Understanding how habitat, landscape context, and human disturbance influence local species-specific deer density provides evidence informing strategic management of increasing deer populations. Across an extensive (187 km²) heterogeneous forest-mosaic landscape in eastern England, spatially explicit density surface models of roe deer Capreolus capreolus and introduced muntjac Muntiacus reevesi were calibrated by thermal imaging distance sampling (recording 1590 and 400 muntjac and roe deer groups, respectively, on 567 km of driven transects). Models related deer density to local habitat composition, recreational intensity, and deer density (roe deer models controlled for muntjac density and vice versa) at a local grain across 1162 composite transect segments, incorporating geographical coordinates accounting for spatial autocorrelation. Abundance of both species was lower in localities with more grasslands (inter-quartile, IQ, effect size: roe −2.9 deer/km²; muntjac −2.9 deer/km²). Roe abundance (mean = 7 deer/km², SD = 6) was greater in localities with more young stands (IQ effect size, + 1.3 deer/km²) and lower at localities with more recreationists (−1.1 deer/km²). Muntjac density (mean = 21 deer/km², SD = 10) was greater in localities with more recreationists (+ 2.4 deer/km²), with more mature (≥ 46 years) stands (+ 1.5 deer/km²), or calcareous soil (+ 7.1 deer/km²). Comparison of models incorporating candidate variables and models comprising geographical coordinates only shows candidate variables to be weak predictors of deer densities. Adapting forest management to manipulate habitat and recreational access may influence local deer densities, but only subtly: effect sizes are not sufficient to mitigate deer impacts through planting vulnerable tree crops in areas avoided by deer. Effective culling remains the most viable management option.

了解栖息地、景观环境和人类干扰如何影响当地特定物种的鹿密度为增加鹿种群的战略管理提供了证据。在英格兰东部广阔的（187 公里²）异质森林马赛克景观中，狍狍 Capreolus capreolus和引入的麂Muntiacus reevesi 的空间显性密度表面模型通过热成像距离采样进行校准（分别在 567 公里的驱动横断面上记录了 1590 和 400 只麂和狍群）。模型将鹿密度与当地栖息地组成、娱乐强度和鹿密度（狍子模型控制为麂皮密度，反之亦然）在 1162 个复合横断面的当地谷物中相关联，并结合了地理坐标来解释空间自相关。在草原较多的地区，这两种物种的丰度较低（四分位间距，智商，效应大小：狍 -2.9 鹿/km ²；麂 -2.9 鹿/km ²）。在幼林较多的地区，獐子丰度（平均值 = 7 头鹿/km ²，SD = 6）更大（IQ 效应大小，+ 1.3 头鹿/km ²) 并且在休闲爱好者较多的地方更低（-1.1 头鹿/km ²）。麂密度（平均值 = 21 头鹿/km ²，SD = 10）在具有更多游憩者（+ 2.4 头鹿/km ²）、更成熟（≥ 46 年）林分（+ 1.5 头鹿/km ²）或石灰质土壤（+ 7.1 鹿/km ²）。包含候选变量的模型和包含地理坐标的模型的比较仅显示候选变量是鹿密度的弱预测因子。调整森林管理以操纵栖息地和娱乐活动可能会影响当地的鹿密度，但只是微妙的：通过在鹿避开的地区种植脆弱的树木作物，效果大小不足以减轻鹿的影响。有效的剔除仍然是最可行的管理选择。