| Abstract: |
For many radiometer-based parameter retrievals, brightness temperature observations have to be separated into soil and vegetation contributions [1]. This is commonly done by solving the widely applied τ-ω model for vegetation optical depth τ, effective scattering albedo ω and soil reflectivity r [2]. The precise estimation of these parameters is critical for soil moisture estimation and retrieval of biophysical properties of the vegetation canopy, like biomass or vegetation water content [3,4].
In this research contribution a physical framework is established to combine vegetation optical depth (retrieved with the τ-ω model) and information about canopy thickness (derived from ICESat lidar data). The above-named framework is applied to derive on physics basis the vegetation loss coefficients and invert them for canopy penetration depths, which provide complementary information on canopy radiative transfer characteristics with respect to vegetation optical depth. This expands the description of attenuation from absolute values towards relative terms with respect to vegetation height. The described framework will be applied on L-, C- and X-band data.
For this study, almost two years of global SMAP L-band and AMSR2 C- and X-band brightness temperatures were processed and analyzed. Vegetation heights were acquired from ICESat lidar measurements [5] and assumed to be constant over the study period. The τ-ω model was solved using an optimization based multi-temporal dual channel algorithm (MT-DCA) originally established for L-band in [6]. Vegetation parameters (τ,ω) were retrieved for all three frequencies. However adapting the MT-DCA Algorithm to higher frequencies (C- and X-band) remains a point of active research.
From the L-Band retrieval results, we find physically meaningful patterns for τ with maximum attenuation for dense tropical and boreal forest. Focusing on forest areas for further analysis, penetration depth results indicate that for L-band a sufficiently strong soil signal can be obtained even under tropical vegetation. For the C- and X-band retrievals, we find meaningful τ values for selected areas of interest despite the parameters within the τ-ω model become more effective with respect to L-Band [2,6]. It reveals that penetration depths generally decrease with increasing frequency, while differences in penetration depth for the selected areas of interest remained relatively similar. Temporal dynamics of τ are generally in agreement with the annual growth cycle for all three frequencies. This shows that L-, C- and X-band can be used to monitor plant canopy growth, while each frequency is sensitive to a different canopy volume due to varying penetration.
[1] E.G. Njoku and D. Entekhabi, “Passive microwave remote sensing of soil moisture,” Journal of Hydrology, 184, pp. 101-129, October 1996.
[2] T.J. Jackson, T. Schmugge and J.R. Wang, “Passive microwave remote sensing of soil moisture under vegetation canopies,” Water Resources Research, 18, pp. 1137-1142, August 1982.
[3] S. Paloscia, E. Santi, P. Pampaloni and S. Pettinato, “Multifrequency microwave emission for estimating optical depth and vegetation biomass,” 2016 IEEE International Geoscience and Remote Sensing Symposium (IGARSS), Beijing, pp. 5296-5299, July 2016.
[4] P. O'Neill, A. Joseph, P. Srivastava, M. Cosh and R. Lang, "Seasonal parameterizations of the tau-omega model using the ComRAD ground-based SMAP simulator", 2014 IEEE Geoscience and Remote Sensing Symposium, Quebec City, pp. 2423-2426, 2014.
[5] M. Simard, N. Pinto, J.B. Fisher, A. Baccini, “Mapping forest canopy height globally with spaceborne lidar,” Journal of Geophysical Research: Biogeosciences, 116, pp. 2156-2202, November 2011.
[6] A. G. Konings, M. Piles, N. Das, D. Entekhabi, “L-band vegetation optical depth and effective scattering albedo estimation from SMAP,” Remote Sensing of the Environment, 198, pp. 460-470, July 2017.
|
