Summary
One of the issues that needs to be addressed in managing feedlots is the possibility of cattle deaths due to heat stress brought on by adverse weather conditions. One tool for managing heat stress is to forecast stress inducing conditions for a prescribed future period. In the summer of 2001-02, Katestone Environmental developed a forecasting system for MLA to predict a cattle heat stress index out to 6 days ahead for four sites in Queensland and New South Wales. Meteorological data were obtained on a daily basis from the on-site meteorological stations and the nearest Bureau of Meteorology automatic weather station (AWS). The Temperature Humidity Index (THI, an indicator of heat stress) was calculated from these data and made available to feedlot operators.
The forecasting study was expanded over the summer of 2002-03 to incorporate a Heat Load Index (HLI) developed specifically for feedlot cattle and to extend the coverage to 14 sites across eastern Australia. The service was expanded for the 2003-04 summer period with the addition of Katanning (Western Australia), again in 2004-05 to include Charlton in Victoria and also to incorporate a revised HLI algorithm and the Accumulated Heat Load Unit (AHLU). In 2005, the service was again expanded to include the site at Cessnock, NSW. The present study (2006-07) includes the following 17 sites:
Queensland – Amberley, Emerald, Miles, Oakey, Roma, Warwick;
New South Wales – Albury, Armidale, Cessnock, Griffith, Hay, Moree, Tamworth, Yanco; South Australia – Clare;
Western Australia – Katanning; and Victoria – Charlton.
Key issues
The key issues in implementing a viable feedlot weather forecasting system include:
(a) Identification of primary and derived meteorological parameters that indicate excessive heat load in cattle.
(b) Selection of methodology for predicting primary and derived parameters at AWS locations for a suitable time horizon.
(c) Development of a forecasting software system for predicting feedlot conditions.
(d) Making the forecasting results available to all feedlot operators on a daily basis.
At the outset, the following constraints were identified:
Bureau of Meteorology AWS sites are not generally in close proximity to feedlots and this limits the utility of forecasts made from these sites. Most AWS sites are situated near significant populations or industrial regions and as such only 17 sites were identified to be in close proximity to feedlot operations.
The Bureau of Meteorology’s weather forecast model data (LAPS and GASP), necessary to conduct a forecast, is only stored by the Bureau of Meteorology when requested. Therefore the models created for the recently added sites (viz. Cessnock and Charlton) were based on a small amount of historical LAPS/GASP data, which can affect model performance.
It was found that the most effective technology for making the forecasts available to feedlot operators was through the World Wide Web. The advantages are that the data can be presented in a way which is easily interpreted and is readily accessible by all feedlots.
Selected methodology
The following methodology was adopted following discussions between MLA and Katestone Environmental on the most viable options:
Utilise fully the information from the nearest AWS maintained by the BoM.
Calculate the key parameters at a fine time resolution out to 6 days ahead.
Transfer forecasts to a web site on a daily basis.
Software system to include automatic model retraining as more data become available.
The forecasts were based on the models generated during the previous study conducted by Katestone Environmental for MLA. See Appendix A for a description of the models.
Forecast performance
The main factors that affect the HLI (and AHLU) are temperature, relative humidity (obtained from the dew point) and wind speed. There was good agreement between the forecast temperature and dew point and the observed quantities, however, the wind speed forecasting performance was relatively poor.
In terms of forecasting the heat stress category, it should be noted that the categories are broad – the low risk category ranges from 0 to 20 AHLUs, the higher risk categories extend over 30 and 50 AHLUs. Therefore, although agreement between the forecast and observed AHLU values might be poor, these would fall into the same heat stress category, giving better performance in predicting the category in contrast to forecasting individual AHLU values.
Recommendations
If a future forecasting system is to include more sites, we would recommend ample warning of the sites of interest so we can request that the Bureau of Meteorology store the LAPS/GASP information for these regions. Having a larger database of information from which to conduct the forecasts would improve forecast performance in the initial months.
As heat stress management in cattle is an ongoing area of research, future projects should include up to date methods for calculating heat stress parameters on cattle and reporting these on a regular basis.
