The three MCE’s preformed provided insight into where Lyall’s mariposa lily could be reintroduced across the Southern Okanagan Basin. When comparing the three images produced by the MCE’s and the WLC, it is apparent that they show a somewhat similar suitability across the study area. There are no significant changes in suitability across the ecosection, but small local changes were produced by the differing WLC’s. Because the land use and biogeoclimatic zone factors were not weighted extremely high in their respective MCE’s, the suitability results do not show extreme increases in the overall suitability. This is why the final images do not show any significant changes in areas with the highest possible suitability (255). Local changes do occur. Since all of the factors are considered to be important for the analysis, factor weights were kept above 0.1243. The process of assigning suitability scores (as seen in the Methodology section) and subsequent weights to factors is a very subjective.
To find the total size of the areas with a suitability of 255, a reclass was preformed on the three MCE suitability images. The AREA module was used to calculate area and it was found that the evenly weighted MCE had around 11 hectares with a suitability of 255, the biogeoclimatic zone weighted MCE yielded around 17 hectares, while the land use MCE had 18 hectares of land with a suitability of 255. It is apparent that the two more heavily weighted factors produced increases in the total amount of suitable land (value of 255). All the MCE results show high suitability in the East Chopaka area, where the current mariposa lily populations exist. It is the triangular shaped area in the southwest corner of the ecosection.
The total amount of areas with the highest suitability was calculated to show how much area with suitability score of 255 actually covered this ecosection. For the purposes of this project, these areas will serve as the best places to plant Lyall’s mariposa lily. Because setting a suitability threshold provides extremely discrete and constricting results, areas in the higher range of suitability can also be said to be suitable. With this in mind, the MCE results for the Even Weights would be a good base for which to begin reintroduction planning because it takes into account all factors and shows a large number of areas with high suitability (from 240 to 255).
Knowing that there does exit areas where Lyall’s mariposa lily could be reintroduced gives hope for many plant species who are threatened or endangered. The Multi-Criteria Evaluation preformed for this analysis shows viable areas where the mariposa lily can be reintroduced across the southern portion of the Southern Okanagan Basin ecosection. Further research needs to be conducted into the ownership of land, whether Crown or private, and if the areas considered are available for the species to be planted in.
- Digitizing the forest cover from a picture was required due to the fact that the most current paper map available for South Okanagan forest cover was from 1958. The digitizing process led to a highly subjective representation of the forest polygons. Since Google Maps was used as a reference, forest cover polygons were made in Arc by looking at Google Map and digitizing areas based on the assumption that the dark green was forest. This process did not provide a highly accurate representation of forest cover, but did allow this important land use category to be included in the analysis.
- An important habitat characteristic of the mariposa lily is the soil type on which this plant species grows. Sadly, soil type was not included in the analysis. Soil data for the South Okanagan was not available, although significant time was spent attempting to find data for this area. This factor would have been extremely useful for this analysis.
- Finding a proper DEM was a difficult. Three DEM’s were analyzed to determine if they were sufficient to use. The final and proper DEM was not found until quite late. More time should have been spent critically assessing each DEM and making sure it was able to be used. DEM manipulation was quite time costly in both ArcGIS and Idrisi.
- Using the rangeland data provided an accurate representation of suitable grass-ford meadows. The problem with this data is that it did not detail rangeland which was Crown land or of private ownership. The assumption was made that all rangeland used was available as areas to reintroduce the lily. If the data was available, Crown and private rangeland could have been created as separate layers and assigned different suitability values.
- The resolution chosen for this analysis could have been higher. With the resolution set at 75m, the resulting vector layers were suitable enough for the analysis. A major error associated with this was the rasterization of the road layer. This layer had to have its representation in ArcGIS changed to be a larger line. The vector road line itself was very small, and since this constraint was needed for the analysis, the line had to be made larger so that rasterization could take place. This led to an inaccurate representation of the road layer. Knowing this, the resolution for the analysis should have been higher, possibly 30m and fewer inaccuracies might have been present. At the same time, rasterizing all vector layers with a 75m resolution did lead to a loss of detail.
- The process of assigning suitability scores and factor weights was based solely on the opinion of the project creator, Laura Dilley. This is the nature of a Muli-Criteria Evaluation, as the user has control over the suitability values assigned to the raster layers and the final weighting of the factors. Different standardization values and different factor weights would have produced different final suitability images, which would then have been interpreted differently. Therefore, it can be said that any MCE analysis like this one is highly subjective.
- The steps designed for this analysis did provide desired results (finding suitable areas), although the results were somewhat constricted. Using the ecosection to confine the area of study, in the end, was found to constrict the results and possibilities of finding proper suitability across the southern portion of the South Okanagan. If this project was to be done again, a larger area would have been chosen as the study region.
