CHAPTER II
METHODS

All plots were mapped by Walt Thies at the start of this study (personal communication 2003), and the spatial locations of the data gathered in 2003 were located on these maps. The extent of scale of the study is medium, with each site covering about 50,000 m2. Each plot is a 0.02 ha circular measurement plot surrounded by a 0.04 ha treatment plot. Soil measures were taken in the 0.02 ha plots. Subplots are located within the 0.02 ha plot by degrees and distance from plot center (see Chapter I). The greatest distance between plots is 350 m, the shortest distance between plots is 22 m.

Maps of the study areas were rasterized then spatially located, using the appropriate coordinate system and projection, onto existing ortho photos and digital elevation models (DEMs) downloaded via the internet. Vector shapefiles of boundaries and points for plot centers and subplots were located on the maps using the ArcMap (ESRI 2001) editing function. Buffers corresponding to the circular plot sizes were created from the plot centers. Geodatabases were created for each site in Microsoft Access (Microsoft Corporation 2002) for the measured attributes of the plot and subplot locations. Plot data were averaged or composited from subplot samples. In the case of nitrogen concentration, the composited plot samples were the only data acquired due to the cost of CHN analysis. For other data such as bulk density and forest floor depth and weight, subplot samples and locations could be used. At each site there are three spatially located subplots on each plot. The geodatabases for all variables measured for each plot or subplot were joined or added to the shapefile attribute tables and a layer was created for each feature (plots, subplots, buffers, a tree layer for Hoodsport). These layers were assembled to make a map document for each site. Figure 3.1 shows detail of a Hoodsport map with thematic layers displayed.

Deterministic and geostatistical surfaces representing predicted values of attributes at unmeasured locations were created using Geostatistical Analyst (ESRI 2001). Default software settings were accepted for lag size when creating semivariograms for predicted values. GIS software makes it possible to create three dimensional displays. This is often not a practical use of the data, but can make for dramatic displays and presentations. Three-dimensional maps visualizing the sites and differing data were created in ArcGis 3D Analyst (ESRI 2001).