As a primary factor determining microwave radiative emission and scattering properties

soil surface roughness plays an important role in microwave remote sensing studies.Since microwave backscatter is also influenced by dielectric property and penetration depth

retrieval of soil roughness using solely microwave remote sensing data faces great difficulties.In this study

soil surface roughness is estimated using multiangular optical remote sensing data in order to provide useful information to microwave remote sensing studies.First

a geometric-optical model is established to simulate bidirectional reflectance for bare soil surface under different roughness conditions.The model assumes that soil aggregates or clods are semi-ellipsoids with circular bottom

and a number of aggregates of clods distribute randomly within a pixel.Then the model is inverted using multiangular reflectance to retrieve structural parameters of each soil pixel.After that

root-mean-square height

as an indicator of soil surface roughness

can be estimated from structural parameters of the model.This method is validated using in situ multiangular reflectance observations.Results demonstrate that estimation of soil surface roughness using multi angular remote sensing data is an effective approach.