Modified Tau-Omega Model for Moderately to Densely Vegetated Landscapes
October 26, 2012 15:26:14
Description of Problem
Soil moisture (SM) is recognized as an important component of the water, energy, and carbon cycles at the interface between the Earth’s surface and atmosphere, yet it is difficult to measure globally using traditional in situ techniques. Several planned microwave space missions, most notably ESA’s Soil Moisture Ocean Salinity (SMOS) mission (launched November 2009) and NASA’s Soil Moisture Active Passive (SMAP) mission (to be launched 2014/2015), are focusing on obtaining accurate SM information over as much of the Earth’s land surface as possible. However, current baseline retrieval algorithms for SMOS and candidate retrieval algorithms for SMAP are based on an easily implemented but theoretically simple zero-order radiative transfer (RT) approach which includes components from the soil and vegetation, but ignores vegetation scattering except for the effect of the scatterers in the attenuation of the emission through the vegetation canopy. This approach essentially places a limit on the density of the vegetation through which SM can be accurately retrieved. Our proposed work involves the development of a new SM retrieval model which could potentially overcome this limitation and thus could be used with SMAP and SMOS data to increase the accuracy and reliability of SM products over moderately to densely vegetated landscapes.
Both SMOS and SMAP have mission requirements to retrieve SM to an accuracy of 0.04 cm3/cm3 through vegetation water content (VWC) of 5 kg/m2. These missions are expected to meet their requirement for SM retrieval accuracy using the heritage tau-omega model (zero-order RT solution) approach over approximately 65 % of the Earth’s land surface where the VWC does not exceed 5 kg/m2. As the density of vegetation increases, sensitivity to the underlying SM begins to degrade significantly and errors in the retrieved SM increase accordingly. Thus, knowledge of L-band vegetation features appears to be of great importance when the tau-omega approach is applied to dense vegetation (i.e. forest, mature corn, etc.) where scattering from branches and trunks (or stalks in the case of corn) is likely to be very important.
Our proposed new model is a first-order scattering RT model for microwave radiometry of vegetation at L-band. The model is based on an iterative solution (successive order of scattering) of the RT equations up to the first-order. This formulation adds a new scattering term to the tau-omega model. The additional term represents emission by particles in the vegetation layer and emission by the ground that is scattered once by particles in the layer. The resulting model represents an improvement over the standard zero-order solution since it accounts for the scattered vegetation and ground radiation that can have a pronounced effect on the observed emissivity and subsequent SM retrieval. Although the new approach would add another parameter to the list of unknowns in the inversion procedure to retrieve SM from microwave measurements, it has the advantage that the formula relating SM is physically-based, and as a result, should be more robust under varying conditions.
Scientific Objectives and Approach
Global measurements and interpretation of SM products might be best accomplished by a combination of ground-based and spaceborne techniques. The first stage of our research will be to focus on the ground-based ComRAD radiometer data because of the extensive heritage that this type of observation has in SM applications. Building on our SM investigations using ComRAD observations, we will evaluate how the proposed new algorithm could be applied to spaceborne data for densely vegetated areas which are normally beyond the baseline SM retrieval range of such space missions. To test the performance of the new model against the candidate SMAP algorithms in the years prior to the SMAP launch, we plan to use SMOS data as proxy SMAP data.
In the proposed research, a first-order RT model will be developed to investigate the relationship between scattering mechanisms within vegetation canopies (with large scatterers) and the microwave brightness temperature. The model will then be used to perform a physical analysis of the scattered and emitted radiation from vegetated terrain. The main goal of the proposed work is to more accurately account for vegetation canopy scattering by modifying the basic tau-omega model to include the first-order scattering term, and then to assess the performance and operational usefulness of both versions of the tau-omega model in retrieving SM. Specifically, the proposed work involves the following approaches:
To determine the impact of scattering at L-band on the microwave retrieval of SM under vegetation using existing ground-based measured field data,
To obtain some insights into the first-order scattering solution regarding possible emission/scattering mechanisms,
To assess whether modifications are necessary to the tau-omega model in terms of form or quantification of parameters,
To determine appropriate values for vegetation parameters which produce the most accurate SM retrievals under moderately and densely vegetated terrain,
To implement and to assess the usefulness of the new modeling approach to improve the accuracy of large scale SM retrievals through different types of vegetated landscapes, where the VWC exceeds 5 kg/m2.
A first-order radiative transfer model was developed to more accurately account for vegetation canopy scattering by modifying the basic zero-order RT (tau−omega) model, which is used in soil moisture retrieval algorithms.
Effective tree scattering and opacity parameters were evaluated with theoretical definitions of these parameters for forest canopies and a new effective scattering albedo formulation is developed for soil moisture retrievals.
Antenna illumination effects such as platform height on the measurements of the backscattering coefficient by a ground-based scatterometer (which plays an important role in development and validation of soil moisture products) were demonstrated.
Refereed Journal Publications
M. Kurum, P. E. O’Neill, R. H. Lang, A. T. Joseph, T. J. Jackson and M. H. Cosh, “Effective tree scattering and opacity at L-band”, in press, Remote Sensing of Environment, vol. 118, pp. 1-9, March 2012
M. Kurum, P. E. O’Neill, R. H. Lang, M. H. Cosh, A. T. Joseph, and T. J. Jackson, “Impact of forest litter on forest emission at L-band: modeling the forest floor”, in press, IEEE Transaction on Geosciences and Remote Sensing, vol. 50, no. 4, pp. 1-14, April 2012.
M. Kurum, R. H. Lang, P. E. O’Neill, A. T. Joseph, T. J. Jackson and M. H. Cosh, “A first-order radiative transfer model for microwave radiometry of forest canopies at L-band”, IEEE Transactions on Geoscience and Remote Sensing, vol. 49, no. 9, pp. 3167 – 3179, September 2011 (cover article).
Other Publications and Conferences
M. Kurum and P. E. O’Neill, “Backscatter measurements over vegetation by ground-based microwave radars,” in Proceedings, XXXth URSI General Assembly and Scientific Symposium, Istanbul, Turkey, August 13 – 20, 2011 (invited).
M. Kurum, P. E. O’Neill, R. H. Lang, A. T. Joseph, M. H. Cosh, and T. J. Jackson, “Effective tree scattering at L-band”, in Proceedings, International Geosciences and Remote Sensing Symposium, Vancouver, Canada, July 24 – 29, 2011.