Snow depth estimation accuracy using a dual-interface GPS-IR model with experimental results

Abstract

Measuring snow depth using the GPS interferometric reflectometry is an active microwave remote sensing technique and an emerging approach because of its relatively large spatial coverage and high temporal sampling capability. The current geodetic GPS networks are capable of measuring snow depth in the vicinity of the antenna installation at no additional hardware cost. However, the performance is constrained by the geodetic GPS antenna which was originally designed to minimize the reception of the reflected signal. In our prior work, we proposed a horizontally polarized antenna which has equal gain for both direct and reflected signals and tested its performance for a single snow event. In order to comprehensively assess its performance, we set up a horizontally polarized snow monitor (HPSM) using the improved antenna at Marshall, Colorado, USA, over the 2013–2014 water year. The data from the HPSM clearly shows that the proposed design has high sensitivity to even very light snowfalls. However, some anomalies are observed from the HPSM measurements, which tend to either overestimate or underestimate the actual snow depth. We explain the observed measurement anomalies by replacing the traditional air–snow single-interface model with an air–snow–soil dual-interface model. The effectiveness of the new model is validated by comparing the simulated results to the HPSM measurements. Utilizing the dual-interface model, we simulate the error curve for snow depth estimation given various snow depths and snow permittivities. The error curve shows that the estimation biases can be observed only for shallow snow with a relatively small permittivity value.