introduction:
FIGURE 1A. The Study Area
Background
The study area, Mesa del Huracan (Hurricane Mesa), is located in the Chihuahua Province of Mexico (Figure 1A), along the northern edge of the well known Copper Canyon, near Cuidad Madera. The study by Flores-Garnica (2001) utilized the standard interpolation methods of the software available at that time to estimate forest fire fuel loads, towards generating a comprehensive forest fire hazard model. This study was one of the first to apply geostatistical modeling (Kriging) to Forestry Science (Flores-Garnica et al. 2006). Since ArcView did not have the current capabilities of the Geostatistical Analyst Extension (e.g. Kriging methods), other software options, such as S-Plus and GS+, were utilized to fill this gap.
Most spatial statistics studies apply software options such as GS+ and S-Plus, the Geostatistical Library (GSLIB) open source software from Stanford University, or increasingly the R open source statistical software from the University of Auckland, New Zealand. Open source software, although free, often has the problem of poor documentation. The wide availability of ESRI products with documentation and support readily available should make Geostatistical Analyst a viable option for organizations that already use ArcGIS 9.X.
Data Properties
The data set from Mesa del Huracan, Chihuahua, Mexico consists of a mapbase polygon, and a shapefile with an attribute table that contains sampling locations and descriptive data such as the types and amounts of forest fire fuels, shrub and crown heights, and trunk diameters. The study area comprises a section of the Mesa del Huracan located within a private commercial forest the Ejido El Largo y Anexos, encompassing 1400 hectares (~3460 acres). The elevation ranges from 2189-2524 meters, with most of the samples occurring between 2300-2400 meters. The region has been projected onto the UTM Zone 12 NAD 83 Grid, with a resolution of 90 meters (i.e. cell size = 90m X 90m). Figure 2 shows the SE to NW trend in elevation, which may be a factor in the final interpolation models.
Figure 2. Elevation Estimates based upon Ordinary Kriging. This method yields a slight overestimate of the maximum elevation observed. 10 meter contours are shown.
There were 554 individual circular sampling sites, each 1000 m2 in area, and sampled in concentric zones of 80 m2 and 400 m2 (Flores-Garnica 2001). Fuel types, as defined by the U.S Forest Service (Peterson et al. 2005) were Duff (partially decayed organic materials), Wood (fallen trees and branches), Herbs (wildflowers, lichens, moss), Litter (leaves and needles) and Live Fuels (shrubs and saplings). These data were reclassified into 1-Hr, 10-Hr, 100-Hr and Live Woody Fuels (Table 1) according to the lag time. Fuel lag time is the amount of time necessary for a fuel to lose ~63% of its moisture content when a change in the environment has occurred (Utah State University 2005).
