Türkiye üzerinde radar ve yağış istasyon ağını kullanarak iyileştirilmiş yüksek çözünürlüklü radar yağış haritalarının tahmini

Meteorological weather radars can provide spatio-temporally high resolution precipitation estimates. Nevertheless, these estimates are prone to systematic and random errors due to the indirect nature of the measurement algorithm of radars. Gauge-based observations are known to be complementary data for reducing the radar-based estimation errors. In this study, the precipitation data retrieved from seventeen C-band meteorological weather radars in Turkey are merged with the station-based precipitation observations to obtain a spatially-continous high accuracy quantitative precipitation estimates (QPE). In contrast to the previous studies which focused on limited methodology, region or number of events, this study impliments, investigates and validates four gauge adjustment [Mean Field Bias (MFB), Local Multiplicative Bias (LMB), Local Additive Bias (LAB), Local Mixed Bias (LMIB)] and four time-independent [Multiple Linear Regression (MLR), Artificial Neural Network (ANN), Cumulative Distribution Function (CDF), Z-R Matching (Z-R)] bias correction methods over all the operating radars during the years 2014-2018. The relative performances of these bias correction methodologies were compared both spatially and temporally in training, and validation datasets. Among these methodologies, a consistent algorithm (LAB) that generally results in the highest QPE performance was used in producing a high-resolution composite precipitation map. The datasets and maps produced in this study can be used as a significant input and contribution for future hydrology and water resources studies. Keywords: Radar Precipitation, Bias Correction, Precipitaiton Estimation