Characterization of fire effects on forest ecosystems in the Tillamook Forest, Oregon

Graduation date: 1998Presentation date: 1997-07-11From the 1920's through 1951 several severe fires occurred in the predominantly conifer forest ecosystems of the northern Oregon Coast Range. Of the 211,151 ha. of mapped area, 57 percent was burned. The effects of frequent fires with high severity on forest ecosystems over time at the landscape level is not fully understood. A reconstruction of fire history was conducted to help investigate the effects of fire severity, frequency, and area extent on distribution of postfire tree regeneration, species composition, and stand tree size, as well as on current species composition and stand tree size. I hypothesized that: 1) vegetation patterns (1950's and 1988) would vary with\ud time, because the persistence of disturbance effects (fire, logging, reforestation) on\ud forest vegetative responses varied, and 2) environmental controls (topography, soil,\ud climate) would become the primary influences when disturbance events were absent. In this study historical maps, sketches and notes were used to reconstruct spatial and temporal patterns of fires from the 1920's to 1951 and to identify unburned patches on a Geographic Information System. Relationships between fire regime and postfire and current vegetative patterns were tested. Constructing precise spatial data layers from early maps, produced before the availability of aerial photography or satellite image, was difficult. Historical map accuracy and quality were variable and poor by present day standards. Geographic reference points were used to transform inappropriate map scales. The reconstruction of spatial data was used to characterize spatial patterns of historic fires: my estimates of burn areas were similar to estimates in the literature. To reduce questionable data along fire and vegetation patch boundaries for hypothesis testing, an exclusion approach was used. Data within a 100 m width of fire and vegetation type boundary lines were called a fuzzy zone and removed from raw data. The distribution of various attributes in the reduced data was similar to the distribution of the complete data set. Regression analysis examined the effects of fire, logging, reforestation,\ud topography, climate, and soil type on vegetation patterns. Patterns of postfire (1950's) species composition, tree regeneration and tree size (DBH) were associated with the effects of fires, as well as influences of logging and soil type. Indices of fire occurrences (reflecting the time variation and severity of fires) frequently correlated to the 1950's vegetation patterns. The number of fires (frequency) did not cause great differences in vegetation patterns. Current (1988) species composition and tree size (after absence of fire for more than three decades) were correlated more with terrain variables. Plant succession also influenced the current vegetation patterns. Neither the date or number of fires caused marked differences in distribution of species and tree size, except large conifers were found in areas missed by fires. Postfire and current vegetation patterns were correlated with soil types which reflect the influence of topographic and climatic characteristics. However, historic fires occurred frequently on some soil types. Fires have a confounding influence on soil type. This confounding influence of fire on soil type cannot be avoided. Reforestation efforts appeared to have little influence on the postfire and current vegetation patterns. I inferred that the short time period of reforestation effects did not show its importance on the 1950's vegetation landscape. Although regression analysis results did not support my hypothesis, by 1988, reforested area in the northern Coast Range had increased since 1950's. Most of the large fire-open patches became mixed forest in about 3 decades may still relate to the force of reforestation. On private unburned ownerships conifer forests in 1955 were subsequently cut and replaced by mostly mixed forest by 1988. The analysis supported the hypotheses that soil type, aspect and plant succession were dominant influences on current (1988) vegetation patterns, while forest disturbances such as fire and logging were important influences on the immediate postfire (1950's) patterns. The results not only interpret the relationship between historic disturbances and vegetation distribution, but may also serve as a useful background for the management of the future forest landscape