...
  • 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Pure and Applied Geophysics >Spatial Interpolation of Ewert's Index of Continentality in Poland
【24h】

Spatial Interpolation of Ewert's Index of Continentality in Poland

机译：埃夫特大陆展望率的空间插值

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

The article presents methodological considerations on the spatial interpolation of Ewert's index of continentality for Poland. The primary objective was to perform spatial interpolation and generate maps of the index combined with selection of an optimal interpolation method and validation of the use of the decision tree proposed by Szymanowski et al. (Meteorol Z 22:577-585, 2013). The analysis involved four selected years and a multi-year average of the period 1981-2010 and was based on data from 111 meteorological stations. Three regression models: multiple linear regression (MLR), geographically weighted regression (GWR), and mixed geographically weighted regression were used in the analysis as well as extensions of two of them to the residual kriging form. The regression models were compared demonstrating a better fit of the local model and, hence, the non-stationarity of the spatial process. However, the decisive role in the selection of the interpolator was assigned to the possibility of extension of the regression model to residual kriging. A key element here is the autocorrelation of the regression residuals, which proved to be significant for MLR and irrelevant for GWR. This resulted in exclusion of geographically weighted regression kriging from further analysis. The multiple linear regression kriging was found as the optimal interpolator. This was confirmed by cross validation combined with an analysis of improvement of the model in accordance with the criterion of the mean absolute error (MAE). The results obtained facilitate modification of the scheme of selection of an optimal interpolator and development of guidelines for automation of interpolation of Ewert's index of continentality for Poland.
机译：本文提出了关于埃韦尔特队的波兰大陆指数的空间插值的方法论思考。主要目标是执行空间插值并生成指数的地图，结合选择了Szymanowski等人提出的决策树的最佳插值方法和验证。 （Meteorol Z 22：577-585,2013）。分析涉及四个选定的年份和1981-2010期间的多年平均值，基于111个气象站的数据。三个回归模型：多元线性回归（MLR），地理加权回归（GWR），以及混合地理加权回归用于分析以及其中两种延伸到残留的克里格形式。比较回归模型，并比较了本地模型的更好拟合，因此，空间过程的非公平性。然而，在选择内插器中的决定性作用被分配给延伸回归模型以剩余克里格的可能性。这里的一个关键元素是回归残差的自相关，这证明了MLR和对GWR无关的重要性。这导致从进一步分析中排除地理加权回归克里格。多元线性回归克里格被发现为最佳插值。通过交叉验证确认，根据平均绝对误差（MAE）的标准，交叉验证与模型的改进分析。得到的结果促进了选择最佳插值者的选择方案以及开发埃韦尔特为波兰大不年产权指数的内插自动化指南。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号