Estimation of soil moisture of a high Andean wetland ecosystem (Bofedal) with geo-radar data and In-Situ measurements, Ayacucho - Peru

Abstract

High Andean ecosystems within microbasins serve as crucial areas for water recharge, containing both surface and subsurface moisture. However, these ecosystems are currently under threat due to overgrazing, degradation, and the impacts of climate change. The objective is to validate the subsoil moisture of bofedal estimated using ground-penetrating radar (GPR) data in comparison to in-situ measurements obtained with a soil moisture meter (SMM) in the Apacheta microbasin of the Ayacucho region. The validation method involves comparing soil moisture values obtained with the SMM, with the estimated dielectric permittivity (DP) values from GPR surveys along four transects (T) in a bofedal. Reflected wave amplitude data are converted to DP values to identify water pockets ($70<DP<81$) and saturated soil moisture ($10<DP<40$). An analysis of the determination coefficient R² and the Kappa index $\left(\kappa \right)$ was conducted between both groups of bofedal subsoil moisture data along the four surveyed transects at depths ranging from 0 to 24 cm that contain water and saturated moisture. T1 contains a volume of $1,16m^3$ (47.85 %), T2 has $0.98m^3$ (46.6 %), T3 lacks water (40.8 %), and T4 holds $0.63m^3$ (31.45 %). The correlation of DP data with SMM for T1 ($R^2=0.801)$, T2 ($R^2=0.949)$, T3 ($R^2=0.837)$ y T4 ($R^2=0.842)$ implies that the SMM measurements significantly explain the estimated DP. Moreover, the kappa test demonstrated good agreement reliability between both observations made with GPR and SMM, with $\kappa =0.763;\left[95\%CI:0.471-1.055\right]$, indicating that the GPR method for measuring subsoil moisture is acceptable with an 87.5% confidence level.