| 引用本文: | 邱中齐,周琳琳,刘红娟,等.玉米农田生态系统蒸散发模型参数优化[J].灌溉排水学报,0,():-. |
| qiuzhongqi,zhoulinlin,liuhongjuan,et al.玉米农田生态系统蒸散发模型参数优化[J].灌溉排水学报,0,():-. |
|
| 摘要: |
| 【目的】在不能根据实测值得到具体模型参数的地域，对经验参数进行优化，提高区域蒸散发模型的精度。【方法】通过黑河流域-生态水文过程综合遥感试验水文气象观测数据集中大满超级站气象要素梯度观测系统的数据，研究玉米农田生态系统下的蒸散发模型优化问题，采用差分进化自适应算法，以潜热通量和感热通量为优化目标，引入能量闭合因子，对模型参数优化，核心思想为贝叶斯理论，通过构造多条马尔科夫链来估计参数的后验信息；引入传统评价指标，决定系数（R2）、线性回归斜率、均方根误差(RMSE)、一致性指数(IA)、纳什系数（NSE），对Shuttleworth-Wallace原模型和优化后模型潜热通量和感热通量的模拟性能进行评价。【结果】模型校准期，优化后模型相对于Shuttleworth-Wallace原模型在模拟潜热通量时，均方根误差降低52.46%，一致性指数提高17.3%；优化后模型模拟潜热通量时，纳什系数达到0.82。在模拟感热通量时，优化后模型相对于Shuttleworth-Wallace原模型的评价指数提高不明显。模型验证期，优化后模型相对于Shuttleworth-Wallace原模型在模拟潜热通量时，均方根误差降低50.51%，一致性指数提高14.46%；优化后模型模拟潜热通量时，纳什系数达到0.80。在模拟感热通量时，优化后模型相对于Shuttleworth-Wallace原模型的评价指数提高不明显。【结论】不能根据实测值得到具体模型参数的地域，根据条件相似获得该区域的经验参数，将经验参数代入Shuttleworth-Wallace原模型中模拟该区域蒸散发，效果不理想，模型评价数据较差；相对于Shuttleworth-Wallace原模型的评价数据，优化方案对经验参数进行优化，优化后模型对潜热通量的模拟有了较大的提高，但是对感热通量模拟性能提升的效果不明显。 |
| 关键词: 参数优化;蒸散发模型;能量闭合; |
| DOI： |
| 分类号:S181 |
| 基金项目:国家自然科学基金项目（面上项目，重点项目，重大项目） |
|
| Parameter Optimization of Evapotranspiration Modelfor Maize Field Ecosystem |
|
qiuzhongqi1, zhoulinlin1, liuhongjuan1, tianqianglong1, zhaozijing1, zhangxiaomei2, weiguoxiao1
|
|
1.College of Earth and Environmental Sciences, Lanzhou University, Lanzhou;2.Agricultural comprehensive service center of Taipingdian Town, Huining County
|
| Abstract: |
| Abstract:【Objective】To optimize the empirical parameters and improve the accuracy of Regional Evapotranspiration Model in areas where the specific model parameters cannot be obtained according to the measured values.【Method】 Based on the data of meteorological element gradient observation system of Daman super station in the comprehensive remote sensing test of ecological hydrological process in Heihe River Basin, the optimization of Evapotranspiration Model under corn farmland ecosystem is studied. Differential evolution adaptive algorithm is used to optimize the parameters of the model by introducing energy unclosed factor for multi-objective of latent heat flux and sensible heat flux. Bayesian theory is the core idea. Multiple Markov chains are constructed to estimate the posterior information of parameters; By introducing traditional evaluation indexes, such as coefficient of determination (R2), linear regression slope, root mean square error (RMSE), consistency index (IA) and Nash coefficient (NSE), the simulation performance of latent heat flux and sensible heat flux of Shuttleworth-Wallace original model and optimized model was evaluated.【Result】During the model calibration period, compared with the original Shuttleworth Wallace model, the root mean square error of the optimized model is reduced by 52.46% and the consistency index is increased by 17.3%; When the optimized model simulates the latent heat flux, the Nash coefficient reaches 0.82. When simulating sensible heat flux, the evaluation index of the optimized model is not significantly improved compared with the original Shuttleworth Wallace model. During the model validation period, compared with the original Shuttleworth Wallace model, the root mean square error of the optimized model is reduced by 50.51% and the consistency index is increased by 14.46%; When the optimized model simulates the latent heat flux, the Nash coefficient reaches 0.80. When simulating sensible heat flux, the evaluation index of the optimized model is not significantly improved compared with the original Shuttleworth Wallace model.【Conclusion】In the area where the specific model parameters cannot be obtained according to the measured values, the empirical parameters of the area are obtained according to the similar conditions, and the empirical parameters are substituted into the original Shuttleworth Wallace model to simulate the evapotranspiration of the area, the effect is not ideal, and the model evaluation data is poor; Compared with the evaluation data of the original Shuttleworth Wallace model, the optimization scheme optimizes the empirical parameters. The simulation of latent heat flux has been greatly improved by the optimized model, but the simulation performance of sensible heat flux has not been improved to a great extent. |
| Key words: parameter optimization; evapotranspiration model; energy closure |
