Retrieval of temporal Leaf Area Index data sets for grassland in Central Kazakhstan using satellite data and in situ measurements

Leaf area index (LAI) is a biophysical variable that have importance in climate, weather, and ecological studies. LAI has been effectively estimated from remote sensing measurements. This paper reports about an approach for determining temporal data sets of LAI over a wide area in Central Kazakhstan using fine-resolution imagery of Landsat ETM+ and coarse resolution 10-day time-series of the normalized difference vegetation index (NDVI) from SPOT VEGETATION sensor over the period of 1998-2007. The LAI algorithm used in this study bases on the physical radiative transfer model which establishes a relationship between LAI and given pattern of surface reflectance, view-illumination conditions, optical properties of vegetation and other parameters. Optical properties of vegetation cover were represented by a number of factors including the extinction coefficient (k), the canopy projection coefficient (G) and the clumping index ( ) which are strongly dependent on vegetation type. The coefficients G and were computed from the measurements of crown architecture and hemispherical photographs during field surveys. NDVI and a linear mixture model were applied to calculate fractional vegetation cover . LAI was calculated using a non-linear relationship to and than compared with ground truth measurements made in 20 plots using hemispherical photography at the peak of growing season in June, 2008. Calculated LAI, corrected with a measured clumping index ( ), was highly correlated with measured LAI (R² = 0.75, p < 0.01). This approach was used to produce a 30-m resolution LAI map of grassland. The procedure was then applied to the SPOT VEGETATION data set, where the influence of view-illumination conditions on the extinction coefficient was simulated by a view angle geometry model that incorporates the solar zenith angle and the sensor viewing angle as the input variables. Histograms of resulting LAI distributions and descriptive statistics at different spatial resolutions are compared. LAI spatial distribution at coarse resolution was similar to that obtained at fine resolution and remained close to being normally distributed.
The work was carried out by logistical and technical support of "Drylands Management Project" founded by the govenment of Kazakhstan.