Future climate rasters, land cover predictors and sampling effort grids for invasive species distribution modelling in Europe using the wiSDM v.2 workflow

This dataset contains a set of future climate rasters derived from the CHELSA v.2.1 bioclim dataset at a spatial resolution of 1 × 1km (Karger et al., 2021; Brun et al., 2022). Climate rasters are available for two future periods (2041-2070 and 2071-2100), and for each period, projections are provided for three Shared Socioeconomic Pathways (SSP1-2.6, SSP3-7.0, and SSP5-8.5). Each layer represents the mean values across five different global circulation models (GFDL-ESM4, IPSL-CM6A-LR, MPI-ESM1-2-HR, MRI-ESM2-0, and UKESM1-0-LL), which were standardized relative to the corresponding present-day climate layer by subtracting the present-day mean and dividing by the present-day standard deviation. In addition to future climate layers, this dataset contains a set of land cover predictors. Land cover information was extracted from the Corine Land Cover 2018 raster (version U2018_CLC2018_V2020_20u1), a 100m resolution dataset in which each cell represents a specific land-cover category. These cells were aggregated into 1 × 1 km grid cells to quantify the percentage cover of broader land-use classes (see Table 1 for an overview). Grid cells containing more than 5% NA values or more than 50% sea or ocean cells were assigned NA, while cells with a value of zero indicate the absence of a given land-cover class. Next to these Corine land cover predictors, three freshwater-related predictors were derived from the EU-Hydro dataset (Copernicus): log_total_water_length.tif: Captures the total length of linear water features (e.g., rivers, canals, and drainage ditches) within each grid cell. This quantifies narrow but widespread water elements that polygon mapping often misses, particularly in agricultural landscapes or riparian networks. The output is provided as a log-transformed raster layer. proportion_total_water_polygon_cover.tif: Describes how much of each grid cell’s surface is covered by broader water bodies such as lakes, ponds, reservoirs or wide river sections. Values range from 0 to 1, where 0.25 indicates that 25% of the pixel is covered by water. log_distance_to_water.tif: Represents the distance to the nearest pixel that contains water. Areas that directly touch a riverbank or lake shoreline have values near zero, while more distant, upland areas have larger values. It expresses the physical access to water across the landscape and is presented in a log-transformed format. To mitigate the effects of uneven sampling effort in species distribution modelling, we provide taxonomic sampling effort grids that quantify large-scale spatial variation in sampling intensity. These grids can be used to guide the selection of background points or pseudo-absences, for example by reducing or excluding areas with low sampling effort (Phillips et al., 2009; Elith et al., 2010; Barbet-Massin et al., 2012). To generate the sampling effort grids, we downloaded all available GBIF occurrences for each taxonomic group from 1974 to 2024. Records were filtered to retain only those with valid geographic coordinates and no reported geospatial issues. Sampling grids were then constructed by using CHELSA climate grids as the basis for a global 1° spatial reference raster in WGS84. We mapped each occurrence record to a 1° raster cell and counted the total number of records within each cell. The resulting values represent taxon-specific sampling effort, with higher values indicating more intensely surveyed regions. To improve interpretability and reduce the influence of highly surveyed regions, values are provided in a log-transformed format using log(1 + N), where N is the number of occurrences in that cell. These grids are used as input data for the wiSDM v.2 species distribution modelling workflow of invasive alien species designed for the Tracking Invasive Alien Species (TrIAS) project. Table 1. Overview of Corine land cover predictors and the classes they comprise. Land use predictor Corine label Agriculture Non-irrigated arable land; Permanently irrigated land; Rice fields; Vineyards; Fruit trees and berry plantations; Olive groves; Pastures; Annual crops associated with permanent crops; Complex cultivation patterns; Land principally occupied by agriculture, with significant areas of natural vegetation; Agro-forestry areas Artificial Continuous urban fabric; Discontinuous urban fabric; Industrial or commercial units; Road and rail networks and associated land; Port areas; Airports; Mineral extraction sites; Dump sites; Construction sites; Green urban areas; Sport and leisure facilities Coastal_wetland Salt marshes; Salines; Intertidal flats Coniferous_forest Coniferous forest Deciduous_forest Broad-leaved forest Inland_wetland Inland marshes; peat bogs Mixed_forest Mixed forest Shrub_and_herbaceous Natural grasslands; Moors and heathland; Sclerophyllous vegetation; Transitional woodland-shrub 1. Future climate layers Layers are named as follows: scaled_layer_CHELSA_abbreviation_period_scenario.tif, with period being either 2041-2070 or 2071-2100, and scenario being one of the following: ssp126, ssp370, and ssp585. Variable (abbreviation). Mean annual air temperature (meantemp_1) Temperature seasonality (temp_seasonality_4) Mean daily maximum air temperature of the warmest month (maxTmpWarmestMon_5) Mean daily minimum air temperature of the coldest month (minTmpColdestMon_6) Annual range of air temperature (temp_annRange_7) Annual precipitation amount (annPrecip_12) Precipitation amount of the wettest month (precipWettestMon_13) Precipitation amount of the driest month (precipDriestMon_14) Precipitation seasonality (precip_Seasonality_15) 2. Land cover predictors Agriculture.tif Artificial.tif Coastal_wetland.tif Coniferous_forest.tif Deciduous_forest.tif Inland_wetland.tif Mixed_forest.tif Shrub_and_herbaceous.tif log_distance_to_water.tif log_total_water_length.tif proportion_total_water_polygon_cover.tif 3. Taxonomic sampling effort grids log_amphibians_1degree_layer.tif log_birds_1degree_layer.tif log_fish_1degree_layer.tif log_hydrozoa_1degree_layer.tif log_insects_1degree_layer.tif log_malacostraca_1degree_layer.tif log_mammals_1degree_layer.tif log_mollusca_1degree_layer.tif log_plants_1degree_layer.tif log_reptiles_1degree_layer.tif本数据集包含一套基于CHELSA v.2.1生物气候数据集（Karger等，2021；Brun等，2022）生成的未来气候栅格，空间分辨率为1×1km。该气候栅格覆盖两个未来时段（2041-2070年与2071-2100年），每个时段均提供三种共享社会经济路径（Shared Socioeconomic Pathways, SSP1-2.6、SSP3-7.0及SSP5-8.5）的预测结果。每个栅格图层代表5种全球环流模型（Global Circulation Models, GFDL-ESM4、IPSL-CM6A-LR、MPI-ESM1-2-HR、MRI-ESM2-0及UKESM1-0-LL）的平均值，且通过减去当前气候均值并除以当前气候标准差的方式，相对于对应当代气候图层完成标准化。 除未来气候图层外，本数据集还包含一套土地覆盖预测变量。土地覆盖信息提取自Corine Land Cover（Corine土地覆盖）2018栅格（版本号U2018_CLC2018_V2020_20u1），该数据集分辨率为100m，每个栅格单元代表一种特定土地覆盖类别。我们将其聚合至1×1km网格单元，以量化更宽泛土地利用类别的覆盖百分比（详见表1概述）。若栅格单元的NA值占比超过5%或海洋/海洋类单元格占比超过50%，则将该单元赋值为NA；若单元值为0，则代表无对应土地覆盖类别。除上述Corine土地覆盖预测变量外，还从EU-Hydro数据集（Copernicus）衍生得到3种与淡水相关的预测变量： 1. log_total_water_length.tif：表征每个栅格单元内线性水体（如河流、运河及排水沟）的总长度。该变量可量化多边形制图常遗漏的狭窄但分布广泛的水体要素，尤其在农业景观或河岸网络中。输出结果为经过对数转换的栅格图层。 2. proportion_total_water_polygon_cover.tif：描述每个栅格单元的地表被大型水体（如湖泊、池塘、水库或宽阔河段）覆盖的比例，取值范围为0至1，例如0.25代表该像素25%的区域被水体覆盖。 3. log_distance_to_water.tif：代表至最近含水栅格单元的距离。直接毗邻河岸或湖岸的区域取值接近0，而距离更远的高地区域取值更大。该变量表征景观中水体的物理可达性，同样以对数转换格式呈现。 为缓解物种分布建模（Species Distribution Modelling）中采样强度不均的影响，本数据集提供分类群采样强度栅格，用于量化采样强度的大尺度空间变异。此类栅格可用于指导背景点或伪缺失点的选取，例如通过降低或排除低采样强度区域的权重（Phillips等，2009；Elith等，2010；Barbet-Massin等，2012）。 为生成采样强度栅格，我们下载了1974年至2024年间各分类群在GBIF（Global Biodiversity Information Facility）平台上的所有可用出现记录。对记录进行过滤，仅保留具有有效地理坐标且无空间异常的记录。随后以CHELSA气候栅格为基础，构建WGS84坐标系下的全球1°空间参考栅格，将每条出现记录映射至对应的1°栅格单元，并统计每个单元内的记录总数。最终取值代表类群特异性采样强度，取值越高代表该区域调查越密集。为提升可解释性并降低高采样强度区域的影响，所有取值均采用log(1+N)的对数转换格式，其中N为该单元内的物种出现记录数。 本数据集的栅格可作为wiSDM v.2外来入侵物种分布建模工作流的输入数据，该工作流专为追踪外来入侵物种（Tracking Invasive Alien Species, TrIAS）项目开发。 表1 Corine土地覆盖预测变量及其包含的类别 | 土地利用预测变量 | Corine标签 | | ---- | ---- | | 农业（Agriculture） | 非灌溉耕地、永久灌溉耕地、稻田、葡萄园、果树与浆果种植园、橄榄园、牧场、与永久作物伴生的一年生作物、复合种植模式、以农业为主且带有大面积自然植被的土地、农林业区域 | | 人工地表（Artificial） | 连续城市建成区、非连续城市建成区、工业或商业用地、道路与铁路网络及附属用地、港口区域、机场、采矿场、垃圾堆放场、建筑工地、城市绿地、体育与休闲设施 | | 滨海湿地（Coastal_wetland） | 盐沼、盐田、潮间带滩涂 | | 针叶林（Coniferous_forest） | 针叶林 | | 阔叶林（Deciduous_forest） | 阔叶林 | | 内陆湿地（Inland_wetland） | 内陆沼泽、泥炭地 | | 混交林（Mixed_forest） | 混交林 | | 灌丛与草本植被（Shrub_and_herbaceous） | 天然草原、沼泽与灌丛、硬叶植被、过渡性林灌地带 | 1. 未来气候图层 图层命名格式为：scaled_layer_CHELSA_缩写_时段_情景.tif，其中时段可选2041-2070或2071-2100，情景可选以下三种：ssp126、ssp370及ssp585。 变量（缩写）如下： - 年平均气温（meantemp_1） - 温度季节变化（temp_seasonality_4） - 最热月平均日最高气温（maxTmpWarmestMon_5） - 最冷月平均日最低气温（minTmpColdestMon_6） - 气温年较差（temp_annRange_7） - 年降水量（annPrecip_12） - 最湿月降水量（precipWettestMon_13） - 最干月降水量（precipDriestMon_14） - 降水季节变化（precip_Seasonality_15） 2. 土地覆盖预测变量 Agriculture.tif Artificial.tif Coastal_wetland.tif Coniferous_forest.tif Deciduous_forest.tif Inland_wetland.tif Mixed_forest.tif Shrub_and_herbaceous.tif log_distance_to_water.tif log_total_water_length.tif proportion_total_water_polygon_cover.tif 3. 分类群采样强度栅格 log_amphibians_1degree_layer.tif log_birds_1degree_layer.tif log_fish_1degree_layer.tif log_hydrozoa_1degree_layer.tif log_insects_1degree_layer.tif log_malacostraca_1degree_layer.tif log_mammals_1degree_layer.tif log_mollusca_1degree_layer.tif log_plants_1degree_layer.tif log_reptiles_1degree_layer.tif