DEVELOPING A METHODOLOGY FOR ESTIMATING GRASSLAND VARIABLES WITH REMOTELY SENSED DATA

Remotely sensed data in a digital format can be used to estimate several variables which indicate the status of pasture. An empirical modelling methodology is developed through three studies, in North Derbyshire, England; in South Australia and in New South Wales, Australia using airborne ATM data, SPOT HRV data and Landsat MSS data respectively. The collection and processing of the ground data and the remotely sensed data which are used in the regression models are discussed. The use of regression techniques in this type of analysis, where errors occur in the data, is considered. Green leaf area index, green biomass and green cover were successfully estimated but the ratio vegetation indices were not good indicators of brown vegetation.