摘要
气温是一个重要的气象参数,为了给农业研究提供更加精准且物理意义完善的气温空间分布,建立分站分月的月平均气温物理优化模型。通过融合遥感数据合成比辐射率与气温模拟结果迭代优化,克服以往模型中气温回代的缺陷,对模型进行优化,生成了2007年我国1 km×1 km分辨率的月平均气温空间分布图。分析表明,模拟结果能较好地反映气温的宏观分布趋势和局地分布特征。误差分析结果表明,迭代优化后的年平均绝对误差为0.72℃,年平均均方根误差为1.01℃。相较于IDW插值法,优化模型模拟结果精度更高,影响气温的各项因子在结果中显示的特征更明显;相对于前人的隐式统计模型,参数物理意义更加完善,精度略有提升。为气温空间分布模拟提供了一种新思路,对科学指导农业生产、合理利用农业资源具有一定的参考意义。
Temperature is a significant meteorological parameter. Physical optimization temperature model for the monthly average temperature based on Month-Station was built to provide the more accurate temperature distribution for the agricultural research. The model was optimized by the remote sensed data and the iterative optimization. Results of 2007 monthly mean temperature characteristic of 1 km × 1 km spatial resolution were showed in this paper. The results showed that the simulation results can well reflect the temperature macroscopic distribution and the local characteristics. The error analysis results showed that the annual temperature MABE is 0. 72℃ and the annual temperature RMSE is 1. 01℃. The optimization temperature model had higher accuracy than IDW interpolation method,and more obvious what each factor expressed. Compared with the previous implicit statistical model,the physical meaning ofthe parameters was more perfect and the precision was improved slightly. It provided a new idea for the simulation of temperature,which had certain reference value for agricultural production and agricultural resources.
出处
《科技通报》
2018年第9期30-36,59,共8页
Bulletin of Science and Technology
基金
国家自然科学基金项目(41330529)
江苏省第四期“333高层次人才培养工程”科研项目(BRA2014373)
关键词
气温
分布式模拟
模型优化
temperature
distributed Simulation
model optimization
