Machine Learning Reveals Hidden Bias in ERA5 Cloud Heights Over Earth's Third Pole

作者：WEI ZHAO · 发表于：Figshare · 年份：2026 · DOI：10.6084/m9.figshare.32018046 · 被引用次数：1 · 研究领域：Meteorology、Remote sensing、Environmental science、Computer science、Geography

The two-step machine learning framework utilized three years of ground-based lidar observations (October 2021–December 2024), high-confidence CALIPSO satellite retrievals, and ERA5 reanalysis data to mitigate systematic biases in cloud base height (CBH) estimates over the Tibetan Plateau. The dataset includes ground-based lidar-derived CBH, CALIPSO-derived CBH, and the following ERA5 meteorological variables used for calculations: specific humidity (q), temperature (t), cloud ice water content (ciwc), cloud liquid water content (clwc), fraction of cloud cover (cc), surface attributes (including elevation and land cover types), as well as geographic coordinates (latitude and longitude). These variables enable readers to reproduce the ERA5 refinement process as well as the main figures and key results presented in the manuscript.