Precipitation is a key component of the water cycle, but its strong spatial and temporal variability makes it difficult to quantify accurately. Ground-based point observations provide direct measurements but have limited spatial representativeness, while weather radar and satellite remote-sensing retrievals extend precipitation estimates over larger areas but remain indirect and sensitive to assumptions about hydrometeor microphysics. For rainfall, a major source of uncertainty is the variability of the drop size distribution (DSD), which controls both bulk rainfall properties and the signals measured by weather radars.

This thesis develops open research software and data infrastructure to improve the observational basis, retrieval, and analysis of rainfall. It first addresses the fragmentation of disdrometer DSD observations through the development of DISDRODB, an international scientific initiative aimed at collecting, standardizing, processing, and analyzing DSD measurements acquired with different instruments by many institutions. This provides a reproducible, analysis-ready dataset for studying DSD variability across scales and for supporting the development of radar algorithms to estimate rainfall microphysical properties.

Building on this observational foundation, the thesis investigates how DSD variability affects polarimetric and dual-frequency radar signatures, and explores observation-driven, non-parametric approaches for retrieving rainfall and DSD properties. The results show that these approaches outperform conventional parametric retrievals under both idealized and real conditions. Their potential for broader, possibly global, application is assessed for both ground-based and spaceborne radar observations.

The thesis also addresses the technical barriers that limit reproducible multi-sensor precipitation research by developing GPM-API, an open-source Python package for accessing, manipulating, visualizing, and analyzing products of the Global Precipitation Measurement (GPM) mission.

Overall, this work contributes towards a more open, scalable, productive, and community-oriented ecosystem for rainfall science, connecting disdrometer observations, radar measurements, and satellite precipitation datasets. Beyond improving access to and analysis of precipitation measurements, it introduces and evaluates a new approach for improving radar-based rainfall retrievals. By increasing the availability, consistency, and usability of DSD observations, it also lays the foundation for addressing long-standing questions on DSD variability.