McKenzie, D, Peterson DL, and Agee, JK. “Fire frequency in the interior Columbia River Basin: Building regional models from fire history data.” Ecological Applications 10 (2000):1497-1516. (reviewed by Carissa Sharp)
summary:
It is important to be able to predict the effect of fires on forests and ecosystems. Vegetation is constrained by climactic variables, but changes in fire regimes could profoundly alter vegetation patterns. This article attempts to address these issues by creating simulation models that would predict vegetation change, incorporating fire effects.
The scientists accumulated fire history data in order to study the fire frequency of the region, a difficult task because fire regimes in Western North America tend to vary according to temperature and moisture. This paper presents statistical models for patterns of fire frequency in the interior Columbia River Basin. They used empirical methods to test relationships between fire frequency and vegetation types as well as environmental gradients. The fire history data vary in quality; thus they made two models, one using all of the data that they could possibly use, and one that used only the more authoritative sources. They also made models that were based on data about types of vegetation (tree-based models).
The goal was to extrapolate local data to make it pertain to the entire region, for example discovering that the fire return interval was found to be less strongly correlated with precipitation at high elevations than at low elevations. They found that while tree-based models have a stronger statistical fit locally, only regression models can be extrapolated for regional predictions. The regression models provide a more accurate estimation of fire return intervals than were previously available, and provide a new kind of model that has not been used before. These models are very important in the applications of ecosystem management.
The discussion of the limitations of the study include the heterogeneity of the region which makes it impossible to make completely accurate models, and the fact that the models could be improved by having a larger data set and more accurate data.
critique:
This article, from a peer-reviewed journal, focuses very heavily on statistics, as the attempt is to create regression and tree-based models for fire frequencies in the interior Columbia River Basin. There is detailed description of statistical methods used and the rationale behind using those methods. However, although the article is quite technical, it is still fairly accessible to the non-scientist.
Concepts are explained in a manner that is easy to understand, and for the most part the article focuses on the “big picture” which helps the reader understand the focus and general principles underlying the study. Additionally, the article includes charts and figures that show the area as well as the general principles the scientists are covering. These include maps of the area with highlighted national parks, national forests, and fire history sites, charts of types and specific species of vegetation, and detailed output maps with prediction of fire return rates.
This article has implications for the study of the Columbia River watershed because fires, just like any ecological variable, have a huge effect on the health of an entire ecosystem, including the vegetation of a forest as well as the streams and rivers that run through it.
Robert D. Clark Honors College, University of Oregon
HC 441: Science Colloquium, Columbia River Ecology
Fall term, 2005
Send e-mail to the Webmaster