Monthly gridded precipitation databases performance evaluation in North Patagonia, Argentina

Precipitation records in North Patagonia (Argentina) are scarce, which hinders climate research. This research aims to assess the performance of four novel local monthly datasets together with three commonly used global datasets for North Patagonia to evaluate the advantages and disadvantages of each one. First, four different local observed-interpolated datasets were built, two using the Angular Distance Weighted (ADW) method and two using ordinary Kriging. In addition, the global datasets CRU, ERA5-Land, and GPCC were evaluated. To assess the performance of the precipitation datasets, four metrics were used to evaluate the systematic errors (bias), the mean errors, the representation of time variations, and the representation of the probability density function. The ERA5-Land with a correction factor stands out as the best global dataset and it also presents the overall best representation of the probability density function (PDF). The built dataset with ADW using a precipitation index presents the overall best performance, especially in representing the time variations. Even though ADW presents an overall better performance, ERA5-Land with a correction factor presents a better performance in terms of errors in the southern region (south of 40°S). The novel dataset is freely available through the link provided in the conclusions section.