Methodological and Operational Problems
The material being created and studied in this website is not the truth about the world. They are not true, however they may be a close representation of the real world. The study is basically a stylized representation of how I tried to represent an environmental issue with the aid of GIS. Although great care was taken to minimize errors, many have occurred in this analysis. The following is a summary of the errors and mistakes that could not be avoided.
Methodological Problems:
The main problem encountered with this project was collecting the weather data. After meeting with Owen Hertzman, I found that the best weather maps were from the UBC Atmosphere and Science Program created by Roland Stull. However; the maps available were only of the current week in November and I could not find any historical or averaged data regarding the wind. A very important assumption was made that the weather maps chosen were random throughout the last month and represent a typical month in the fall in Vancouver. Perhaps given more time and a more efficient digitizing method, more samples could be analyzed over a whole season. Further studies must be completed to investigate the weather in other seasons to test the true viability of a wind turbine, as the fall is not even our windiest season.
Only one constraint and 2 factors were used in the MCE. There are many other constraints and factors that would influence the project, e.g. wind direction (although here, assumed to not matter), precipitation, access to a power line or large battery storage, temperature (batteries reduce storage and diminish in the cold), economic feasibility in each area in the GVRD etc...
Also, throughout the use of the MCE, the control points were subjectively assigned and may not hold precise relationships.
Operational Problems:
As stated before, the digitizing process was tedious and brought up more issues of uncertainty in the results. The more points used, the more precise the data; however, having a time limit, and no budget, this process had to be completed faster so a lot of accuracy was lost. The distance around each polygon was decreased a little but, and overall this may have decreased the total suitable area.
Software frustrations:
Throughout my data acquisition, i realized that most of the data I needed was vector or just a shapefile. I had to use ArcMap for a large portion of my project. Once finished with ArcMap, the data was in shapefile format. Using a Vector data set in a raster based program, points and lines are generalized and attributes are lost in the conversion. IDRISI did not assign values to polygons correctly when importing from shapefiles. Instead, it assigned the polygon (or line, or point - depending on what coverage was used) ID's to the polygons in IDRISI vector layers. This required many additional steps in reclassing the data.