Too much of a good thing? Supplementing current species observations with fossil data to assess climate change vulnerability via ecological niche models

Ecological niche models (ENMs) are a powerful tool in ecological research and conservation planning. Since ENMs provide probability maps of suitable areas under environmental change, they may assist in designing conservation actions and addressing conservation priorities. However, ENMs are usually implemented by learning the species climatic preferences from their current geographic distribution, which leaves them vulnerable to the issue of niche truncation issues, as if comes with non-climatic limits to the current species distribution posed by e.g. anthropic activities and settlements, and is bound to assume that species are at equilibrium with their environments. These problems might be alleviated by the inclusion of fossil occurrences, which refer to moments during species evolution when such limits were absent, and a larger fraction of the species fundamental niche was probably explored. Here, we combined current and fossil occurrence data for 38 medium-large mammal species of cons..., , , # Data from: Too much of a good thing? Supplementing current species observations with fossil data to assess climate change vulnerability via ecological niche models

## Description of the data and file structure

#### Data folder contains:

1. **Geopackages**
* Geopackage_BEYER_CHELSA: In this folder are the geopackages for the 28 species achieving AUC > 0.7 in both modern and full ENMs. In each geopackage are coordinates and filtered climatic variable values (once VIF < 5 was applied) for occurrence and background points. The geopackages are loaded in all the scripts.
2. **Variables**
* CHELSA_10km: These are the variables used for present and 2080 climates. They are loaded in following scripts:
* 1_Modern_ENM_training
* 1b_Modern_ENMs_temporal_block
* 2_Full_ENM_training
* 5_RANDOM_FOREST
* Krapp_change: These are the variables used for past climates. They are loaded in following scripts:
* 3_NICHE_OVERLAP
...

生态位模型（Ecological Niche Models, ENMs）是生态学研究与保护规划中的有力工具。其可输出环境变化下的适宜分布区概率图谱，能够辅助制定保护行动方案、明确保护优先级。然而，现有生态位模型通常基于物种当前的地理分布学习其气候偏好，这使其极易受生态位截断问题影响：若当前物种分布受到人类活动、定居点等非气候因子限制，模型会默认物种已与其所处环境达到平衡状态。通过纳入化石出现记录可缓解上述问题——化石记录代表了物种演化进程中不存在此类非气候限制的时期，此时物种大概率已探索了其基础生态位的更大范围。本研究整合了38种中大型哺乳动物的当前分布与化石出现数据，数据来源为《过犹不及？通过生态位模型结合当前物种观测与化石数据评估气候变化脆弱性》。

## 数据与文件结构说明
数据文件夹包含以下内容：
1. **地理数据包（Geopackage）**
* Geopackage_BEYER_CHELSA：该文件夹收纳了28个物种的地理数据包，这些物种在现代与全数据集生态位模型中均取得了受试者工作特征曲线下面积（Area Under Curve, AUC）>0.7的表现。每个地理数据包中包含经方差膨胀因子（Variance Inflation Factor, VIF）<5阈值筛选后的出现点与背景点的坐标及气候变量值，所有分析脚本均需加载此类地理数据包。
2. **环境变量**
* CHELSA_10km：该数据集为当前及2080年气候情景下所用的环境变量，需在以下脚本中加载：
* 1_Modern_ENM_training：1_现代生态位模型训练
* 1b_Modern_ENMs_temporal_block：1b_现代生态位模型时间分区训练
* 2_Full_ENM_training：2_全数据集生态位模型训练
* 5_RANDOM_FOREST：5_随机森林模型训练
* Krapp_change：该数据集为古气候情景下所用的环境变量，需在以下脚本中加载：
* 3_NICHE_OVERLAP：3_生态位重叠度计算
……