Akademik Veri Yönetim Sistemi
THEORETICAL AND APPLIED CLIMATOLOGY, sa.1-2, ss.103-117, 2022 (SCI-Expanded, Scopus)
Over the past several decades, weather radar has become an indispensable tool in flood forecasting studies, especially in ungauged regions. However, the rainfall extremes with radar data may contain biases, and this can be removed by using the information provided by rain gauges. For this purpose, a procedure that covers additive and multiplicative error correction methods with three vector norms, taxicab norm (L-1), Euclidean norm (L-2), and maximum norm (L-infinity), is proposed for adjusting radar data. The top 26 flood events with 20 stations in a 5-year database of radar data are used to quantify the overall agreement between radar and gauge datasets. After finding the best model and norm for adjusting the error in radar data, the correction factor in the model is analyzed with the aspects of the season, station height, rainfall amount, and distance between radar location and rain gauge station. Results show that the multiplicative error model with L-1 norm minimization is the best choice given all criteria. It is noted that as the systematic error in the radar data is nonlinear, the additive factors are inconclusive in rainfall variability. Moreover, it is found out that the values in multiplicative constant evaluation are more apparent with grouping done for the season, rainfall amount, and distance, but a consistent result is not found for the grouping done for the station height.