基于水碳耦合过程的区域作物模型构建与干旱缺水影响评估

Development of Process-based Regional Crop Model and Assessment of Drought and Water-deficit Impacts

摘要:

干旱频发背景下,农业减产风险持续增加,我国农业抗旱用水管理与粮食安全生产面临极大挑战。现有研究通过构建干旱指标表征农业干旱,无法反映干旱对作物生长与耗水过程的影响,且干旱指数阈值往往难以有效量化。在此背景下,本文以冀中南地区夏玉米和冬小麦为研究对象,集成考虑干旱胁迫的作物生长模块、作物光合蒸腾耦合模块及农田水量平衡模块,构建考虑干旱影响过程的区域尺度作物模型,模拟干旱影响下作物生长、蒸散发（ET）过程以及最终产量的时空动态变化。结果表明,模型能较好模拟不同作物生育期叶面积指数（LAI）变化过程,决定系数（R2）为0.93~0.94,归一化均方根误差（NRMSE）为0.12~0.16,纳西效率系数（NSE）为0.59~0.71;ET模拟精度同样较高,R2为0.981~0.989,NRMSE为0.07~0.11,夏玉米率定期和验证期NSE分别为0.70和0.58,冬小麦分别为0.81和0.91;与地市统计产量相比,夏玉米模拟产量相对误差小于20%,冬小麦小于8.0%。在模型验证基础上,开展了干旱年份灌溉不足导致的干旱缺水情景模拟：夏玉米生育期内7月阶段性干旱对作物ET与产量的影响有限;冬小麦生育期对缺水更为敏感,灌溉减少导致区域平均产量明显下降。构建的区域作物模型在有限参数条件下,能有效刻画干旱年份不同缺水情景下作物LAI、ET及产量的时空动态变化,可为区域农业干旱影响评估及水资源调控提供技术支撑和理论依据。

Abstract:

Against the background of increasingly frequent drought events, the risk of agricultural yield loss has continued to increase, posing significant challenges to agricultural drought water management and food security in China. Existing studies have mainly characterized agricultural drought using drought indices, which failed to represent the impacts of drought on crop growth and evapotranspiration processes, while the threshold values of drought indices were often difficult to determine accurately. Therefore, summer maize and winter wheat in the southern Hebei Plain were selected as research objects. A regional-scale crop model considering drought impact processes was constructed by integrating a drought stress crop growth module, a coupled crop photosynthesis-transpiration module, and a field water balance module. The model was used to simulate the spatiotemporal dynamics of crop growth, evapotranspiration (ET), and final yield under drought conditions. The results showed that the model reproduced the variation in leaf area index (LAI) well during the growing seasons of different crops, with coefficient of determination (R2) values ranging from 0.93 to 0.94, normalized root mean square error (NRMSE) values ranging from 0.12 to 0.16, and Nash-Sutcliffe efficiency (NSE) values ranging from 0.59 to 0.71. The simulation accuracy of evapotranspiration (ET) was also high, with R2 values ranging from 0.981 to 0.989 and NRMSE values ranging from 0.07 to 0.11. For summer maize, NSE values were 0.70 and 0.58 during the calibration and validation periods, respectively, while the corresponding values for winter wheat were 0.81 and 0.91. Compared with municipal statistical yields, the relative errors of simulated yields were less than 20% for summer maize and less than 8.0% for winter wheat. Based on model validation, drought water-deficit scenarios caused by reduced irrigation in drought years were further simulated. Stage-specific drought during July had limited effects on evapotranspiration and grain yield of summer maize, whereas winter wheat was more sensitive to water deficit during its growing season, and reduced irrigation resulted in a marked decrease in regional average grain yield. The regional-scale crop model developed effectively captured the spatiotemporal dynamics of LAI, ET, and crop yield under different water-deficit scenarios in drought years with a limited number of model parameters. The proposed model could provide technical support and a theoretical basis for regional agricultural drought impact assessment and water resources management.