Spatially and temporally evaluating the skill of seasonal forecasting systems to aid with on-farm decision making

For full details of this procurement, including links to relevant documents and information on Questions and Answers, visit:

https://grdc.com.au/research/applying-and-reporting/current-procurement/open-tenders/spatially-and-temporally-evaluating-the-skill-of-seasonal-forecasting-systems-to-aid-with-on-farm-decision-making

Background

Climate-variability has a significant impact on farm-business profitability. While notable improvements have been made in seasonal forecasting with general circulation models (GCMs), growers and producers are frustrated by the inherently variable skill within and between models from one season to the next and in different locations (i.e. different grid cells for which forecasts are issued). A recently completed Rural R&D for profit project on Improving Seasonal Forecasting (RDC00014) completed an assessment of seasonal forecast skill for a small number of GCMs where forecasts were issued at key decision points for aiding with on-farm decision making. The models evaluated showed considerable discrepancies in skill when forecasting rainfall terciles based on a statistical analysis of forecast ‘hit-rate’.

The GRDC and the other participating sectors within the Managing Climate Variability Program (Meat and Livestock Australia, Cotton Research and Development Corporation, Sugar Research Australia and AgriFutures) are seeking to build upon previous and current research assessing the comparative skill of different seasonal forecast systems (i.e. model x grid cell x time of the year the forecast is issued) to aid with on-farm decision making. While it’s important for growers and producers to have access to robust, quantitative assessments of seasonal forecast skill, it’s equally important for that information to be well contextualised and communicated in a manner that accounts for any limitations with research methodologies. To demonstrate the benefits of integrating that information within decision making processes on-farm, a series of case studies identifying the potential economic benefits from integrating seasonal forecasts with crop and/or production system models to guide strategic decision making (e.g. crop choice) is a supplementary focus area within the project.

It’s also envisaged that the knowledge generated in the project may lead to the identification of locations and/or conditions under which different models consistently perform better than others, which could in-turn create a foundation for future targeted R&D based on the comparative strengths and weakness of different seasonal forecasting systems and their relative potential to aid on-farm decision making.

This a co-investment from GRDC and the other sectors participating within the Managing Climate Variability Program. It is anticipated that the successful applicant will be contracted via GRDC’s Standard Two Party Research Agreement.

Applicants should note that the list of seasonal forecast systems provided in output 1 is intended as a guide (i.e. not a definitive or exclusive list) and will likely be subject to negotiation with GRDC and MCV partners.

Investment description

Australian growers and producers lack access to quantitative, contextualised comparisons of seasonal forecast skill for rainfall and temperature for different seasonal forecast systems. This project will develop that knowledge base, extend it to growers and producers, and demonstrate the value of seasonal forecasting for strategic decision making on-farm via simulated case studies linking crop and livestock modelling tools with seasonal forecast information.

Expected Outcome

By December 2022, at least 20% of Australian growers and producers are increasing profit by an average of $8/ha/yr (indexed against 2018 benchmarking data) via greater knowledge of how and when to value 3-6 month seasonal forecasts for rainfall and temperature terciles (wet, dry, average) based on assessment of forecast skill for different models at different times of the year and in different locations.

Expected Outputs

Output 1

By February 2020, produce quantitative comparisons of forecast skill for rainfall and temperature with 3 and 6-month lead times (where applicable) for each of the seasonal forecast systems listed below*. The comparisons should be spatially and temporally based, available for each forecast grid cell within the Australian grain, red meat, cotton, rice and sugar growing regions, and their skill contextualised for the ability to aid strategic grower and producer decision making on-farm based on the stat­e of the El Nino Southern Oscillation and Indian Ocean Dipole in any given year.

• ECMWF – seasonal forecast system 5
• BoM – ACCESS-S1
• JAMSTEC - SINTEX-F
• UK Meteorological office – GloSea5
• The North American Multi-modal Ensemble
• NASA’s Goddard Earth Observing System Atmosphere-Ocean General Circulation Model version S2S_2.1 (USA)
• NCEP - Climate Forecasting System version 2
• Beijing Climate Center Climate system model version 1.1
• Copernicus C3S seasonal forecast system
• Queensland government SOI phase system (only evaluate in Queensland)
• DPIRD statistical forecast system (only evaluate in WA)

* Applicants should note that the list of seasonal forecast systems provided in output 1 is intended as a guide (i.e. not a definitive or exclusive list) and will likely be subject to negotiation with GRDC and MCV partners.

Output 2

By March 2020, for each year of the project develop a communication, extension and industry engagement plan with the aid of the Managing Climate Variability Programthat enables growers, producers and their advisors to A) understand and contextualise the spatial and temporal skill of available seasonal forecast systems for their particular regions, and B) demonstrate the potential benefit of integrating crop and livestock production models with seasonal forecasts to guide strategic on-farm decision making.

Output 3

By December 2020, produce a technical report detailing known biases and errors for each dynamical model in each forecast grid-cell applicable to the grains, red meat, cotton, rice and sugar industries across Australia during the key (i.e. 2 to 3) windows applicable to support on-farm decision making. The key windows and decision points will be specified via sector specific reference groups established by the participating RDCs, and link to existing projects such as the Rural R&D for Profit Project “Forewarned is Forearmed” where appropriate

Output 4.A)

By March 2021, via two simulated case studies in each GRDC region, demonstrate the potential economic benefit to Australian grain growers from linking crop models to the best performing seasonal forecast system at each respective case study site to inform profitable decision making at sowing via optimised crop-type, plant density and variety selection.

Output 4.B)

By March 2021, via three simulated case studies (exact industries and locations to be agreed in consultation with MCV partners), demonstrate the potential economic benefit to growers/producers from linking production models to the best performing seasonal forecast system at each case study site.

For full details please visit:

https://grdc.com.au/research/applying-and-reporting/current-procurement/open-tenders/spatially-and-temporally-evaluating-the-skill-of-seasonal-forecasting-systems-to-aid-with-on-farm-decision-making