Forest land is the essential and important natural resource that provides strong support for human survival and development. Research on forest land changes at the county level about its characteristics, rules, and sp...Forest land is the essential and important natural resource that provides strong support for human survival and development. Research on forest land changes at the county level about its characteristics, rules, and spatial patterns is, therefore, important for regional resource protection and the sustainable development of the social economy. In this study we selected the GIS and Geoda software package to explore the spatial disparities of forest land changes at the Beijing-Tianjin-Hebei area county level, based on the global and local spatial autocorrelation analyses of exploratory spatial data. The results show that： 1） during 1985-2000, the global spatial autocorrelation of forest land change is significant in the study area. The global Moran＇s I value is 0.3122 for the entire time period and indicates significant positive spatial correlation （p 〈 0.05）. Moran＇s I value of forest land change decreases from 0.3084 at the time stage I to 0.3024 at the time stage II; 2） the spatial clustering characteristics of forest land changes appear on the whole in Beijing- Tianjin-Hebei area. Moran＇s 1 value decreases from the time stage I to time stage II, which means that trend of spatial clustering of forest land change is weakened in the Beijing-Tianjin-Hebei area; 3） the grid map of the local Moran＇s I for each county reflects local spatial homo- geneity of forest land change, which means that spatial clustering about regions of high value and low value is especially significant. The regions with ＂High-High＂ correlation are mainly located in the north hilly area. However, the regions with ＂Low-Low＂ correlation were distributed in the middle of the study area. Therefore, protection strategies and concrete measures should be put in place for each regional cluster in the study area.展开更多
There are various occasions where simple, ordinary, and universal kriging techniques may find themselves incapable of performing spatial prediction directly or efficiently. One type of application concerns quantificat...There are various occasions where simple, ordinary, and universal kriging techniques may find themselves incapable of performing spatial prediction directly or efficiently. One type of application concerns quantification of cumulative distribution function (CDF) or probability of occurrences of categorical variables over space. The other is related to optimal use of co-variation inherent to multiple regionalized variables as well as spatial correlation in spatial prediction. This paper extends geostatistics from the realm of kriging with uni-variate and continuous regionalized variables to the territory of indicator and multivariate kriging, where it is of ultimate importance to perform non-parametric estimation of probability distributions and spatial prediction based on co-regionalization and multiple data sources, respectively.展开更多
文摘Forest land is the essential and important natural resource that provides strong support for human survival and development. Research on forest land changes at the county level about its characteristics, rules, and spatial patterns is, therefore, important for regional resource protection and the sustainable development of the social economy. In this study we selected the GIS and Geoda software package to explore the spatial disparities of forest land changes at the Beijing-Tianjin-Hebei area county level, based on the global and local spatial autocorrelation analyses of exploratory spatial data. The results show that： 1） during 1985-2000, the global spatial autocorrelation of forest land change is significant in the study area. The global Moran＇s I value is 0.3122 for the entire time period and indicates significant positive spatial correlation （p 〈 0.05）. Moran＇s I value of forest land change decreases from 0.3084 at the time stage I to 0.3024 at the time stage II; 2） the spatial clustering characteristics of forest land changes appear on the whole in Beijing- Tianjin-Hebei area. Moran＇s 1 value decreases from the time stage I to time stage II, which means that trend of spatial clustering of forest land change is weakened in the Beijing-Tianjin-Hebei area; 3） the grid map of the local Moran＇s I for each county reflects local spatial homo- geneity of forest land change, which means that spatial clustering about regions of high value and low value is especially significant. The regions with ＂High-High＂ correlation are mainly located in the north hilly area. However, the regions with ＂Low-Low＂ correlation were distributed in the middle of the study area. Therefore, protection strategies and concrete measures should be put in place for each regional cluster in the study area.
基金Supported by the National 973 Program of China (No. 2007CB714402-5)
文摘There are various occasions where simple, ordinary, and universal kriging techniques may find themselves incapable of performing spatial prediction directly or efficiently. One type of application concerns quantification of cumulative distribution function (CDF) or probability of occurrences of categorical variables over space. The other is related to optimal use of co-variation inherent to multiple regionalized variables as well as spatial correlation in spatial prediction. This paper extends geostatistics from the realm of kriging with uni-variate and continuous regionalized variables to the territory of indicator and multivariate kriging, where it is of ultimate importance to perform non-parametric estimation of probability distributions and spatial prediction based on co-regionalization and multiple data sources, respectively.
