灌排系统水转化模拟模型研究

Study of water transformation simulation model of irrigation and drainage system

受自然与人为因素综合影响,灌区水转化过程复杂多变。以湖北漳河水库灌区为例,分别从水平和垂直方向描述典型灌排系统的水转化过程,以水量平衡原理为基础,研发灌排系统边界处置、稻作区田间多层土壤、塘坝及排水沟道之间多个水转化模拟子模块,由此构建一种客观描述灌排系统边界半封闭结构及其水平输送、垂直运动与渠-田-沟-塘水转化过程的耦合模拟模型,解决了应用流域分布式水文模型求解灌区水转化过程中面临的灌区边界不确定、分层土壤简化处理等问题。利用漳河灌区谭店灌排系统2014-2015年水平衡测试观测数据对该模型进行检验,采用拉丁超立方抽样法对模型参数进行分层抽样,并利用偏相关法分析参数敏感性。通过模型计算可得,支沟率定期和验证期模拟结果的复相关系数分别为0.83、0.70,Ens系数分别为0.81、0.68,斗沟率定期和验证期模拟结果的复相关系数分别为0.79、0.68,Ens系数分别为0.73、0.62,结果表明该模型适用于灌排系统水转化过程模拟。与SWAT模型对比,对于同一研究区域,复相关系数由0.11提高到0.73,Ens系数由-0.71提高到0.70,进一步验证了该模型的先进性,为灌排系统尺度水转化模拟和机制研究提供了新方法。

The process of water transformation in irrigation areas is complicated due to the influence of natural and artificial factors.Taking the irrigation area of Zhanghe Reservoir in Hubei Province as a case study,we described the water transformation process of a typical irrigation and drainage system from both vertical and horizontal perspectives.Based on the principle of water balance,we established a simulation model that was composed of multiple modules for boundary treatment and simulation of the water transformation between multiple soil layers of the rice field,ponds and ditches.The model was calibrated and verified based on the data of Tandian irrigation and drainage system in Zhanghe district from 2014 to 2015.After sampling the model parameters using Latin hypercube,we conducted sensitivity analysis of the parameters using partial correlation analysis.The model calculation showed that the multiple correlation coefficients of the simulation results for the branch ditches in calibration and validation periods were 0.83 and 0.70 respectively,and the Ens coefficients were 0.81 and 0.68 respectively;the multiple correlation coefficients of the simulation results for the tertiary ditches in calibration and validation periods were 0.79 and 0.68,and the Ens coefficients were 0.73 and 0.62,indicating that the model was suitable for simulating the water transformation process of irrigation and drainage systems.As compared with the SWAT model,for the same study area,the proposed model improved the multiple correlation coefficient from 0.11 to 0.73 and the Ens coefficient from-0.71 to 0.70,proving the proposed model was superior to the SWAT model.This study provided valuable information for the study of water transformation in irrigation and drainage systems.