Assessment of the Aerosol, Albedo, and Surface Type EDRs

1
 Assessment of the Aerosol, Albedo, and Surface
Type EDRs  

• Crystal Schaaf
• Alan Strahler, Feng Gao, Mark Friedl, Shunlin
  Liang
• Department of Geography, Boston University
• Department of Geography, University of Maryland
• Nov 4-6, 2003 

2
Land EDR/CDRs to be Assessed

• Aerosol EDR
• Atmospheric Correction
• Albedo EDR
• Surface Type EDR
• Interactions
• Lyapustin, Torres, Vermote, Privette, Stamnes,
  Maslanik, Loveland, Ranson, Wolfe

3
PI and CoI Roles

• PI Crystal Schaaf
• Project administration, albedo assessment
• CoI Shunlin Liang
• Aerosol and atmospheric correction assessment,
  bright target albedo estimation
• CoI Feng Gao
• Dark target albedo estimation
• CoI Mark Friedl
• Surface type assessment
• CoI Alan Strahler
• Surface type and albedo assessments

4
Approach Aerosol

• Aerosol retrieval over land ATBD
• relies on the dark-object approach.
• does not work over non-vegetated surfaces
• Assessment
• Further analysis of aerosol ATBD on the
  underlying assumptions and empirical
  relationships
• Extensive simulations and sensitivity study using
  independent packages
• Explore any new alternative algorithms
• Spatial and temporal analysis of both MODIS and
  MISR aerosol products to understand how algorithm
  performs
• Validation using Aeronet observations

5
Approach Albedo

• ATBD relies on 2 algorithms
• Dark Pixel Sub-Algorithm (DPSA)
• Modis BRDF/Albedo algorithm (Schaaf et al., 2002)
• Bright Pixel Sub-Algorithm (BPSA)
• TOA-BOA (Liang, 2003)
• Both algorithms are computed globally
• Assessment
• Sensitivity of daily DPSA to inputs, BRDFs, N2B
• Comparisons between DPSA and BPSA
• Determine situations for the best use of each
  algorithm
• Investigate the role of multi-satellites
• Demonstrations based on Aqua and Terra
• Validations and comparisons
• BSRN

6
DPSA Albedo
MODIS Shortwave White-Sky Albedo June 10-25, 2003
(2000161)
Terra-only
Aqua plus Terra
0.0 0.25
7
DPSA Albedo
MODIS Shortwave White-Sky Albedo Quality Flags
June 10-25, 2003 (2000161)
Terra-only
Aqua plus Terra
Green Best Quality Blue Moderate Quality
Red Poor Quality BlackFill
8
BPSA Albedo

Color compositing imagery of MODIS data acquired
on May 7, 2000 over China (A) and three broadband
albedos using a bright surface albedo retrieval
algorithm (Liang, 2003) total shortwave (B),
visible (C) and near-IR (D).
9
Approach Surface Type

• Training database
• UMD database
• STEP Training database
• 2100 sites (monitored and updated)
• ATBD relies on decision tree classifier
• C5.0 (commercial)
• MODIS Land Cover Algorithm (Friedl et al., 2002)
• OCI (freeware)
• Assessment
• Effects of input accuracies
• MODIS relies on NBARs
• Output accuracies and confidences
• Feature selection
• Ancillary schemes desired
• Consistency with cover per type

10
Surface Type MODIS MOD12 IGBP
MODIS data from Nov 00Oct 01
11
Surface Type IGBP Training Sites
12
Evaluation Approach

• MODTRAN simulations
• MODIS
• Aerosol-gtAtmospheric Correction-gtBRDF-gtAlbedo-gtNBA
  R
• Other remotely sensed data (MISR etc.)
• Field data

13
Deliverables

• Algorithm descriptions
• Validation plans
• Quality Assessments