Discussion

In order to properly implement any wildland fire behavior or fuel model, it is first necessary to understand several underlying principles and assumptions that underlie the analysis. These considerations are specific to this study, however, they may apply to any study area.

The relatively low variation in the study area’s elevation and relief can result in analysis outputs that are relatively fuel-dominated. While it is acknowledged that the study area’s topography can be described as “gently rolling” in the northwestern and glacially-affected southern subregions, it must be understood that this relief occurs at too small a spatial scale to provide meaningful analysis outputs across the entirety of the study area.

VFRDB sampling locations were located along roadways and trails that were accessible by vehicle rather than at randomly distributed forested locations throughout the study area. Although it would have been ideal to randomize the location of sampling points throughout the study area, it was quickly realized that the logistics of physically accessing random forested points throughout the study area would be a difficult task. 71% of Rhode Island’s forested land is privately owned and identifying the landowner to obtain permission would have extended the project beyond the end date of the grant. Also, since a “brute force” method of improving accuracy was required, more points in any forested location served to enhance the analysis. Finally, since off-street parking is not available at the vast majority of VFRDB sampled points, safety dictated the location of many VFRDB sampling points.

Sampled locations near roads may skew classification to higher fuel loadings, and therefore, more volatile fuel models. Since road edges provide a natural window to sunlight, denser understory is to be expected in some locations due to the lack of canopy cover. To combat this potential over estimation of 1-hour and 10 hour live and dead fuel classes, measurements were taken during both leaf-on and leaf-off seasons (spring-summer and fall-winter, respectively). Also, whenever possible locations 50 to 100m off the road would be physically accessed. This “interior” vantage point minimized the edge effect of the road and allowed the field surveyor to determine if a VFRDB point taken at the roads edge would accurately reflect the forest stand it depicts.

These analyses represent “potential” for fire behavior. In order for an actual fire to ignite and propagate, climactic conditions must be favorable. A primary reason that the study area has not sustained an uncontrolled burn in recent history is due in part to humid weather conditions characterized by high relative humidity levels and short durations between precipitation events.

Over 70% of the State’s forest land is composed of deciduous forest (RIGIS land use code 310). As such, many observations in the VFRDB were collected in deciduous forest polygons. A table of fuel models observed and land use forest code can be found in figure 7 in the "downloads" section, or in the interactive maps section.