Back to R&D main

Heat load forecast service – 2009/2010

Project start date: 01 October 2009
Project end date: 28 June 2010
Publication date: 01 August 2010
Livestock species: Grainfed cattle
Relevant regions: National
Download Report (4.4 MB)


The Feedlot Cattle Heat Load Forecast Service has been operational for 9 years and has expanded to cover 91 locations across Australia. The service has been providing feedlot operators with warnings of impending adverse weather conditions that could lead to excessive heat loads for feedlot cattle. As the service progresses into its tenth year of operation some limitations to the system have been identified by MLA and feedlot operators, these wereReliability of daily upload of forecasts to the website has become an issue with the expansion of the forecasting to over ninety sites
Poor performance at sites not located near a Bureau of Meteorology (BoM) Automatic
Weather Station (AWS) site or located between two AWS sites giving conflicting forecasts AWS site locations not being representative of feedlot conditions

The first limitation was due to the web site being hosted by a third party. Forecast data had to be uploaded from Katestone Environmental before the forecast was made available to feedlot operators. As the number of sites progressively increased over time the amount of data being transferred became too large for the third party to process and the connection was lost. To address this issue Katestone Environmental will now be hosting the website in house on a dedicated web server attached to an onsite data store. The system will also be backed up offsite, meaning that even in the event of catastrophic failure of the Katestone server the forecast will still be available.
The fact that the current system is limited to BoM AWS sites is inherent in its design. The use of statistical models was necessary during the original systems development to downscale the relevant meteorological parameters from model data supplied by the BoM. These downscaled data then required training to determine the line of best fit for the multiple parameters, meaning that the data are then only representative of the location where the training data was measured and not where the actual feedlot is located, for which the forecast is intended.
Katestone Environmental developed a dynamical model capable of addressing this issue by simulating the environmental parameters used to calculate the HLI, for the entire Australian continent, on a low resolution grid with a resolution of 25 km and two high resolution grids at 9 km. These grids provide forecasts for any location in Australia to within 4.5 km for the high resolution grid and 12.5 km for the coarse grid.
During the 2009-2010 forecast season development of the dynamical model began and performance testing was carried out on two test periods. This initial developmental stage was undertaken to establish the dynamical models limitations and directions for improvement and to compare its performance against the current system. The results indicate that the dynamical model performs as well as the current system in predicting the HLI and associated meteorological parameters. While the current system is at the end of its development capacity the new system shows substantial scope for improvement.

More information

Project manager: Des Rinehart
Primary researcher: Katestone Environmental Pty Ltd