Our research aims to clarify the causes and predictability of the variability of the atmospheric water cycle in a climate context, primarily focusing on the interactions between land /vegetation, fire/aerosols, and clouds and rainfall, and connections with the adjacent oceans. We use a wide range of observations to diagnose the dominant mechanisms in the above mentioned issues and to determine their representations and sources of uncertainties in climate models. Our works were cited by articles in Nature, Science, the Proceedings of the National Academy of Sciences, the Bulletin of the American Meteorological Society, Science News, and popular media such as Wall Street Journal, NewsWeek, and CNN Discovery Channel. In addition, We have developed a seasonal prediction system for summer rainfall anomalies over the US Great Plains, now regularly used by the Texas Water Development Board (TWDB) and Texas State Drought Preparedness Council to support water resource planning[1]. We have also developed a prototype seasonal prediction for winter rainfall anomalies over California.

The Congo Basin, in Central Africa, covers only 10% of the continent's landmass but plays a crucial role in global rainfall patterns, supplying 30% of Africa’s water resources. This rainfall is essential for agriculture, river transport, socio-economic stability, and sustaining the world's second-largest rainforest. More than 80% of the region’s rainfall comes from mesoscale convective systems (MCSs)—large thunderstorm clusters—far exceeding their contributions in other tropical areas. However, the Congo Basin remains the least studied tropical region due to sparse in-situ observations. Existing research suggests that MCSs in this region exhibit unique meteorological relationships, challenging current understanding.

Our newly awarded NSF project (AGS-2404970, 2024-2027, $473,998), "A Process-Level Understanding of Mesoscale Convective Processes over the Congo Basin Using the Model for Prediction Across Scales (MPAS)," leverages a state-of-the-art cloud-resolving model to study MCS initiation, development, and propagation. During December 2023–January 2024, the Congo Basin experienced one of its most severe floods in 60 years. Our study analyzed MCS events from November 21–25, 2023, which contributed to increased runoff and flooding. Using the MPAS model initialized with ERA5 reanalysis data, we tracked MCSs with the Tracking Algorithm for Mesoscale Convective Systems (TAMS). Despite inconsistencies between satellite data and ERA5 due to sparse gauge stations and low resolution, MPAS effectively captured key meteorological structures such as updrafts, wind shear, and cold pools, demonstrating its value in data-sparse regions. Our findings highlight the critical role of topography, with a mountainous region in southern Congo experiencing more intense MCSs than the northern lowlands.

MPAS outputs and processed observational data are freely available for researchers and students worldwide. Please use the provided link to access the data. Citation: Siyu, Z. (2025). Data for NSF project 2404970 "A Process-Level Understanding of Mesoscale Convective Processes over the Congo Basin Using the Model for Prediction Across Scales (MPAS)" [Data set]. Zenodo. https://doi.org/10.5281/zenodo.14968711