Mapping leaf area index over semi-desert and steppe biomes in kazakhstan using satellite imagery and ground measurements
Propastin, Pavel
Kappas, Martin
8, 1: 75 - 92
Propastin, Pavel; Kappas, Martin, 2009: Mapping leaf area index over semi-desert and steppe biomes in kazakhstan using satellite imagery and ground measurements. In: EARSeL eProceedings, Band 8, 1: 75 - 92, DOI: 10.23689/fidgeo-2661.
|
View/
|
Maps of peak seasonal leaf area index (LAI) were produced using the Normalised Difference Vegetation Index (NDVI) from SPOT VEGETATION (VEG) satellite at 1 km resolution over a large region in the semi-arid zone of Kazakhstan. Ground measurements of LAI were acquired using indirect and direct techniques across a 150·150 km2 large region. A Landsat Enhanced Thematic Mapper (ETM+) scene at 30 m spatial resolution was used to locate ground sites and to facilitate spatial scaling to 1 km pixels. A high-resolution LAI map retrieved from the Landsat ETM+ data was aggregated to 1 km resolution and afterwards used as reference data. The methods tested for transfer function between ETM+ LAI and SPOT-VEG were ordinary least squares (OLS) regres-sion, non-linear regression, and reduced major axis (RMA) regression. In this paper, final maps of peak season LAI at a 1 km resolution are presented after an assessment of their accuracy using the aggregated ETM+ LAI scene. The most appropriate results were attained by RMA. Advantages and shortages of the used regression approaches were analysed and discussed. Errors were mostly caused by uncertainties in co-registration of Landsat ETM+ and SPOT-VEG images as it was demonstrated by a pixel degradation experiment. The methodology presented in this paper can serve as a basis for generation of medium- and coarse-resolution LAI satellite products for wide areas of Central Asia and Kazakhstan. The study exposed a general transferability of the de-veloped model for LAI estimations at coarser scales. The 1000 m SPOT-VEG model has proved to be fully suitable for utilising with the SPOT-VEG data with resolution of 2 km.
Statistik:
View StatisticsCollection
- Geologie [927]