The spectral response of pastures in an intensively managed dairy system

All grazing-based industries require information on their feed resources in order to manage them optimally. Gathering this information through traditional methods for measuring pasture biomass is time-consuming and error-prone, resulting in increased interest in remotely-sensed methods. Remote sensing used to monitor feed resources in farming systems differs from remote sensing of systems such as forestry because of how the time-scale of management practices impacts on the growth rate and accumulation patterns of biomass. Also, in operational systems, designed for near real-time delivery to end-users of quantitative pasture measurements, we are restricted to the commercially available broad-band high-resolution sensors. The goal of this paper is to understand how remotely-sensed observations of pastures in an intensively managed dairy system change in relation to intensive management practices, so that better image analysis and ground-validation methods can be developed for measuring and monitoring such systems. At two dates in the growing season we examined high-resolution (SPOT-5 and Ikonos) images of an intensively managed perennial dairy farm in Victoria (Australia). We showed that the observed spectral response in the images varied with the length of time since the paddock was grazed, consistent with the re-growth of pastures post-grazing. The operational remote sensing of pastures is often restricted by the range of spectral bands that are available on broad-band sensors. However, these results suggest that when choosing a vegetation index for intensively managed dairy pastures it should incorporate the short-wave infrared (SWIR) band to improve observations of recently grazed pastures and tune analyses based on the spectral response