2010-2015年中国1km年平均气温数据

本数据集各要素通过建立NCEP/NCAR中国区域的再分析资料数据与植被指数、地形之间的回归树模型，降尺度得到。计算方法如下：基于气候要素和NDVI、地形之间的密切关系，首先建立50km尺度上的NCEP/NCAR平均气温和NDVI、DEM、地理位置之间的回归树模型，再将建立的模型应用到1km尺度的DEM和NDVI，得到1km尺度的中国平均气温。具体的降尺度算法描述如下：（1）将分辨率为1km×1km的DEM和NDVI数据合成到分辨率为50km，合成方法为求取每一个50km栅格内的所有1km像元值的算术平均值；（2）建立50km尺度上的NCEP/NCAR平均气温与DEM、NDVI以及地理位置之间的回归树模型；（3）将（2）中建立的模型应用于高分辨率的DEM、NDVI数据，得到1km×1km分辨率的平均气温数据；（4）第（3）步中得到的合成到50km尺度，合成方法同（1），求得与NCEP/NCAR平均气温的差值，得到模型计算的残差；（5）利用样条插值方法将分辨率为50km的残差插值到1km×1km分辨率，得到1km尺度的残差值；（6）将（5）中得到的1km尺度的残差值与（3）中得到的1km尺度平均气温相加，得到最终的1km分辨率的中国年平均气温降尺度预测结果。

Each element of this dataset is obtained through downscaling based on a regression tree model constructed between the NCEP/NCAR reanalysis data over China and vegetation indices (NDVI), terrain data (DEM) as well as geographic location. The detailed calculation procedure is as follows: Based on the close correlation between climatic elements, NDVI and terrain, a regression tree model is first established between the 50 km-scale NCEP/NCAR mean air temperature, NDVI, DEM and geographic location. Then the established model is applied to 1 km-scale DEM and NDVI data to obtain 1 km-scale mean air temperature over China. The specific downscaling algorithm is detailed below:
(1) Resample the 1 km × 1 km resolution DEM and NDVI data to 50 km resolution by calculating the arithmetic mean of all 1 km pixels within each 50 km grid cell;
(2) Establish a regression tree model between the 50 km-scale NCEP/NCAR mean air temperature, DEM, NDVI and geographic location;
(3) Apply the model established in step (2) to the high-resolution DEM and NDVI data to obtain 1 km × 1 km resolution mean air temperature data;
(4) Resample the 1 km-scale mean air temperature data obtained in step (3) to 50 km resolution using the same method as step (1), and calculate the difference between the resampled data and the original NCEP/NCAR mean air temperature to obtain the model calculation residual;
(5) Interpolate the 50 km-resolution residual data to 1 km × 1 km resolution via spline interpolation to obtain 1 km-scale residual values;
(6) Add the 1 km-scale residual values obtained in step (5) to the 1 km-scale mean air temperature data obtained in step (3) to obtain the final 1 km-resolution downscaling prediction result of China's annual mean air temperature.

搜集汇总

背景与挑战

背景概述

该数据集提供了2010-2015年中国1km空间分辨率的年平均气温数据，基于NCEP/NCAR再分析资料，通过回归树模型结合NDVI、地形和地理位置进行降尺度处理生成。数据以GeoTiff栅格格式存储，共包含6个文件，总容量为468.09 MB，适用于高精度气候分析和研究。