Methodological and Operational Problems
After conducting the analysis on this project, I discovered that there were many problems inherent in the project, that could have made the results better and more accurate. First of all, the criteria I used to conduct the analysis were sort of arbitrary; although all would be important for logging, there may have been some important criteria that were left out. An important criteria that was omitted is the fact that logging would only be allowed in certain areas. The land could be already reserved for parks or other uses and would not be allowed to be logged by the government. There is even a chance that the areas I have grouped together may not be contiguous and may be broken by roads, streams, railroads, etc. The logging company would therefore have to apply for a permit for the area I narrowed down in my analysis in order to discover whether logging was allowed there.
Another major problem is the fact that the data I have only applies to the area of the Grand River Basin and does not have any information on its environs. Roads, streams and cities in other areas of Ontario bordering on the basin could have skewed the results. This is not as big a problem as it may seem, since this is a watershed and therefore should be surrounded by mountains or highlands on all sides. That would suggest that few if any streams would cross the boundary and that few cities would be found on the immediate borders (cities are much more frequently found in valleys).
The criteria used to narrrow down the land in terms of tree species was also a bit arbitrary. Although the criteria used to apply to the usual environments of hemlock and white pine, there would be many other factors that would affect their distribution. Even the data used, drainage, soil slope and soil characteristics for agriculture, seem to be categorized fairly arbitrarily (and is not very detailed) and therefore the data for these may not be completely accurate. Still, this analysis simply serves to narrow down the areas to those most likely to contain those two tree species. The boolean layer created from this doesn't eliminate too much of the area either and therefore shouldn't affect the results too much. Also, the assumption that hemlock and white pine are the most commercially-viable species is an educated one, though other species may be accessible for logging as well.
Data from some of the other layers may be incomplete as well. The layers of rivers/lakes and roads (and maybe even cities) are undoubtably incomplete since at this scale they cannot possibly contain all the little details contained within the region. More detailed data would be necessary to make the analysis more complete. This may greatly affect the data, especially if a lot of small streams are added to the map, limiting the area that it would be possible to log.
The fractions used for the different criteria for the multiple-criteria evaluation were a bit arbitary, although their values are educated ones. Furthermore, the minimum values for certain criteria (area, distance to streams) perhaps should have been made into boolean layers, though the magnitude of these minimums are extremely small compared to the overall scale of the map making them fairly insignificant for this evaluation. The city criteria however fits its objective, since it probably would be possible to log within 2 kilometers of a city (as long as the site's not in the viewshed), it just wouldn't be very desirable and might require cutting through red tape. Also, the function used for the fuzzy evaluation of each criteria is not perfect. I used linear increasing or decreasing for all the functions, but sigmoidal or J-shaped curves could just as easily been more correct for certain criteria. Regardless, none of these three functions would probably have completely reflected the ideal function for each of the variables (linear was in other words probably as good an approximation as any).
Finally there is inherent error in all layers of GIS. Nothing can be perfect, especially at the scale at which the data for this area is collected. The elevation model for instance, though probably fairly accurate, undoubtably contains small errors and imperfections.
Back to Top | Back to Main Menu | Previous Page