ANSWER模型评估新疆咸水灌溉棉花产量与效益

利用咸水或微咸水进行农田灌溉是缓解中国新疆地区农业水资源供需矛盾从而保障当地棉花产业可持续发展的主要途径之一。为了明确不同咸水灌溉措施对棉花产量及经济效益的影响，该研究通过2 a的棉花膜下滴灌大田试验和文献检索获取了新疆9个不同试验地点的土壤、作物及灌溉等数据资料，评估作物产量-水盐胁迫响应分析模型（ANalytical Salt WatER，ANSWER）在新疆棉花产量评估中的适用性和可靠性，并结合经济收支平衡方法，模拟分析不同咸水灌溉措施（包括不同灌溉定额和灌溉水电导率的组合）对棉花产量与经济效益的影响。采用决定系数（R2）、均方根误差（root mean squared error，RMSE）、相对均方根误差（relative root mean squared error，RRMSE）评价模型精度。结果表明，在9个不同试验地点，ANSWER模型均可较准确地估算棉花的相对产量，其估算值与实测值之间的R2≥0.54，RMSE≤0.14，RRMSE≤0.16；不同试验地点，优化获得的各个模型生物参数（与棉花根系吸水的水盐胁迫响应相关的参数）差异较小，变异系数的绝对值处于0.08～0.37之间；基于不同试验地点优化的各生物参数均值估算各地的棉花相对产量，其与实测值仍然吻合良好（R2 为0.59，RMSE为 0.06，RRMSE为 0.07）；此外，当灌溉水电导率一定时，棉花净收益随灌溉定额增加呈先增后降的趋势，净收益达到峰值所需的灌溉定额随灌溉水电导率升高而迅速增加；当灌溉水电导率不大于10 dS/m时，通过加大供水量均可获得与淡水灌溉相当的净收益。研究可为新疆地区棉花产量与效益评估以及咸水资源合理开发利用提供理论依据。

Estimating the yields and profits of saline water irrigated cotton in Xinjiang based on ANSWER model

Abstract: Slightly saline water can be an important way to alleviate the current contradiction between the water supply and the demand in the Xinjiang irrigation area, particularly for the sustainable development of the local cotton industry. In this study, a systematic estimation was made to clarify the effects of different combinations of the amount (I) and electrical conductivity (EI) of irrigation water on the yields and profits of cotton. A two-year field experiment was also carried out under film-mulched drip irrigation in the nine experimental sites. Some indicators were then collected, including the data regarding soil (e.g. soil texture and bulk density), crop (e.g. cotton yield), and irrigation management (e.g. I and EI). The applicability and reliability were then verified on a biological-physical model-ANSWER (ANalytical Salt WatER) in the cotton yield. Combined with the Break-even method, The ANSWER was applied to investigate and evaluate the effects of different combinations of I and EI on the yield and profit of cotton using scenario analysis. Three statistical indices were employed to evaluate the performance of the model, such as the determination coefficient (R2), root mean squared error (RMSE) and relative root mean squared error (RRMSE). The results show that the ANSWER fully estimated the relative yields of cotton in the different experimental sites and irrigation treatments with the various levels of I and/or EI. The R2 values between the simulated and measured relative yields were consistently greater than 0.54, while the RMSE and RRMSE were not more than 0.14 and 0.16, respectively. There were relatively small differences in the optimized values of each biological parameter in the ANSWER (representing the physiological response of crop root water uptake to water-salinity stresses) among all experimental sites, with the absolute Coefficients of Variation (CV) value ranging from 0.08 to 0.37 and a mean of 0.19. An individual parameter presented a large degree of variation (i.e. CV = 0.37), but little sensitivity for the estimation of relative yield. Furthermore, the combination of the mean (MN) of originally optimized biological parameters for different experimental sites was also used to estimate the relative yields of cotton for each specific experimental site. The estimated results were comparable to the measurements, with an R2 of 0.59, an RMSE of 0.06, and an RRMSE of 0.07. Scenario analysis showed that the cotton profit first increased and then decreased with increasing I under a given EI, and that the I required for reaching or being close to the maximum profit at least increased rapidly with increasing EI. When the EI was limited to 10 dS/m, the resulting profit by saline water irrigation still shared the potential to approach or even exceed the level of profit resulting from the freshwater irrigation only by means of sufficient saline water replenishment. However, when the EI was above 15 dS/m, there was no profit, due to the excessive salt stress induced by the low yield. Moreover, if a relative yield of 0.80 was desired or expected, the EI should be set to less than 17 dS/m, and meanwhile, the relative irrigation amount (Ir) should be set to larger than 0.8. Overall, the Ir linked to the peak value of profit can be expected to serve as an appropriate value of irrigation amount to guide the irrigation management under a specific EI, in order to balance the relationships of high profit, water-saving, and sustainable use of saline water. By contrast, it is very necessary to provide some measures (e.g. physical, chemical, or biological) for the soil salt removal or reduction to avoid or relieve the excessive accumulation in the root zone. The finding can also provide the theoretical basis to evaluate the cotton yield and profit in the rational development and utilization of saline water resources in Xinjiang of Western China.