Vegetation predicts soil shear strength in Arctic Soils: Ground-based and remote sensing techniques

Abstract

Soil shear strength (SSS) is an important soil attribute that is influenced by vegetation. If aboveground biomass estimates can be used to predict soil shear strength, it would greatly enhance our ability to estimate SSS across large areas. Using data collected from 24 plots in Alaska, we analyzed the relationship between soil shear strength and ground-collected vegetation attributes and remotely sensed (RS) variables. We constructed both univariate and multivariate models to assess the predictive capabilities of the vegetation and RS variables. Total trees and total conifers were significant predictors of SSS, with a negative relationship existing between total trees/total conifers and SSS. Graminoid cover (%) was positively correlated with soil shear strength and was also a significant predictor of SSS. Of the RS variables, the bands B1 (0.443 μm), B2 (0.490 μm), and B3 (0.560 μm) from the Sentinel 2 satellite system were all significant predictors of SSS. A multivariate model improved model fit over the simple univariate models, with an R2 = 0.46. We have both demonstrated a connection between SSS and aboveground vegetation attributes for areas within interior Alaska and that it is possible to link SSS to RS variables using a multivariate model.