How Consistent Are Extreme Precipitation Trends in Tanzania? Evidence from CHIRPS and TAMSAT

The consistency of extreme precipitation trends across datasets remains uncertain in Tanzania. Daily Climate Hazards Center InfraRed Precipitation with Stations (CHIRPS v3.0) and Tropical Applications of Meteorology using SATellite data and ground-based observation (TAMSAT v3.1) estimates for 1983–2025 were compared across bimodal and unimodal rainfall regimes. Annual and seasonal trends were assessed using the Theil–Sen estimator and Mann–Kendall test with selective trend-free pre-whitening and grid-cell false discovery rate correction. Both datasets showed significant national increases in days with rainfall ≥ 20 mm, rainfall accumulated on wet days exceeding the reference-period 95th-percentile threshold, and maximum daily rainfall, at 7.5–9.2%, 10.5–10.9%, and 3.5–3.9% per decade, respectively. Neither showed significant national trends in maximum five-day rainfall or consecutive wet days. Spatial agreement was limited: after correction, significant increases in upper-tail rainfall remained in 24.8% of CHIRPS cells and 1.8% of TAMSAT cells. Seasonal agreement was clearest during October–December in the bimodal region. National significance classifications were stable across pre-whitening options, although some changed with the analysis period. These findings indicate increasing heavy rainfall in both datasets at the national scale; without station validation, they do not confirm observed changes, and local patterns remain uncertain. Station validation is needed before local use.