Vulnerability assessment of the multi- sector North American bison Bison bison management system to climate change

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2021 The Authors. People and Nature published by John Wiley & Sons Ltd on behalf of British Ecological Society 1Center of Excellence for Bison Studies, South Dakota State University, Rapid City, SD, USA 2Department of Ecology and Conservation Biology, Texas A&M University, College Station, TX, USA 3School of Education, Colorado State University, Fort Collins, CO, USA 4Department of Rangeland, Wildlife, and Fisheries Management, Texas A&M University, College Station, TX, USA


| INTRODUC TI ON
Grasslands and bison Bison bison are co-evolutionarily dependents, each facing challenges from climate change and land use in the 21st century. Yet each may facilitate the conservation success of the other. Grasslands of the Great Plains once encompassed 2.8 million km 2 at the time of European settlement representing 14% of the landmass of the United States and Canada combined (Johnsgard, 2003;Licht, 1997). Since then, Great Plains grasslands have undergone agricultural intensification, especially in the eastern portion of the region (i.e. tallgrass and mixed grass prairies) and today are 90% privately owned (Holechek et al., 2011). Historic populations of bison numbered between 30 and 60 million prior to 1868 (Flores, 1991;Hornaday, 1889) and predominately inhabited the Great Plains ( Figure 1). North American bison conservation has overcome an incredible hurdle of restoring populations to nearly 400,000 from fewer than 1,000 individuals in 1884 (Stoneberg Holt, 2018). However, the International Union for Conservation of Nature, Red List assessment, considers bison 'near threatened', because their assessment relies solely on publicly owned (i.e. Federal and State) 'conservation' bison herds (Aune et al., 2017). Publicly owned bison populations have remained static around 30,000 bison since the 1930s because the extent of public lands has not expanded, especially not in the Great Plains (Gates et al., 2010). In addition, the few existing public herds are positioned at the margins of the historical bison range (i.e. in and near the Rocky Mountains) and each have populations that are considered too small (i.e. fewer than 1,000 individuals) for long-term conservation success (Sanderson et al., 2008). Meanwhile, the bison meta-population approaches 400,000, of which nearly 85% of those are privately owned and considered 'production' herds (Gates et al., 2010;United States Department of Agriculture, 2016), yet many of these 'production' herds satisfy 'conservation' guidelines established in the 'Vermejo Statement' about bison conservation (Sanderson et al., 2008). This suggests that private bison stewardship and private land conservation (Drescher & Brenner, 2018;Kamal et al., 2015;Lueck, 2002) is essential to maintain this iconic species and its grassland habitats.
Bison conservation on private lands is unique because this native wildlife species can be managed as livestock in some respects (Aune & Plumb, 2019;Ranglack & du Toit, 2015;Redford et al., 2016). For example, they can be bought and sold, and vaccinated as are livestock, but genetic manipulation and use of artificial insemination is allowed only for research purposes (National Bison Association, 2013). The duality of this conservation and production system exists at the intersection of the conventional North American Model for wildlife management with their respective mandates (Clark & Milloy, 2014;Geist et al., 2001;Peterson & Nelson, 2017) and conventional large livestock production (Byrd et al., 2017). Bison and their management transcend multiple sectors of ownership including public, private, tribal and non-governmental organizations (NGOs). We refer to this K E Y W O R D S bison conservation, bison production, ecosystem services, Great Plains, private land conservation, social-ecological systems, vulnerability scoping diagram, wildlife conservation and management F I G U R E 1 Map of North America regions and regions and prairie types investigated; regions: northern (blue), central (orange), southern (red) and eastern (green). Prairie types: shortgrass prairie (stippled), mixed grass prairie (diagonal hatch) and tallgrass prairie (cross hatch); collectively the prairies represent the Great Plains. The absence of a prairie type indicates 'other' ecosystem type. Historic bison range prior to 1868 is indicated by thick black outline (Hornaday, 1889) ty multi-sector combination of land and bison ownership as the bison management system (BMS).
Private land conservation has proven challenging because of the tension that resides between societal benefits realized from non-provisioning ecosystem services (e.g. supporting, regulating and cultural) and private market-based provisioning services optimized on private property. Procedures to balance private ownership with the costs of conservation have remained elusive (Drescher & Brenner, 2018;Kamal et al., 2015). Consequently, the potential for private bison ownership to contribute to future bison conservation in the absence of collaboration with other sectors is unknown. Effective collaboration among sectors may prove necessary for the BMS to adapt to climate change and increasing climate variability (Klemm et al., 2020a).
Accelerating climate change throughout the Great Plains in the 21st century (Wuebbles et al., 2017) may represent the next major challenge to bison survival. Climate change directly affects bison by increasing thermal stress (Martin & Barboza, 2020b) and decreasing forage and water availability. Indirect consequences of climate change include increasing distribution and intensity of parasites (Kutz et al., 2005;Morgan & Wall, 2009;Patz et al., 2000) and several diseases (Janardhan et al., 2010) that are known to reduce reproductive success (Fuller et al., 2007). These stresses have been estimated to collectively reduce bison body size by 50% if global temperature warms by 4°C near the end of the 21st century (Martin & Barboza, 2020a;Martin et al., 2018). Furthermore, climate change, particularly warming and drought, may have cofounding impacts on sustainability of the remaining grasslands of the Great Plains by altering intensification rates of agriculture, land use and woody plant encroachment (Allred et al., 2013;Bowler et al., 2020;Klemm et al., 2020b;Knapp et al., 1999).
A vulnerability assessment of the BMS to increasing climate variability and change throughout the 21st century was conducted to further clarify the challenges that bison conservation may face in future climates. Data referencing the exposure, sensitivity and adaptive capacity of both private and public bison managers were collected via an email survey. Data describing these three components were analysed with a vulnerability scoping diagram (VSD; Supporting Information Figure S1) and framework as described by Adger (2006) andPolsky et al. (2007). Exposure characterizes the stressors or hazards that may threaten a system, especially coupled human and natural systems.
Sensitivity characterizes the rate and magnitude of ecological or economic impact on these systems. Adaptive capacity characterizes the societal responses that mitigate system sensitivity to stressors and hazards. The VSD provides a holistic approach to incorporate elements of social connections, biological responses, climatological drivers and decision making for assessing first-, second-and third-order effects of climate change (Supporting Information Figure S1).

| Survey instrument
Our VSD relied on participant responses to a survey questionnaire we developed which contained 68 total questions-divided into 18 measures, two measures for each of the nine components that represent three dimensions of vulnerability-exposure, sensitivity and adaptive capacity (Table 1). We also queried stakeholders about their environmental values, attitudes and management practices.
Total vulnerability scores are derived from the 11 questions measuring exposure, 23 questions measuring sensitivity and 15 questions measuring adaptive capacity (Table 1; Supporting Information   Table S2). To test whether survey participants could understand the scale items (i.e. face validity), a pretest was conducted with a 10-person focus group comprised of private and public bison herd management officials (Martin, 2020). Although the dimensions of the VSD are consistent with Polsky et al. (2007), the components, measures and survey questions are tailored to the bison management system (BMS). Additionally, to determine underlying drivers of differences in adaptive capacity, we measure environmental values, attitudes and management practices of participant perceptions.
Survey responses were recorded on a Likert scale from 0 to 5-equating to no response, strongly disagree, somewhat disagree, neutral, somewhat agree, strongly agree-or on a scale of agreement with three points (agree-neutral-disagree) or two points (positive or negative perceptions). Focal topics included elements of management philosophy, diversity of income, land and animal health monitoring, management practices, perceived value of economic and ecological factors, quality of life and career experiences and personal characteristics. Attitudes towards various ecological and management practices, such as using prescribed burning or diversifying livestock species were also surveyed. Respondents were also questioned about their perceptions and/or observations of climate over the last 10 years, such as warming mean summer temperature or shifting calf survival rates on their operations.
Responses from the VSD indicating more vulnerability were valued higher (e.g. 'Do you provide water to your animals?'-Yes or no; where yes is scored as 0 and no as 1). An alternative procedure was developed to interpret and visualize the results of the VSD. All non-personal trait questions were standardized to a 10-point scale for subsequent analyses with high scores to each question indicating higher vulnerability. We collated the respective questions within each measure of the VSD to record the median score. Building up to the scores of each dimension, we collated the respective measures within each dimension of the VSD to record the median score.
High scores (> 6) indicate increasingly lower vulnerability, at the individual, sector and system level, and are depicted on the outer edge of a 10-point radar plot (Adger, 2000(Adger, , 2006. The overall vulnerability score is calculated as the mean of the three median dimension scores of exposure, sensitivity and adaptive capacity.

| Respondents
We collected 156 responses from an estimated pool of 1,049 bison managers (National Agricultural Statistics Service, 2019) for a response rate of 15%. We removed 24 incomplete or unqualified responses for an analytical set of 132 responses. The median time spent by each respondent on the survey was 10.4 min. We classified respondents by sector, region, ecosystem, operation size and education level. Respondents were asked if they represent private, public or non-governmental organization bison herds. Demographic attributes of respondents are presented in Table 2.
An independent two-sample t test using groups was conducted to determine the difference among various attitudes and practices of public and NGO sectors. There were no significant differences between public and NGO sectors in their environmental values, attitudes and practices (albeit, each group has low n; Supporting Information Table S1). As an example, their attitudes towards economic diversification are not significantly different (μ = 3.6 and 3.8; σ = 1.7 and 0.8; p = 0.82). We present additional Supporting Information in the Supplemental Information about ecosystem and regional variance using Pearson's correlation coefficients, of which do not significantly vary, yet indicate that sector variance is similar for the public and NGO sectors (Supporting Information Figures S2-S4). As such, the public and NGO sectors were combined into one group because of overlap in common values, attitudes and practices that was compared to the private sector in subsequent analyses. Most responses were from the private sector (121% or 92%) with 5 (4%) responses from NGO, and 6 (5%) from managers of public herds-for a combined public/NGO sector of 11 (9%) responses. This sample reflects the proportional ownership of bison in North America: 81% private, 5% NGO, 8% public and 5% tribal.   (Valliant & Dever, 2018). We used t tests to compare means between private and public/NGO sectors with α at p < 0.05.

| RE SULTS
Total vulnerability, on a scale of 0 to 10-10 being the most vulnera-

| Exposure, sensitivity and adaptive capacity
Adaptive capacity was the only dimension of vulnerability that varied significantly between the two sectors ( Figure 3). Public/ NGO respondents had less adaptive capacity (median = 6.0 ± 1.8;

F I G U R E 2
Radar plot of median vulnerability scores derived from a vulnerability scoping diagram (VSD; Supporting Information Figure S1) for private (red solid line, n = 121) and public/NGO (blue dash line, n = 11) bison mangers. At the top of the VSD (12 o'clock position) with the average of all measures indicated as 'Overall'. Clockwise from 'Overall' include the three dimensions (separated by thick black lines radiating from the centre) of sensitivity, adaptive capacity and exposure. Each of the six measures within each dimension is indicated as a spoke on the VSD wheel. Each measure is comprised of various survey questions (see Table 1). Low scores near the centre are less vulnerable than high scores near the edge. Symbols: ‡ moderately or highly vulnerable (i.e. ≥5), and ** significantly different between sectors (p ≤ 0.05). Abbreviations: Exp. and Edu., experience and education; Bison Pop., bison population; Env. Variation, environmental variation; and Info. Exchange, information exchange 2.5-9.3; ≈highly vulnerable) than private sector respondents (median = 4.0 ± 2.1; 0-8.6; ≈moderately vulnerable; Table 3).
Overall, exposure (4.2 ± 1.2) was perceived slightly less than sensitivity (3.8 ± 1.6). Means of exposure, sensitivity and adaptive capacity did not differ significantly across regions and ecosystem types (Supporting Information Figure S4).
Between public/NGO and private sectors, the largest differences in adaptive capacity were based on four measures: information exchange, external revenue, access to leased grazing lands and management plans (Figure 3). Information exchange-monitoring vegetation and wildlife diversity-was more prevalent in public/NGO sector (p ≤ 0.001), while external revenue-investment income and subsidies for sustaining bison production (p ≤ 0.05)-and leasing of grazing lands were more prevalent in the private sector (p ≤ 0.001). Management plans, specifically having or creating drought contingency plans, were more prevalent in the private than the public/NGO sector (p ≤ 0.04).

| Educational level
An increase in formal education of managers corresponded to a lower vulnerability score. Estimated scores were 4.7 for those without F I G U R E 3 Correlation between adaptive capacity and over overall vulnerability scores for private (n = 119) and public/NGO (n = 11) sectors with kernel density plots illustrating overlapping distribution. Adaptive capacity scores differ significantly (p ≤ 0.005) between sectors, with private and public/NGO sector scores of 4.0 (moderately vulnerable) and 6.0 (highly vulnerable) respectively (

| Values, attitudes and practices
Adaptive capacity is the primary vulnerability dimension that differed between private and public/NGO sectors. This may primarily be a consequence of the administrative mandates that limit management activities of the public/NGO sector compared to the private sector. Overall, private and public/NGO bison managers share similar values and attitudes towards ecological integrity values ( Figure 4) and economic diversification ( Figure 5). Ecological integrity values was endorsed by 90% of public/NGO and 86% of private managers ( Figure 4). Only 60% of public/NGO managers valued economic diversification, which was endorsed by 81% of private managers ( Figure 5).
However, when presented with various attitudes and practices for ecological and economic management techniques, consensus of affirmation declined within and between the sectors as described below.

F I G U R E 4 Bison manager ecological values, attitudes and practices by sector
Economic attitudes were less consistent than ecological attitudes in both sectors ( Figure 5) Survey responses indicated that private bison managers were 'moderately vulnerable' and exhibit 'moderate adaptive capacity' to climate change (Table 3), whereas previous studies have described cattlemen as 'highly vulnerable' (Joyce et al., 2013;Williamson et al., 2012;Wilmer et al., 2018). However, sole reliance on private bison managers is insufficient for a comprehensive bison and grassland conservation strategy. We found the public/NGO sector to be 'moderately vulnerable' to climate change, which was greater than the private sector (Table 3). The private and public/NGO sectors share common values and attitudes for bison conservation even though management practices vary (Figures 4 and 5).
Variation in adaptive capacity among the private and public/ NGO sectors (Figure 2) originated from differences in management practices specific to each sector (Figures 4 and 5), not from ecoregional variation (Supporting Information Figures S2-S4). This is to be anticipated given the economic concern of private managers and the agency and federal mandates prescribed for public/NGO sectors.
However, tensions between production and conservation goals may be surmountable based on their shared environmental values and attitudes and high level of education. Higher levels of education of bison managers may be key to reducing vulnerability in two ways: (a) education increases the ability to seek, sort and apply new information from multiple sources (Supporting Information Table S4), and (b) education also increases the ability to generate external income that diversifies revenue streams (Supporting Information Figure S5).
Education facilitates information exchange through social networks and associations beyond the bison system, which may be critical to recruitment of new private sector managers and the development of additional adaptive capacity (Table 1).
Here, we propose that the BMS should be further organized into a bison management coalition to collective mitigate and adapt to accelerating challenges throughout the region, especially climate change. Formal comprehensive integration of the robust, nascent multi-sector BMS into a bison management coalition could potentially resolve tensions between the conservation and production goals and collectively reduce system wide vulnerability. For example, enhanced collaboration may promote development of regional drought adaptation plans that seek to strategically destock and restock bison populations between and among sectors and regions.
This strategy may minimize non-harvest, lethal culling practices and promote non-lethal, translocation culling practices that potentially enhance and share operational costs by increasing external revenue for the public/NGO sectors. Bison conservation goals may be enhanced by the transfer of animal genetics from public herds to non-public (i.e. private, NGO and Tribal) herds that may increase both meta-population size and bison production. This represents the implementation of conservation-oriented bison strategies that conserve unique genetic traits in public herds while sustaining the meta-population.
Examples currently exist for inter-sector and inter-agency management coordination (Table 4). These examples illustrate the extent and complexity of current interactions among sectors that may serve as the foundation for a more formal bison management coalition.
Cross-sector coordination may greatly enhance bison population regulation by facilitating animal distribution and improve development and implementation of management plans and strategies.
Bison exchange among sectors will inevitably increase the potential for disease and parasite transmission, but a system wide monitoring program could minimize regional exposures. Moreover, monitor-

ACK N OWLED G EM ENTS
We thank Rachel A. Short, Clint Patterson, Jim I. Mead and Jason B.
West for reading and critiquing all or parts of this article and study design. J.M.M. thanks the comradery and support of 'The Write Stuff' interdisciplinary writing group. We thank Jim Matheson and Dave Carter of the National Bison Association for clarifying bison rancher access to various disaster relief programs. We are grateful for the feedback from two anonymous reviewers that greatly enhanced the quality and clarity of the manuscript. We are also thankful to the respondents that participated in our survey.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.