Accurately predicting water table depth is an important basis for water resources management in irrigation areas. Based on hysteresis and nonlinearity of groundwater in time series, a CAR-SVM water table depth prediction model was developed by integrating multivariate time series controlled auto-regressive(CAR)and support vector machine(SVM). To improve the performance of the model in freezing-thawing period, a water table depth fitting model, i.e. CAR-SVM(T-TF)model, was established for seasonal freezing-thawing irrigation district. Simulation results indicate that the performance of the CAR-SVM(T-TF)model with ambient temperature effect in the freezing-thawing period is better than either the CAR-SVM(T)model with ambient temperature effect of the whole year or the CAR-SVM without any ambient temperature effect. The CAR-SVM(T-TF)model was applied to predict the water table depth in Hetao Irrigation District. The results demonstrate that the coefficient of multiple determination, R2, is 0.954 and 0.973 in validation period and freezing-thawing period, respectively, and all the RMSE in different periods are less than 0.09 m. suggesting a relatively high accuracy. The 3-order CAR-SVM(T-TF)model structure obtained from Hetao Irrigation District as a whole was used to simulate the water table depths in five irrigation areas in the district. The model has a good applicability in each area, and the predicted water table depths are closed to the measurements. Specially, R2 is more than 0.90, RMSE is less than 0.10 m, and BIAS is less than 0.04 in the freezing-thawing period in each area.
ZHAO Tian-Xing,ZHU Yan-*,YANG Jin-Zhong et al. CAR-SVM model based water table depth prediction in seasonal freezing-thawing areas[J]. Journal of Drainage and Irrigation Machinery Engin, 2018, 36(11): 1180-1186.
YUE Weifeng, GAO Hongyong, CHEN Aiping, et al. Study on conjunctive use of water resources in Hetao Irrigation District of Inner Mongolia based on GAMS simulation and optimization[J]. South-to-north water transfers and water science & technology, 2013,11(3):12-16.(in Chinese)
HUANG Yifan, LIU Junmin, JIANG Peng, et al. Research for groundwater dynamic and forecasting of Jinghui Irrigation Zone based on Modflow[J]. Research of soil and water conservation, 2014,21(2):273-278.(in Chinese)
WANG Mengmeng, DAI Lingquan, DAI Huichao, et al. Support vector regression based model for predicting water level of Dongting Lake[J]. Journal of drainage and irrigation machinery engineering, 2017,35(11):954-961.(in Chinese)
SURYANARAYANA C, SUDHEER C, MAHAMMOOD V, et al. An integrated wavelet-support vector machine for groundwater level prediction in Visakhapatnam, India[J]. Neurocomputing, 2014,145:324-335.
EBRAHIMI H, RAJAEE T. Simulation of groundwater level variations using wavelet combined with neural network, linear regression and support vector machine[J]. Global and planetary change, 2017,148:181-191.
ZHANG Zhanyu, LIANG Zhenhua, FENG Baoping, et al. Groundwater level forecast based on principal component analysis and multivariate time series model[J]. Advances in water science, 2017,28(3):415-420.(in Chinese)
SUJAY RAGHAVENDRA N, DEKA P C. Support vector machine applications in the field of hydrology: a review[J]. Applied soft computing, 2014,19:372-386.
YANG Wenyuan, HAO Peijing, ZHU Yan, et al. Groundwater dynamics forecast under conjunctive use of groundwater and surface water in seasonal freezing and thawing area[J]. Transactions of the CSAE, 2017,33(4):137-145.(in Chinese)
WANG Xiao,YANG Peiling,SU Yanping. Response of greenhouse tomatoes to buried depth and low irrigation limit of small flow drip irrigation[J]. Journal of drainage and irrigation machinery engineering, 2018, 36(10): 937-942.(in Chinese)
LIU Yanwei,WANG Shuying,TU Xinglei, et al. Characteristic analysis of dry-wet condition and temperature trend in Yuanmou dry-hot valley(DHV)in resent 60 years[J]. Journal of drainage and irrigation machinery engineering, 2018, 36(2): 172-178.(in Chinese)