基于主成分分析的畦灌质量评价

针对畦灌灌水质量影响因素和评价指标较多造成分析和评价困难的问题，基于大田灌水试验数据，使用主成分分析方法将各评价指标的加权和作为灌水质量综合主成分指标，并利用该指标评价和分析灌水质量。结果表明，灌水质量综合主成分可以代表96.74%的灌水质量变异信息，且服从正态分布，具有较好的代表性与客观性，可用于畦灌灌水质量的评价。利用灌水质量综合主成分作为指标分析表明，畦长和改水成数对灌水质量有显著的影响，各因子对灌水质量的影响从大到小依次为改水成数、畦长、坡度、单宽流量、灌水定额、入渗系数、粗糙系数和入渗指数，并且影响灌水质量的因子之间存在相互拮抗作用或相互协同的作用。

Evaluation of border irrigation performance based on principal component analyses

Abstract: The numerous influential factors and evaluation indexes of border irrigation performance usually result in big difficulties in the analysis and evaluation of such irrigation. Establishing a concise, representative and credible evaluation indicator of irrigation performance is a basic and key problem. Based on field experimental data, the Elliott and Walker's Two-Point Method of WinSRFR4.1 model was used to simulate the process of irrigation according to the field irrigation experiment in the Guanzhong Plain in Shaanxi Province, the method uses the two observed advance times to set up two mass balance equations, which verified the model and parameter selected. On this basis, the weighted sum of different evaluation indices was used as a comprehensive index for the evaluation of border irrigation performance through principle component analysis. Then, this index was used to evaluate and analyze the performance of border irrigation in the Guanzhong Plain in Shaanxi Province. Results showed that WinSRFR4.1 model effectively simulated the hydraulic processes, that is, the advance of the surface water ?ow, although there were some points that deviated from the fitting curve due to the change in hydraulic conditions after the water supply stopped. The error resulting from simulated advance time and observed advance time was less than 1.0 min among the twenty-five irrigation events. This indicates that WinSRFR4.1 model is able to evaluate border irrigation performance. The comprehensive principle component index of border irrigation performance followed a normal distribution and could represent 96.74% of variation in irrigation performance. It is proved to be a representative and objective index and could be used as an index for the evaluation of overall performance of border irrigation. Analyses with this index showed that both border length and inflow cutoff have significant effects on border irrigation performance. The order of influential factors based on their decreasing influences on border irrigation performance is as: inflow cutoff, border length, slope, flux per width, irrigation quota, infiltration coefficient, roughness coefficient, and infiltration index. With comprehensive consideration of the relationship among irrigation quality, irrigation quota and the border length, ncreasing irrigation quota can improve the irrigation performance while large irrigation quota may also lead to bad irrigation performance as the increase of border length. There were mutual antagonistic or synergistic effects among the influential factors above. The method proposed in this paper can much better reflect the actual irrigation performance of border irrigation and provide a proper irrigation scheme with theoretical foundation and technological support.