HMRFS-TP：青藏高原逐日无云积雪数据集(2002-2025)

基于长时间序列MODIS积雪产品，采用隐马尔可夫随机场（Hidden Markov Random Field, HMRF）建模框架，制备了青藏高原2002-2025年空间分辨率为500 m的逐日无云积雪数据集(HMRFS–TP)。该建模框架将MODIS积雪产品的光谱信息、时空背景信息，以及环境相关信息以最优形式进行整合，不仅填补了云层遮挡引起的数据空缺，而且提高了原始MODIS积雪产品的精度。特别地，本数据集在环境背景信息中引入了太阳辐射能量对积雪分布的影响，有效改进了地形复杂山区的积雪识别精度。通过与实测雪深、Landsat-8 OLI识别的积雪分布对比分析，本数据集精度依次为98.29%和91.36%，并且在积雪转化期、海拔较高、太阳辐射较多的阳坡提升效果显著。本数据集改善了原始MODIS积雪产品时空不连续和在地形复杂山区精度较低的问题，能为青藏高原气候变化研究和水资源管理提供重要的数据基础。

Based on long-time-series MODIS snow cover products and using the Hidden Markov Random Field (HMRF) modeling framework, a daily cloud-free snow cover dataset (HMRFS–TP) with a spatial resolution of 500 m for the Qinghai-Tibet Plateau during 2002–2024 was developed. This modeling framework optimally integrates the spectral information, spatiotemporal background information, and environment-related information of MODIS snow cover products, which not only fills the data gaps caused by cloud occlusion but also improves the accuracy of the original MODIS snow cover products. Specifically, this dataset incorporates the impact of solar radiation energy on snow cover distribution into the environmental background information, effectively improving the snow cover recognition accuracy in mountainous areas with complex terrain. Through comparative analysis with in-situ snow depth measurements and snow cover distributions identified by Landsat-8 OLI, the accuracy of this dataset reached 98.29% and 91.36% respectively, with significant performance improvements during snow cover transition periods, at high-altitude sunny slopes with high solar radiation. This dataset addresses the issues of spatiotemporal discontinuity and low accuracy in complex-terrain mountainous areas of the original MODIS snow cover products, providing an important data foundation for climate change research and water resource management on the Qinghai-Tibet Plateau.

搜集汇总

背景与挑战

背景概述

该数据集是基于MODIS积雪产品，采用隐马尔可夫随机场（HMRF）建模框架生成的青藏高原2002年至2024年逐日无云积雪数据，空间分辨率为500米。它通过整合光谱、时空和环境信息，有效填补了云层遮挡的数据空缺，并提高了积雪识别精度，特别是在地形复杂山区，精度分别达到98.29%（与实测雪深对比）和91.36%（与Landsat-8 OLI对比），为青藏高原气候变化研究和水资源管理提供了重要数据基础。