 Natural Disturbances |
|
Natural Disturbances |
|
Fire Factor Database
The Fire Factor Database is a spatially explicit set of biotic, abiotic, and human factors covering the northern Lake States region. Developed as part of the Great Lakes Assessment Fire Project , the database is intended to serve as a companion database for the Lake States Fire Database.
Layers of the Fire Factor Database are spatial co-registered with the gridded framework used to analyze the Lake States Fire Database, to permit direct overlay-based comparison. Data for each layer were collected for the entire final study area used in the fire project. They are described below.
Biotic factors have considerable influence on fire patterns of the northern Lake States region (Frissell 1973, Vogl 1970, Frelich and Lorimer 1991, Loope 1991, Whitney 1986). We coarsened the one-kilometer land cover data set of McGhie et al. (1996) using grids of 2 km, 3 km, 5 km, and 10 km. Wherever possible, the majority land cover within a cell was chosen when coarsening; in cases of a tie, a random land cover was chosen from within the cell. This approach resulted in a consistent land cover proportion across scales.
Distance to Nonforest was derived directly from the land cover data. A side-by-side inspection of the maps of data for all fires larger than or equal to 1 acre and of land cover indicate that more fires appear to happen on or near Nonforest. To see the possible effect of how far away each cell was from a Nonforest pixel, we interpreted the land cover data at each scale to create Distance to Nonforest layers at 10 km, 5 km, 3 km, and 2 km spatial resolutions.
Abiotic factors, such as weather and soil conditions, are known to influence fire patterns (e.g. Whitney 1986, Heinselman 1973, Haines et al. 1983) and have been used in fire prediction models (Martell et al. 1987, 1989, Todd and Kourtz 1991, Vega Garcia et al. 1995).
We computed Available Water Capacity (AWC) from the STATSGO soil data set (USDA, 1994). Defined as "the volume of water that should be available to plants if the soil ... were at field capacity", this data set is intended to serve as a proxy for soil conditions throughout the northern Lake States region. To determine the value for a cell in the data set, where higher numbers generally indicate more clayey soils, we weighted the estimated AWC for each soil layer by the thickness of the layer, and the proportion of that soil type within the soil association at a given point.
Using data provided by the ZedX corporation (Boalsburg, PA), we developed maps of 30-year averages for 1961-1990 for monthly maximum temperature, monthly minimum temperature, and monthly precipitation. ZedX provided data at a 1 km resolution, and we coarsened those coverages to 2 km, 3 km, 5 km, and 10 km data sets. Here we show three sets believed to be related to fire frequency at the 10 km resolution: Average March Precipitation, Average June Precipitation, and Average Maximum August Temperature.
These factors are those driven by the human influence on the study area. Human factors affect fire patterns in the Lake States (e.g., Haines et al. 1975, USDA 1987, Frelich and Lorimer 1991). Several studies have explained or predicted fire risk using human factors such as the proximity to the nearest road (Chou et al. 1993, Chuvieco and Congalton 1989), population density (Donoghue and Main 1985), and distance to nearest town (Vega Garcia et al. 1995). Because human factors appear to be a major influence upon the fire regime of the northern Lake States region, a large set of human factors was created for this study.
We used block-level data from the 1990 U.S. census to create several layers representing human factors across the study area. Blocks, the finest granularity available from public census data, were aggregated within the study area to produce images at the 2 km, 3 km, 5 km, and 10 km resolutions. Example sets produced for the Fire Factor Database include Population Density, Number of Housing Units, and Number of Rental Units.
We used United States Geological Survey 1:100,000 Digital Line Graph information to create road and railroad coverages for the entire northern Lake States region. Using our gridded framework, we used these coverages to create maps for both Rail Density and Road Density.
