DISCUSSION
Study Conclusions
The results of this project, while not statistically significant, suggest possibilities for future research. The apparent potential correlation between glacier topography and terminus behaviour indicates that it may be possible to group BC glaciers as decisively as Pelto and Hedlund (2001) grouped North Cascades glaciers, although BC glaciers may not fall into the same patterns of terminus behaviour. Slope-area plots for stream channels appeared to suggest distinct patterns of channel morphology, although in such a limited study as this it is impossible to identify patterns with a high degree of confidence. With a larger sample size, more definitive patterns may emerge. A better method of quantifying morphological variability is important, and would involve a more detailed analysis with more measurement points to delineate process domains, as described in Brardinoni and Hassan (2006). Such a detailed analysis was outside the time scope of a term project, but it would be crucial to further analysis. Even with a highly simplified method of estimating channel morphology variability, a potential correlation with glacier type appears, giving promise for future studies.
In terms of spatial analysis, delineating glacier areas using georeferenced Google Earth images proved to be a useful, relatively fast method of representing the most recent glacier extents. The vector-based contour method of extracting slope-area data is more accurate than the raster-based method using the ArcMap Watershed tool, but it is extremely time-consuming and is a somewhat subjective process.
The primary issue that emerged during this project was the extremely time-consuming nature of hand-digitizing watersheds and extracting slope-area information. Due to this, the project scope was significantly downscaled from its original intent, and digitization was performed using 30 m contours, which sacrificed detail. There is also a degree of error in such a subjective method of data extraction, although given the tools available at present it appears to be less error-prone than using the ArcMap Watershed approach. Future analysis may involve development of a model to be run through SAGA or R.
Due to time constraints and problems with data quantity and quality, it was not possible to perform a more detailed classification of morphology types. Future analysis should follow the methods for delineating process domains described in Brardinoni and Hassan (2006) to achieve a more accurate, quantitative analysis.
Potential issues exist with the classification of BC glacier types. First, BC glaciers may not follow the same patterns of terminus behaviour as North Cascades glaciers, despite the high degree of regional synchronicity. The topographic correlation with glacier terminus behaviour described by Pelto and Hedlund (2001) may not hold true for BC glaciers. Second, for this study, terminus behaviour was inferred for most glaciers, whereas for the North Cascades study, terminus behaviour was confirmed. There may be additional errors of estimation, such as using precipitation as snow as a stand-in for accumulation. Also, the PAS measurement itself is an interpolation from the Climate BC model, so it is not a confirmed number.
Finally, a more comprehensive multi-criteria
analysis of all potential controls on stream morphology and glacier
type is necessary. This project is a simplified attempt at initial
analysis and description, not an effort to describe a full model.
However, despite the errors and limitations of this project, it
suggests potential paths of future analysis.