North Dakota and potentially beyond.



Our study area encompassed approximately 2800 kmJ, primarily within Billings, Dunn, and McKenzie counties, North Dakota, U.S.A. (47[degrees]43'N latitude, 103[degrees]24'W longitude; Fig. 1). Much of the study area was comprised of the Little Missouri Badlands Region (Badlands) but also included the Killdeer Mountains. The Badlands are characterized by a highly variable landscape of clay slopes, steep canyons, buttes, and bottomlands carved by the Little Missouri River (Hagen et al., 2005). Elevation ranges from 570 m to 710 m above mean sea level (Hagen el al., 2005). The Badlands are vegetated, primarily on north and east facing slopes, with stands of Rocky Mountain juniper (Juniperus scopulorum) and green ash (Fraxinus pennsylvanica); riparian areas generally contain stands of cottonwood (Populus deltoides). Shortgrass prairie is dominant on southern and western slopes, plateaus, and bottomlands (Dyke et al., 2015).

East of and adjacent to the Badlands are the Killdeer Mountains, a 60 [km.sup.2] elevated region rising 213-305 m above the surrounding prairie, to a maximum elevation of 1010 m above mean sea level (Hagen et al., 2005). The vegetation of the Killdeer Mountains consists of deciduous woodlands, dominated by burr oak (Quercus macrocarpa), quaking aspen (Populus tremuloides), and green ash, interspersed with open areas of shortgrass prairie and some rocky escarpments (Hagen et al., 2005). Our study area was a mixture of public (56%) and private land (44%). Annual precipitation averaged 16.2 cm, and mean monthly temperatures ranged from a low of -9.2 C in January to a high of 20.8 C in July (National Centers for Environmental Information, 2013). Cattle grazing was the most common land use; however, oil and gas development has increased dramatically in recent years (Dyke et al., 2015). Primary prey available to mountain lions throughout the region include mule deer (Odocoileus hemionus), white-tailed deer (O. virginianus), elk (Cervus elaphus), and bighorn sheep (Ovis canadensis). Secondary prey included porcupine (Erethizon dorsatum), beaver (Castor canadensis), turkey (Meleagris gallopavo), raccoon (Procyon lotor), several species of rodents and lagomorphs, and domestic livestock (e.g., cattle, horses, goats; Wilckens et al.,

2016). Mountain lions are classified as a furbearer in North Dakota with a regulated hunting season (September through March). The state is divided into two management zones (Zone 1 and Zone 2) with a season harvest limit in place for Zone 1, whereas Zone 2 is deemed 'unsuitable habitat' and has no season harvest limit. Our study area was located within the boundaries of mountain lion management Zone 1 (NDGFD, 2017).


From 2012-2016 we captured mountain lions via foothold traps and foothold cable restraints at established bait sites (Logan et al, 1999). We ear-tagged and fitted real-time Global Positioning Systems (GPS) collars (G2110E, Advanced Telemetry Systems, Isanti, Minnesota, U.S.A.) on captured subadult (dispersal until 3 y) and adult (>3 y) mountain lions, whereas we only ear-tagged kittens (dependent on mother; see Johnson et al., 2017 and Wilckens et al, 2016 for detailed capture methods). Our capture and handling methods followed the guidelines of the American Society of Mammalogists (Sikes et al, 2016) and were approved by the Institutional Animal Care and Use Committee at South Dakota State University (Approval numbers 11-080Aand 14-094A).


We estimated annual survival of mountain lions using a known fate analysis with the logit-link function in Program MARK (White and Burnham, 1999). We created a monthly encounter history for each radio-collared adult or independent subadult mountain lion beginning at initial capture date and continuing until mortality, collar failure, or calendar year end. Individuals were assigned an age covariate (subadult or adult) which was updated if an individual survived and transitioned into the adult cohort in a given year. Collar failures were right censored, but individuals were re- entered into the analysis if recaptured. We developed a series of a priori models to investigate the effects of sex, age (subadult or adult), hunting season (early season, late season [use of hounds allowed], or early and late season combined), and calendar year on mountain lion annual survival.


We recorded data on cause-specific mortalities of mountain lions by investigating collar mortality signals immediately upon detection and from carcass necropsies of harvested or otherwise killed mountain lions. All mountain lions killed (e.g., legal harvest, depredation, vehicle-collisions) in North Dakota were required by law to be reported to the North Dakota Game and Fish Department.


To build our population model, we analyzed age-at-harvest data and radio-collar data using statistical population reconstruction (SPR; Gove et al, 2002; Skalski et al, 2005). This technique estimates historical abundance based upon a joint likelihood model analyzing age-at-harvest data in concert with at least one source of auxiliary information. The general form of the joint likelihood model is

[L.sub.-Joint] = [L.sub.Age-at-harvest] x [L.sub.Auxiliary] x [L.sub.Reporting]

where [L.sub.Age-at-harvest] a likelihood model that describes the cohort data within the age-at-harvest matrix as a function of survival and harvest parameters; [L.sub.Auxiliary] is a likelihood model used to estimate one or more of the abundance, survival, or harvest parameters; and [L.sub.Reporting] is a likelihood model that describes the probability a harvested animal is reported and included in the age-at-harvest data (Gove et al, 2002; Skalski et al, 2005). We used program PopRecon 2.0.26 (Clawson et al, 2017; Lady and Skalski, 2015) to estimate the likelihood models and parameters.

Our age-at-harvest data consisted of all reported mountain lion mortalities (legal harvests and nonharvest mortalities [e.g., illegal take, depredation, vehicle collision]) that occurred in North Dakota, including those on the Fort Berthold Reservation, between March 2005 and February 2017. We included Zone 2 mortalities to maximize our sample size but doing so required the assumption all mortalities from Zone 2 originated from the North Dakota population and could be included in historical cohorts. Although nearly all mountain lions were aged to year of birth via tooth cementum analysis (Matson's Laboratory, Manhattan, Montana, U.S.A.), the age-at-harvest matrix was sparse when encompassing all ages (i.e., age 0 to age 16). Therefore, we pooled the adult data into a 3+ year age class and assumed harvest and survival probabilities were similar among the 3+ age classes and sexes. Pooling adult data has been shown to have negligible effects on SPR model performance (Skalski et al, 2012a) particularly when demographic parameters do not differ within the adult age class (Gast, 2012).

Age-at-harvest likelihood.--We modeled time as two discrete periods: hunting season (September through February) and nonhunting season (March through August). A necessary assumption with this structure was natural mortality was negligible during the hunting season, and this assumption was supported by data on mountain lions from the Black Hills of South Dakota (Juarez, 2014). Harvest probability (P) refers to the probability an individual is legally harvested during the hunting season. Conversely, survival probability (S) should approximate natural survival, or the probability an individual survives the nonhunting season. Therefore, the probability an animal survives from the beginning of a hunting season in year i to the beginning of the hunting season the following year is (1 - P) * S.

In SPR models hunter effort data are necessary to estimate harvest probabilities (Gast et al, 2013). The relationship between hunter effort and harvest probability is:

[mathematical expression not reproducible]

where P is the harvest probability, c is an estimated harvest vulnerability coefficient, y is an optional random effects parameter that allows for interannual variation in the relationship between hunter effort and harvest probability, and f is the supplied annual estimate of annual hunter effort (Clawson, 2015). However, we lacked detailed information regarding annual hunter effort. We approximated annual hunter effort by dividing the annual Zone 1 hunting season limit by the total number of statewide mortalities for each year. The Zone 1 limit varied throughout the years modeled, from no more than five legal harvests allowed during the first 3 y, up to 21 legal harvests allowed during four of the final 5 y, with the final year reduced to no more than 15 legal harvests allowed. Approximating hunter effort in this way resulted in higher levels of hunter effort in years when the hunting season limit was not met and lower levels of hunter effort when the hunting season limit was met or exceeded. This seemed appropriate as total hunter effort should have increased with every additional day the hunting season remained open due to the hunting season limit not being met; conversely, total hunter effort should have been lower in years when the hunting season limit was met in a short time. Additionally, this approximation of hunter effort represented higher catch-per-unit-effort in years when more mountain lions were killed outside of the hunting season or in Zone 2. We assumed this higher catch-per-unit-effort reflected higher mountain lion densities because out-of-season and Zone 2 effort (i.e., vehicle collisions, depredation, chance encounters) likely did not change significantly annually. True hunter effort was likely variable among and within years and was likely highly influenced by yearly hunting season limits and regulations, population size, and weather/tracking conditions (Wilckens, 2014).

Auxiliary likelihood.--Because age-at-harvest data alone cannot estimate the necessary demographic parameters needed in SPR, auxiliary field studies are needed to provide the missing information (Gove et al., 2002). We used 5 y of mark-recapture data to fulfill the requirement of independent auxiliary data. From this information we used the number of tagged independent mountain lions alive at the beginning of a hunting season and the number of those same individuals legally harvested to estimate harvest probability P. To estimate survival probability S, we used the number of tagged independent mountain lions alive at the end of a hunting season and the number of those same individuals that survived until the start of the next hunting season (Clawson et al., 2017).

Reporting likelihood.--All mountain lions killed (e.g., legal harvest, depredation, vehicle-collisions) in North Dakota are required by law to be reported to the North Dakota Game and Fish Department, yet a few database entries were missing age information. Therefore, we estimated the annual reporting rates as year-specific using the total number of reported and aged mountain lions and the number of known mortalities. It is likely a few additional mortalities occurred (i.e., illegal harvest, hunter wounding loss) that were never reported (Wilckens, 2014), but this information cannot be known and was not used in the model.

We developed six a priori models (Table 4) to investigate the effects of varying harvest and survival probability configurations based upon those provided by Gove et al. (2002), our own knowledge of the system, and those possible in Program PopRecon at the time (Clawson et al., 2017). We ranked candidate models using Akaike's Information Criterion corrected for small sample sizes (AICc; Bumham and Anderson, 2002). Output from the best-fitting models was analyzed for realism, and models were excluded from further consideration if output was biologically unrealistic (Skalski et al., 2012b). Finally, we used point-deletion techniques to assess the stability of the abundance estimates by removing both the first 2 and first 4 y of harvest data and then comparing abundance estimates and trends among the three datasets (Skalski et al., 2012b).


Between 2011 and 2016, we live-captured 31 mountain lions, 13 of which were resident adults (6 male (M), 7 female (F)), six were considered dispersing subadults (3 M, 3 F), and 12 were dependent kittens (~10 mo.: 2 M, 3 F; ~1 mo.: 6 M, 1 F). All captured adult and subadult mountain lions were collared and ear-tagged (n = 19). Two of the dependent kittens were collared, whereas the other 10 were ear-tagged only. Two subadults (1 M, 1 F) transitioned to adults during the study, and two of the noncollared dependent kittens (1 M, IF) transitioned to subadults during the study (known via sightings and/or harvest). In addition to animals captured in North Dakota, two subadult males previously captured and marked in Montana (R. Matchett, Charles M. Russell National Wildlife Refuge, pers. comm.) also were located within our study area; both individuals were included in our documentation of cause-specific mortalities, whereas only one was included in our survival analysis (given the individual was ear-tagged but not collared only mortality could be known with certainty).

We estimated survival with data collected from 20 individuals (9 M, 11 F) captured from 2011-2016, along with one subadult male mountain lion collared in Montana, resulting in 245 total monthly encounter histories. The top-ranked model for the known fate analysis included year, sex, and late hunting season as covariates (Table 1). Age of the individual was not influential in the analysis. However, we considered the second-ranked model (sex and late hunting season) as a competing model because it was <2 [AIC.sub.c] ([DELTA][AIC.sub.c] = 1.9653) from the top-ranked model and carried considerable [AIC.sub.c] weight ([DELTA]AICc weight = 0.2085). Together, the top two models carried the majority of [AIC.sub.c] weight (combined [AIC.sub.c] weight = 0.7656). Overall annual survival estimates from the two competing models were nearly identical (annual survival = 0.456 and 0.455, SE = 0.107 and 0.102, respectively). Because the year variable was the only difference between the competing models, we chose to exclude it from future consideration because: (1) annual survival estimates were nearly identical between models, (2) the year variable beta estimate included zero (beta = 0.0015, 95% CI = -0.0001-0.0031), and (3) the observed effect was likely due to changes in sample sizes between years rather than being biological meaningful. Therefore, we considered the model with sex and late season as the top model and used it alone to estimate survival. The late season variable had a strong negative influence on survival (beta = -3.1325, 95% CI = -4.4161-1.8488). Sex-specific survival was estimated at 0.589 (95% CI = 0.338-0.800) for females and 0.259 (95% CI = 0.089-0.555) for males. Over the 5 y period, the average annual survival rate estimated using known fate data was 0.456 (95% CI = 0.264-0.661).

We recorded 17 mortalities of marked mountain lions over the 5 y period; 10 legal hunter harvests, two illegal harvests, three depredation removals, and two vehicle collisions (Table 2). In addition we documented the probable failures of two litters of dependent marked kittens (n = 7) due to the radio-collared females being legally harvested.

Our age-at-harvest matrix consisted of 12 y of data (2005-2017) and contained a total of 189 mortalities (legal harvests [n = 129] and nonharvest mortalities [n = 60]), including two marked individuals known to have immigrated from Montana into North Dakota. Within the dataset 183 entries were aged via cementum annuli (Table 3). Our auxiliary survival information had 5 y (2012-2017) of mark-recapture data, with a total of 40 yearly entries, 36 of which survived. Our auxiliary harvest information had 5 y (2012-2017) of mark-recapture data, with a total of 37 yearly entries, 13 of which were harvested. We were unable to incorporate random harvest or survival effects into any models due to program instability and low sample sizes.

We considered the MpyaS model as our top model because it was >2 AICc ([DELTA]AICc = 6.7772) from the next best model (Table 4). This model structure assumed natural survival (i.e., the probability an individual survives the nonhunting season) was constant across years and age classes, whereas harvest probabilities varied by year and age class. Natural survival was estimated at 0.893 (95% CI = 0.796-0.990) across all age classes and years. Harvest probabilities varied by year and age class (Table 5) and ranged from a minimum of 0.103 (95% CI = 0.054-0.152) for the age 0 class to a maximum of 0.535 for the age 3+ age class (95% CI = 0.355-0.715). Annual abundance estimates from the top model ranged from low of 27 total mountain lions (95% CI = 1-52) in 2005-2006 to a high of 165 total mountain lions (95% CI = 89-241) in 2011-2012 (Table 6, Fig. 2). Finally, the model produced similar abundance estimates with both the full and reduced datasets, indicating the model was robust to the number of years of data used (Skalski et al., 2012b; Fig. 3).


The overall 5 y annual survival rate we documented (0.456) was below the survival estimates documented in several other hunted populations, including South Dakota (0.64, Jansen, 2011), Montana (0.65, Robinson et al., 2014), the Pacific Northwest (0.59, Lambert et al, 2006), Washington (0.56, Cooley et al, 2009; 0.60, Robinson et al, 2008), Arizona (0.62, Cunningham et al., 2001), Utah (0.64, Stoner et al, 2006), and Alberta (0.67, Knopff et al, 2010). The North Dakota population was thought to be in decline during this same period, consistent with several studies documenting similarly low annual survival rates (Cooley et al., 2009; Knopff et al., 2010; Lambert et al., 2006; Stoner et al, 2006).

Prior research has suggested mountain lion mortalities due to hunting are not compensated by a reduction in natural deaths or increased vital rates but rather through immigration of subadults from nearby populations (Cooley et al, 2009; Cooley et al, 2011; Robinson et al, 2014). Immigration into North Dakota has been previously documented (Juarez et al, 2016; Thompson and Jenks, 2010; Wilckens, 2014). However, the number of immigrating mountain lions (particularly females) likely is low due to harvest of neighboring populations in South Dakota and Montana (Jansen, 2011), as well as the vast expanse of agricultural and grassland landscapes found between populations, which results in a geographically semi-isolated Badlands population (NDGFD, 2006). Therefore, this population should be considered particularly at risk of over-exploitation.

Hunter harvest was the leading cause of mortality in our study (Table 2), consistent with the findings of many other studies across North America (Jansen, 2011; Lambert et al., 2006; Logan and Sweanor, 2001; Robinson et al., 2008; Stoner et al., 2006). More broadly, all mortalities of marked mountain lions between 2012 and 2016 were human-induced, including hunter harvest (n = 10), depredation (n = 3), illegal harvest (n = 2), and vehicle collision (n = 2). Furthermore, two of the hunter-harvested mountain lions were radio-collared females with dependent kittens which we had ear-tagged at the natal den sites during the summer months. While it is possible one or both litters could have already been lost for a myriad of reasons, they surely would not have survived the loss of the mother. This further supports the low estimated survival rates and overall population decline we documented and demonstrates the potential population ramifications caused by the loss of even a few reproductively active females, particularly in a relatively small and isolated population.

The top-ranked survival model in our analysis included the late hunting season (hound use permitted) and sex covariates. Use of hounds is widely considered to be the most effective tool for hunting mountain lions (CMGWG, 2005). During the time period in which we calculated survival, the late hunting season (hound use permitted) began immediately following the deer gun season (~Nov 30) which typically coincided with the arrival of snow. Furthermore, the Badlands have an extensive network of roads as a result of oil and gas development. This combination of hound use, snow, and road density translates into a much-improved ability for hunters to locate, bay, and harvest mountain lions during the late hunting season (Dawn, 2002), and this is reflected in our survival analysis.

Determining the influence of sex on mountain lion survival is difficult because of several factors influencing sex-specific harvest probabilities. Anderson et al. (2009) suggested that because male mountain lions generally exhibit larger daily movements, they become more susceptible to harvest, particularly by hunters aided by hounds. However, it is reasonable to think females may be more susceptible to boot hunters, predator callers, and opportunistic take because females generally make up a larger proportion of the population on the landscape (Logan and Sweanor, 2001). Moreover, sex may be misidentified, particularly by inexperienced hunters (Dawn, 2002), or nearly impossible to determine if the mountain lion bays in a hole or cave (often the case in North Dakota). Finally, changes in hunting regulations can impact sex-specific harvest probabilities. During the last year of our study, the Zone 1 total hunting limit was lowered from 21 to 15, and a female limit of three was implemented during the late hunting season (NDGFD, 2017). This could have led to hunters avoiding harvest of female mountain lions to keep the hunting season open as long as possible, and personal communication with hunters supported this notion. During that year five of the six marked mountain lions were female, likely resulting in increased survival estimates. Our result of females having higher survival rates than males also has been documented in studies in Washington (Cooley et al., 2009) and Montana (Robinson et al., 2014).

Our population model indicated an increasing trend in abundance until 2011-2012 when the trend reversed (Fig. 2). A sharp decrease in abundance followed for several years, and although the decreasing trend continued, the population seemed to level off in the most recent years. The sharp drop in abundance coincided with the largest number of both legal harvests (n = 17) and nonharvest mortalities (n = 16) recorded in a single year (2011-2012) in North Dakota. In the years that followed, the number of legal harvests remained at similar levels while nonharvest mortalities decreased to five or less annually (Fig. 4). While speculative it seems reasonable the additional nonharvest mortalities, along with legal harvest, were the primary drivers of the decline. Because the nonharvest mortalities have been reduced, the population could be approaching a stable equilibrium maintained by current levels of mortality. Additionally, this relationship lends support to the idea that in the effective absence of immigration nonharvest mortalities in this population may be mostly density dependent.

The harvest and survival probabilities estimated by our SPR model support the low annual survival rate estimated via known fate data. The estimated natural nonharvest mortality rate of ~0.11 agrees with research conducted in Montana (Robinson et al., 2015). However, the relatively high estimated harvest probabilities result in low annual survival estimates, particularly among the older adult age classes. This agreement between methods adds confidence to both results. Additionally, both analyses indicate mortality during the hunting season was the primary contributor to low annual survival rates.

All models are simplified representations of reality. Therefore, it is important to address their limitations, particularly when applying a technique such as SPR, which had only recently begun to be used in wildlife population modeling. Perhaps our most important consideration was the small yearly sample sizes within our age-at-harvest dataset. While there are no published guidelines for SPR data minimum requirements, our dataset was considered sparse (M. Clawson, University of Washington, pers. comm.). As such we stressed the importance of realistic and supportable model output based upon the biology of mountain lions, our knowledge of the system, and comparison with our concurrent and previous (Wilckens, 2014) research results (Skalski et al., 2012b). One direct consequence of sparse data we observed were four occasions of zeroes within our age-at-harvest data due to no mortalities reported within that particular year and age-class. Because abundance (N) is calculated as N = h/P (where h = number harvested and P = harvest probability) within each year and age class, a zero in the observed harvest translates to a zero in the corresponding year's age class. This is an unrealistic result and leads to underestimation of total abundance within that year, particularly if the zero occurs in the younger, larger age classes. The four zeroes within our age-at-harvest data occurred in three separate years negatively biasing total abundance for those years (Table 6).

Another limitation of our data was the lack of detailed hunter effort information. Some survey data had been collected by the state (NDGFD, 2017) but was too coarse to be of use in modeling. Changes in hunting season structure and harvest limits over the years increased the difficulty of approximating annual hunter effort. Furthermore, different methods (e.g., use of hounds, predator calling, chance encounters) employed by hunters of varying skill and experience, coupled with changes in hunting efficiency and perceived novelty of hunting mountain lions in a new state, all contribute to a complex reality of hunter effort. Nonetheless, we believe our technique of estimating hunter effort was the best available avenue. In any event we were required to assume we approximated annual hunter effort in an appropriate and meaningful way. More refined estimation of annual hunter effort in the future would certainly increase model performance and accuracy.

A major assumption of SPR modeling is that survival and harvest processes are modeled correctly and estimated without bias (Skalski et al., 2005). Our top model configured natural survival to be constant across years with harvest probabilities that varied by both year and age class (Table 4). We are comfortable with this modeling of the harvest and survival processes, as hunter harvest was the leading mortality source of marked mountain lions during our study. This should translate to relatively high survival rates during the nonharvest period, and the model estimated them at 0.893 across years and age-classes. The similarity between years and age-classes seems plausible as well, despite the number of nonharvest mortalities varying through the years, because most of this mortality was attributed to the relatively random processes of vehicle collisions, illegal take, depredation, and natural causes. In other words the absolute number of nonharvest mortalities may increase if population abundance increases and may decline along with a drop in abundance, all the while the actual natural survival probability of any given individual remains consistent throughout the years. The variability in harvest probabilities can be attributed to annual differences in hunting season structure, hunting season limits, true population abundance, weather, and true hunter effort. Overall, harvest probabilities increased with age class and as population abundance declined. As expected individuals in the first age-class had the lowest harvest probabilities, due to being dependent young and not available for legal harvest (NDGFD, 2017). The higher harvest probabilities estimated for subadult and adult mountain lions agreed with the overall survival documented in our study (~0.46), mostly as a result of hunter harvest. If such high levels of adult mortality continue, the population structure could shift towards younger individuals undoubtedly hindering reproduction and potentially increasing human-mountain lion conflicts (Lambert et al., 2006; Robinson et al., 2008).

Another drawback of our model was the inability to differentiate between male and female individuals. We documented female mountain lion annual survival rates substantially higher than males (0.589 vs. 0.259, respectively), and this certainly has implications for estimates of recruitment within the model. However, at the time of analysis, the PopRecon program we used to estimate our SPR likelihoods was unable to account for these types of differences. Similarly, we were unable to successfully incorporate a random effects term within the harvest processes due to program instability. This was likely due to our limited sample sizes within years. Incorporating sex-specific differences, as well as adding random effects in harvest vulnerability (Gast, 2012), would certainly lead to improved model performance and realism.

Finally, SPR assumes the population being modeled is closed with regard to immigration and emigration. Violating this assumption would likely lead to overestimation of population abundance. If a harvested individual immigrated from a different source population, they would be falsely included in historical cohort abundance estimates. Similarly, some individuals may emigrate out of the population before entering the 3+ age class but would be erroneously included. Both immigration and emigration of mountain lions has been documented in the North Dakota population (Juarez et al., 2016; Thompson and Jenks, 2010; Wilckens, 2014); indeed, we included two individual mountain lions which immigrated into North Dakota in our SPR analysis. It remains unlikely that movement of individuals in and out of this population occurred at levels high enough to greatly influence our SPR abundance estimates.


This work represents the most comprehensive assessment conducted to date of the characteristics of this unique population of mountain lions. As such it will not only be critical in establishing responsible management strategies for the North Dakota mountain lion population, but our results also have the potential to positively influence management well beyond the borders of North Dakota as mountain lions continue to expand eastward in the coming decades. Collectively, our results indicate that while the North Dakota Badlands region has been supporting a relatively small population of mountain lions, the population appears to have experienced a period of decline, attributable to human-induced mortality, before becoming more stable in recent years. A reduction in mortality, especially among adult females (Robinson et al., 2014), would allow the population to further stabilize and perhaps even increase. Based on our survival analysis results, this may be accomplished through a reduction in annual hunter harvest. However, a critical point is that although this population should be considered particularly at-risk of over-exploitation, we see no reason that it cannot remain a viable population while sustaining a conservative level of hunter harvest.

Future work should focus on refining and improving the SPR population model presented here, because as this SPR modeling technique evolves, the inputs should evolve with it. Specifically, managers in the state could work to improve estimates of annual hunter effort. This may be in the form of more targeted surveys of mountain lion hunters or perhaps a formula incorporating several important variables to estimate yearly hunter effort. Variables could be weighted and potentially include items such as annual hunting season limit, number of days with ideal snow conditions, timing of first snow, or the number of days the late-hunting season remains open. Surrogate measures for opportunistic early-season mountain lion hunters and predator callers could be incorporated as well by using metrics such as the number of deer and elk tags issued in the Badlands hunting units. Additionally, as development work continues on the PopRecon program, the incorporation of important sex-specific differences in survival and recruitment, as well as random harvest effects, into the model may soon be a reality. By taking advantage of these future improvements, managers could continually increase upon the accuracy and precision of the population model. In the meantime management should focus on adjusting annual mortality to achieve and maintain a desired population level. Given our results managers in the state now have the information needed to make scientifically-informed decisions regarding the management of this unique population.

Acknowledgments.--We thank project technicians S. Dyer and P. Ryan for their tireless help with capture efforts. We thank Dr. D. Grove for knowledge and assistance with captures. C. Kochanny, and J. Roth were of great assistance with trap transmitter and radio collar maintenance and troubleshooting, respectively. We thank USDA Wildlife Services for supplying additional baits and we thank North Dakota Department of Transportation for their help locating vehicle-killed deer. We appreciate the commitment and access to land offered by the USDA Forest Service, National Park Service, North Dakota Department of Trust Lands, North Dakota Game and Fish Department, and numerous landowners. Dr. E. Michel, N. Martorelli and Dr. M. Clawson were of great help with statistical analyses. We thank S. M. Murphy and two anonymous reviewers for reviewing and improving previous versions of this manuscript. Funding for this project was provided by Federal Aid to Wildlife Restoration administered through the North Dakota Game and Fish Department.


ANDERSON JR., C. R., K. H. KNOPFF, M. G.JALKOTZY, AND M. S. BOYCE. 2009. Cougar management in North America, p. 41-54. In: M. Hornocker and S. Negri (eds.). Cougar ecology and conservation. The University of Chicago Press, Chicago, Illinois, U.S.A.

BAILEY, V. 1920. A biological survey of North Dakota. North American Fauna, No. 49. US Department of Agriculture, Bureau of Biological Survey, Washington D.C., U.S.A.

BURNHAM, K. P. AND D. R. ANDERSON. 2002. Model selection and multimodel inference: a practical information-theoretic approach. Second edition. Springer-Verlag, New York, U.S.A.

CLAWSON, M. V. 2015. Management application of statistical population reconstruction to wild game populations. Dissertation, University of Washington, Seattle, Washington, U.S.A. 159 p.

CLAWSON, M. V.,J. R. SKALSKI, J. M. LADY, C. A. HAGEN, J.J. MILLSPAUCH, D. BUDEAU, AND J. P. SF.VERSON. 2017. Performing statistical population reconstruction using program PopRecon 2.0. Wildl. Soc. Bull., 41 (3):581--589.

COOLEY, H. S., R. B. WIELGUS, G. M. KOEHLER, H. S. ROBINSON, AND B. T. MALETZKE. 2009. Does hunting regulate cougar populations? A test of the compensatory mortality hypothesis. Ecology, 90 (10):2913-2921.

--. K. D. BUNNELL, D. C. STONER, AND M. L. WOLFE. 2011. Population management: cougar hunting, p. 111-134. In: J.A. Jenks (ed.). Managing cougars in North America. Western Association of Fish and Wildlife Agencies and Jack H. Berryman Institute, Utah Stale University, Logan, Utah, U.S. A.

COUGAR MANAGEMENT GUIDELINES WORKING GROUP (CMGWG). 2005. Cougar management guidelines. Wild Futures, Bainbridge Island, Washington, U.S.A.

CUNNINGHAM, S. C., W. 15. BALLARD, AND H. A. WHITLAW. 2001. Age structure, survival, and mortality of mountain lions in southeastern Arizona. Southwest Nat., 46(1):70-80.

DAWN, D. 2002. Management of cougars (Puma concolor) in the western United Slates. Thesis, San Jose State University, San Jose, California, U.S.A. 55 p.

DYKE, S. R., S. K. JOHNSON, AND P. T. ISAKSON. 2015. North Dakota state wildlife action plan. North Dakota Game and Fish Department, Bismarck, North Dakota, U.S.A.

FEGSKE, D. M., D. J. THOMPSON, AND J. A. JENKS. 2011. Cougar ecology and natural history, p. 15-40 In:]. A. Jenks (ed.). Managing cougars in North America. Western Association of Fish and Wildlife Agencies and Jack H. Berryman Institute, Utah Slate University, Logan, Utah, U.S.A.

GAST, C. M. 2012. Fixed and random effects models and multistage estimation procedures for statistical population reconstructions. Dissertation, University of Washington, Seattle, Washington, U.S.A. 350 p.

GAST, C. M., J. R. SKALSKI, J. L. ISABELLE, AND M. V. CLAWSON. 2013. Random effects models and multistage estimation procedures for statistical population reconstruction of small game populations. PLoS ONES(6): e(55244. doi: 10.1371/journal.pone.0065244

GOVE, N. E., J. R. SKALSKI, P. ZAGER, AND R. L. TOWNSEND. 2002. Statistical models for population reconstruction using age-at-harvest data. J Wildl. Manage., 66(2):310-320.

HAGEN, S., P. T. ISAKSON, AND S. R. DYKE. 2005. North Dakota comprehensive wildlife conservation strategy. North Dakota (lame and Fish Department, Bismarck, North Dakota, U.S.A.

JANSEN, B. D. 2011. Anthropogenic factors affecting mountain lions in the Black Hills of South Dakota. Dissertation, South Dakota State University, Brookings, South Dakota, U.S.A. 145 p.

JENKS, J. A. 2018. Mountain lions of the Black Hills: history and ecology'. John Hopkins University Press, Baltimore, Maryland, U.S.A.

JOHNSON, R. D., J. A. JENKS, AND S. A. TUCKER. 2017. Use of satellite trap transmitters in capturing mountain lions. Prairie Nat., 49(1):31-3fi.

JUAREZ, R. L. 2014. Estimating population size and temporal genetic variation of cougars in the Black Hills, South Dakota. Thesis. South Dakota Slate University, Brookings, South Dakota, U.S.A. 84 p.

--, M. K. SCHWARTZ, K. L. PILGRIM, D.J. THOMPSON, S. A. TUCKER, J. B. SMITH, AND J. A. JENKS. 2016. Assessing temporal genetic variation in a cougar population: influence of harvest and neighboring populations. Conserv. Genet., 17:379-388.

KNOPEF, K H., A. A. KNOPFF, A. KORTELLO, AND M. S. BOYCE. 2010. Cougar kill rate and prey composition in a multiprey system. J. Wildl. Manage., 74(7): 1435-1447.

LADY, J. AND J. R. SKALSKI. PopRecon version 2.0.26 user manual. 2015. Columbia Basin Research, School of Aquatic and Fisheries Sciences, University of Washington, Seattle, Washington, U.S.A.

LAMBERT, C. M. S., R. B. WIELGUS, H. S. ROBINSON, D. D. KATNIK, H. S. CRUICKSHANK, R. CLARKE, AND J. ALMACK. 2006. Cougar population dynamics and viability in the Pacific Northwest. J. Wildl. Manage., 70(1):246-254.

LARUE, M. A. AND C. K. NIELSEN. 2008. Modelling potential dispersal corridors for cougars in midwestern North America using least-cost path methods. Ecol. Model., 212:372-381.

--AND C. K. NIELSEN. 2016. Population viability for recolonizing cougars in midwestern North America. Ecol. Model., 321:121-129.

LOGAN, K. A. AND L. L. SWEANOR. 2001. Desert Puma: evolutionary ecology and conservation of an enduring carnivore. Hornocker Wildlife Institute, Island Press, Washington, DC, U.S.A.

--, L. L. SWEANOR, J. F. SMITH, AND M. G. HORNOCKER. 1999. Capturing pumas with foot-hold snares. Wildl. Soc. Bull., 27 (1) :201-208.

NATIONAL CENTERS FOR ENVIRONMENTAL INFORMATION. 2013. Summary of monthly normals 1981-2010. National Oceanic and Atmospheric Administration, Ashville, North Carolina, U.S.A.

NORTH DAKOTA GAME AND FISH DEPARTMENT. 2006. Status of mountain lions (Puma concolor) in North Dakota: a report to the Legislative Council. North Dakota Game and Fish Department, Bismarck, North Dakota, U.S.A. 68 p.

--. 2017. Status of mountain lion management in North Dakota. North Dakota Game and Fish Department, Bismarck, North Dakota, LT.S.A. 18 p.

ROBINSON, H. S., R. B. WIELGUS, H. S. COOLEY, AND S. W. COOLEY. 2008. Sink populations in carnivore management: cougar demography and immigration in a hunted population. Ecol. Appl., 18(4):1028-1037.

--, R. DESIMONE, C. HARTWAY. J. A. GUDF., M. J. THOMPSON, M. S. MITCHELL, AND M. HEBBLEWHITE. 2014. A test of the compensatory mortality hypothesis in mountain lions: a management experiment in west-central Montana. J. Wildl. Manage., 78(5):791-807.

--, T. RUTH, J. A. GUDF., D. CHOATE, R. DESIMONE, M. HEBBLEWHITE, K. KUNKEL, M. R. MATCHETT, M. S. MITCHELL, K. MURPHY, AND J. WILLIAMS. 2015. Linking resource selection and mortality modeling for population estimation of mountain lions in Montana. Ecol. Model., 312:11-25.

SIKES, R. S., and the Animal Care and Use Committee of the American Society of Mammalogists. 2016. Guidelines of the American Society of Mammalogists for the use of wild mammals in research and education. J. Mammal., 97(3):663-688.

SKALSKI, J. R., J. J. MILLSPAUGH, AND M. V. CLAWSON. 2012a. Comparison of statistical population reconstruction using full and pooled adult age-class data. PLoS ONE, 7(3): e33910.

--, M. V. CLAWSON, AND J. J. MILLSPAUGH. 2012b. Model evaluation in statistical population reconstruction. Wildlife Biol., 18(3):225-234.

--, K. E. RVDING, AND J. J. MILLSPAUGH. 2005. Wildlife demography: Analysis of sex, age, and count data. Elsevier Academic Press, Burlington. Massachusetts, U.S.A.

STONER, D. C., M. L. WOLFE, AND D. M. CHOATE. 2006. Cougar exploitation levels in Utah: implications for demographic structure, population recovery, and metapopulation dynamics. J. Wildl. Manage., 70(6):1588-1600.

THOMPSON, D. J. AND J. A. JENKS. 2010. Dispersal movements of subadult cougars from the Black Hills: the notions of range expansion and recolonization. Ecosphere, 1:1-11.

WHITE, G. C. AND K. P. BURNHAM. 1999. Program Mark: survival estimation from populations of marked animals. Bird Study, 46:120-139.

WILCKENS, D. T. 2014. Ecology of mountain lions (Puma concolor) in the North Dakota Badlands: population dynamics and prey use. Thesis. South Dakota State University, Brookings, South Dakota, U.S.A. 92 p.

--, J. B. SMITH, S. A. TUCKER, D. J. THOMPSON, AND J. A.JENKS. 2016. Mountain lion (Puma concolor) feeding behavior in the Little Missouri Badlands, North Dakota. J. Mammal., 97(2):373-385.

Submitted 16 April 2018

Accepted 4 December 2018

RANDY D.JOHNSON (1) and JONATHAN A. JENKS Department of Natural Resource Management, South Dakota State University, Brookings 57007

STEPHANIE A. TUCKER North Dakota Came and Fish Department, Bismarck 58501


DAVID T. WILCKENS New Mexico Department of Came and Fish, Santa Fe 87507

(1) Corresponding Author present address: South Dakota Department of Game, Fish, and Parks, Sioux Falls 57107; E-mail:

Caption: FIG. 1.--The study area was located within the Little Missouri Badlands of western North Dakota, U.S.A., from 2012-2016, and was entirely within mountain lion Management Zone 1 (NDGFD, 2017; inset map)

Caption: FIG. 2.--Annual estimates of mountain lion (Puma concoloi) population abundance and associated 95% confidence intervals in North Dakota, U.S.A., from 2005-2017, calculated using age-at-harvest data and statistical population reconstruction (SPR) methods (Gove et al., 2002, Skalski et al. 2005)

Caption: FIG. 3.--Annual estimates of total mountain lion (Puma concolor) population abundance from SPR population model, North Dakota, U.S.A. between March 2005--February 2017, with varying levels of historical age-at-harvest data removed

Caption: Fig. 4.--Annual estimates of total mountain lion (Puma concolor) population abundance from SPR population model along with the number of mortalities resulting from legal harvest and nonharvest (e.g., illegal harvest, vehicle collision, depredation) in North Dakota, U.S.A. between March 2005 February 2017
TABLE 1.--The top five models used in the known-fate survival analysis
to estimate mountain lion (Puma concolor) annual survival in the
North Dakota Badlands, U.S.A., 2012-2016

Model                 K    [AIC.sub.c]     [DELTA]

Year (a) + Sex (b)    4      88.5336       0.0000
+ Late season (c)

Sex (b) + Late        3      90.4989       1.9653
season (c)

Late season (c)       2      91.5460       3.0124

Year (a) + Late       3      91.7828       3.2492
season (c)

Monthly              12     102.0063       13.4727

Model                [AIC.sub.c]     Model      Evidence
                       weight      likelihood    ratio

Year (a) + Sex (b)     0.5571        1.0000      1.0000
+ Late season (c)

Sex (b) + Late         0.2085        0.3743      2.6719
season (c)

Late season (c)        0.1235        0.2218      4.5109

Year (a) + Late        0.1097        0.1970      5.0784
season (c)

Monthly                0.0006        0.0012     928.5000

(a) Year = calendar year (2012, 2013, 2014, 2015, 2016)

(b) Sex = sex of individual mountain lion (male, female)

(c) Late season = mountain lion hunting season in NDGFD Zone
1 when use of hounds was permitted

(d) Monthly = survival varies by month

TABLE 2.--Cause-specific mortalities (n = 17) of marked mountain lions
(Puma concolor) in the North Dakota Badlands, U.S.A., 2012-2016

                           Male             Female

Cause                Adult   Subadult   Adult   Subadult

Hunter harvest         3        3         3        1
Illegal harvest        0        1         1        0
Depredation            2        1         0        0
Vehicle collision      1        0         1        0

Cause                Total   (% of all mortality)

Hunter harvest        10            (58%)
Illegal harvest        2            (12%)
Depredation            3            (18%)
Vehicle collision      2            (12%)

TABLE 3.--Age-at-harvest data and breakdown of all known mortalities
used in the statistical population reconstruction (SPR) of the
mountain lion (Puma concolor) population in North Dakota, U.S.A.,

                                  Age class

Model year            0-1 yr   1-2 yr   2-3 yr   3+ yr

Mar. 2005-Feb. 2006     1        1        2        1
Mar. 2006-Feb. 2007     3        2        0        4
Mar. 2007-Feb. 2008     y        4        5        3
Mar. 2008-Feb. 2009     3        0        4        2
Mar. 2009-Feb. 2010     1        2        2        7
Mar. 2010-Feb. 2011     6        3        3        8
Mar. 2011-Feb. 2012     3        5        9       16
Mar. 2012-Feb. 2013     4        3        4        8
Mar. 2013-Feb. 2014     1        3        5       11
Mar. 2014-Feb. 2015     2        2        3        8
Mar. 2015-Feb. 2016     3        3        5        7
Mar. 2016-Feb. 2017     0        3        0        4
Total                   29       31       42      79

Model year            # Legal    # Non-harvest      Total
                      harvests    mortalities    mortalities

Mar. 2005-Feb. 2006      5             0              5
Mar. 2006-Feb. 2007      5             4              9
Mar. 2007-Feb. 2008      5             9             14
Mar. 2008-Feb. 2009      8             2             10
Mar. 2009-Feb. 2010      11            1             12
Mar. 2010-Feb. 2011      12            8             20
Mar. 2011-Feb. 2012      17           16             33
Mar. 2012-Feb. 2013      14            8             22
Mar. 2013-Feb. 2014      15            5             20
Mar. 2014-Feb. 2015      13                          15
Mar. 2015-Feb. 2016      13            5             18
Mar. 2016-Feb. 2017      11            0             11
Total                   129           60             189

TABLE 4.--The statistical population reconstruction (SPR) model
structures used to estimate mountain lion (Puma concolor) population
abundance in North Dakota, U.S.A., 2005-2017

Model         K    [AIC.sub.c]     [DELTA]

MpvaS (1)    17     204.1518          0
MpyS (2)     14     21(5.2298      12.0780
MpS (3)      14     216.7956       12.6438
MpyaSa (4)   20     217.3412       13.1894
MpySa (5)    17     230.2298       26.0780
MpSa (6)     17     230.9192       26.7674

Model        [AIC.sub.c]     Model      Evidence
               weight      likelihood    ratio

MpvaS (1)      0.9942        1.0000      1.0000
MpyS (2)       0.0024        0.0024     414.2500
MpS (3)        0.0018        0.0018     552.3333
MpyaSa (4)     0.0014        0.0014     710.1429
MpySa (5)      0.0000        0.0000     460,064
MpSa (6)       0.0000        0.0000     649,379

(1) Assumed natural survival was constant across years and age
classes, and harvest probabilities varied by year and age class

(2) Assumed natural survival was constant across years and age
classes, and harvest probabilities varied by year but not age class

(3) Assumed both natural survival and harvest probabilities were
constant over time and across age classes

(4) Assumed age- specific natural survival was constant across
years, and age- and year-specific harvest probabilities

(5) Assumed age-specific natural survival was constant across
years, and harvest probabilities varied by year but not age class

(6) Assumed age-specific natural survival was constant across
years, and harvest probabilities that were constant across years
and age classes

TABLE 5.--Year-specific harvest (P) and natural survival (S)
probabilities estimated by the top model in the statistical
population reconstruction (SPR) of the mountain lion (Puma
concolor) population abundance in North Dakota, U.S.A., 2005-2017.
Overall annual survival for each year and age-class can by
calculated as (1 - P)*S

               Harvest probabilities (P)

                  Age class (years)

                 0-l             1-2

Year         Mean     SE     Mean     SE
(Mar. 1-
Feb. 28)

2005-0206    0.103   0.025   0.151   0.031
2006-0207    0.103   0.025   0.151   0.031
2007-2008    0.103   0.025   0.151   0.031
2008-2009    0.103   0.025   0.151   0.031
2009-2010    0.076   0.019   0.112   0.024
2010-2011    0.087   0.021   0.127   0.027
2011-2012    0.086   0.021   0.126   0.026
2012-2013    0.151   0.035   0.217   0.043
2013-2014    0.142   0.033   0.205   0.041
2014-2015    0.162   0.038   0.232   0.045
2015-2016    0.162   0.038   0.232   0.045
2016-2017    0.138   0.033   0.200   0.040

              Harvest probabilities (P)         Survival
                                             probabilities (S)

                    Age class (years)

                  2-3             3+         All age classes

Year         Mean     SE     Mean     SE     Mean     SE
(Mar. 1-
Feb. 28)

2005-0206    0.263   0.043   0.377   0.076   0.893   0.050
2006-0207    0.263   0.043   0.377   0.076   0.893   0.050
2007-2008    0.263   0.043   0.377   0.076   0.893   0.050
2008-2009    0.263   0.043   0.377   0.076   0.893   0.050
2009-2010    0.199   0.034   0.291   0.063   0.893   0.050
2010-2011    0.225   0.038   0.326   0.069   0.893   0.050
2011-2012    0.222   0.038   0.323   0.068   0.893   0.050
2012-2013    0.368   0.056   0.508   0.090   0.893   0.050
2013-2014    0.348   0.053   0.485   0.088   0.893   0.050
2014-2015    0.390   0.058   0.535   0.092   0.893   0.050
2015-2016    0.390   0.058   0.535   0.092   0.893   0.050
2016-2017    0.341   0.053   0.476   0.088   0.893   0.050

TABLE 6.--Annual estimates of mountain lion (Puma concolor)
population abundance in North Dakota, U.S.A., from 2005-2017,
calculated using age-at-harvest data and statistical population
reconstruction (SPR) methods (Gove et al., 2002, Skalski et al,
2005). Underlined estimates indicate unrealistic abundances of 0.0
due to a zero observed in the corresponding spot in the
age-at-harvest matrix, resulting in a negative bias for the total
abundance estimate for that year

                  Age class (years)

War (Mar 1-    0-1    1-2    2-3     3+
Feb 28)

2005-2006       9.7    6.6    7.6    2.7
2006-2007      29.0   13.3   0.0#   10.6
2007-2008      19.4   26.5   19.0    8.0
2008-2009      32.7   0.0#   17.1    6.0
2009-2010      13.1   17.8   10.0   24.0
2010-2011      69.1   23.6   13.3   27.6
2011-2012      34.9   39.7   40.5   49.6
2012-2013      31.2   16.2   12.8   20.8
2013-2014       7.1   14.7   14.4   22.7
2014-2015      12.4    8.6    7.7   15.0
2015-2016      18.6   12.9   12.8   13.1
2016-2017      0.0#   25.3   0.0#   14.2

War (Mar 1-    SPR total   Lower    Upper
Feb 28)         annual     95% CI   95% CI

2005-2006        26.6        0.9     52.3
2006-2007        52.9       11.3     94.5
2007-2008        72.8       27.9    117.7
2008-2009        55.8       10.9    100.7
2009-2010        65.0       20.5    109.5
2010-2011       133.6       55.9    211.3
2011-2012       164.6       88.5    240.7
2012-2013        81.1       36.4    125.8
2013-2014        58.8       30.0     87.6
2014-2015        43.7       19.1     68.3
2015-2016        57.4       26.9     87.9
2016-2017        39.5        3.4     75.6

Note: Unrealistic abundances of 0.0
due to a zero observed in the corresponding spot in the
age-at-harvest matrix, resulting in a negative bias for the total
abundance estimate for that year are indicated with #.
COPYRIGHT 2019 University of Notre Dame, Department of Biological Sciences
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2019 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Johnson, Randy D.; Jenks, Jonathan A.; Tucker, Stephanie A.; Wilckens, David T.
Publication:The American Midland Naturalist
Article Type:Report
Geographic Code:1U4ND
Date:Apr 1, 2019
Previous Article:Resource Availability and the Potential for Bison Introduction in a Landscape Mosaic.
Next Article:A Comparison of Bat Activity in a Managed Central Hardwood Forest.

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters