Continuous Fields Tree Cover Project MODIS

Available at:

http://glcfapp.umiacs.umd.edu:8080/esdi/index.jsp

Data reference:

University of Maryland, Global Land Cover Facility (GLCF)

Year:

2003

  Version:   1.0

Available data products:

Products

Percentage cover for Treecover, bare, and herbaceous for each continent
Percent cover of life form (woody vegetation, herbaceous vegetation, or bare ground), leaf type (needleleaf or broadleaf) and leaf longevity (evergreen or deciduous)
Global and continental (Africa, Oceania, Eurasia, North America, South America) products
MODIS Vegetation Continuous Fields product (MOD44)
  Products
  Projection

Africa, Oceania, Eurasia, North America, South America

  Interrupted Goode Homolosine
  Geographic (Lat/Long)

Scale/spatial resolution:

0,5km

Classification system:

Proportional cover scale
IGBP based land cover classification

Remote sensing data:

MODIS/TERRA (2000-2001)

  Interpretation process:
  Automated classification procedure using a regression tree algorithm
 
Using high-resolution imagery (Landsat ETM+, IKONOS) to derive global training data
 
Using  training data as dependent variable, predicted by independent variables in form of
  annual MODIS metrics
 
Employing these training data and phenological metrics with a regression tree to derive
  global percent cover
 
Outputs from the regression tree are further modified by stepwise regression and bias
  adjustment

Projection:

Interrupted Goode Homolosine and Geographic (Lat/Long)
  Validation:
  Interpretation of field data along with IKONOS and Landsat ETM+ data
 
Establishment of test areas with size of an ETM+ image
 
Crown cover maps of IKONOS images are binned to ETM+ cells to map percentage crown
  cover for 30m pixel
 
The retrieved ETM+ crown cover map is averaged to a 500m resolution to validate the
  MODIS map
  Objective:
  To support change detection studies at the global scale
 
To continue with and improve the continuous fields product series, formerly retrieved from
  AVHRR data now using more detailed spatial/spectral characteristics of the MODIS data
  File Format:
  GeoTIFF

Data release policy:

Open access

Costs:

For free

  Updates:
  intended
  Contact:
  John Townshend, University of Maryland (jt59@umail.umd.edu)