Evaluating CML Rainfall Processing Methods on a Countrywide Scale in Zambia

Abstract The relevance of rainfall estimation for hydrological applications is universal. In Zambia, like many countries in the Global South, obtaining reliable rainfall information is particularly challenging due to sparse observational networks. Commercial microwave links (CMLs) are widely available and can improve the situation. We estimate and evaluate hourly rainfall amounts derived from attenuation data of more than 400 CMLs distributed across the country during 1 month of the rainy season 2022. We test several approaches of CML processing each of which has specific advantages and disadvantages in this setting. Integrated MultisatellitE Retrievals for Global Precipitation Measurement (GPM) (IMERG) satellite products and rain gauges of the Trans-African Hydrometeorological Observatory (TAHMO) are used as reference. Results show that the estimation quality of CML-derived rainfall amounts varies strongly and depends on CML hardware characteristics and network topology. Due to the significant differences between TAHMO and IMERG data, the choice of reference data does, however, have an impact on the validation of the CML-derived results. Nevertheless, we find that there are, in particular, differences between rural and urban CMLs. While many of the former display noisy signals and are hence difficult to use, the latter allow for meaningful map interpolation, which we show for the city of Lusaka. Given the scarcity of reliable rainfall observation data in the Global South, CMLs provide useful additional high-resolution rainfall information, potentially available in real time, that can support flood forecasting and agricultural management.