| 苏婉怡,董莉霞,李广,袁建钰.苏婉怡等：基于NSGA-II算法的APSIM模型春小麦产量与土壤N2O排放参数多目标优化[J].干旱地区农业研究,2026,(3):244~254 |
| 苏婉怡等：基于NSGA-II算法的APSIM模型春小麦产量与土壤N2O排放参数多目标优化 |
| Multi\|objective optimization of spring wheat yield and soil N2O emission parameters based on APSIM model using NSGA-II algorithm |
| |
| DOI：10.7606/j.issn.1000-7601.2026.03.23 |
| 中文关键词: 春小麦 土壤N2O排放 APSIM模型 EFAST方法 NSGA-II多目标优化算法 敏感性分析 |
| 英文关键词:spring wheat soil N2O emission APSIM model EFAST method NSGA-II multi\|objective optimization algorithm sensitivity analysis |
| 基金项目:国家自然科学基金(32360438)；甘肃省自然科学基金(25JRRA356)；甘肃省高校教师创新基金(2025A-094)；甘肃省科技重大专项计划项目(25ZDFA011)；中央引导地方科技发展资金项目(25ZYJA035) |
|
| 摘要点击次数: 266 |
| 全文下载次数: 15 |
| 中文摘要: |
| 为估算西北半干旱区春小麦产量与土壤N2O排放的响应特征，提升APSIM模型在西北半干旱区春小麦产量与土壤N2O排放模拟中的精度和适用性，首先采用扩展傅里叶振幅敏感性测试方法(EFAST)对模型内部参数进行敏感性分析，筛选出小麦品种和土壤中硝化和反硝化反应影响最显著的关键参数，再利用非支配排序遗传算法II(NSGA-II)对筛选出的参数进行多目标优化。结果表明：敏感性分析筛选出的关键参数包括小麦品种关键参数，即潜在籽粒灌浆速率、每克茎籽粒数量、从萌发到开花的积温、对春化作用的敏感性、对光周期的敏感性；土壤中硝化和反硝化反应关键参数，即反硝化土壤水因子的幂项、硝化过程中氮以N2O形式排出的比例、反硝化系数、最大硝化速率、50% KMax的NH4浓度。优化后的APSIM模型精度显著提高，产量的均方根误差从121.31 kg·hm-2降至64.79 kg·hm-2，归一化均方根误差从10.05%降至5.37%，小麦产量模型的决定系数R2从0.74提升至0.92；N2O排放的均方根误差从32.46 μg·m-2·h-1降至7.16 μg·m-2·h-1，归一化均方根误差从13.92%降至3.07%，N2O排放模型的决定系数从0.65提升至0.82。优化参数在不同施肥水平（0、55、110、220 kg·hm-2）下均表现出良好的泛化能力，验证了模型的稳健性和适用性，为APSIM模型在当地应用与参数校准提供了科学依据。 |
| 英文摘要: |
| To estimate the response characteristics of spring wheat yield and soil N2O emissions in the semi\|arid region of Northwest China, and to enhance the accuracy and applicability of the APSIM model in simulating spring wheat yield and soil N2O emissions in this region, firstly, the extended fourier amplitude sensitivity test (EFAST) method was used to conduct sensitivity analysis on the internal parameters of the model, and the key parameters that have the most significant impact on nitrification and denitrification reactions in wheat varieties and soil were selected. Then, the non dominated sorting genetic algorithm II (NSGA-II) was used to perform multi\|objective optimization on the selected parameters. The results showed that the key parameters selected by sensitivity analysis included the key parameters of wheat varieties: potential grain filling rate, number of grains per gram of stem, accumulated temperature from germination to flowering, sensitivity to vernalization, and sensitivity to photoperiod. The key parameters of nitrification and denitrification reactions in soil include the power term of the denitrification soil water factor, the proportion of nitrogen discharged in the form of N2O during nitrification, the denitrification coefficient, the maximum nitrification rate, and the NH4 concentration of half of KMax(KNH4). The accuracy of the optimized APSIM model has significantly improved. The root mean square error of the yield decreased from 121.31 kg·hm-2 to 64.79 kg·hm-2, the normalized root mean square error decreased from 10.05% to 5.37%, and the determination coefficient of the wheat yield model increased from 0.74 to 0.92. The root mean square error of N2O emissions decreased from 32.46 μg·m-2·h-1 to 7.16 μg·m-2·h-1, the normalized root mean square error decreased from 13.92% to 3.07%, and the determination coefficient of the N2O emission model increased from 0.65 to 0.82. The optimized parameters showed good generalization ability at different fertilization levels (0, 55, 110, and 220 kg·hm-1), providing a scientific basis for the local application and parameter calibration of the APSIM model. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| | |