Comparison of chaparral regrowth patterns between Santa Ana wind-driven and non-Santa Ana fire areas.

摘要

Wildfires are a common occurrence in California shrublands and island forests. Fire has a fundamental role in maintaining the ecosystem functions in chaparral where fire intensity and severity play important roles in the regeneration of species. In San Diego, the Cedar Fire that occurred in the fall of 2003 was unique in that one side was burned with wildfire fueled by dry, strong easterly Santa Ana winds that later died down, burning the remainder of the area under a mild westerly wind, allowing fuel-fed conditions.;The objective of this study was to understand the connection between vegetation type and structure and environmental response to extreme fire events by analyzing life form regrowth in chaparral communities from the Santa Ana wind driven, Santa Ana backing, and non-Santa Ana fire types. Environmental factors of slope angle, aspect, elevation and soils were investigated in an effort to isolate shrub regrowth patterns. Fire burn characteristics, anthropogenic disturbance, fire history, and moisture availability were also analyzed to identify additional factors that may have influenced shrub regrowth. Shrub extents before the fire and six year after the fire were examined per slope aspect, slope angle, elevation, and fire characteristic categories.;The closed canopy and natural features of the chaparral environment make ground based mapping very difficult. Remote sensing data and methods can be very helpful to evaluate the health of the vegetation and condition of the watershed for flood, erosion, and fire control. This study used high spatial resolution aerial imagery and a machine learning algorithm with a spatial contextual classifier to map three different areas from within the Cedar Fire perimeter. Geographic information science (GIS), field mapping, and image interpretation methods were used to identify vegetation samples for the classification and accuracy assessment of the vegetation maps. Object-based image samples were selected for the classifier from shrub, subshrub, herbaceous vegetation, bare soil and rock. A second set of samples was selected for the accuracy assessment using field methods then confirmed on the imagery. Class accuracies ranged from a low of 77% for herbs/subshrubs to over 98% for shrubs. The high accuracy of the classified maps gave us confidence in the class extents used for the shrub regrowth analysis.;Results indicated substantial differences in the type and amount of vegetation regrowth in the three study sites. Prior to the Cedar Fire, vegetation in all study sites consisted of 75%-85% true shrubs. Six years after the fire, shrub cover in the Santa Ana wind driven site was substantially less extensive (54%) then the other two sites (71% and 66%). Relative change in class proportions was notable. The SA site had up to twice the decrease in shrub extent of the non-Santa Ana study site on south/southwest facing slopes and on slope angles over 45°. Fire burn characteristics of residence time, intensity and severity may have influenced shrub recovery. Site specific attributes of precipitation, moisture availability, human disturbance, and fire history may also have influenced shrub recovery at each site.