Characterizing Land Surfaces from MISR Measurements Michel M. Verstraete 1 , with contributions from Bernard Pinty 1 , Nadine Gobron 1 , Jean-Luc Widlowski 1 and David J. Diner 2 1 Institute for Environment and Sustainability (IES) Joint Research Centre, Ispra (VA), Italy 2 NASA Jet Propulsion Laboratory, Pasadena, CA, USA ISSAOS course in L’Aquila Thursday August 29, 2002
38
Embed
Characterizing Land Surfaces from MISR Measurements Michel M. Verstraete 1, with contributions from Bernard Pinty 1, Nadine Gobron 1, Jean-Luc Widlowski.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Characterizing Land Surfaces from MISR Measurements
Characterizing Land Surfaces from MISR Measurements
Michel M. Verstraete1, with contributions from Bernard Pinty1, Nadine Gobron1,
Jean-Luc Widlowski1 and David J. Diner2
1Institute for Environment and Sustainability (IES)Joint Research Centre, Ispra (VA), Italy
2NASA Jet Propulsion Laboratory, Pasadena, CA, USA
ISSAOS course in L’AquilaThursday August 29, 2002
OutlineOutline
• MISR and multiangular ToA observations• Surface anisotropy primer• Multiangular reflectance nomenclature• Interlude: The RPV GUI tool• MISR standard surface products• MISR non-standard surface products• Case study: AirMISR and surface
heterogeneity
Overview of MISROverview of MISR
• 9 cameras at ±70.5, ±60, ±45.6, ±26.1, 0°
• Each camera at 446, 558, 672, and 866 nm
• Spatial resolution: 275 m (250 m nadir)
• Global mode: Full res. nadir and red, 1.1 km otherwise
• Local mode: Full resolution all cameras and all bands
Gulf coast wetlands along thePascagoula, Mobile-Tensaw,and Escambia Rivers arespectrally similar to surroundingvegetation but have a distinctiveangular signature.
• Solar illumination is highly directional, especially under clear skies
• All surfaces, natural or artificial, and in particular water, soils, vegetation, snow and ice, are anisotropic
• Surface anisotropy is controlled by the structure and optical properties of the observed media
• Reflectance of geophysical media is bidirectional• Specular reflectance and hot spot, Lambertian panel• Atmospheric constituents also interact anisotropically
with the radiation fields (Rayleigh, Mie scattering)• Anisotropy is itself a spectrally-dependent property
An anisotropy primer (2)An anisotropy primer (2)
• Imaging instruments with a small IFOV sample the reflectance of the surface-atmosphere system in the direction of the sensor, measure the hemispherical-conical reflectance of the geophysical system
• These measurements thus depend on the particular geometry of illumination and observation at the time of acquisition
all sensors, including ‘nadir-looking’, are affected applications that do not exploit anisotropy must
nevertheless account for these effects unique information on the observed media (e.g.,
structural characteristics) can be derived from observations of these angular variations
Illumination and observation geometryIllumination and observation geometry
Illumination direction:Ω0 = [θ0, φ0]
Observation direction:Ω = [θ, φ]
μ0 = cos θ0
μ = cos θ
Ref: Vogt and Verstraete (2002) RPV IDL tool
Nomenclature (1)Nomenclature (1)
Ref: Nicodemus et al. (1977) NBS Monograph
Incoming Outgoing
Nomenclature (2)Nomenclature (2)
BRDF: Bidirectional Reflectance Distribution Function, Units: [sr -1], non-measurable
BRF: Bidirectional Reflectance Factor, BRDF normalized by equivalent reflectance of a Lambertian surface, non-dimensional, measurable in the laboratory as a biconical reflectance factor
HCRF: Hemispherical Conical Reflectance Factor, Units: [N/D], common measurement
Ref: Nicodemus et al. (1977) NBS Monograph
Nomenclature (3)Nomenclature (3)
HDRF: Hemispherical Directional Reflectance Factor, single integral of BRDF on the incoming directions (i.e., direct + diffuse illumination)
DHR: Directional Hemispherical Reflectance, single integral of BRDF on the outgoing directions (“black sky albedo”)
BHR: Bi-Hemispherical Reflectance (also known as albedo or “white sky albedo”), double integral of BRDF
Ref: Nicodemus et al. (1977) NBS Monograph
Families of BRF modelsFamilies of BRF models
• 3-D ray-tracing or radiosity models simulate the reflectance of realistic heterogeneous scenes (computationally expensive)
• 1-D turbid medium models simulate the reflectance of homogeneous scenes (computationally inexpensive)
• Parametric models represent the shape of the BRDF function without providing a physical explanation (computationally extremely fast)
• Appropriate inversion techniques should be selected (e.g., LUT for 3-D, non-linear iterative for 1-D, linear scheme for parametric)
Anisotropy of heterogeneous systemsAnisotropy of heterogeneous systems
• 3-D radiation transfer models (e.g., Monte Carlo ray tracing) are required to simulate heterogeneous systems
• All plant elements are represented explicitly• Input variables: size, shape, orientation and optical
properties of each individual object
Ref: Govaerts and Verstraete (1999) IEEE TGRS
Anisotropy of homogeneous systemsAnisotropy of homogeneous systems
• Plant canopies are represented as a ‘cloud’ of scatterers of finite dimension
• 1-D’ (vertical) radiation transfer models can simulate horizontally homogeneous systems, hot spot
• Input variables: number, size and orientation of leaves, leaf and soil optical properties, canopy height
Ref: Gobron et al. (1997) JGR
RPV parametric modelRPV parametric model
Ref: Rahman et al. (1993) JGR
),,,,( 00 λλλλλ ρρρ cHG
v kF ΘΩΩ=
),,(),,(),,(),,,,( 0302010 λλλλλλ ρθθρ cvHG
vvcHG
v ffkfkF ΩΩΘΩΩ=ΘΩΩ
λ
λ
θθθθθθ λ k
v
kv
v kf −
−
+= 1
0
10
01 )cos(cos
)cos(cos),,(
2/32
2
02 ])(cos21[
)(1),,(
HGHG
HGHG
v gf
λλ
λλ Θ+Θ+
Θ−=ΘΩΩ
Gf c
cv +−
+=ΩΩ11
1),,( 03λ
λρρ
Interlude…Interlude…
• Play with the RPV-GUI tool
MISR standard surface productsMISR standard surface products
• Surface BRF, albedo (BHR and DHR), and DHR-based NDVI: Provisional since Aug. 2, 2002
• LAI/FPAR and DHR-PAR: Beta since Sep. 2002• Martonchik et al. (1998) ‘Determination of land and
ocean reflective, radiative, and biophysical properties using multiangle imaging’, IEEE TGARS, 36, 1266-1281 (1998)
• Knyazikhin et al. (1998) ‘Estimation of vegetation canopy leaf area index and fraction of absorbed photosynthetically active radiation from MISR data’, J. Geophys. Res., 103, 32,239-32,256.
• Diner et al. (1999) ‘MISR Level 2 Surface Retrieval Algorithm Theoretical Basis’, JPL D-11401, Rev. D.
MISR LEVEL 2 RETRIEVALSSUA PAN, BOTSWANAAugust 27, 2000
Retrieved BRF
Blue 46º aft BRFGreen 46º aft BRFRed 46º aft BRF
Retrieved BRF
Nadir BRF46º forward BRF (backscatter)46º aftward BRF (forward scatter)
Credit: NASA/JPL MISR Team
MISR LEVEL 2 RETRIEVALSSUA PAN, BOTSWANAAugust 27, 2000
Retrieved DHR
Blue DHRGreen DHRRed DHR
Retrieved DHR
Green DHRRed DHRNIR DHR
Credit: NASA/JPL MISR Team
DHR over Southern Africa, 6-week mosaic during SAFARI 2000(Paths: 165-183)
RGB = NIR, Red, GreenRGB = Red, Green, Blue
Credit: NASA/JPL MISR Team
MISR non-standard surface productsMISR non-standard surface products
• VEGAS: FAPAR, canopy structure and heterogeneity
• Pinty et al. (2002) 'Uniqueness of Multiangular Measurements Part 1: An Indicator of Subpixel Surface Heterogeneity from MISR', IEEE TGRS, MISR Special Issue, in print.
• Gobron et al. (2002) 'Uniqueness of Multiangular Measurements Part 2: Joint Retrieval of Vegetation Structure and Photosynthetic Activity from MISR', IEEE TGRS, MISR Special Issue, in print.
• Widlowski et al. (2001) 'Characterization of Surface Heterogeneity Detected at the MISR/TERRA Subpixel Scale', GRL, 28, 4639-4642.
An Optimized FAPAR algorithm (1)An Optimized FAPAR algorithm (1)
• Traditional vegetation indices (e.g., NDVI) are subject to numerous perturbing influences, including atmospheric, soil and directional effects
• Elements exist to improve surface retrievals: canopy and atmospheric radiation models, bidirectional reflectance models, multispectral and multidirectional observations
• Exploit these tools to simultaneously rectify observations for aerosol scattering and directional effects, using MISR blue band and multiple cameras
• Estimate FAPAR from these rectified channels• Similar algorithms for SeaWiFS, MERIS, GLI and
VEGETATION (consistency and continuity)
Credit: Gobron et al. (2002) IEEE TGRS
An Optimized FAPAR algorithm (2)An Optimized FAPAR algorithm (2)
• Construct large look-up table of simulated spectral and directional reflectances for a variety of surfaces and atmospheres (ToC, ToA and associated FAPAR training data sets)
• Express rectified red and nir channels as polynomials of (blue, red) and (blue, nir) measurements, and optimize coefficients so that rectified values match simulated ToC values
• Express FAPAR as polynomial of (rect. red, rect. nir) bands, and optimize coefficients so that values match simulations
• Apply polynomials to actual MISR data
Credit: Gobron et al. (2002) IEEE TGRS
Monitoring FAPAR with MISR (1)Monitoring FAPAR with MISR (1)
• Denmark• Composite for Sep.
2000 (2 acquisitions)• Spatial resolution:
~300 m• Input: blue, red and
NIR at nadir• Algorithm: VEGAS
Credit: Gobron et al. (2002) IEEE TGRS
Monitoring FAPAR with MISR (2)Monitoring FAPAR with MISR (2)