Estimating pasture land cover in the New England region of Northern New South Wales

摘要

Land cover across the southern Australian temperate agricultural region comprises primarily of native pasture; introduced improved pastures and crops for livestock production and also perennial remnant vegetation. A feed-base pasture audit was carriedout throughout southern Australia commencing mid-year 2011 (Donald and Burge 2012; Donald et al. 2012). The purpose of the audit was to map and analyse information obtained about the pasture feed-base for livestock production by surveying Statistical Local Areas (SLAs) across the southern states. The purpose of this Feed-Base audit was to survey pastures within agricultural NSW, Victoria, Tasmania, South Australia and Southwestern Australia, collate these data into an organised database, and prepare ashort report and summarise by tabulating and mapping pasture species abundance and distribution. Data collected were based on "desk-top estimates" by state district agronomists and agricultural consultants. In this paper a method using satellite imageryis described on how more objective assessments of pasture types can be provided as a means to discriminate between the SLA's major pasture classes far more objectively than by visual assessment. Satellite remote sensing may be used to define landcover classes for large regional areas. A number of procedures have been developed to discriminate between pastures, crop and woody vegetation (for example Hill etal. 1997, Emelyanova etal. 2008). In the Hill study (Hill et al. 1997) NOAA AVHRR NDVI provided spatial land cover maps of pasture cover at 1 km resolution. The classifications results in that study showed that satellite information may be used to help in the interpretation of pasture survey results, and in turn, the survey data can provide some validation data for the pasture types ascribed to the remotely sensed classes.