Advancing Forest Health Assessment through Multi-Method Geophysical Surveys: Case Study in Snodgrass, Colorado.

This project is part of a larger project, SFA watershed function, which is poised to develop new conceptualizations and insights, as well as novel approaches for characterizing and predicting aggregated watershed hydro-biogeochemical responses to abrupt perturbations. This research topic aims to enhance forest health assessment by combining multiple geophysical methods with machine learning techniques. First, the Normalized Difference Vegetation Index (NDVI) timeseries data are used to detect potentially vulnerable and interesting areas within the forest. Upon identification of these critical areas, we will deploy a grid pattern of passive seismic sensors. These will predominantly be positioned in the areas pinpointed through the NDVI analysis, but will also include other selected locations throughout the site to capture the heterogeneity. To do so, we would like to use 50-60 Smart Solo IGU-16HR 3C sensors to acquire the ambient seismic across the site. The smart solo sensors will be installed in networks of about 250x125m with a 25m spacing between each sensors. HV spectral ratio will be computed form the data to have the 3D mapping of the bedrock. In a second time, cross-correlation between each pair of sensors will enable to perform a surface wave tomography of the subsurface at each network locations. To do so, the nodes configuration would be a sampling rate of 500Hz, a gain of 25dB and minimum phase. In addtition to these acquistion, In conjunction with the seismic data collection, we have a single Electrical Resistivity Tomography (ERT) device at our disposal. Given the limitation of having only one ERT device, we will sequentially position it along five selected lines within the identified regions of interest. The combined data will provide a comprehensive view of the subsurface properties, fortifying the capacity of our machine learning models to generate accurate forest health indicators.