基于站点观测数据的气温空间化方法评述

基于统计学的插值方法是地理学、生态学领域研究气温空间化的主要方法之一,对获取精细化气温数据进行生态模拟具有重要意义。结合国内外气温空间插值的主要研究成果,对常用气温空间化方法进行了归纳、对比,探讨各种方法的适用性和不足之处,从而为涉及气温空间化的具体研究提供一定的参考,并探讨了各类方法优化的方向。不同方法的对比分析结果表明:各种气温空间化方法各有所长,在具体的应用中都取得过较好的效果,但并不存在普适性的方法,在实际应用时必须针对研究区域具体的地理特征进行方法适用性验证或对各类方法中的具体参数进行改进,才能实现区域气温的空间最优化模拟。根据气温场的物理分布特征,结合GIS技术,考虑地形等更多的相关因子以提高气温分布微观细节的模拟精度是未来重要的发展趋势。

Review on spatial interpolation methods of temperature data from meteorological stations

Spatial interpolation is an important method for creating spatial representation of temperature in geographic and ecological research and is important for supplying fine resolution temperature data for ecological models. This paper reviews existing spatial interpolation research of meteorological factors and compares a number of interpolation methods, including global interpolators (trend surfaces and regression models), local interpolators (inverse distance weighting, gradient plus inverse distance squares method, PRISM, splines, ANUSPLIN), geostatistical methods (Ordinary Kriging, Co-kriging), and mixed methods (combined global, local, and geostatistical methods). These methods are commonly used for the spatial interpolation of temperature data. The aim of this study is to explore the suitability and inadequacies of these methods in order to provide references for future research involving spatial interpolation of temperature data. It also attempts to explore ways to improve the application of the various methods. The comparison of these methods shows that each method has its own strength in particular applications. There is no universal method suitable for all practical applications. In practice, specific geographical characteristics of the study area must be considered and tests should be done to determine the suitability of specific methods. In order to achieve optimal interpolation result of regional temperature, parameters of the methods should be adapted based on actual geographical conditions. Global interpolation and geostatistical methods can be applied to study global trends. Local interpolation based on distance similarity principle does not apply to global trends simulation. Mixed methods are able to combine advantages of global interpolation, local interpolation, and geostatistics, and improve the simulation accuracy. Mixed methods and PRISM and ANUSPLIN are more suitable for application under complex terrain conditions. In future research, integration of various temperature spatial interpolation methods will improve, and more mixed methods combining global, local, and geostatistical methods will be created. Methods based on the physical distribution characteristics of temperature and combined with GIS technology will be prevalent. In order to improve the simulation accuracy of temperature in microscopic details, introduction of additional factors, such as terrain, will be an important future trend.