Mapping Climatological Bare Soil Albedos over the Contiguous United States Using MODIS Data

Tao He,Shunlin Liang,Yi Peng,Feng Gao

doi:10.3390/rs11060666

Abstract

Surface bare soil albedo is an important variable in climate modeling studies and satellite-based retrievals of land-surface properties. In this study, we used multiyear 500 m albedo products from the Moderate Resolution Imaging Spectroradiometer (MODIS) to derive the bare soil albedo for seven spectral bands and three broadbands over the contiguous United States (CONUS). The soil line based on red and green spectral signatures derived from MODIS data was used as the basis to detect and extract bare soil albedo. A comparison against bare soil albedo derived from 30 m Landsat data has been made, showing that the MODIS bare soil albedo had a bias of 0.003 and a root-mean-square-error (RMSE) of 0.036. We found that the bare soil albedo was negatively correlated with soil moisture from the Advanced Microwave Scanning Radiometer-Earth Observing System (AMSR-E), with a relatively stable exponential relationship reflecting the darkening effect that moisture has on most soils. However, quantification of the relationship between bare soil albedo and soil moisture still needs to be improved through simultaneous and instantaneous measurements at a finer spatial resolution. Statistics of the multiyear climatological bare soil albedos calculated using soil types and the International Geosphere-Biosphere Programme (IGBP) land cover types suggest that: Land cover type is a better indicator for determining the magnitude of bare soil albedos for the vegetated areas, as the vegetation density is correlated with soil moisture; and soil type is a better indicator for determining the slope of soil lines over sparsely vegetated areas, as it contains information of the soil texture, roughness, and composition. The generated bare soil albedo can be applied to improve the parameterization of surface energy budget in climate and remote sensing models as well as the retrieval accuracy of some satellite products.

Highlights

Bare soil albedo has been widely used in climate models and remote sensing estimates of surface energy balance as one key component of the surface albedo by determining the amount of solar radiation reflected and absorbed at the Earth’s surface [1,2]
Significant differences in soil albedo were found in comparisons of land surface models and satellite products [12], which lead to substantially different estimations of surface energy balance and hydrologic budget partitioning [18]
The purpose of this study is to build a 500 m bare soil albedo dataset over the contiguous United States (CONUS) on a pixel-basis from multi-year Moderate Resolution Imaging Spectroradiometer (MODIS) data based on the soil line feature and to explore the relationship between bare soil albedo and other surface properties such as soil moisture, soil type, and vegetation type

Summary

Introduction

Bare soil albedo has been widely used in climate models and remote sensing estimates of surface energy balance as one key component of the surface albedo by determining the amount of solar radiation reflected and absorbed at the Earth’s surface [1,2]. Oversimplified soil albedo parameterization has been reported to introduce uncertainties both in climate modeling [12,13] and satellite LAI/fPAR products [14,15,16,17]. Significant differences in soil albedo were found in comparisons of land surface models and satellite products [12], which lead to substantially different estimations of surface energy balance and hydrologic budget partitioning [18]

Objectives

Methods

Findings

Conclusion