Search

Precipitation prediction is central in hydrology and water resources planning and management. This paper introduces a semi-empirical predictive model to predict monthly precipitation and compares its predictive skill with those of machine learning (ML) methods. The stochastic method presented herein estimates monthly precipitation with one-step-ahead prediction properties. The ML predictive skill of the algorithms is evaluated by predicting monthly precipitation relying on the statistical association between precipitation and environmental and topographic factors. The semi-empirical predictive model features non-negative matrix factorization (NMF) for investigating the influence of multiple predictor variables on precipitation. The semi-empirical predictive model's parameters are optimized with the hybrid genetic algorithm (GA) and Levenberg-Marquardt algorithm (LM), or GALMA, yielding a validated model with high predictive skill. The methodologies are illustrated with data from Hubei Province, China, which comprise 27 meteorological station datasets from 1988–2017. The empirical results provide valuable insights for developing semi-empirical rainfall prediction models. [ABSTRACT FROM AUTHOR]

Evapotranspiration (ET) is a fundamental component of the hydrological cycle. It is also the main mode of irrigation and precipitation loss in cultivated lands. Accurate estimation of ET is therefore important for sustainable water resources management, including monitoring droughts and simulating crop production. Whereas precipitation and runoff are measured with reasonable accuracy, ET is difficult to quantify especially at the basin/regional scale. Remote sensing (RS) energy balance models are therefore the commonly preferred methods for estimating ET on large spatial scales. The ETWatch model is one such RS-based model that interactively links several standard RS-algorithms to estimate ET. Another promising alternative method for estimating ET is the terrestrial-based water balance approach. Observations of the Gravity Recovery and Climate Experiment (GRACE) allow sufficient closure of terrestrial water balance. In combination with measured precipitation and runoff data, GRACE data can be used to estimate ET. This paper compares ET estimated by the RS-based ETWatch model with that derived as a residual of the terrestrial-based water balance technique (GRACE plus measured precipitation data) for the approx. 319 000-km2 Hai River Basin (HRB) in North China. About 63 consecutive months of data, spanning December 2002 to February 2008, are used in the study. The amplitudes and phases of the RS-based ETWatch-estimated ET match favourably those of the terrestrial-based water balance technique. The average seasonal (12.7 mm) and annual (16.6 mm) RMSE is less than 10% of the estimated ET. This implies that the RS-based ETWatch model reliably estimates ET at large spatial scales. Time series plots show that the amplitudes of the hydrological components of precipitation, ET and total water storage change are highest in summer and lowest in winter. The GRACE-derived total water storage change shows overall storage loss in the basin. This method could be used not only to fine-tune RS-based ET models, but also to monitor basin/regional water storage change in real-time conditions. Citation Moiwo, J. P., Yang, Y., Yan, N. & Wu, B. (2011) Comparison of evapotranspiration estimated by ETWatch with that derived from combined GRACE and measured precipitation data in Hai River Basin, North China. Hydrol. Sci. J. 56(2), 249-267. [ABSTRACT FROM AUTHOR]