Spatially Complete Global Surface Albedos Derived from Terra/MODIS Data

Michael D. King1, Eric G. Moody1,2, Steven Platnick1, and Crystal Schaaf3

1NASA 2L3-Communications

Goddard Space Flight Center GSI 3Boston University

Operational MODIS surface albedo data product (MOD43B3/MCD43B3) 0.47, 0.56, 0.67, 0.86, 1.24, 1.64, 2.1 m , 0.3-5.0, 0.7-5.0, 0.3-0.7 m 1 km spatial resolution (sinusoidal projection stored as 10 x 10 tiles) 16-day periodicity (001, 017, , 353) Limitations Spatial and temporal gaps due to cloud cover and seasonal snow Motivation Ancillary input for ground-based (AERONET), airborne, and satellite remote sensing Land surface and climate modeling Global change research projects

Conditioned MOD43B3 Albedo Maps

l = 0.858 m

IGBP Ecosystem Classification (MOD12Q1)

0 Water 1 Evergreen Needleleaf Forest 2 Evergreen Broadleaf Forest

6 Closed Shrublands 7 Open Shrublands 8 Woody Savannas

12 Croplands 13 Urban and Built-Up 14 Cropland/Natural Veg. Mosaic

3 Deciduous Needleleaf Forest

4 Deciduous Broadleaf Forest 5 Mixed Forests

9 Savannas
10 Grasslands 11 Permanent Wetlands

15 Snow and Ice

16 Barren or Sparsely Vegetated

Equal Area IGBP Ecosystem Classification Statistics

Grasslands 7.30% Savanna 6.63% Permanent Snow 11.31%

Barren / Desert 13.00%

Wetlands 0.33% Cropland 10.09%

Woody Savanna 6.32%

Open Shrubs 18.86% Closed Shrubs 0.63%

Other Veg 13.93 %

Urban 0.17% Crop Mosaic 3.34% Evrg. Needle Forest Evrg. Broad Forest 4.19% 10.38%

Mixed Forest Decd. Broad Forest Decd. Needle Forest 4.85% 1.40% 1.19%

General Methodology
Compute regional ecosystem statistics

0.5, 1-5, 10, 10x20, 10x30 box sizes

Obtain pixel-level and regional ecosystem statistical phenological trends

Curves will have different magnitudes Shapes will be consistent

Impose shape of curves onto pixel level data

Only 10-15% of pixels have adequate coverage to obtain phenology Maintains pixel level detail with appropriate phenology
Select curve that has best representation

Representation of existing pixel data Completeness of trend Bias towards smaller statistical box sizes
Fill missing values with selected curve

Processing Preparation
Prepare MOD43B3 data

Mosaic SIN data onto 1-minute climate modeling grid Apply MOD43B3 QA
Remove lesser quality pixels Remove seasonal snow

Prepare MOD12Q1 data

Mosaic SIN data onto 1-minute climate modeling grid

Compute regional ecosystem statistics

CMG grids at 0.5, 1-5, 10, 10x20, 10x30 A priori computation

Phenological Curves for Deciduous Broadleaf Forest

Vermont, USA
Phenological Curves

Phenological Curves with Offset Applied

Continental United States

July 12-27, 2002

Seasonal Snow Methodology

Cloud and snow cover obscure full decay state Over hemisphere average of high latitudes

Unique ecosystem and wavelength extrema percent change Compute % change from pixels with adequate representation
For each pixel/statistical curve

Pin winter endpoints with value computed from Percent change Summer extrema
Then apply General Methodology

Phenological Curve for Mixed Forest

Northern Russia

Europe and Western Asia

December 3-18, 2002

Persistent Cloud Methodology

Clouds obscure trends over large regions

Usually full growth stage is obscured Even 10x30 boxes may not observe complete temporal trend
Compute 1 statistical curve per ecosystem class

5-15 Latitude belts Centered around pixels latitude Yearly phenological behavior curves Instead of 2 half year curves
Impose shape of curve onto existing pixel data

Missing Season
Asian Monsoon

Indian Subcontinent during Monsoon

June 10-26, 2002

Limited Yearly Data

Tropical Cloudiness

South America during Rainy Season

June 10-26, 2002

Africa in the Presence of Persistent Clouds

December 3-18, 2002

Spatially Complete Albedo Maps

l = 0.858 m

Spectral Variability by Ecosystem Class

June 26 - July 11, 2001

Temporal Variability by Ecosystem Class

January - July 2001
Cropland Ecosystem

Urban Ecosystem

Available Products
2001 (v3) and 2002 (v4) spatially complete albedo maps Maps and QA stored in 1-minute rectangular coordinates First seven MODIS wavelengths and 3 broadband White- and black-sky albedo values 2001 (v3) and 2002 (v4) statistics of filled albedo maps 0.5, 1, 2, 3, 4, 5, 10 box sizes Statistics computed in boxes and by ecosystem class in boxes
Currently processing 2000-2004 (v4)

Statistics of snow albedos Hemispheric averages of MOD43B3 validated data Separated by ecosystem class and NISE wet/dry classification IGBP ecosystem classification map Maps and QA stored in 1-minute rectangular coordinates

Land Surface Albedo Data Distribution

modis-atmos.gsfc.nasa.gov
Acquiring data

Anonymous FTP Via website Direct link Available Albedo maps Albedo QA maps Albedo statistics Ecosystem map Snow albedo statistics NDVI Images and movies

Spectral Albedo of Snow

Used near real-time ice and snow extent (NISE) dataset

Distinguishes land snow and sea ice (away from coastal regions) Identifies wet vs dry snow
Projected onto an equal-area 1 angle grid

Aggregate snow albedo from MOD43B3 product

Surface albedo flagged as snow Aggregate only snow pixels whose composite NISE snow type is >90% is flagged as either wet or dry snow in any 16-day period Hemispherical multiyear statistics Separate spectral albedo by ecosystem (MOD12Q1)
Results represent average snow conditions

Additional sources of variability include snow depth, snow age, grain size, contamination (soot), and, in the case of black-sky albedo, solar zenith angle

Northern Hemisphere Multiyear Average (2000-2004)

Snow Albedo by Forest Ecosystems

Snow Albedo for Sparse Vegetation Ecosystems

Northern Hemisphere Multiyear Average (2000-2004)

Spatially Complete White-Sky Albedo

January 1-16, 2002
Snow-free

0.8

0.6

Snow-covered

0.4

0.2

0.0

Surface Albedo (0.86 m)

Summary and Conclusions

Spatially complete surface albedo datasets have been generated

Uses high-quality operational MODIS surface albedo dataset Imposes phenological curve and ecosystem-dependent variability White- and black-sky albedos produced for 7 spectral bands and 3 broadbands
See modis-atmos.gsfc.nasa.gov for data access and further descriptions, and

Moody, E. G., M. D. King, S., Platnick, C. B. Schaaf, and F. Gao, 2005: Spatially complete global spectral surface albedos: Value-added datasets derived from Terra MODIS land products. IEEE Trans. Geosci. Remote Sens., 43, 144-158.
Spectral albedo of snow

Hemispheric averages of MOD43B3 validated data Separated by ecosystem class and NISE wet/dry classification Additional variability due to snow depth, age, grain size, and contamination not accessible from MODIS data alone, and hence not incorporated