Conclusion | Avalanche

Without going out into the field to observe real-world avalanche conditions and comparing them with the results of this analysis, no real conclusions can be made. So far, this project is purely a proof of concept, however, with more refinement, implementation of the daily functionality of the analysis, and a whole lot of comparison against actual avalanche conditions, a project like this could serve as a supplement to the CAC's existing avalanche forecasts.

There are infinitely many factors that affect whether or not an avalanche will occur, so forecasting is incredibly difficult and complex. One of the strengths of this project is that it does not focus on the wind, snowfall, and temperature information that can produce wildly variable avalanche conditions. Instead, this project considers snowpack stability forecast information already produced by a very qualified organization of forecasters and combines those forecasts with the constants of natural terrain that help to define avalanche start zones.

Despite relying on a trusted forecast from the CAC, challenges still exist in the analysis process of this project. For example, it is difficult to decide which generalizations are safe to make, and which are not. Of course, it is better to generalize on the side of safety, but whether it is OK to combine avalanche size with likelihood before conducting the rest of the analysis or whether it needs to be a more complicated aspect of the analysis remains to be seen.

The eventual future of this project is fortunate in that the Spearhead Range is very well traveled by many visitors each day, and observations for comparison with the results of such analysis would be easily available, and could even be crowd-sourced.

Finally, for the project to function as a daily application that can be run and repeated on a daily basis, it will need to exist entirely as a self-contained Python script, and will need to be far more robust than my current system, where much of the analysis and reclassification was still done "by hand" in ArcMap rather than automatically.

In all, this project has identified a possible route for visualizing avalanche forecasts, and has identified some of the problems and difficulties that face such a project. These difficulties come in two forms: accuracy and technical difficulty. First, avalanche forecasting is incredibly difficult, and it would be extremely important to make sure this method of visulaizing the forecast is robust and consistent with the CAC's daily forecast. Second, the way the CAC provides data for application use is somewhat unhelpful—the data is available in a difficult to parse XML file which takes considerable thought to read, let alone use in a meaningful way. I am certain someone with more coding experience and a better understanding of Python than I could figure out a way to do it, but unfortunately, those people are expensive and hard to come by for the moment.

I hope to some day continue the project, but for now, I unfortunately have to leave it in its current state: as a proof of concept.