Utilization of Multi-Temporal Microwave Remote Sensing Data within a Geostatistical Regionalization Approach for the Derivation of Soil Texture

Philip Marzahn,Swen Meyer

doi:10.3390/rs12162660

Abstract

Land Surface Models (LSM) have become indispensable tools to quantify water and nutrient fluxes in support of land management strategies or the prediction of climate change impacts. However, the utilization of LSM requires soil and vegetation parameters, which are seldom available in high spatial distribution or in an appropriate temporal frequency. As shown in recent studies, the quality of these model input parameters, especially the spatial heterogeneity and temporal variability of soil parameters, has a strong effect on LSM simulations. This paper assesses the potential of microwave remote sensing data for retrieving soil physical properties such as soil texture. Microwave remote sensing is able to penetrate in an imaged media (soil, vegetation), thus being capable of retrieving information beneath such a surface. In this study, airborne remote sensing data acquired at 1.3 GHz and in different polarization is utilized in conjunction with geostatistics to retrieve information about soil texture. The developed approach is validated with in-situ data from different field campaigns carried out over the TERENO test-site “North-Eastern German Lowland Observatorium”. With the proposed approach a high accuracy of the retrieved soil texture with a mean RMSE of 2.42 (Mass-%) could be achieved outperforming classical deterministic and geostatistical approaches.

Highlights

Climate Change is present and affects many ecological, social and economic developments
Airborne remote sensing data acquired at 1.3 GHz and in different polarization is utilized in conjunction with geostatistics to retrieve information about soil texture
While the digital elevation models (DEM) and the derived topographic indices reflect the obvious topographical field situation with smooth terrain, the coefficient of variation of the backscatter values reveals several additional features that do not follow the topographic features in that area

Summary

Introduction

Climate Change is present and affects many ecological, social and economic developments. A new challenge to solve in climate change research is the prediction of large scale climate change to develop suitable adaption strategies [1]. The development of management and adaption strategies implies the utilization of robust Land Surface Models (LSM), as well as reliable model input parameters [2,3]. The robustness of the utilized LSM is defined by the physics behind it, but still, it depends directly on the quality and accuracy of the model input parameters [4,5,6]. While standard vegetation parameters can nowadays be derived reliably from remote sensing data, the availability of high-quality soil data as model input parameters on different scales is often becoming the limiting factor in pedo-hydrological modeling attempts. Conventional soil maps cannot feasibly satisfy those requirements and are not available in the desired scale or resolution [6,7,8]

Methods

Results

Conclusion