摘要
该文论述了根据1951—2000年中国逐月温度和降水量数据采用样条函数插值实现气候时空数据数字化和可视化扩展的过程。对来自气象站的逐月气候数据利用样条函数插值拟合气候空间曲面,在DEM支持下建立气候变量的空间栅格数据文件,用于GIS对气候变量的查询、成图和二次开发。结果表明插值效果因样本数量、样本空间分布和不同的气候变量会有所不同,较多的样本数量有助于插值质量的提高。由栅格数据文件转换而成的逐月气候图形格式文件可以通过Java编程的浏览器实现气候时空动态变化的直观表达,从而使历史气候数据得到有效扩展、集成和应用。
Surfaces of monthly mean temperature and rainfall were fitted with data from meteorological station network recorded in period of 1951-2000 using spline interpolation technique and a spatiotemporal climate database was developed by coupling the fitted spline surfaces to underlying digital elevation model in scale of 1/10 degree to expand the use of long term climate data for climate modeling. Two data formats are available for this database. Gridded files can be used by GIS for data retrieval, climate mapping and secondary climate variables deriving. Graphic files are suitable to visual display of climate change through time with a browsing program developed by Java. The statistical analysis indicated the residual errors of the fitted surfaces were correspondently reduced while the density of original data point increased from 200 up to 700 stations. Thin plate smoothing spline is advantageous to data which is limited either in time period or spatial density.
出处
《地理与地理信息科学》
CSSCI
CSCD
北大核心
2003年第5期27-31,共5页
Geography and Geo-Information Science
