Southeast Alaska is a temperate rainforest, and landslides have been happening here for thousands of years. But shifting rainfall patterns have increased the frequency of landslides, with catastrophic results. In the last 8 years, three different towns in our region have experienced fatal landslides.

After a major landslide in 2015, a number of local people with experience in the outdoors and in various scientific fields noticed a possible correlation between river levels and landslide activity. Our main river has a small watershed, so it responds rapidly to periods of heavy precipitation. Using Python, I was able to investigate the question we all kept coming back to:

Is the correlation between changing river levels and landslide activity strong enough to serve as the basis for predicting landslides?

Using historical readings from a local river gage, a set of conditions was identified that correlate meaningfully with landslide activity. In about 12 events where the river met those conditions, at least 5 were associated with known landslides.

I posted a landslide-risk monitoring tool in 2021, which provides a visual indication of when the river is behaving in a way that correlates with previously known landslide events. This tool is used by many community members to assess the ongoing risk of landslide activity during heavy rain events. Just as importantly, it helps people let go of anxiety when we're experiencing rain that feels heavy, but isn't actually associated with landslide risk.

This talk will show how Python's vast ecosystem supports the entire citizen-scientist lifecycle, from idea to public-facing resource.