%0 Dataset
%T Northern Hemisphere Seasonal Freeze Depth Dataset (1850-2100)
%J National Cryosphere Desert Data Center
%I National Cryosphere Desert Data Center(www.ncdc.ac.cn)
%U http://www.ncdc.ac.cn/portal/metadata/c4a959b9-b07d-47a5-9576-fa040ab0864f
%W NCDC
%R 10.12072/ncdc.nieer.db6394.2024
%A Peng Xiaoqing
%A CHEN Cong
%K Frozen soil;Northern Hemisphere;seasonal freezing depth
%X Based on 1220 long-term time series data of maximum freezing depth stations in the Northern Hemisphere as training data, the XGBoost algorithm is used to construct a soil freezing depth model for the Northern Hemisphere by integrating station data of maximum freezing depth with predictive factors such as temperature, precipitation, freezing index, melting index, snow depth, solar radiation, leaf area index, and soil texture. During the model training process, ten fold cross validation was used to run 300 times and output the optimal model. We simulated and predicted the maximum freezing depth of soil in 22 patterns (SSP126, SSP245, SSP370, SSP585) under different past and future scenarios in the Northern Hemisphere. The average RSME of the model set with 22 modes is 33.15 cm, MAE is 22.96 cm, and R2 is 0.81. The data format is netcdf, with a spatial resolution of approximately 0.5 ° and a temporal resolution of year by year.