Seasonal rainfall and winter crop yield forecasting for southern Australia
Egan, J.P. Mr; Ph: (08) 8688 3424; Fax: (08) 8688 3427; egan.jim@pi.sa.gov.au
Holton, I. Mr; Ph: (08) 8388 6700; Fax (08) 8388 6788; holton7@senet.com.au
Grace, W.J. Dr; Ph: (08) 8201 2648; Fax: (08) 8201 3573; arawg@es.flinders.edu.au
Research organisations:
South Australian Research and Development Institute, GPO Box 397, Adelaide SA 5001;
Bureau of Meteorology, Adelaide Regional Office, PO Box 421, Norwood SA 5067;
Bureau of Meteorology Research Centre, GPO Box 1289K, Melbourne Vic 3001
Collaborators: Balston, Jacqui Ms, Primary Industries South Australia, Adelaide; Abrecht, Doug Dr., Agriculture Western Australia; Project is part of the National Climate Variability Research and Development Program
Sponsors: LWRRDC, Land and Water Resources Research and Development Corporation
Objectives:
1. To refine and validate the seasonal rainfall and winter cereal crop yield forecasting technique development for South Australia;
2. To publish details of the technique in the Australian Meteorological Magazine;
3. To develop prediction equations for other seasonal rainfall criteria;
4. To assess the potential for extending the technique to other agricultural regions of southern Australia.
Methodology:
1. Use the 'cross-validation' statistical technique to validate the seasonal rainfall and crop yield forecasting technique previously developed by I. Holton;
2. Optimise barometric pressure indices and their combination in the prediction equations, and investigate feasibility of developing a single prediction equation for all districts and locations;
3. Publish the results of these statistical analyses in the Australian Meteorological Magazine.
Progress:
Seven basic forecasting models (or forecast indices) were developed using a data set of barometric pressure, upper level geopotential height and rainfall data from 1957 to 1983 inclusive (27 years total). These models were then tested for forecasting skill on rainfall and crop yield data for the period 1984 to 1996 (10 to 13 years, depending on the availability of most recent data), using cross-validation routines. Skill for climate forecasts was measured relative to forecasts based on climatology (i.e. a forecast of median rainfall for the period).
These analyses show that the Holton models have good skill in forecasting growing season rainfall, seasonal rainfall, and wheat crop yields over much of southern Australia. The models are particularly accurate over the crop growing regions of SA, Victoria and parts of southern NSW, with correlation coefficients (r) of 0.7 to over 0.9 over most of these regions. Some of these forecasting models are available for use in November the year prior to the growing season, while updated forecasting models using later information are available in April, May and June of the current growing season.
A paper on the models has been accepted for publication in Australian Meteorological Magazine in late 1998.
Keywords: cereals; grain crops; rain; winter; crop yield; yield forecasting; prediction; seasonal variation
Period: starting date 1996-07; completion date 1997-03
Status: completed
Publications:
Holton, I. (1996). Seasonal and yearly prediction of rainfall and crop yields in Australia. Proceedings of the Second Australian Conference on Agricultural Meteorology, 1-4 October 1996, The University of Queensland, pp. 127-129.
Holton, I. and Egan, J.P. (1997). Developing better forecasts for seasonal rainfall and crop yields in southern Australia. Proceedings of Farming Systems Developments 1997 Workshop, Adelaide, March 1997, pp. 189-191.
