H33D-1642
A Hybrid Model for ET (Evapotranspiration) Forecasting Based on EEMD for ET-based Water Resources Management

Wednesday, 16 December 2015
Poster Hall (Moscone South)
Aijun Guo and Yimin Wang, Xi'an University of Technology, Xi'an, China
Abstract:
 It is widely known that the water shortage is much more serious in some regions or basins under changing environment. In this paper, ET-based (evapotranspiration-based) water resources management (ET-WRM) model is purposed for land use planning and water resources management, which mainly focus on generalized ET, i.e. agricultural, industrial, domestic, and ecological water consumption, to achieve high efficient use of water resource. To accurately predict the ET, it is decomposed to several intrinsic mode functions and one residue by the ensemble empirical mode decomposition (EEMD) method and forecasted by a hybrid model combined artificial neural network (ANN), support vector machine (SVM) and autoregressive integrated moving average (ARIMA). The model is applied in Ningxia Hui Autonomous Region, China, a typical region of water resources shortage, and the results show that: (i) the pass rate of prediction obtained by the modified hybrid forecasting model reaches up to 93% in the study area, which shows higher accuracy than applying any one method of them singly. The predicted ET in programming year, 2020, will reach to 3.48 billion m³. (ii) in the study area, to achieve water-saving goal, agricultural water-saving measures should be taken in the future, due to the existing phenomenon of low irrigation water use efficiency and wide planting area of high water-consuming crops. (iii) the water-saving volume in agriculture, industry and domestic in 2020 should be reached to 1.48×108m³, 1.11×108m³and 0.61×108m³, to balance the available water supply and future water consumption, compared with the baseline year, 2011. (iv) for water-saving in agriculture, the adjustment of planting structure, irrigation scheduling, agricultural activities and engineering measures is the main measure.