Modeling Fence Location and Density at a Regional Scale for Use in Wildlife Management

Barbed and woven wire fences, common structures across western North America, act as impediments to wildlife movements. In particular, fencing influences pronghorn (Antilocapra americana) daily and seasonal movements, as well as modifying habitat selection. Because of fencing's impacts to pronghorn and other wildlife, it is a potentially important factor in both wildlife movement and habitat selection models. At this time, no geospatial fencing data is available at regional scales. Consequently, we constructed a regional fence model using a series of land tenure assumptions for the Hi-Line region of northern Montana – an area consisting of 13 counties over 103,400 km2. Randomized 3.2 km long transects (n = 738) on both paved and unpaved roads were driven to collect information on habitat, fence densities and fence type. Using GIS, we constructed a fence location and a density model incorporating ownership, size, neighboring parcels, township boundaries and roads. Local knowledge of land ownership and land use assisted in improving the final models. We predict there is greater than 263,300 km of fencing in the Hi-Line region, with a maximum density of 6.8 km of fencing per km2 and mean density of 2.4 km of fencing per km2. Using field data to assess model accuracy, Cohen's Kappa was measured at 0.40. On-the-ground fence modification or removal could be prioritized by identifying high fence densities in critical wildlife areas such as pronghorn migratory pathways or sage grouse lekking habitat. Such novel fence data can assist wildlife and land managers to assess effects of anthropogenic features to wildlife at various scales; which in turn may help conserve declining grassland species and overall ecological functionality.

带刺铁丝网与编织铁丝网是北美西部广泛分布的常见构筑物，会对野生动物的移动形成阻隔。具体而言，围栏会影响叉角羚（Antilocapra americana）的日常与季节性移动行为，同时改变其生境选择模式。鉴于围栏对叉角羚及其他野生动物的影响，其是野生动物移动与生境选择模型中潜在的关键影响因子。当前尚无区域尺度下的地理空间围栏数据可用。为此，我们基于一系列土地权属假设，构建了针对蒙大拿州北部高线（Hi-Line）区域的区域围栏模型——该区域涵盖13个县，总面积超过103400平方千米。我们沿铺装与非铺装道路布设了总计738条长度为3.2千米的随机样带（n=738），开展野外行车调查以收集生境、围栏密度与围栏类型相关数据。借助地理信息系统（GIS），我们整合土地权属、地块规模、相邻地块、乡镇边界与道路等要素，构建了围栏位置与围栏密度模型。结合当地土地权属与土地利用相关的本土认知，进一步优化了最终模型。我们预测，高线区域内的围栏总长度超过263300千米，最大围栏密度为6.8千米/平方千米，平均围栏密度为2.4千米/平方千米。通过野外实测数据评估模型精度时，科恩卡帕系数（Cohen's Kappa）得分为0.40。通过识别叉角羚迁徙通道、艾草松鸡求偶场等关键野生动物栖息地内的高围栏密度区域，可优先开展野外围栏改造或移除工作。这类新颖的围栏数据集可帮助野生动物与土地管理者在不同尺度下评估人为活动设施对野生动物的影响，进而助力衰退草原物种的保护与整体生态系统功能的维持。