Did you know that you can use land surface roughness to map out landslides? See this thread for tools and methods to do just that in #python. These are the results from a 6 month @NSF funded internship I conducted at the @utahgeological survey. (thread)

@SeanLahusen and Adam Booth (PSU) have written extensively on this topic. And Matteo Berti wrote an excellent review article on the subject: https://www.sciencedirect.com/science/article/pii/S0169555X12004862

My goal while at the @utahgeological survey was to leverage the extensive #Lidar datasets within the state to test different methods of measuring roughness and machine learning approaches to classifying rough topography as landslide or not landslide.

I settled on using several different approaches to measuring land surface roughness:
1) standard deviation of slope
2) root-mean squared height
3) continuous wavelet transform
3) directional cosine eigen vector

Figure below is roughness map of landslides in Utah.

Using the Receiver Operating Characteristic (ROC) curve, I was able to properly classify landslide terrain at >70% effectiveness (correctly identified), see example ROC curve below.

All of the code to run through and calculate land surface roughness, and classify landslides can be found here:

https://github.com/utah-geological-survey/RuffSurfFinder
Its all written in Python and has a reasonably sized manual. Take a look if #landslides, #lidar, or #python are up your alley.

This project was supported by the @NSF #Intern program, which I cannot recommend highly enough. I know several other students who've taken part and recommend you to look into it if it sounds right for you: https://www.nsf.gov/pubs/2021/nsf21013/nsf21013.pdf

Also this thread comes full circle, as I learned about the program 1.5 years ago from @SedimentStarved on twitter! During my 6 months @utahgeological, I made connections throughout the State and found a career opportunity with the @USGS_UT. A great and successful program!