Reducing the Risk of Avalanches with GIS and Machine Learning

Cooperstein, M. (2019). Reducing the risk of avalanches with GIS and machine learning. ArcNews. https://www.esri.com/about/newsroom/arcnews/reducing-the-risk-of-avalanches-with-gis-and-machine-learning/

This article details the use of python based machine learning algorithms that process existing and newly collected field data to predict and display avalanche paths relative to the current weather and topography. Using geoprocessing tools, scientists at the Colorado Avalanche Information Center calculate the elevation and angle of all slopes within the 28,000 square miles of terrain they monitor. Once their GI System processes that data, it outputs feature layers that can be displayed and queried.

Threshold value is an important concept for this work and for avalanche science in general. The threshold value is that point at which conditions cause the release of an avalanche. Python scripts written by scientists at CAIC determine threshold values as XML using historic data from each avalanche path combined with present weather and snowpack conditions. Then, they can then color code avalanche paths that are predicted to reach threshold within 12, 24, and 36 hours. While this technique requires a substantial amount of field data, in my opinion it represents a substantial step forward for avalanche mitigation efforts in the transportation sector. It will almost certainly decrease the economic impact of avalanche related road closures, and has the potential to save lives. It also brings with it a more efficient and streamlined way for CAIC scientists to answer their own questions without having to consult GIS technicians, the author notes.

Since it comes from an Esri publication, the information presented in this article should be reviewed with caution. However, it provided me an excellent and succinct overview of how Esri software is being used for avalanche monitoring and control work along Colorado's roadways. Furthermore, since I am a novice when it comes to the use of Python scripting within GI systems, this article helped identify concepts that I need to study to gain a better understanding of the capabilities of GI software in the field of avalanche mapping/research.