Core/ Periphery Approach to Modeling

The core/ periphery approach attempts to model vegetation by identifying unique and non-unique environmental conditions.  The unique areas might be considered to define the core region of each vegetation class where unique combinations of environmental conditions occur (using specific temperature, precipitation and elevation values).  These core areas should provide a clear distinction among vegetation classes.  Peripheral areas outside the core region described by slightly different environmental conditions that aren’t as unique as the environmental conditions that describe the core area should represent potential transition zones between vegetation classes.
 
 

Figure 8: The Core/ Periphery Approach – A profile of a distribution of Pine Oak, Grassland, and Desert Scrub, illustrates the core/ periphery concept.  Each vegetation type has a core and periphery area.  The core areas are associated with a unique combination of environmental values.  The peripheral areas are transition areas between vegetation types and are associated with actual vegetation shared between two or more vegetation types and slightly different environmental values that are not unique.





Clustering with Imagine: ERDAS Imagine was used to cluster the data to delineate the core and periphery areas of each vegetation class and help prediction of the potential vegetation distribution.  Imagine was selected because of its geospatial clustering capabilities and to explore the potential of using a remote sensing software traditionally used for land cover classification to classify potential vegetation.  Twenty classifications or model runs were performed differing in either the variables used or the number of clusters.  Classifications were then compared with Rzedowski’s Potential Vegetation of Mexico to assess the correspondence or accuracy.
 

Using Higher vs. Lower Resolution Data Sets: Three sets of the same variables were assembled using the same resolution for topographic factors but differing actual and apparent resolutions for the climatic variables.  The classifications compared the use of higher vs. lower resolution data sets for predicting the regional-scale distribution of potential vegetation.
 

Accuracy Assessment: An accuracy assessment was performed to analyze the similarity or correspondence between the classified grids and Rzedowski’s Potential Vegetation map.  Several remote sensing techniques and statistics were used to assess accuracy including error matrices, producer’s and user's accuracy, overall accuracy and the kappa statistic.
 

Fieldwork: Fieldwork was conducted around Durango, Mexico to test my hypothesis that there are core and peripheral areas for each vegetation class, that the core areas should exhibit the clearest distinction between vegetation classes, and that peripheral areas should be associated with transition zones between vegetation classes.
 
 
 
 

home         Methodology         Modeling Potential Vegetation Distribution