Estimating leaf area index in savanna vegetation using remote sensing and inverse modelling
Bowyer, P 2005, Estimating leaf area index in savanna vegetation using remote sensing and inverse modelling , PhD thesis, University of Salford.
Download (64MB) | Preview
Leaf area index (LAI), defined as the one sided green leaf area per unit ground area, is a key parameter in ecosystem process models. Owing to the large area of the earth's surface that they occupy, savanna ecosystems represent the third largest terrestrial carbon sink. There is considerable uncertainty however, as to the functioning of these ecosystems, particularly as they respond to land cover changes. Consequently, ecosystem process models constitute one of the best methods available for investigating the effect this may have on terrestrial carbon cycling. If these models are to be used over large areas however, they need to be parameterised. This thesis develops a methodology to estimate LAI in savanna ecosystems, using remotely sensed earth observation (EO) data, laboratory bidirectional reflectance measurements (BRDF), physically based canopy reflectance models (CRMs), and artificial neural networks (ANN). First, the scattering behaviour of Kalahari soils was characterised, by making laboratory BRDF measurements. Soils were shown to be highly non-Lambertian. These measurements were then used to parameterise three different CRMs. Modelled reflectances were assessed with respect to Landsat ETM+ and Terra-MODIS reflectances. Results showed that a 1-D turbid medium provided the closest fit to the measurements. A series of model sensitivity analyses (SA) were performed, and it was shown that reflectance in the red and shortwave infrared displayed greatest sensitivity to LAI, sensitivity in the near-infrared was negligible. Model inversions were performed with ANN and different waveband combinations, and LAI was estimated. The results showed that LAI could be estimated with high accuracy, an RMSE of 0.3 1, and 0.18, from ETM+ and MODIS measurements, respectively. These results were promising, and with further improvements to models, coupled with more accurate input data, will see the use of EO data play an increasingly important role in understanding the functioning of these savanna ecosystems.
|Item Type:||Thesis (PhD)|
|Uncontrolled Keywords:||Ecology, Africa, Savannah, Ecosystems|
|Themes:||Subjects / Themes > Q Science > QK Botany|
Subjects / Themes > Q Science > QH Natural history > QH001 General, inc. conservation, geographical distribution
Subjects / Themes > Q Science > QH Natural history > QH301 Biology
Subjects outside of the University Themes
|Schools:||Colleges and Schools > College of Science & Technology > School of the Built Environment|
|Depositing User:||Institutional Repository|
|Date Deposited:||08 Sep 2009 15:42|
|Last Modified:||07 Apr 2013 12:38|
Document DownloadsMore statistics for this item...
Actions (login required)
|Edit record (repository staff only)|