Datensatz

Zusammenfassung

The creation of synthetic images of precipitation, temperature, evapotranspiration, and terrestrial water storage is proposed to address the gaps in satellite data availability prior to the year 2000 and extend the data to pre-satellite periods. This is necessary to model and manage water resources and evaluate the impact of climate change on hydrological processes in regions with limited data. The synthetic images should closely resemble real satellite images. The approach is based on the relationship between meteorological factors and existing satellite images and the idea that, under similar weather conditions, patterns of specific processes may repeat over time. The ERA5 reanalysis data is used as the meteorological predictor, and a K-Nearest Neighbor algorithm with a process-specific similarity metric is applied to generate the synthetic images. The method is tested in the Volta River Basin in West Africa where water resources are critically impacted by climate change. The synthetic images are input into a spatially-distributed hydrological model for calibration and validation, and their quality is assessed by their ability to reproduce historical streamflow time series. The goal of this testing phase is to improve the generation technique and produce synthetic images that closely approximate unobserved processes and improve the accuracy of the modeling.