METHODOLOGICAL PROBLEMS
The quantitative results of average altitude of treelines, and the varying results depending on aspect, can be assumed to be quantitatively meaningful, with fairly high accuracy. The original TRIM data quality is reasonably high, and the analytic transformations do not result in any increase in inaccuracy. However, possible sources of errors are:
- subjectivity of spatially determining the treeline by manual editing
- imprecise definition in the data how wooded areas are defined (e.g. is an individual 2.5 m high tree at 2400 m elevation a "tree" or a "shrub") => fuzzy boundary between wooded and non-wooded areas
- the small size of the two study areas (analyzing larger areas would lead to more accurate results)
However, as a matter of fact, the scope of this project only allows me to produce strongly approximate results in the case of determining environmental factors, i.e. I can only show very general trends in terms of temperature and winter precipitation. The models that I developed are much too simple to be any more than a visual aid to understanding the correlation of treeline elevation and those factors. No absolute quantitative conclusions should be derived from this analysis.
A range of factors has been listed in the Background Research section. The shortcoming of my model is that only two of those factors are being assessed in the analysis (temperature and precipitation). And even the individual factors have complexities of their own. For example, temperature lapse rates vary locally due to the following factors:
- lapse rate increases with elevation
- lapse rates vary with latitude
- slope orientation (aspect) affects lapse rate
- lapse rates vary with season of year
- lapse rates vary diurnally (usually higher at night)
The only factor that I took into account when determining the lapse rates for the two study areas was seasonal variation of lapse rates. The other factors are not incorporated.
Another example of inaccuracy is the linear "interpolation" (too simplistic to be called real interpolation) method I employed for the precipitation surface. Precipitation definitely does not increase in a linear fashion. According to climatic literature, there is, for instance a high point, at which precipitation starts to decrease again with increasing altitude. As a result mountain slopes generally receive more precipitation than their peaks. This high point of precipitation varies from location to location. However, generally, it is the case that precipitation increases with elevation, at least in the range of treeline elevation. I acquired all existing precipitation data for the regions around my study areas, but these were simply not enough. It is impossible to model an accurate precipitation surface without having an appropriately large number of points of known precipitation that can be used for interpolation procedures. In wilderness areas like these, we lack appropriate amount of climatic data.
