Application and appraisal of a visual estimation technique for composition and yield sampling of grass-legume pastures in the wet tropics of north-eastern Australia

Abstract

A visual estimation technique (SVET) was applied to assess the yield and botanical composition of the rapidly growing, highly seasonal and morphologically diverse pastures of the wet tropics. The technique uses the comparative yield method for total dry matter yield, and the framework of the dry-weight-ranking procedure for botanical composition. Widespread use of cumulative ranking and divided ranking gave a close numerical description of each quadrat. Methods of visual estimation based on dry-weight ranking were evaluated for pastures dominated by sown grass and including a sown legume as a minor but important component. Although the usual dry weight ranking method showed large departures from the known composition, SVET was able to assess the pastures with small overall departures from results obtained by hand-sorting. The full theoretical potential of SVET could not be attained in the field, where ranks had to be estimated. Nevertheless, with sufficient quadrats and trained operators, SVET can assess composition and yield of these tropical pastures with acceptable precision. The relationship between total yield and visual rating was linear but the regression equation for data quadrats departed from its corresponding calibration equation in either slope or coincidence in 11 of 34 cases.