1 HABITAT SELECTION BY MOOSE IN AN EMERGENT LOW-DENSITY EDGE POPULATION David W. Kramer1, Kenneth F. Kellner2, Joseph W. Hinton3, Michael S. Muthersbaugh4, Jerrold L. Belant2, and Jacqueline L. Frair1 1State University of New York, College of Environmental Science and Forestry, Roosevelt Wild Life Station, 1 Forestry Drive, Syracuse, NY 13210; 2Michigan State University, Department of Fisheries and Wildlife, 480 Wilson Road, East Lansing, MI 48824; 3Wolf Conservation Center, 7 Buck Run St, South Salem, NY 10590; 4New York Department of Environmental Conservation, 625 Broadway, Albany, NY 12207 ABSTRACT: The Adirondack Park in northern New York contains about 700 moose (Alces alces) that persist as a low-density population (0.03 moose/km2) that occurs along the periphery of the moose’s southern range in the eastern United States. As part of a comprehensive effort to evaluate the status of the New York moose population, we fitted 26 moose with GPS collars during 2015–2017 and assessed summer (June–August) and winter (December–March) resource selection to understand moose space use and potentially limiting factors (e.g., climate, forage availability). Home ranges (x = 22 km2) predominately contained deciduous forest, including managed forest stands recently harvested for timber. During summer moose did not exhibit variation in selection among years suggesting that ade- quate forage may be available across the landscape regardless of habitat type. Moose resource selec- tion within home ranges was most variable during winter, and moose selected areas of managed timber during the most severe winters. Observed habitat selection highlights the potential of direct and indirect interactions with white-tailed deer (Odocoileus virginianus), given that deer in the Adirondack Park forage in areas selected by moose such as those with regenerative timber. Because white-tailed deer are an intermediate host for two fatal moose parasites (Parelaphostrongylus tenuis and Fascioloides magna), increase in habitat overlap between moose and deer could be detrimental to the long-term health of the New York moose population. Additionally, the dependence of both moose and white-tailed deer on regenerating forest stands for optimal forage could put commercial stands at risk of regenerative failure. ALCES VOL. 60: 1–17 (2024) Key Words: Adirondack Park, Alces alces, habitat selection, home range, large ungulates, moose Moose (Alces alces) occurred in the State of New York from the Pleistocene until the late 19th century, and prehistorical evidence sug- gest that moose were present in the northern part of the state, north of the Mohawk River (Fischer 1955, Ritchie 1969, Ritchie and Funk 1973), corresponding largely with the present-day Adirondack Park and Forest Preserve. Moose were extirpated from New York by 1861 due to intense forest manage- ment, timber extraction and unregulated hunting (Grant 1894). Several failed reintroductions occurred in northern New York between 1870 and 1902 (Colvin 1880, Wish 1902, Barnham 1909, Bump 1940). By the late 1950s, transient moose from neigh- boring states and provinces (Massachusetts, Quebec and Vermont) occasionally ventured into areas of New York (Severinghaus and Jackson 1970) with moose becoming a per- manent resident of northern New York by the 1980s (Hicks 1986, Hicks and McGowan 1992, Garner and Porter 1990, Hickey 2008). Today, the Adirondack Park (hereafter the Park) HABITAT SELECTION BY MOOSE ALCES VOL. 60, 2024 2 population is estimated to be approximately 700 moose and persists as a low-density population (0.29 moose/km2) along the moose’s southern geographical extent (Hinton et al. 2022a). Generally, moose exist in clustered pockets of high-quality habitat in the northern region of the Park, resulting in localized area with moose densities of 0.09–0.12 moose/km2 (Hinton et al 2022a). Commercial timber harvests have long been identified as an important tool for cre- ating optimal habitat for moose in North America (Raymond et al. 1996, Bergeron et al. 2011, Andreozzi et al. 2014, 2016, Peterson et al. 2020, 2022). Young regener- ating forests created post-harvest (<20 years; Peek et al. 1976) often provide preferred for- age that is both high in abundance and high in nutritional quality (Rea and Gillingham 2001, Peterson et al. 2020), and can be used as a means of thermoregulation (Renecker and Hudson 1986, Thompson et al. 2021). Though regenerating forests can provide most of the summer forage (Peterson et al. 2022), moose habitat selection can still vary seasonally. Moose have been documented to transition between regenerating deciduous and conifer forests in winter (Courtois et al. 2002, Andreozzi et al. 2016) and then increasing their use of wetlands and decidu- ous stands and decreasing their use of conif- erous cover in warmer months (MacCracken et al. 1997, Laforge et al. 2016, Teitelbaum et al. 2021). Despite the abundance of optimal forag- ing habitat via young forests, neighboring New England moose populations have expe- rienced recent declines. These population declines have highlighted the importance of identifying resource use beyond the tradi- tional lens of intraspecific resource competi- tion, resource availability, and predation (Jones et al. 2017, 2019; Debow et al. 2021). Detrimental parasites, such as winter tick (Dermacentor albipictus), brainworm (Parelaphostrongylus tenuis) and liver fluke (Fascioloides magna), can potentially thrive in areas of optimal moose habitat. This shift in focus on parasite mediated population responses may be correlated with the north- ward expansion of white-tailed deer (Odocoileus virginianus) over the past few decades (Dawe and Boutin 2016, Ditmer et al. 2020). White-tailed deer are aptly suited to fill the ecological niche created by commercial harvest for the similar reasons as moose (Côté et al. 2004). Concerningly, white-tailed deer act as a primary host for two of the detrimental parasites (P. tenuis, F. magna; Vanderwaal et al. 2015, Vannatta and Moen 2016). This results in forage abun- dance and habitat quality mediating parasite dynamics through effects on moose and white-tailed deer host densities and space use (Lankester and Foreyt 2011, Healy et al. 2018). By identifying preferred habitat and environmental conditions, managers can focus monitoring in areas which may present the greatest likelihood of parasitic impacts through habitat niche overlap between deer and moose. Our primary objective was to deter- mine if moose in the Park select for early successional habitat created by forest management. Broadly, we hypothesized moose would preferentially use certain landcover types, and selection would vary seasonally. We defined four competing predictions: (1) moose do not select for recently managed forested habitat (i.e., early successional, or regenerating stands); (2) moose select for recently managed forest regardless of harvest regime; (3) moose selection of managed forest differs by harvest regime (interme- diate or overstory removal); and (4) moose selection differs by both harvest regime and forest type (deciduous, conifer). We tested these hypotheses in winter and summer and in multiple years with ALCES VOL. 60, 2024 HABITAT SELECTION BY MOOSE 3 variable winter severity. In addition, as a secondary objective we examined if moose selected for wetland areas due to the abundance of forage and utility for thermoregulation that those habitats provide. STUDY AREA We studied moose resource selection in the Adirondack Park (24,281 km2; Figure 1) in northern New York. Elevations range from 100 m in low-lying lake shores to about 1600 m. The Park consists of large glacial Fig. 1. Map of the Adirondack Park boundary in northern New York, USA. Simple hatch pattern denotes areas of conservation easements and black dots are locations of collared moose from 2015 to 2017. The inset map is a representation of the state of New York, USA. HABITAT SELECTION BY MOOSE ALCES VOL. 60, 2024 4 valleys that gradually rise in elevation to the High Peaks region in the east-central part of the Park. Average monthly temperatures ranged from -9°C in winter to 18°C in sum- mer. The region received an average of 100 cm of rainfall precipitation a year, with an additional 290 cm of snowfall annually (Jenkins and Keal 2004). The Park included public (61%) and pri- vate (39%) lands, where all public lands were protected by Article XIV of the New York State Constitution as “forever wild for- est” which prohibits any resource extraction (i.e., timber harvest) or development (N.Y. Const. art. XIV, § 1) on public lands regard- less of purpose. The Park comprises a patch- work of the late seral stage northern boreal ecosystem interspersed with temperate deciduous forests and large peatland com- plexes. Lower elevations with fertile soils support diverse tree species dominated by American beech (Fagus grandifolia), yellow birch (Betula allegheniensis), paper birch (B. papyrifera), sugar maple (Acer saccha- rum) and red maple (A. rubrum). Higher ele- vations are more coniferous, dominated by species such as red spruce (Picea rubens), balsam fir (Abies balsamea), white pine (Pinus strobus) and eastern hemlock (Tsuga canadensis; Jenkins and Keal 2004, Peterson et al. 2020). About 25% of forested private lands in the Park (13% of all Park land) are enrolled in the New York State Conservation Easement Program (NYSDEC 2022). Private proper- ties enrolled in an easement participate in a structured forest management program, which allows for timber harvest and other associated activities. Lands subjected to tim- ber management are predominately com- posed of marketable timber species such as sugar maple, red maple, red oak (Quercus rubra), white ash (Fraxinus americana), black cherry (Prunus serotina) and white pine; Peterson et al. 2020). Timber management methods include shelterwood removal, overstory removal, single tree selection and salvage thinning. Because pub- lic land acquisition continues within the Park, portions of public land could have been exposed to resource extraction immediately prior to acquisition by New York State. METHODS Moose Capture and Health Assessment In 2015–2017, we captured 26 adult and sub- adult moose (3 male, 23 female) by net-gun fired from helicopters by Native Range Capture Services during January when snow conditions provided increased visibility and limited potential injury to moose. Given the overall low density of moose population within the Park (see Hinton et al. 2022), the majority of moose were captured in the north- ern portions of the Park where the population density was the greatest to increase the effi- ciency of capture efforts. We fitted captured moose with Iridium Global Positioning System (GPS) collars (BASIC Iridium Track M 3D, Lotek Wireless, Newmarket, ON; or TGW-4670-3, Telonics, Mesa, AZ) and esti- mated their ages based on a visual inspection of tooth wear and body size. Moose were released at capture sites. We programmed col- lars to attempt a GPS location every 2 hours for 2 years, and collars achieved a mean loca- tion rate of 98.7 ± 1.1% (Peterson et al. 2020). Animal capture and handling protocols met American Society of Mammalogists recom- mended guidelines (Sikes et al. 2016) and were approved by the State University of New York College of Environmental Science and Forestry Animal Care and Use Committee (Protocol #140901). Data Analysis We analyzed moose 3rd-order resource selec- tion for years 2016—2019, comparing used and available locations within the home ALCES VOL. 60, 2024 HABITAT SELECTION BY MOOSE 5 ranges of individual moose (Johnson 1980). Used locations were the GPS fixes described above, with data truncated to remove the first two weeks following capture or last two weeks prior to collar release or death (if applicable). We defined availability as the cumulative home range of each moose, cal- culated using all available GPS data from 2015-2019 and Brownian bridge movement models (Horne et al. 2007) using the R (R Core Team 2022) package adehabitatHR (Calenge 2006). We randomly sampled available locations for each moose equal to the number of used locations. We used spatial rasters of the study area with the following land cover categories: decidu- ous forest, conifer forest, mixed forest, grass- land, and wetland. Forested cells (e.g., deciduous, conifer and mixed) were further classified as either mature forest, intermediate removal, or overstory removal (Kramer et al. 2022), resulting in 11 landcover classes (Table 1). The land cover raster combined National Land Cover Database (NLCD) data (Yang et al. 2018), Landsat 8 satellite imagery, and timber management polygons obtained from timber companies (Kramer et al. 2022). We obtained rasters for years 2015 – 2018. We calculated the minimum distance from each used and available location to each of the land cover types using the R package raster (Hijmans 2021). We also calculated the dis- tance to the nearest managed forest of either type, and the nearest intermediate removal and overstory, regardless of forest type. For each used and available point, we derived landcover covariates from the most recent raster layer available (e.g., some points were given values from the previous calendar year). Table 1. Landcover classifications and source data used to assess habitat selection of moose in northern New York, USA. All canopy removals have occurred within the past 20 years. Land cover class Definition Deciduous Forest (Mature)*# Forest that is 75% or more of deciduous trees; <30% canopy removal Deciduous Forest (Intermediate Removal)*# Forest that is 75% or more of deciduous trees; 30%-60% canopy Deciduous Forest (Overstory Removal)*# Forest that is 75% or more of deciduous trees; >60% canopy Conifer Forest (Mature)*# Forest that is 75% or more of coniferous trees; <30% canopy removal Conifer Forest (Intermediate Removal)*# Forest that is 75% or more of coniferous trees; 30%-60% canopy Conifer Forest (Overstory Removal)*# Forest that is 75% or more of coniferous trees; <60% canopy removal Mixed Forest (Mature)*# Forest that is neither deciduous nor coniferous trees that are >75% of total tree cover; <30% canopy removal Mixed Forest (Intermediate Removal)*# Forest that is neither deciduous nor coniferous trees that are >75% of total tree cover; 30%-60% canopy Mixed Forest (Overstory Removal)*# Forest that is neither deciduous nor coniferous trees that are >75% of total tree cover; <60% canopy removal Grassland* Landcover dominated by upland grasses or forbs Wetland+ Any land which is annually subject to periodic or continual inundation of water which are either (a) one acre or more in size or (b) located adjacent to a free-flowing body of water Data Sources:*Yang et al. 2018, #Kramer et al. 2022, +Adirondack Park Agency 2004 HABITAT SELECTION BY MOOSE ALCES VOL. 60, 2024 6 To assess the annual variability in winter conditions we calculated an index of winter severity for white-tailed deer (Verme 1968) using snow depth (NOHRSC 2004) and daily mean temperatures (Daly et al. 1994). Winter severity was the sum of the total number of days (November 1–March 31) where snow depths were greater than 38 cen- timeters and total number of days where the average daily temperature was below -17.8°C. Winter severity for winter 2017- 2018 was extended to April due to an abnor- mal late season extreme weather event, resulting in a winter severity measured from 1 November 2017–April 30 2018 (Table 2). We acknowledge the potential that extend- ing the sampling window only for 2018 could induce a bias, however by not incorpo- rating the abnormal late season blizzard into our calculations, we would have failed to accurately quantify the overall influence of climatic variation on moose selection. We used logistic regression models to determine the effects of land cover, winter severity and forest management on moose habitat selection. We quantified selection for each season-year combination in separate analyses, and defined winter as December- March, and summer as June-August. Accordingly, we quantified habitat selection for winter and summer separately for December 2015-August 2019, resulting in 8 total analyses (4 years x 2 seasons). Within each analysis, we compared four candidate models, corresponding to our four compet- ing hypotheses: (1) no forest management covariates (NULL); (2) distance to nearest forest management (intermediate or over- story removal) regardless of forest type (MAN); (3) distance to intermediate removal and overstory removal regardless of forest type (TYPE); and (4) distance to each possi- ble forest management-forest type combina- tion (TYPE x FOR). All four models included covariates for distance to natural cover types including grassland, wetland, and mature deciduous, conifer, and mixed forest. All four models also included random intercepts and slopes by moose ID to account for individual variation in selection (Muff et al. 2020). All models were fit using the R package glmmTMB (Brooks et al. 2017). Within each analysis we ranked the four can- didate models using AIC (Burnham and Anderson 2002) and calculated the area under the receiver operating characteristic curve (AUC) on the full dataset as a measure of each model’s predictive power (Cumming 2000). RESULTS We estimated home ranges for 26 moose during December 2015–August 2019 using an average of 7,532 GPS locations per moose (SD = 3,411; range = 1,824–5,986). Annual moose home ranges averaged 21.6 km2 Table 2. The average winter severity index (WSI) and standard deviation (SD) for northern New York, USA from November 1–March 31, from four consecutive winters (2015-2019). The value is the total number of days where snow depths were greater than 38 centimeters and total number of days where the average daily temperature was below -17.8°C. Winter severity for winter 2017-2018 measured from 1 November 2017–April 30, 2018. Year WSI SD 2015-2016 13 4 2016-2017 39 29 2017-2018 61 35 2018-2019 69 39 ALCES VOL. 60, 2024 HABITAT SELECTION BY MOOSE 7 (SD = 11.0; range = 3.8–54.4). There was considerable spatial overlap between indi- vidual study animals, with only two collared individuals not exhibiting any range overlap with another collared moose. Composition of land cover within home ranges varied among moose (Figure 2). Among natural (i.e., non-management) land cover types, deciduous forest was the most common land cover type in home ranges, followed by wet- land and coniferous forest. Managed/har- vested forests made up a smaller portion of moose home ranges, with intermediate removal of deciduous forest being the most common. We fit four winter season resource selec- tion models (yearly 2016-2019; 21, 22, 10 and 6 moose, respectively) using an average Fig. 2. Land cover composition in moose all-season 95% home ranges in Adirondack Park, NY, 2016- 2019. The boxplots represent the interquartile interval, where 50% of the data is found within the bounds of the box, with a line representing the mean value. The vertical lines represent the upper and lower quartiles (>75% and <25% of the data, respectively), and dots represent outliers. The top panel shows non-management cover types (Decid = deciduous forest, Mixed = deciduous/conifer forest, Conif = conifer forest, Grass = open grass, Wet = wetlands) and the bottom panel shows forest cover with recent management (int = intermediate removal, over = overstory removal). HABITAT SELECTION BY MOOSE ALCES VOL. 60, 2024 8 of 17,629 locations per moose (range 8,231 – 25,186). We fit three summer resource selection models (yearly 2016- 2018; 19, 14 and 6 moose, respectively) using an average of 13,191 locations per moose (range 5,467 – 20,389). We did not fit a model in summer 2019 due to a small sam- ple size (768 locations; 1 moose). Across all year and season combina- tions, the top-ranked and clearly superior model was model TYPE x FOR which included effects of distances to individual forest type – management combinations (Tables 3,4). Top-ranked models had AUC values ranging from 0.83 to 0.97. Because these AUC values were calculated from the full dataset without subsetting, they likely overestimate the accuracy of the model predicting for new data, but still indicate adequate predictive power. Moose selection for early successional forest habitat varied by season, year, and forest type. In winter, there was little evi- dence that moose selected for or against habitat close to forest stands with recent intermediate removal within their home ranges. The exception was 2019, when moose selected for habitat further from this management type in deciduous forest (Figure 3). In 2016 (a relatively mild win- ter), selection for overstory removal var- ied by forest type, with selection for areas close to overstory removal in mixed forest and selection against overstory removal in deciduous forest. In years 2018 and 2019 that were characterized by above average Table 3. Candidate resource selection models for moose in winter (December–March) 2016-2019 in Adirondack Park, NY. Candidate models were NULL (no forest management parameters); MAN (distance to nearest forest management); TYPE (distance to nearest intermediate removal and nearest overstory removal); and TYPExFOR (distance to nearest intermediate and overstory removal by forest type). All models also included parameters for distance to nearest natural habitat types. K is the number of free parameters in the model, AIC represents the mathematical evaluation of model fit, ΔAIC is the difference between the best fit candidate model from those provided and AUC is the measure of how different a model prediction is from random chance. Year Model K AIC ΔAIC AUC 2016 TYPExFOR 24 113,063.8 0.0 0.88 TYPE 16 126,463.4 13,399.5 0.83 MAN 14 135,275.9 2,2212.0 0.79 NULL 12 138,250.5 25,186.6 0.77 2017 TYPExFOR 24 92,642.4 0.0 0.92 TYPE 16 114,179.5 21,537.0 0.87 MAN 14 123,819.7 31,177.3 0.82 NULL 12 130,198.6 37,556.2 0.80 2018 TYPExFOR 24 50,094.6 0.0 0.96 TYPE 16 60,425.5 10,331.0 0.93 MAN 14 64,744.9 14,650.3 0.91 NULL 12 67,719.6 17,625.1 0.90 2019 TYPExFOR 24 34,672.0 0.0 0.97 TYPE 16 37,896.5 3,224.5 0.96 MAN 14 41,042.8 6,370.8 0.94 NULL 12 41,358.2 6,686.2 0.94 ALCES VOL. 60, 2024 HABITAT SELECTION BY MOOSE 9 winter severity, there was a general pat- tern of selection for habitat close to overstory removal regimes in all forest types. However, the associated parame- ter estimates were accompanied with large uncertainty, possibly due to high variability in selection among individual moose and smaller sample sizes in these years. During summers, moose had a mixed response to intermediate removal harvest (Figure 4). In 2016, there was no strong selection within home ranges for or against intermediate removal. In 2017, moose selected habitat further from intermediate removal in coniferous forest, and in 2018, moose selected for habitat further from all three forest types but particularly deciduous forest. In 2017 and 2018, moose selected for habitat close to overstory removal harvests in most forest types. As with managed forest habitat, there was no consistent yearly or seasonal pattern in moose selection for wetland habitat (Figure 5). In 2016-2018, moose did not select for or against areas close to wetland habitat in either season. In 2019, moose selected for habitat closer to wetlands during the winter. DISCUSSION The average home range size (21.6 km2) for New York moose was similar to ranges observed in regional populations in Maine (Leptich and Gilbert 1989). This suggests that the population has not surpassed resource availability and will continue to persist in localized, higher-density sub-populations concentrated around suitable habitat. Home ranges predominately contained deciduous forest and included forested stands that had received recent timber harvest (i.e., shelterwood harvests, overstory removal) similar to home range compositions in neighboring states (Wattles and DeStefano 2013, Andreozzi et al. 2016). Observed Table 4. Comparison of candidate resource selection models for moose in summer (June – August) 2016- 2018 in Adirondack Park, NY. Candidate models included NULL (no forest management parameters); MAN (distance to nearest forest management); TYPE (distance to nearest intermediate removal and nearest overstory removal); and TYPExFOR (distance to nearest intermediate and overstory removal by forest type). All models also included parameters for distance to nearest natural habitat types. K is the number of free parameters in the model, AIC represents the mathematical evaluation of model fit, ΔAIC is the difference between the best fit candidate model from those provided and AUC is the measure of how different a model prediction is from random chance. Year Model K AIC ΔAIC AUC 2016 TYPExFOR 24 111,194.4 0.0 0.83 TYPE 16 117,819.3 6,624.8 0.78 MAN 14 121,096.1 9,901.7 0.76 NULL 12 123,677.3 12,482.9 0.74 2017 TYPExFOR 24 77,424.0 0.0 0.87 TYPE 16 81,253.8 3,829.8 0.84 MAN 14 84,990.9 7,566.9 0.81 NULL 12 87,005.3 9,581.3 0.79 2018 TYPExFOR 24 32,915.8 0.0 0.94 TYPE 16 34,690.9 1,775.1 0.92 MAN 14 36158.4 3,242.6 0.91 NULL 12 36621.8 3,706.1 0.90 HABITAT SELECTION BY MOOSE ALCES VOL. 60, 2024 10 non-migratory behavior of New York moose, coupled with high spatial overlap among indi- viduals and previous work on forage avail- ability and nutrition (Peterson et al. 2020, Kramer et al. 2022, Peterson et al. 2022) sug- gest that moose in the Park are at or below the landscape-level population capacity. Moose resource selection within home ranges varied during years with severe winter conditions, with moose selecting for overstory removals during the harsh winters in 2018-2019. Timber management treat- ments in the Park are often conducted at small scales (x̅ = 4.2 hectares) due to the lim- its on timber extraction on public land and the matrix of land ownership in the Park. Moose may use more heterogenous patches of managed and non-managed forest during Fig. 3. Moose selection for combinations of forest and management type from the top-ranked moose resource selection models in winter (December – March) of 2016 – 2019 in Adirondack Park in New York, USA. Covariates are the minimum distance to the nearest forest/management type; therefore, positive logit-scale coefficient values indicate moose were selecting against (i.e., for areas further away from) that forest/management type, and vice-versa. ALCES VOL. 60, 2024 HABITAT SELECTION BY MOOSE 11 severe weather to reduce energetic demands during periods of thermal stress (summer) or impeded movements (winters) (Poole and Stuart-Smith 2005, Andreozzi et al. 2016). However, our measure for winter severity may not be ideal to measure the impacts of winter stressors on moose given that it was initially derived for white-tailed deer. It is possible to modify the index to meet moose thresholds more aptly, but it is unlikely that anywhere in New York would meet those conditions, given our location on the south- ern extent of moose range. There was no indication that moose selected wetlands within their home ranges during summer, which was unexpected given that moose typically derive a portion of the summer diet from aquatic plants and that wetlands can facilitate thermoregulation during warm summer conditions, especially along the southern range extent for moose (Broders et al. 2012, Morris 2014, Teitelbaum Fig. 4. Moose selection for combinations of forest and management type from the top-ranked moose resource selection models in summer (June – August) of 2016 – 2018 in Adirondack Park in New York, USA. Covariates are the minimum distance to the nearest forest/management type; therefore, positive logit-scale coefficient values indicate moose were selecting against (i.e., for areas further away from) that forest/management type, and vice-versa. There was no 2019 model due to a small sample size. HABITAT SELECTION BY MOOSE ALCES VOL. 60, 2024 12 et al. 2021, Tischler et al. 2022). It is likely that wetlands are not a limiting resource for moose in the region. Regional beaver activ- ity is high (Zevin 2022), and the lack of selection may indicate that moose are ran- domly encountering wetlands at a rate that may be sufficient to meet their foraging and thermoregulatory needs. Our findings align with the concerns identified by private landowners and timber managers that the current distribution and density of the New York moose population may be negatively impacting commercial timber harvest through disproportionate selection of managed forests (Connelly et al. 2020). When paired with previous research on browse selection and population distribu- tion (see Peterson et al. 2020, Hinton et al. 2022a, Peterson et al. 2022), our findings indicate that the status of the current moose population is influenced by the availability of commercial working forests. Monitored moose in our study showed little selection for or against unmanaged forests, which largely occur on public lands and are the dominant habitat on the landscape, account- ing for half of all moose home ranges. Alternatively, managed forests (i.e., private commercial forest) were selected for during summer and winter. The result is that private owners of working forests generate the hab- itat which supports the majority of the New York moose population, but they thus expe- rience the negative effects of moose brows- ing (ex, damage to regenerating timber stands). It is important to acknowledge the potential for overlap in the ecological niche and population distribution between moose and white-tailed deer (Whitlaw and Lankester 1994, Post and Stenseth 2002). Early successional habitat, which can be Fig. 5. Moose selection for wetland habitat from the top-ranked moose resource selection models in winter (December – March) and summer (June – August) of 2016 – 2018 in Adirondack Park in New York, USA. Covariates are the minimum distance to the nearest wetland; therefore, positive logit-scale coefficient values indicate moose were selecting against (i.e., for areas further away from) wetlands, and vice-versa. There was no summer 2019 model due to a small sample size. ALCES VOL. 60, 2024 HABITAT SELECTION BY MOOSE 13 found in timber stands 5-10 years post-man- agement, are important sources of forage for deer and often constitute a large portion their home range when available (Aycrigg and Porter 1997, Beier and McCullough 1990, Quinn et al. 2013, Lesser et al. 2019). The Adirondack landscape is predominately mature forest, with smaller pockets of com- mercial forest, creating a landscape where deer may be more likely to congregate in those commercial forests to access preferred forage. Although deer densities in northern New York are lower than other parts of the state, deer and moose co-occurrence is likely occurring (Whitlaw and Lankester 1994, NYSDEC 2021, Hinton et al. 2022b). The amount of spatial overlap between moose and deer has likely increased with time in New York because milder winters have resulted in deer population expansion northward while moose population distribu- tion has expanded (Teitelbaum et al. 2021; Hinton et al. 2022 a,b). Historically, severe winters influenced deer populations in upstate New York through winter die-off events or by encouraging deer to congregate in pockets of dense thermal cover (i.e., deer yards) at low elevations (Tierson et al. 1985, Hurst and Porter 2008). The increasing fre- quency of mild winters in the region due to climate change has allowed deer to persist year-round in areas that were typically sea- sonally inhospitable, while also lessening the frequency of mortality events. The result is a potential increase in year-round interac- tions between white-tailed deer and moose and an increased opportunity for parasitic transmissions via feces and gastropod vec- tors (Lankester 2002, Vanderwaal et al. 2015, Vannatta and Moen 2016, Hinton et al. 2022a). There are already indications that the frequency of parasitic infection in moose has increased over the past 10-15 years (Kevin Hynes, NYSDEC; personal comm.) in New York. The shared preference for managed for- est habitat creates a conundrum for wildlife managers responsible for managing moose and deer populations. Wildlife managers in New York have limited tools available to manage moose. Currently, the state has no regulated hunting of moose; predation of moose and deer is limited; and forest man- agement is restricted to private lands, with public forests protected from commercial timber extraction. Managers have more tools for managing the white-tailed deer popula- tion through regulated hunting seasons and depredation permits. However, portions of the Park are remote and the region has some of lowest hunter and deer harvest densities in the state (NYSDEC 2021). There is poten- tial for agency managers to work with com- mercial forest owners to develop long-term management plans that focus on the distribu- tion and timing of forest management that would potentially lessen the impacts of moose browsing but also to mitigate the increasing overlap between white-tailed deer and moose. In light of apparent moose population declines in various New England states (i.e., Maine, New Hampshire and Vermont), future moose research in New York should elucidate patterns of overlap between moose and white-tailed deer, examine landscape parasite prevalence, seek to identify poten- tial management actions should the moose population in the Park begin to decline, and work with commercial foresters to develop long-term management plans for working forests. ACKNOWLEDGEMENTS We thank our partners at SUNY-ESF, NYSDEC, and Cornell University for their assistance. Funding was provided by SUNY- ESF and NYSDEC (Federal Aid in Wildlife Restoration Grant W-173-G). We thank E. Bergman, R. Harris, and two anonymous HABITAT SELECTION BY MOOSE ALCES VOL. 60, 2024 14 reviewers for their constructure and helpful reviews. LITERATURE CITED AdirondAck pArk Agency. 2004. Wetlands effects database and GIS for the Adirondack Park. Available online: h t t p s : / / a p a . n y . g o v / R e s e a r c h / ParkwideGISFinalReport.pdf. Date Accessed January 1, 2024 Aycrigg, J. L., and W. F. porter. 1997. Sociospatial dynamics of white-tailed deer in the Central Adirondack Mountains, New York. Journal of Mammalogy 78: 468-482. Andreozzi, H. A., p. J. pekins, and M. L. LAngLAis. 2014. Impact of moose brows- ing on forest regeneration in northeast Vermont. Alces 50: 67-79. Andreozzi, H. A., p. J. pekins, and L. e. kAntAr. 2016. Using aerial survey observations to identify winter habitat use of moose in northern Maine. Alces 52: 41-53. BArnHAM, J. s. 1909. Report of the chief game protector for 1909 in state of New York. Fish and Game Commission Annual Report. 407pp. Beier, p., and d. r. MccuLLougH. 1990. Factors influencing white-tailed deer activity patterns and habitat use. Wildlife Monographs 109: 3-51. Bergeron, d. H., p. J. pekins, H. F. Jones, and W. B. LeAk. 2011. Moose browsing and forest regeneration: A case study in north- ern New Hampshire. Alces 47:39-51. Broders, H. g., A. B. cooMBs, and J. r. MccArron. 2012. Ecothermic responses of moose (Alces alces) to thermoregula- tory stress on mainland Nova Scotia. Alces 48: 53-61. Brooks, M. e., k. kristensen, k. J. VAn BentHeM, A. MAgnusson, c. W. Berg, A. nieLsen, H. J. skAug, M. MAcHLer, and B. M. BoLker. 2017. glmmTMB balances speed and flexibility among packages for zero-inflated generalized linear mixed modeling. The R Journal 9:378-400. BuMp, g. 1940. The introduction and trans- portation of game birds and animals into the state of New York. North American Wildlife Conference 5: 409-412. BurnHAM k.p., and d. r. Anderson. 2002. Model selection and multimodel infer- ence: A practical information-theoretic approach (2nd ed.). Springer, New York. cALenge, c. 2006. The package adehabitat for the R software: tool for the analysis of space and habitat use by animals. Ecological Modelling 197: 516-519. coLVin, V. 1880. Topographic survey of the Adirondack region of New York. 7thAn- nual Report. Weed, Parsons and Co., Albany, NY. côté, s. d., t. p. rooney, J. treMBLAy, c. dussAuLt, and d. M. WALLer. 2004. Ecological impacts of deer overabun- dance. Annual Review of Ecology, Evolution, and Systematics 35:113-147. conneLLy, n. A., B. LAuBer, r. c. stedMAn, and H. kretser. 2020. Attitudes towards moose among large private forestland owners and managers in northern New York. CCSS Series No. 20-1. https:// ecommons.cornell.edu/items/f781a98a- 4c11-43a5-ac73-7003a5cbb4c5. Date Accessed January 1, 2024 courtois, r., c. dussAuLt, F. potVin, and g. dAigLe. 2002. Habitat selection by moose (Alces alces) in clear-cut land- scapes. Alces 38: 177-192. cuMMing, g.s. 2000. Using between-model comparisons to fine-tune linear models of species ranges. Journal of Biogeography 27: 441–445. dALy, c, r. p. neiLson, and d. L. pHiLLips. 1994. A statistical-topographic model for mapping climatological precipitation over mountainous terrain. Journal of Applied Meteorology and Climatology 33: 140-158. dAWe, k. L., and s. Boutin. 2016. Climage change is the primary driver of white- tailed deer (Odocoileus virginianus) https://apa.ny.gov/Research/ParkwideGISFinalReport.pdf https://apa.ny.gov/Research/ParkwideGISFinalReport.pdf ALCES VOL. 60, 2024 HABITAT SELECTION BY MOOSE 15 range expansion at the northern extent of its range; land use is secondary. Ecology and Evolution 18: 6435-6451. deBoW, J., J. BLouin, e. rosenBLAtt, c. ALexAnder, k. gieder, W. cottreLL, J. MurdocH, and t. donoVAn. 2021. Effects of winter ticks and internal para- sites on moose survival in Vermont, USA. Journal of Wildlife Management 85: 1423-1439. ditMer, M. A., A. M. McgrAW, L. corniceLLi, J. d. Forester, p. J. MAHoney, r. A. Moen, s. p. stApLeton. V. st-Louis, k. VAnderWAAL, and M. cArstensen. 2020. Using movement ecology to investigate meningeal worm risk in moose, Alces alces. Journal of Mammalogy 101: 589-603. FiscHer, d. W. 1955. Prehistoric mammals of New York. The New York Conservationist 9:18-22. gArner, d., and W. porter. 1990. Movements and seasonal home ranges of bull moose in a pioneering Adirondack population. Alces 26: 80-85. grAnt, M. 1894. The vanishing of moose and their extermination in the Adirondacks. Century Magazine. 47: 345-356. HeALy, c., p. J. pekins, L. kAntAr, r. g. congALton, and s. AtALLAH. 2018. Selective habitat use by moose during critical periods in the winter tick life cycle. Alces 54: 85-100. Hickey, L. 2008. Assessing re-colonization of moose in New York with HIS models. Alces 44: 117-126. Hicks, A. 1986. The history and current sta- tus of moose in New York. Alces 22: 245-252. _______, and e. McgoWAn. 1992. Restoration of moose in northern New York state: environmental impact state- ment. New York Department of Environmental Conservation. 62 pp. HiJMAns r. 2021. Raster: geographic data analysis and modeling. R package ver- sion 3.5-2. https://CRAN.R-project.org/ package=raster. Hinton, J., r. WHeAt, p. scHuette, J. Hurst, d. krAMer, J. stickLes, and J. FrAir. 2022a. Challenges and opportunities for robust population monitoring of moose along their southern range in eastern North America. Journal of Wildlife Management 86 https://doi.org/10.1002/ jwmg.22213 _______, J. Hurst, d. krAMer, J. stickLes, and J. FrAir. 2022b. A model-based esti- mate of winter distribution and abun- dance of white-tailed deer in the Adirondack Park. PLoS ONE 17: e0273707. https://doi.org/10.1371/jour- nal.pone.0273707 Horne, J. s., e. o. gArton, s. M. krone, and J. s. LeWise. 2007. Analyzing ani- mal movements using Brownian bridges. Ecology 88: 2354-2363. Hurst, J. e., and W. F. porter. 2008. Evaluation of shifts in white-tailed deer winter yards in the Adirondack region of New York. Journal of Wildlife Management 72: 367-375. Jenkins, J., and A. keAL. 2004. The Adirondack atlas: a geographic portrait of the Adirondack Park. Syracuse University Press, Syracuse, NY, USA. Jones, H., p. J. pekins, L. e. kAntAr, M. o’neAL, and d. eLLingWood. 2017. Fecundity and summer calf survival of moose during 3 successive years of win- ter tick epizootics. Alces 53: 85-98. Jones, H., p. J. pekins, L. e. kAntAr, i. sidor, d. eLLingWood, A. LicHtenWALner, and M. o’neAL. 2019. Mortality assessment of moose (Alces alces) calves during successive years of winter tick (Dermacentor albipictus) epizootics in New Hampshire and Maine (USA). Canadian Journal of Zoology 97: 22-30. krAMer, d. W., t. J. preByL, n. p. niBBeLink, k. V. MiLLer, A. A. royo, and J. L. FrAir. 2022. Managing moose from home: determining landscape carrying capacity for Alces alces using remote sensing. Forests 13: 150. HABITAT SELECTION BY MOOSE ALCES VOL. 60, 2024 16 LAForge, M. p., n. L. MicHeL, A. L. WHeeLer, and r. k. Brook. 2016. Habitat selec- tion by female moose in the Canadian prairie ecozone. Journal of Wildlife Management 80: 1059-1068. LAnkester, M. W., and W. J. Foreyt. 2011. Moose experimentally infected with giant liver fluke (Fascoiloides magna). Alces 47: 9-15. LepticH, d. J., and J. r. giLBert. 1989. Summer home ranges and habitat use by moose in northern Maine. Journal of Wildlife Management 53: 880-885. Lesser, M. r., M. doVciAk, r. WHeAt, p. curtis, p. sMALLidge, J. Hurst, d. krAMer, M. roBerts, and J. FrAir. 2019. Modelling white-tailed deer impacts on forest regeneration to inform deer man- agement options at landscape scales. Forest Ecology and Management 448: 395-408. JoHnson, d. H. 1980. The comparison of usage and availability measurements for evaluating resource preference. Ecology 61: 65-71. MAccrAcken, J. g., V. V. VAn BALLenBergHe, and J. M. peek. 1997. Habitat relation- ships of moose on the Copper River Delta in coastal south-central Alaska. Wildlife Monographs. 1:3-52. Morris, d. 2014. Aquatic habitat use by North American moose (Alces alces) and associated richness and biomass of submersed and floating-leaved aquatic vegetation in north-central Minnesota. Ph.D. Thesis, Lakehead University, Ontario, Canada. 130pp. MuFF, s., J. signer, and J. FieBerg. 2020. Accounting for individual-specific varia- tion in habitat-selection studies: efficient estimation of mixed-effects models using Bayesian or frequentist computation. Journal of Animal Ecology 89: 80-92. NYSDEC: neW york depArtMent oF enVironMentAL conserVAtion. 2021. Management plan for white-tailed deer in New York State, 2021-2030. Albany, New York. 84pp. ________. 2022. Conservation easements. URL https://dec.ny.gov/nature/for- ests-trees/conservation-easements. Accessed: January 1, 2024 NOHRSC: nAtionAL operAtionAL HydroLogic reMote sensing center. 2004. Snow data assimilation system (SNODAS) data products at NSIDC, Version 1. Boulder Colorado, USA. National Snow and Ice Data Center. https://doi.org/10.7265/N5TB14TC. Accessed: June 1, 2000 peek, J. M, d. L. uricH, and r. J. MAckie. 1976. Moose habitat selection and rela- tionships to forest management in north- eastern Minnesota. Wildlife Monographs 48: 3-65. peterson, s., d. krAMer, J. Hurst, and J. FrAir. 2020. Browse selection by moose in the Adirondack Park, New York. Alces 56: 107-126. _______, _______, _______, d. spALinger, and J. FrAir. 2022. Estimation of bio- mass for moose in the Adirondack Park, New York. Alces 58: 1-30. post, e., and n. c. stensetH. 2002. Large- scale climatic fluctuation and population dynamics of moose and white-tailed deer. Journal of Animal Ecology 67: 537-543. pooLe, k. g., and k. stuArt-sMitH. 2005. Fine scale winter habitat selection by moose in interior montane forests. Alces 41: 1–8. Quinn, A. c. d., d. M. WiLLiAMs, and W. F. porter. 2013. Landscape structure influ- ences space use by white-tailed deer. Journal of Mammalogy. 94:398-407. r core teAM. 2022. R: A language and envi- ronment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https:// www.R-project.org/. rAyMond, k. s., F. A. serVeLLo, B. griFFitH, and W. e. escHHoLz. 1996. Winter for- aging ecology of moose on glypho- sate-treated clearcuts in Maine. The Journal of Wildlife Management. 60:4-753-763. https://doi.org/10.7265/N5TB14TC ALCES VOL. 60, 2024 HABITAT SELECTION BY MOOSE 17 reA, r. V., and M. p. giLLingHAM. 2001. The impact of the timing of brush manage- ment on the nutritional value of woody browse for moose Alces alces. Journal of Applied Ecology 38:710-719. renecker, L. A., and r. J. Hudson. 1986. Seasonal energy expenditures and thermo- regulatory responses of moose. Canadian Journal of Zoology 64: 322-327. ritcHie, W. A. 1969. The archeology of New York state. The Natural History Press. Garden City, NY. 357pp. ________, and r. e. Funk. 1973. Aboriginal settlement patterns in the Northeast. New York State Museum and Science Memoir 20. 378pp. sikes, r. s., and tHe AniMAL cAre And use coMMittee oF tHe AMericAn society oF MAMMALogists. 2016. 2016 Guidelines of the American Society of Mammalogists for the use of wild mam- mals in research and education. Journal of Mammalogy 3:633-688. seVeringHAus, c. W., and L. W. JAckson. 1970. Feasibility of stocking moose in the Adirondacks. New York Fish and Game Journal 17: 19-32. teiteLBAuM, c. s., A. p. k. siren, e. coFFeL, J. r. Foster, J. L. FrAir, J. W. Hinton, r. M. Horton, d. W. krAMer, c. Lesk, c. rAyMond, d. W. WAttLes, k. A. zeLLer, and T. L. MoreLLi. 2021. Habitat use as indicator of adaptive capacity to climate change. Diversity and Distributions 27: 655−667. tierson, W. c., g. F. MAttFeLd, r. W. sAge, and d. F. BeHrend. 1985. Seasonal movements and home ranges of white- tailed deer in the Adirondacks. Journal of Wildlife Management 49: 760-769. tiscHLer, k. B., W. J. seVerud, r. o. peterson, J. A. VuceticH, and J. K. BuMp. 2022. Aquatic areas provide high nitrogen forage for moose (Alces alces) in Isle Royale National Park, Michigan, USA. Alces 58: 75-90. tHoMpson, d. p., J. A. crouse, p. s. BArBozA, M. o. spAtHeLF, A. M. HerBerg, s. d. pArker, and M. A. Morris. 2021. Behavior influences thermoregulation of boreal moose during warm season. Conservation Physiology 9: coaa130, https://doi. org/10.1093/conphys/coaa130 VAnderWAAL, k. L., s. k. WindeLs, B. t. oLson, J. t. VAnnAttA, and R. Moen. 2015. Landscape influence on spatial patterns of meningeal worm and liver fluke infection in white-tailed deer. Parasitology 142: 706-718. VAnnAttA, J. t., and R. Moen. 2016. Giant liver fluke and moose: just a fluke? Alces 52:117-139. VerMe, L. J. 1968. An index of winter weather severity for northern deer. Journal of Wildlife Management 32: 566-574. WAttLes, d. W., and S. desteFAno. 2013. Moose habitat in Massachusetts: assess- ing use at the southern edge of the range. Alces 49: 133-147. WHitLAW, H. A., and M. W. LAnkester. 1994. The co-occurrence of moose, white-tailed deer and Parelaphostrongylus tenuis in Ontario. Canadian Journal of Zoology. 72:819-825. WisH, J. d. 1902. Report of the Secretary of the Commission in: Eighth Report of the Forest, Fish and Game Commission. Albany, NY. 456 pp. yAng, L.M., s.M. Jin, p. dAnieLson, c. HoMer, L. gAss, s. M. Bender, A. cAse, c. costeLLo, J. deWitz, J. Fry, M. Funk, B. grAnneMAn, g. c. Likens, M. rigge, and g. xiAn. 2018. A new generation of the United States national land cover database: requirements, research priori- ties, design, and implementation strate- gies. ISPRS Journal of Photogrammetry and Remote Sensing 146: 108-123. zeVin, r. A. 2022. Modeling the spatial extent and intensity of beaver (Castor canadensis) impacts on stream networks and forest structure in Adirondack State Park, NY. M.S. Thesis, State University of New York, College of Environmental Science and Forestry, Syracuse. 128pp. https://doi.org/10.1093/conphys/coaa130 https://doi.org/10.1093/conphys/coaa130