【摘要】 基于SWAT模型的时变参数和不变参数法可以用来模拟土地利用变化引起的径流效应。这两种方法评估环境水文效应的差异性仍缺乏定量分析。本文以东南沿海山美水库流域为例,在参数敏感性分析基础上,采用自动率定方法,分别构建1995年、2015年土地利用的SWAT模型,并运用时变参数和不变参数法对土地利用变化径流效应进行模拟。结果表明:(1)两期土地利用模型率定的参数集都能很好地模拟流域年、月和日尺度径流过程;纳什效率系数分别大于0.83、0.90和0.66。(2)土地利用变化基本不影响径流模拟参数的敏感性,但对参数取值影响较大;参数取值变化与土地利用变化方式密切相关。(3)与时变参数法相比,不变参数法低估了土地利用变化对年、丰水径流的影响,高估了对枯水径流的影响,时变参数法模拟结果更为合理。如时变参数法模拟的年径流平均增加1.22%,不变参数法为0.23%;时变参数法年最小1 d径流平均减小2.72%,不变参数法为5.07%。研究成果改进了SWAT模型的时变参数法,进一步完善了应用SWAT模型评估土地利用变化对径流影响的模拟方法,可为流域水资源管理提供依据。更多还原

【Abstract】 Both the time-varying parameter method and the invariant parameter method of SWAT model have wide applicability in hydrology to evaluating the effects of runoff incurred by land use changes due to engineered activities over a catchment;however,there was a lack of careful examination of the two methods for the differences in simulating environ-hydrological consequences.In this study,it conducted a case study of the Shanmei Reservoir basin in the southeast coast of China.Two SWAT models of land use of 1995 and 2015 were constructed by separately.Based on sensitivity analysis of parameters and automatic calibration method,it obtained two optimal parameter sets for the two models to runoff simulation.The effects of land use change on runoff were then simulated with the invariant and time-varying parameter methods respectively.Results are as follows:(1) The model with the calibrated optimal parameter sets well simulated the annual,monthly and daily scale runoff processes in the Shanmei Reservoir basin under two-phase land use conditions.The Nash-Sutcliffe efficiency coefficients were bigger than 0.83,0.90 and 0.66 for annual,monthly and daily runoff respectively.(2) Land use change had little effect on the sensitivity of model parameters.The 10 most sensitive parameters were the same for the two land use conditions;just differed slightly in the ranking of the sensitivities,however,the parameter values varied greatly for the two land use conditions.The variation of model parameters was closely related to the way of land use change.(3) There were obvious differences in the simulation of the effects of land use changes on runoff between the time-varying parameter method and the invariable parameter method.Under the same meteorological conditions,the invariant parameter method underestimated the effect of land use change on annual and wet season runoff and overestimated the effect in dry season.For example,the average increase in annual runoff was 1.22% for the time-varying parameter method and 0.23% for the invariant parameter method;the average decrease in annual minimum 1 d flow was 2.72% for the time-varying parameter method and 5.07% for the invariant parameter method.The simulation result with time-varying parameter method is more reasonable and it suggests the modelling of the impact of land use change on runoff using SWAT should consider the change of model parameters with the change of land use.The research improves the simulation method on the impact of land use change on runoff using SWAT model and provides a reference for water resource management.更多还原