全球FY-3/MWRI陆表温度日产品数据集（2011~2024）

地表温度（Land Surface Temperature，LST）是评估水循环、生态过程、气候演变及陆气界面能量交换的关键气象参数，对气象卫星实现精准高频次监测至关重要。然而现有反演算法常难以捕捉昼夜温度变化的细微特征。本研究利用中国风云三号（FY-3）卫星搭载的微波辐射计（Microwave Radiation Imagers，MWRI）独特的全天候观测优势，突破传统观测局限，显著提升了全时段地表温度测量能力。研究旨在开发一种能有效识别近地表冻结与解冻状态的全球通用算法，通过融合多通道亮温数据和多种微波指数，增强不同地表覆盖类型下的测量精度。基于地面实测数据的验证及与ERA5再分析数据、MODIS地表温度产品的对比分析表明，该算法具有高度稳健性。结果显示，除风云三号D星（FY-3D）在14:00的观测时段外，FY-3 MWRI反演地表温度与5厘米土壤温度的相关系数超过0.87，均方根误差（RMSE）约4 K。通过FY-3 MWRI、ERA5和MODIS三套数据集的三重搭配分析估算，全球大部分地区理论不确定度低于4 K。相较于MODIS地表温度产品，FY-3 MWRI呈现的温度日变化幅度更小，其记录的峰值温度相对MODIS存在时间滞后现象，且昼夜温差范围普遍收窄，充分展现了FY-3B/C/D卫星组网观测在刻画昼夜温度循环特征方面的独特优势。

Land Surface Temperature (LST) is a critical meteorological parameter for evaluating water cycles, ecological processes, climate evolution, and energy exchange at the land-atmosphere interface, and it is crucial for accurate and high-frequency monitoring by meteorological satellites. However, existing inversion algorithms often struggle to capture the subtle characteristics of diurnal temperature variations. This study leverages the unique all-weather observation advantages of the Microwave Radiation Imagers (MWRI) onboard China's Fengyun-3 (FY-3) satellites, breaking through the limitations of traditional observations and significantly improving the capability of all-time land surface temperature measurement. The research aims to develop a global general-purpose algorithm that can effectively identify near-surface freezing and thawing states, by fusing multi-channel brightness temperature data and multiple microwave indices to enhance measurement accuracy across different land cover types. Validation based on in-situ ground measurements and comparative analyses with ERA5 reanalysis data and MODIS land surface temperature products demonstrate that this algorithm is highly robust. The results show that, except for the observation period of FY-3D at 14:00, the correlation coefficient between the land surface temperature retrieved by FY-3 MWRI and 5-cm soil temperature exceeds 0.87, with a root mean square error (RMSE) of approximately 4 K. By conducting triple collocation analysis using the three datasets of FY-3 MWRI, ERA5, and MODIS, the theoretical uncertainty in most global regions is estimated to be below 4 K. Compared to the MODIS land surface temperature product, the FY-3 MWRI exhibits a smaller diurnal temperature variation amplitude, the peak temperatures it records have a temporal lag relative to MODIS, and the range of diurnal temperature differences is generally narrowed, fully demonstrating the unique advantages of the FY-3B/C/D satellite network observation in characterizing diurnal temperature cycle characteristics.

搜集汇总

背景与挑战

背景概述

该数据集为全球FY-3/MWRI陆表温度日产品数据集（2011~2024），时间分辨率为日，空间分辨率为1km - 10km，数据大小为5.93 GB，覆盖2014-01-01至2019-12-31。数据集利用FY-3卫星的MWRI仪器，通过多通道亮温数据和微波指数融合，提升了全时段地表温度测量能力，与MODIS产品相比，能更准确地刻画昼夜温度循环特征。