Integrated range maps for North American butterflies derived from expert opinion and relative predicted habitat suitability

Integrated range maps were created by combining expert opinion, in the form of polygons from expert field guides (Glassberg 2017, 2018), with relative predicted habitat suitability from MaxEnt models using data from the Global Biodiversity Information Facility (GBIF) records. For species with limited point occurrence records, we used pseudo-presence points within the expert range. With these resulting integrated range maps, we calculated an estimated range for each species by clipping their range rasters to only cells with predicted suitability >0.20 and converting the remaining cells to a smoothed polygon. We then generated a 'butterfly environment index' which is the mean relative predicted habitat suitability for all species whose ranges encompass a site (i.e. it is the mean suitability of a site for species that may be expected to occur there). 
The integrated range maps were generated to support projects investigating macroscale patterns in butterfly biodiversity in North America, or for species-specific projects that require spatial information for focal species (e.g. where they are likely to occur, estimated extent of occurrence, etc.). The data could also be used for prioritizing regions that may be home to a high diversity of species, or are good habitat for the species that occur within a region. The integrated ranges should be treated as estimates of species distributions, rather than absolute cutoffs in range. The raster values should not be interpreted as probability of occurrence, but rather as relative predicted suitability which also incorporates distance from the expert range to reduce the influence of potentially spurious observations.
<br>
<strong>File list</strong>
<em>README.xml</em> - metadata for all species integrated range maps
<em>integrated-ranges.zip</em> - a compressed folder containing all 750 integrated range maps, including those generated with pseudo-presence points, as rasters in .tiff format
<em>species_richness.tif</em> - a raster of estimated species richness generated by overlaying polygons created at the 0.20 threshold as described in the associated manuscript
<em>continental-BEI.tiff</em> - a raster of estimated relative predicted habitat suitability where values are inclusive of all 750 species (i.e. including species that are not expected to occur in a region) as described in the associated manuscript
<em>local-BEI.tiff </em>- a raster of estimated relative predicted habitat suitability for species that are expected to occur locally (i.e. only including species that <em>are</em> expected to occur in a region) as described in the associated manuscript

本整合分布范围图的构建方法为：结合专家野外指南中的多边形数据（Glassberg 2017、2018）所代表的专家意见，与基于全球生物多样性信息设施（Global Biodiversity Information Facility, GBIF）记录训练得到的最大熵模型（MaxEnt）相对预测生境适宜性结果。针对点位出现记录有限的物种，我们采用其专家分布范围内的伪存在点（pseudo-presence points）进行补充。基于生成的整合分布范围图，我们通过以下步骤计算各物种的估计分布范围：将物种分布栅格仅裁剪至预测适宜性大于0.20的像元，再将剩余像元转换为平滑多边形。随后我们构建了蝴蝶生境指数（butterfly environment index），该指数为所有分布范围覆盖某一地点的物种的相对预测生境适宜性均值，即该地点对于预期在此出现的物种的适宜性均值。

本整合分布范围图的构建目的为支持北美蝴蝶生物多样性宏观格局研究，或需要目标物种空间分布信息的物种特异性研究（例如物种潜在分布区域、估计的分布范围等）。该数据集亦可用于优先划定高物种多样性区域，或识别区域内物种的优质生境。需注意，本整合分布范围仅为物种分布的估计值，而非绝对的分布边界。栅格数值不应被解释为物种出现概率，而是相对预测生境适宜性；该指标同时纳入了与专家分布范围的距离信息，以降低潜在虚假观测的影响。

### 文件清单
- *README.xml*：所有物种整合分布范围图的元数据文件
- *integrated-ranges.zip*：压缩文件夹，包含全部750份整合分布范围图（含通过伪存在点生成的范围图），以.tiff格式的栅格文件存储
- *species_richness.tif*：基于0.20阈值生成的多边形叠加得到的估计物种丰富度栅格文件，详见相关研究论文
- *continental-BEI.tiff*：包含全部750个物种的相对预测生境适宜性估计栅格（即纳入所有物种，包括预期不会在某区域出现的物种），详见相关研究论文
- *local-BEI.tiff*：仅针对预期在当地出现的物种的相对预测生境适宜性估计栅格（即仅纳入该区域预期存在的物种），详见相关研究论文