Advancement of satellite-based rainfall applications for basin-scale hydrologic modeling /

Derin, Yağmur
Accuracy and reliability of hydrological modeling studies heavily depends on quality and availability of precipitation estimates. However hydrological studies in developing countries, especially over complex topography, are limited due to unavailability and scarcity of ground-based networks. This study evaluates three different satellite-based rainfall retrieval algorithms namely, Tropical Rainfall Measuring Mission Multi-satellite Precipitation Analysis (TMPA), NOAA/Climate Prediction Center Morphing Method (CMORPH) and EUMETSAT’s Multi-Sensor Precipitation Estimate (MPE) over topographically complex Western Black Sea Basin in Turkey, using a relatively dense rain gauge network. The results indicated that satellite-based rainfall products significantly underestimated the rainfall in regions characterized by orographic rainfall and overestimated the rainfall in the drier regions with seasonal dependency. Further, a new bias adjustment algorithm has been devised for the satellite-based rainfall products based on the “physiographic similarity” concept. The results showed that proposed bias adjustment algorithm is better suited to regions with complex topography and provided improved results compared to the baseline “inverse distance weighting” method. To evaluate the utility of satellite-based products in hydrologic modeling studies, the MIKE SHE-MIKE 11 integrated fully distributed physically based hydrological model was implemented in the Araç Basin and driven by ground-based and satellite-based precipitation estimates. Model calibration was performed by a constrained calibration approach that is guided by multiple “signature measures” to estimate model parameters in a hydrologically meaningful way rather than using the traditional “statistical” objective functions that largely mask valuable hydrologic information during calibration process. Diagnostic evaluation has the potential to provide a consistent estimates of the parameters of watershed models.