Estimating leaf area index (LAI) of gum tree (Eucalyptus grandis X camaldulensis) using remote sensing imagery and LiCor-2000.

Thesis (M.Sc.)-University of Natal, Pietermaritzburg, 2001.

The use of remotely sensed data to estimate forest attributes involves the acquisition of ground forest data. Recently the acquisition of ground data (field based) to estimate leaf area index (LAI) and biomass are becoming expensive and time consuming. Thus there is a need for an easy but yet effective means of predicting the LAI, which serves as an input to the forest growth prediction models and the quantification of water use by forests. The ability to predict LAI, biomass and eventually water use over a large area remotely using remotely sensed data is sought after by the forestry companies. Remotely sensed LAI values provide the opportunity to gain spatial information on plant biophysical attributes that can be used in spatial growth indices and process based growth models. In this study remotely sensed images were transformed into LAI value estimates, through the use of four vegetation indices (Normalized Difference Vegetation Index (NDVI), Corrected Normalized Difference Vegetation Index (NDVlc), Ratio Vegetation Index (RVI) and Normalized Ratio Vegetation Index (NRVI). Ground based measurements (Destructive Sampling and Leaf Canopy Analyzer) relating to LAI were obtained in order to evaluate the vegetation indices value estimates. All four vegetation indices values correlated significantly with the ground-based measurements, with the NDVI correlating the highest. These results suggested that NDVI is the best in estimating the LAI in Eucalyptus grandis x camaldulensis in the Zululand region with correlation coefficients of 0.78 for destructive sampling and 0.75 for leaf canopy analyzer. Visual inspection of scatter plots suggested that the relations between NDVI and ground based measurements were variable, with R2 values of 0.61 for destructive sampling and 0.55 for Leaf Canopy analyzer. These LAI estimates obtained through remotely sense data showed a great promise in South African estimation of LAI values of Eucalyptus grandis x camaldulensis. Thus water use and biomass can be quantified at a less expensive and time-consuming rate but yet efficiently and effectively.