Conservation status, threats, and information needs of small mammals in Alaska

Despite their diversity and ecological importance, small mammals are under-represented in conservation research relative to other mammals. We evaluated the conservation status of 36 small mammal species in Alaska, U.S.A. using a ranking system that we previously developed, the Alaska Species Ranking System (ASRS). We compared results from the ASRS with NatureServe’s subnational rankings. Finally, we surveyed taxonomic experts to identify recommended conservation actions and research priorities for 5 species of high conservation concern. In general, the ASRS and NatureServe agreed on the rankings of species in the highest and lowest risk categories. Species of highest conservation concern were taxa endemic to the state, including 2 island-endemic shrews, and taxa from the orders Chiroptera and Eulipotyphla. Because the ASRS includes information needs in its assessment, 15 of the 20 species considered lowest concern by NatureServe were considered intermediate concern by the ASRS. In the ASRS, most species (n = 24) were assessed to have low biological vulnerabilities, but high information needs. Population size and trends were unknown for all species; distributional limits and understanding of population dynamics were incomplete for all species except 4. Disease and climate change effects on habitat were perceived as important threats, but affected only 8 species. Taxonomic experts identified addressing data deficiencies and protecting habitat as important conservation actions; they identified monitoring population trends, modeling habitat, and researching species’ genetic diversity and adaptive capacity as high priorities. Conservation assessments that require accurate and current data on population trends or threats may lead to bias against data deficient groups such as small mammals. Our findings demonstrate the importance of accounting for data deficiencies in conservation status ranks to avoid conflation of sparse information with low conservation concern.

and lowest risk categories. Species of highest conservation concern were taxa endemic to the state, 23 including 2 island-endemic shrews, and taxa from the orders Chiroptera and Eulipotyphla. Because 24 the ASRS includes information needs in its assessment, 15 of the 20 species considered lowest 25 concern by NatureServe were considered intermediate concern by the ASRS. In the ASRS, most 26 species (n = 24) were assessed to have low biological vulnerabilities, but high information needs. 27 Population size and trends were unknown for all species; distributional limits and understanding 28 of population dynamics were incomplete for all species except 4. Disease and climate change 29 effects on habitat were perceived as important threats, but affected only 8 species. Taxonomic 30 experts identified addressing data deficiencies and protecting habitat as important conservation 31 actions; they identified monitoring population trends, modeling habitat, and researching species' 32 genetic diversity and adaptive capacity as high priorities. Conservation assessments that require 33 accurate and current data on population trends or threats may lead to bias against data deficient 34 groups such as small mammals. Our findings demonstrate the importance of accounting for data 35 deficiencies in conservation status ranks to avoid conflation of sparse information with low 36 conservation concern. 37 INTRODUCTION 38 Conservation practitioners and natural resource managers are often tasked with prioritizing 39 effort and funding for species based on extirpation risk or vulnerability to threats. To aid with 40 prioritization, practitioners often use ranking systems, such as those developed by the International 41 Union for the Conservation of Nature (IUCN) or NatureServe. Conservation ranking systems 42 assign a status to a species by evaluating and scoring that species across a set of criteria (IUCN 43 2012;Master et al. 2012). Scoring criteria requires data that are accurate, current, and available. 44 When data are scarce or absent, species may receive a special status (e.g., data-deficient for the 45 IUCN or unrankable for NatureServe). In less extreme cases, assessors can score some questions 46 as unknown or select a range of answers to express uncertainty (IUCN 2012;Master et al. 2012). 47 Designations of uncertainty allow assessors to assign conservation status to species that lack the 48 data necessary to reliably score a subset of criteria; however, ranking systems can assign low-risk 49 status to species with unknown population trends or threats if designations of uncertainty have no 50 influence on the rank calculation. Data deficient species present challenges to conservation 51 practitioners because the funding necessary to address data gaps can be difficult to justify for 52 species with low-risk status as compared with species ranked as at-risk because of more complete 53 or accurate data (Jetz and Freckleton 2015). 54 Small mammals (Chiroptera, Eulipotyphla, Rodentia, Lagomorpha) compose over 75% of 55 the Earth's extant mammalian diversity and function as primary consumers, insectivores, vectors 56 of disease, and focal prey species (Ceballos & Brown 1995;Entwistle & Stephenson 2000). 57 Despite their diversity and ecological importance, limited knowledge of population sizes, 58 population and distribution trends, and threats often precludes assessment of their conservation 59 status using traditional ranking systems. Under the IUCN ranking system, 16% of Rodentia species 60 6 combination, the 3 themes effectively assess the risk of regional extirpation of non-endemic 105 species and extinction of endemic species (collectively referred to as "conservation concern"). 106 Conservation concern increases with numeric value in the ASRS and scores can be positive 107 or negative. The 2 questions in Trends are evaluated on a 5-point scale ranging from -10 to 10. A 108 high Trends score indicates currently declining populations or shrinking distributions. Biological 109 Vulnerability comprises 7 questions, which are evaluated on 3-, 4-, or 5-point scales. Three-point 110 scales range from -5 to 5; the others range from -10 to 10. A high score for this theme indicates 111 that the species has several traits (e.g., small population size, restricted range, slow life history, or 112 high ecological specialization) that make it more vulnerable to extirpation. Finally, Action Needs 113 comprises 4 questions that are each evaluated on a 3-point scale, which indicate low (-10), 114 moderate (2), or high needs (10). A high Action Needs score denotes an absence of management 115 and conservation actions, resulting in large information needs. Scores within themes are summed 116 to create a theme score. Each theme score is then categorized to create a final, categorical rank. 117 The ASRS categorizes each theme score as high, unknown, or low; score thresholds for 118 each category are presented in Gotthardt et al. (2012). Categorization results in 9 numerical ranks 119 ranging from I to IX. Numerical ranks are further grouped into one of 4 color categories: red 120 (highest conservation concern, numerical ranks I-II), orange (ranks III-V), yellow (ranks VII and 121 VIII), and blue (ranks VI and IX). Red, Orange, and Yellow indicate high, unknown, or low Trends 122 scores, respectively. Species in these categories also scored high for one or both of the remaining 123 themes, Biological Vulnerability and Action Needs. One exception is the rank Orange III, which 124 indicates a high Trends score but low Biological Vulnerability and Action Needs. Finally, a rank 125 of Blue indicates low Biological Vulnerability and Action Needs, and either an unknown or low 126 Trends score. 127 7 The assessment process for the ASRS begins with a trained assessor conducting a 128 systematic review, preparing a species' account, and conducting an initial assessment to obtain 129 preliminary scores (Fig. 1). The assessor searches primarily for information relevant to Alaskan 130 populations but expands the search to include other populations when information on Alaskan 131 populations is insufficient to assign a score. A second assessor reviews the account and conducts 132 an assessment without consulting the scores of the first assessor. If assessors disagree on a score, 133 they discuss the question and consult additional sources. The assessment is then sent to a 134 taxonomic expert for external review (Fig. 1). Once the review is complete, assessors update scores 135 following the expert's recommendation and finalize the assessment. Species' accounts, along with 136 related conservation resources, are published online where they are publicly available (Fig. 1). 137 timing. An overall threat impact score is calculated using severity and scope scores across all the 144 threats that were identified (Master et al. 2012). 145 S ranks range from 1 to 5. A rank of S1 indicates that the species is critically imperiled in 146 that subnational jurisdiction (typically a state or province). A rank of S5 indicates that the species 147 is secure in that subnational jurisdiction. A range rank (e.g., S3S4) indicates that the status of the 148 species is uncertain within the bounds of the two values. Because the ASRS and NatureServe 149 ranking systems use similar criteria, we used information from the ASRS species' accounts to 150 8 update the S ranks of the 36 small mammal species. Scientific literature and expert opinion 151 informed the Threats assessments, which are not included in the ASRS (see next section). 152

154
In 2019, we surveyed taxonomic experts to obtain their judgment on the most important 155 conservation actions, research priorities, and threats facing 5 species: Glaucomys sabrinus, 156 Marmota broweri, Myotis lucifugus, Ochotona collaris, and Synaptomys borealis. Each of the 157 selected species are of high conservation concern (ADF&G 2015; this paper) and have been the 158 topic of dedicated research projects in the state. For each species, we identified 3 to 7 experts. We 159 defined an expert as a scientist who was directly involved in a research project investigating the 160 species in Alaska. 161 Using an online survey tool, we asked experts a series of 7 questions (Table 1; Droghini et  162 al. 2020). Answers to the first 5 survey questions informed the evaluation of threats for the 5 163 selected species and species with similar ecologies. The last 2 questions asked experts to identify 164 conservation actions to mitigate threats and identify the most critical research needs if they were 165 responsible for allocating a large sum of money (US $10 million) to research activities (Table 1). 166

167
We evaluated the conservation status of 36 small mammal species across 4 orders and 7 168 families. Assessments from both the ASRS and NatureServe ranking systems are available online 169 The median score for Biological Vulnerability was -32, out of a possible minimum of -50. 209 When grouped by order, median scores for Eulipotyphla, Rodentia, and Lagomorpha were low 210 (range: -36 to -32). The median score for Chiroptera was -7, which we consider high. Top-ranking 211 species for Biological Vulnerability were Sorex pribilofensis (theme score = 14), S. jacksoni (8), 212 and Myotis volans (3). 213 Median scores for range size and number of aggregation sites were the lowest possible 214 scores, indicating that most small mammal species are widespread in Alaska (Fig. 2). Variability 215 in scores for these questions was minimal and characterized by the presence of outliers. The 216 median score for population size was -6 (Fig. 2), which is selected if the population size is 217 unknown but suspected to be large (i.e., more than 10,000 individuals; Gotthardt et al. 2012). 218 Median scores for dietary specialization and habitat specialization were 1; because these questions 219 11 are assessed on 3-point scales, a value of 1 indicates moderate specialization (Fig. 2). No species 220 was assessed to have high dietary specialization and few species were assessed to have high habitat 221 specialization (Fig. 2). 222

Management and Conservation Action Needs 223
No species obtained partial scores for any questions in Action Needs. The median score 224 for Action Needs was 24. When grouped by order, Eulipotyphla had the highest median score (32)  225 while Lagomorpha had the lowest (4). Two species received a score of 40, which is the maximum 226 possible score for Action Needs: Sorex minutissimus and S. navigator. Most species had high 227 management needs, indicating that they are not subject to direct management actions, and high 228 monitoring needs, indicating that their population trends are not consistently or extensively 229 monitored (Fig. 3). Species with moderate monitoring needs belonged to one of two families: 230 Vespertilionidae or Leporidae. These species are monitored by state agencies, but data are 231 inadequate to detect trends. 232 Nearly all species (n = 32) had high or moderate inventory needs; thus, knowledge of range 233 limits and habitat associations remains incomplete (Fig. 3). The species with low inventory needs 234 were Lepus americanus, Myodes rutilus, Peromyscus keeni, and Tamiasciurus hudsonicus. These 235 species are widespread, common, and easy to detect or capture in traps. 236 Twenty-three species (64%) had high research needs, reflecting a lack of information on 237 the factors that limit populations. These species included all species endemic to Alaska, all 238 Eulipotyphla, and all Chiroptera with the exception of Myotis lucifugus. 239

240
Two-thirds of the species we assessed (n = 24) received a low threat impact score. Non-241 native disease (i.e., white-nose syndrome) was listed as a threat for all Chiroptera species, though 242 12 impact scores varied by species and ranged from very high to medium-low (Droghini et al. 2020). 243 We considered timber harvest to be a medium-low threat for Chiroptera species largely restricted 244 to Southeast Alaska. We considered habitat alteration due to climate change a high-medium threat 245 for talus specialists. Sorex jacksoni and S. pribilofensis received an impact score of high-low; this 246 score reflects the potentially large, but highly uncertain effects of disturbances on narrowly 247 endemic species. 248 Experts tended to disagree about the severity or scope of threats, which is reflected for 249 some species as ranges in the overall impact scores (Droghini et al. 2020). Experts also disagreed 250 about the timing of climate change related threats, both within and across species, reflecting 251 uncertainty as to whether effects would be expressed in the short-or long-term (Droghini et al. 252 2020). 253

254
We received 23 completed surveys: 5 per species with the exception of Synaptomys 255 borealis, for which we were able to identify only 3 experts. The most commonly suggested 256 conservation actions to mitigate threats were to collect more information and to protect known 257 habitat. When asked to allocate US $10 million to different research topics, experts considered 258 monitoring of population trends, research on genetic diversity and adaptive capacity, and habitat 259 modeling important for all species, with $2 to $3 million devoted to each topic (Fig. 4) on diseases or parasites was judged to warrant comparably little funding (Fig. 4). 264 13

265
The Alaska Species Ranking System (ASRS) explicitly identifies key information needs 266 by assessing the strength of conservation actions around inventory, monitoring, and research. More 267 than 2/3 of small mammal species in Alaska are of high conservation concern in the ASRS, a result 268 driven largely by the lack of information about species' population trends, distributional limits, 269 and population ecology. The prevalence of data deficiencies for small mammals is not unique to 270 Alaska: relative to other mammal groups, a large proportion of small mammal species are listed as 271 data-deficient by the IUCN but are likely threatened (Jetz & Freckleton 2015). 272 Species of the orders Chiroptera and Eulipotyphla and species endemic to Alaska were of 273 particularly high conservation concern. Chiroptera species have low reproductive rates and 274 specific habitat requirements for roosting and hibernating; several Chiroptera species also have Sorex jacksoni and S. pribilofensis, were the highest-ranked species in both the ASRS and 285 NatureServe. Their ranges are restricted to single islands in the Bering Sea; narrowly endemic 286 14 species have greater risk of extinction due to small population sizes, small range sizes, and 287 demographic stochasticity (Hartley and Kunin 2003;Cardillo et al. 2008). 288

289
In general, the ASRS and NatureServe ranking systems agreed on the rankings of species 290 in the highest and lowest risk categories. However, most species that were of intermediate concern 291 by the ASRS were ranked of lowest conservation concern by NatureServe. Both ranking systems 292 recognize these species' low biological vulnerabilities: these species are relatively widespread, 293 presumed common, and have life history traits and ecological preferences that correlate with low 294 extirpation risk. The divergence in conservation ranks largely reflects the importance that the 295 ASRS ascribes towards information needs and data deficiencies; in the NatureServe ranking 296 system, data deficiencies do not weight the score towards greater conservation concern. 297