Comparison of three simplified algorithms for atmospheric corrections of MERIS data over land
Juergen Telaar(1) and Maria von Schönermark(1)
(1) University of Stuttgart, Pfaffenwaldring 31, 70550 Stuttgart, Germany
Today many different algorithms for atmospheric corrections are available. Most of them require detailed knowledge of atmospheric humidity, aerosols, and particles. This report describes two approaches to derive ground reflectances from satellite images without the need for detailed atmospheric observations. The first method makes use of the spectral signature of water, while the second method is based on the spectral signature of dark dense vegetation. The radiative transfer program MODTRAN is used to estimate the functional relationships between the atmospheric radiative properties and the difference between top of atmosphere (TOA) reflectance and ground reflectance for both, water and dark dense vegetation. The third method, based on the assumption of a very clear atmosphere, has been developed for the case of absence of water and dark dense vegetation in the scene. These methods have been extended and adapted for use with MERIS data.
The algorithms are applied to the different spectral channels of MERIS and to MERIS data of different geographic regions. The results show the advantages and disadvantages of the different methods when applied to MERIS data. Based on these results some guidelines are derived to decide which algorithm should be selected for which conditions, i.e. wavelength and geographic region.
The retrieved optical depth is compared to measurements of ground stations to estimate the accuracy of the different algorithms when applied to various conditions. The different methods are furthermore analyzed to determine the effect of modeling errors and the most crucial parameters and uncertainties, i.e. knowledge of the ground reflectance for the first two methods.
The clear water method works well in the infrared, since the water reflectance is known to be very small. But in the visible wave length every lake has its own spectral signature, depending on turbidity, sediments, algae etc. Therefore a simple approach is used to determine the clearest water in the scene. But nevertheless the uncertainties remain large in the visible spectrum for this method.
Since the dark dense vegetation method cannot be directly applied to the infrared channels, some form of extrapolation was necessary. Fortunately there is a strong correlation between the atmospheric radiative properties in the red and the infrared. This correlation is used to extend the dark dense vegetation method to the infrared.
The third method is restricted to very clear atmospheric conditions. It will underestimate the influence of the atmosphere for hazy conditions. This method is advantageous if applied above mountains since the ground level elevation is considered in the algorithm.
Keywords: ESA European
Space Agency - Agence spatiale europeenne,
observation de la terre, earth observation,
satellite remote sensing,
teledetection, geophysique, altimetrie, radar,
chimique atmospherique, geophysics, altimetry, radar,