Satellite derived evidence of whole farmlet and paddock responses to management and climate

Abstract

Satellite imagery was used to assess differences between three treatments in a grazing enterprise systems study of three 53-ha farmlets on the Northern Tablelands of New South Wales, Australia. The study involved a comparison between a typical control farmlet (B) with one with higher levels of sown pasture and soil fertility (A) and one employing intensive rotational grazing (C). Landsat thematic mapper data were used to derive normalised difference vegetation index (NDVI) and spectral class images for eight dates from before the commencement of the farmlet trial (June 2000) to annual spring measurements in September-October of each year from 2000 to 2006 across all paddocks of each farmlet. The Landsat imagery taken before the commencement of the farmlet treatments (June 2000) showed only small differences between the three farmlets, confirming that the allocation of land to the farmlets had been without bias. The assessments using Landsat NDVI in spring over 7 years showed differences in green herbage resulting from the variation in rainfall received over different years as well as differences between the farmlets. The Landsat NDVI images showed increasing and significant differences in pasture greenness over time, especially between Farmlet A and Farmlets B and C. In addition, there were significant differences in pasture spectral classes between Farmlet A and Farmlets B and C, with a significant correlation with higher levels of sown perennial and annual grasses and legumes on Farmlet A. Using different statistical tools, several relationships were found between NDVI and spectral class data and explanatory variables of farmlet, paddock, sowing phase, modelled soil moisture and recent grazing activity. The moderate resolution Landsat data across the entire area of each farmlet proved to be especially useful for assessing pastures within every paddock used in this farmlet study. In addition, moderate resolution imaging spectro radiometer NDVI satellite data were collated for weekly intervals from September 2003 to December 2006 in order to assess seasonal pasture growth patterns on each of the farmlets. These patterns were significantly correlated with a growth index calculated from temperature and available soil moisture, and showed that the growth on the three farmlets was closer to a highly productive reference paddock than a low input, unsown pasture in another reference paddock. The satellite data facilitated the detection of significant differences in pasture botanical composition, soil fertility, grazing management, climate and season. The ready availability of quality remote sensed imagery, combined with the significance of the relationships established, confirms that the technology is a valuable objective tool for both farming systems research and for managing entire farms.