<![CDATA[Flood is one of the most frequent hydrometeorological disasters in Bima, West Nusa Tenggara, Indonesia. Although the C-Band weather radar operated by the Meteorological, Climatological, and Geophysical Agency (BMKG) has been utilized for weather monitoring and early warning, the accuracy of its rainfall estimation products has not been comprehensively validated under the complex topographic conditions of the region. This study aims to evaluate the performance of four weather radar rainfall products, namely Constant Altitude Plan Position Indicator (CAPPI), Column Maximum (CMAX), Rain Tracking (RTR), and Surface Rainfall Intensity (SRI), by comparing radar-derived rainfall estimates with in-situ observations from Automatic Rain Gauge (ARG), Automatic Weather Station (AWS), and Automatic Agroclimate Weather Station (AAWS). The analysis was conducted across three radar distance zones (0–20 km, 20–50 km, and 100–150 km) to investigate the influence of radar range on rainfall estimation accuracy. Model performance was evaluated using Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Receiver Operating Characteristic (ROC), and the Youden Index to determine the optimal rainfall and reflectivity thresholds for flood early warning. The results showed that CAPPI achieved the lowest RMSE (2.578 mm h⁻¹) within the 0–20 km zone, whereas CMAX provided the highest estimation accuracy at greater radar distances. ROC analysis further identified locally optimized rainfall warning thresholds ranging from 0.053 to 0.716 mm h⁻¹, depending on the radar distance zone, thereby improving the discrimination between flood and non-flood events. These findings contribute to improving radar-based rainfall estimation and support the development of a more reliable impact-based flood early warning system in tropical regions with complex terrain.]]>
