A weather reconstruction approach for daily precipitation since 1960s in South America

Long-term weather observations are required to understand past climate and extreme weather events. However, there are a host of factors that affect the measurements and make the data unsuitable for direct use and analysis. In this regard, the proposed research attempts to create a serially complete observed and gridded dataset of daily precipitation in South America, a region with sparse station networks, complex orography (Andes Mountain range) and a diversity of climates (tropical to sub-polar climates). To accomplish this challenging purpose, we will create in a first step a station-based database with high-quality standards using reproducible quality control, gap-filling and homogenisation procedures. In a second step, we will construct a gridded-based dataset by employing weather reconstruction approaches such as analogues plus machine learning together with multiple satellite precipitation products and other remote-sensing- and reanalysis-based variables. Further, we will account for uncertainty in the station- and gridded- based dataset, which is critical for adequately understanding uncertainty in any application modelling chain, especially in complex-sparse terrain regions. Once the gridded data is available, we will evaluate it by analysing extreme events indices in conjunction with other established gridded precipitation products. This analysis will not only evidence the added value of the gridded data but also will enhance the knowledge of high-impact extreme events in South America, particularly over the Andes chain as a whole. Finally, we expect that the data products from the research will be useful for climate science and other geoscientific and operational applications in Earth-system fields in South America. The proposed study will continue previous projects in the tropical Andes (DECADE and CLIMANDES) by the University of Bern as it will expand to the entire continent, providing a wide variety of applications.