Evaluation and spatial downscaling of CRU TS precipitation data in the Philippines

Abstract

This study evaluated and downscaled (using Delta Method) Climate Research Unit time series (CRU TS) monthly precipitation gridded data in the Philippines. Based on the results, raw CRU TS data tends to underestimate (average percent bias = 0.89%) precipitation for most months of the year while downscaled CRU TS showed the opposite (average percent bias = − 2.99%). Overall both raw and downscaled CRU showed acceptable performance when compared with the observed monthly precipitation record. However, downscaled CRU TS data showed better accuracy (lower Mean Absolute Error and Root Mean Squared Error) and better performance (higher Nash–Sutcliffe Efficiency) compared with the raw CRU TS data. On the average, the computed evaluation statistics for downscaled CRU TS data were 79.87 (MAE), 144.56 (RMSE), and 0.43 (NSE) while 87.82 (MAE), 163.69 (RMSE), and 0.30 (NSE) for raw CRU TS.