Characterizing basin-scale precipitation gradients in the Third Pole region using a high-resolution atmospheric simulation-based dataset

Altitudinal precipitation gradient plays an important role in the interpolation of precipitation in the Third Pole (TP) region, where the topography is very complex but in situ data are very sparse. This study proves that the altitude dependence of precipitation in the TP can be reasonably reproduced by a high-resolution atmospheric simulation-based dataset called ERA5_CNN. The precipitation gradients, including both absolute (APGs) and relative gradients (RPGs), for 388 sub-basins of the TP above 2500 m a.s.l. are calculated based on the ERA5_CNN. Results show that most sub-basins have positive precipitation gradients, and negative gradients are mainly found along the Himalayas, the Hengduan Mountains and the western Kunlun. The annual APG and RPG averaged across all sub-basins of the TP are 0.05 mm d(-1) x 100 m(-1) and 4.25 % x 100 m(-1), respectively. The values of the APG are large in wet seasons but small in dry seasons, while the RPG shows opposite variations. Further analyses demonstrate that the RPGs have negative correlations with relative humidity but positive correlations with wind speed, likely because dry air tends to reach saturation at high altitudes, while stronger wind can bring more humid air to high altitudes. In addition, we find that precipitation gradients tend to be positive at small spatial scales compared to those at large scales, mainly because local topography plays a vital role in determining precipitation distribution at small scales. These findings on the spatiotemporal variations of precipitation gradients provide useful information for interpolating precipitation in the TP region.