PALEO-PGEM-Series: a spatial time series of the global climate over the last 5 million years (Plio-Pleistocene)

PALEO-PGEM-Series is a global spatiotemporal dataset of the climate over the last 5 million years, derived from the intermediate complexity atmosphere-ocean general circulation model PLASIM-GENIE using emulation and downscaling techniques (Holden et al. 2019). It consists of estimates of monthly temperature (in degrees Celsius), monthly precipitation (in millimeters), and 17 bioclimatic variables over the last 5 million years every 1,000 years at 1º x 1º spatial resolution.

We provide the mean and the standard deviation of the estimates across the 10 stochastic GCM emulation runs. We suggest that users take advantage of these uncertainty estimates by conducting sensitivity tests to assess how robust results derived using PALEO-PGEM-Series are in the face of emulation uncertainty.

Each .txt file contains 19,233 rows representing geographic grided cells and 5003 columns, the first two with coordinates (longitude and latitude) and the others with the estimates for each time slice, starting from 5,000,000 years and ending with the pre-industrial. Each column is named with “T” followed by a number representing the time in years divided by 1,000 (i.e., T5000 corresponds to time 5,000,000 years ago). These tables can easily be converted into spatial files (e.g., raster) and visualized and explored using GIS libraries. We also provide the data as netcdf raster files, which are included in the pastclim R package (https://doi.org/10.1111/ecog.06481).

We provide an R script to derive the 17 bioclimatic variables from the monthly estimates (code adapted from the function biovars in R package dismo, Hijmans et al. 2021). These bioclimatic variables summarize central tendencies (e.g., annual means), seasonality (e.g., range of monthly means) and extreme conditions (e.g., maximum and minimum monthly means) of relevance for ecological and evolutionary studies. Mean diurnal range (BIO 2) and isothermality (BIO 3) cannot be derived from PALEO-PGEM-Series as they require diurnal cycles. Maximum (BIO 5) and minimum temperature (BIO 6) are modified versions of the original bioclimatic variables, as they are derived from monthly means and thus refer to the maximum and minimum monthly means of temperatures in the year. Consequently, the temperature annual range (BIO 7) is also with reference to monthly means.

PALEO-PGEM-Series是一套覆盖过去500万年的全球时空气候数据集，其基于中等复杂度大气-海洋通用环流模型PLASIM-GENIE，通过仿真（emulation）与降尺度（downscaling）技术生成（Holden等，2019）。该数据集包含过去500万年中以每1000年为时间步长的逐月气温（单位：摄氏度）、逐月降水量（单位：毫米），以及17个生物气候变量的估算值，空间分辨率为1°×1°。

我们提供了10次随机通用环流模式（General Circulation Model, GCM）仿真运行结果的估算值均值与标准差。我们建议用户利用这些不确定性估算值开展敏感性测试，以评估在面临仿真不确定性时，基于PALEO-PGEM-Series得到的研究结果的稳健性。

每个.txt文件包含19233行（代表地理网格单元格）与5003列，前两列为坐标（经度与纬度），其余列对应各时间切片的估算值，时间跨度从500万年前直至工业革命前。每一列的命名格式为“T”后接以千为单位的年份数值（例如，T5000对应500万年前）。此类表格可便捷转换为空间文件（例如栅格（raster）文件），并可通过地理信息系统（Geographic Information System, GIS）库进行可视化与探索。我们同时提供netCDF（网络通用数据格式）栅格格式的数据，该数据已被收录于pastclim R语言包（https://doi.org/10.1111/ecog.06481）。

我们提供了一段R语言脚本，可基于逐月估算值推导得到17个生物气候变量（代码改编自R语言包dismo中的biovars函数，Hijmans等，2021）。这些生物气候变量可概括与生态学及进化生物学研究相关的核心趋势（例如年平均气温）、季节变化特征（例如月均温极差）与极端气候条件（例如月均温的最大值与最小值）。但日均温差（BIO 2）与等温性（BIO 3）无法由PALEO-PGEM-Series推导得到，因为二者需要昼夜循环数据。最高气温（BIO 5）与最低气温（BIO 6）为原始生物气候变量的修正版本，因其基于月均温推导得到，故实际指代的是年内月均温的最大值与最小值。因此，气温年较差（BIO 7）同样基于月均温计算。