Dynamic water level changes in Qinghai Lake from integrating refined ICESat-2 and GEDI altimetry data (2018-2021)

The water levels of inland lakes on the Tibetan Plateau are extremely sensitive to global climate change and can objectively reflect the temporal and spatial changes in local water resources. However, monthly and seasonal variations in lake water levels are difficult to monitor due to the lack of sufficient in situ gauges across moun-tainous areas. Moreover, GEDI products exhibit large uncertainties in mountainous surroundings. In this paper, taking Qinghai Lake as an example, we first refined the raw data of the GEDI and ICESat-2 missions by imple-menting a quality control procedure involving outlier removal tailored to the characteristics of each mission; then, we analyzed the accuracy of each mission, especially targeting factors that affect the water level retrievals in the GEDI products. Third, the bias between the two missions was adjusted by selecting the overlapping or adjacent observation dates. Finally, we constructed dense temporal water-level data by integrating the refined ICESat-2 and GEDI data. Data from water level stations and the DAHITI and Hydroweb datasets were also utilized for validation. The results show that (1) very accurate results can be obtained from the ICESat-2 ATL13 product, and the standard deviations of most observed days are under 0.05 m; (2) the GEDI products derived from al-gorithm 2 can offer more effective footprints than those from algorithm 1, with an improvement of approxi-mately 9.78 %. Moreover, large differences exist among the different GEDI beams, and beams 1 and 2 are recommended for further analysis. Overall, most beams overestimated the lake levels with a bias of 0.264 +/- 0.357 m; (3) the long-time-series water levels showed a mean increasing trend of 0.243 m/yr from 2018 to 2021. The relatively high-water-level periods were distributed mostly in August and September, while the low-water-level periods were distributed mostly in February and March. The combined water levels were very correlated with the DAHITI and Hydroweb datasets, with R values larger than 0.8, and highly consistent with the obser-vations from hydrological stations (the inter-year change range spanned from 0.015 m to 0.327 m, and the intra-year difference range varied from 0.03 m to 0.16 m); and (4) Integrating the GEDI and ICESat-2 missions allowed us to capture the monthly, seasonal and annual dynamics of the lake water level, and the results indicate that the combined dataset presents a valuable resource for hydrological and climatic change studies.

Automated summary

This study optimized the data quality control procedure and integrated GEDI and ICESat-2 data to construct a temporal water-level dataset. The results revealed significant seasonal and inter-annual variations in lake water levels, which showed high consistency with other datasets. The combined dataset provides a valuable resource for hydrological and climatic change studies.