Related Research Areas
Carbon Cycle & Ecosystems

The monitoring and modeling of the terrestrial biosphere within the larger context of climate variability and change studies requires global multi-decadal time series of key variables characteristic of vegetation structure and functioning. Leaf Area Index (LAI) is a key biophysical variable that controls the exchange of energy, mass (e.g. water and CO2) and momentum between the Earth surface and atmosphere. In this study, we aim to generate global 30-m LAI from Landsat surface reflectance data using the radiative transfer theory of canopy spectral invariants which facilitates parameterization of the canopy spectral bidirectional reflectance factor (BRF).

The two phases of the project are as follows:

1) Landsat surface reflectances and land-cover map: Processing of Landsat Global Land Survey (GLS) scenes are performed using LEDAPS for radiometric calibration and atmospheric corrections to derive first-pass surface reflectances (SRs) for year 2005. A next key component was to adjust the SRs of GLS data, which have different dates of acquisition (DOA), to approximate its peak-growing-season (PGS) level when vegetation reaches its maximum greenness. This was performed utilizing the downscaled MODIS NBAR data, which share similar observation geometry w.r.t the Landsat sensor, irrespective of input uncertainties in spectral data. Additionally, a downscaled 30-m land-cover map was created utilizing the MODIS Collection 5 land cover product and GLOBCOVER data.

2) Global Leaf Area Index (LAI) dataset: The adjusted Landsat reflectance to its PGS level (from step 1) and the 30-m land cover map are utilized to retrieve LAI. The radiative transfer theory of canopy spectral invariants provides the required Bidirectional Reflectance Factor (BRF) parameterization via a small set of well-defined measurable variables. The algorithm supports a flexible mode of operation, wherein retrievals of LAI can be obtained based on any number of given bands. Biome-specific Look-up-tables (LUTs) is created, which stores the simulated spectral BRF at the red, NIR and SWIR bands as a function of soil reflectance, LAI and view/azimuth angles. Implementation of a 3-band (Red, NIR and SWIR) and 2-band (Red and NIR) LAI inversion scheme for each 30-m Landsat pixel is performed (Fig. 1a & 1b). Based on results and from previous studies, the SWIR adds a significant bit of information related to background effects (dark/ wet background vs. bright background). Implementation is performed globally for all the GLS 2005 scenes and will be extended to other years (see Fig. 1c for a tile scale implementation).

Figure 1a Figure 1a Figure 1b Figure 1b Figure 1c Figure 1c

Figure 1. Panel (a) shows NDVI vs. retrieved LAI using red, NIR and SWIR. Panel (b) same as (a) but with red and NIR only. Panel (c) shows Landsat LAI for 2005 PGS over northern California.


NDVI vs. LAI using red, NIR and SWIR
17.5 KB
LAI using red and NIR
18.3 KB
Landsat LAI for 2005 PGS
36.6 KB


Everyone's Tags

Popular Resources

Nothing to see here at the moment. Check back later.


Log in to start a discussion.

Project Highlight

7 members

Started: Jun 05, 2010

Last Activity: Jun 05, 2010


What can I do on this project?
  • Any registered users can join
  • Anybody can view this project
  • Any registered users can leave comments
  • Anybody can view comments

New Member

Need help?

Visit our help center