石羊河流域制种玉米咸淡水轮灌模式 的SWAP模型模拟

为了探究石羊河流域制种玉米的咸淡水轮灌模式，利用2014年田间试验观测资料对SWAP模型进行了参数率定和验证，模拟了不同咸淡水轮灌模式下的土壤水盐平衡，并筛选出了较优的咸淡水轮灌模式，预测了较长时期土壤盐分动态变化及制种玉米产量。结果表明：SWAP模型率定与验证过程中，土壤含水量的均方误差(RMSE)值在0.05 cm³·cm^-3以下，平均相对误差（MRE)值在15%以下；土壤含盐量的RMSE值在4.2 mg·cm^-3以下，MRE值在25%以下；制种玉米产量的RMSE值在380 kg·hm^-2以内，MRE值在10%以下，率定和验证后的SWAP模型可用于研究区咸水与淡水灌溉的模拟与预测。3.0 g·L^-1微咸水条件下采用2次淡水、1次咸水和1次淡水、2次咸水的轮灌方式以及6.0 g·L^-1咸水条件下采用2次淡水、1次咸水的轮灌方式为研究区制种玉米的较优轮灌模式，这3种较优咸淡水轮灌模式下的土壤盐分累积量较少，并且能够提高制种玉米的产量。3种较优轮灌模式模拟预测结果显示，在模拟期内土壤含盐量增幅不大，能够达到平稳，不会造成土壤盐渍化，制种玉米减产幅度较小，制种玉米产量能够保持平稳。

SWAP simulation of rotational irrigation models with saline and fresh water for seed maize in Shiyang River Basin

In order to explore rotation irrigation models with saline and fresh water for seed maize in Shiyang River Basin, the SWAP model was calibrated and validated with the field experimental data collected from 2014. The SWAP model was used to simulate soil water and salinity balance under different rotation irrigation models with saline and fresh water. The optimal rotation irrigation models with saline and fresh water for seed maize were screened out by calculating and analyzing soil water and salinity balance. The SWAP model was also predicted long-term effects of soil salt dynamic and seed maize yield. The results showed that the RMSE values were all lower than 0.05 cm³·cm^-3 and the MRE values were lower than 15% in soil water content calibration and validation. The RMSE values were all lower than 4.2 mg·cm^-3 and the MRE values were lower than 25% in soil salinity calibration and validation. The RMSE and MRE values were less than 380 kg·hm^-2 and 10% in seed maize yield calibration and validation, respectively. The SWAP model can be used as a management tool for simulation and prediction of rotation irrigation models with saline and fresh water in the study area. The SWAP model could properly reflect the change tendency of the measured values. The optimal rotation irrigation models were the irrigation of two times of fresh water and one time of saline water, one time of fresh water and two times of saline water under brackish water irrigation of 3.0 g·L^-1, and two times of fresh water and one time of saline water under saline water irrigation of 6.0 g·L^-1. These three optimal rotation irrigation models could achieve the aim of reducing soil salt accumulation in soil and improving seed maize yield. The long-term simulation and prediction indicated that soil salt content and seed maize yield could remain relatively stable for simulation period of time in the optimal irrigation models of seed maize. These three optimal rotation irrigation models would not result in soil secondary salinization and small reduction in seed maize production. The optimal irrigation models have guiding significance to effective irrigation of underground saline water in Shiyang River Basin.