Spring frost damage in northern GRDC region in 2017 – a long term risk management perspective

Call to action/take home messages

  • 2017 will be remembered as a year of severe frost damage across the northern region and we need to learn from this damaging but relatively rare sequence of events.
  • While it is useful to focus on the unique aspects of frost risk, it is important to consider the interaction between frost, heat and water.
  • Recent modelling studies suggest that the northern region suffers the greatest direct impact of frost, but also the greatest indirect impact from strategies to avoid frost.
  • There is useful information from the GRDC National Frost Initiative on genetics, management and environment aspects of reducing frost risk. In this paper we focus on the weather and climate information available for frost risk management.

The GRDC national frost Initiative

Frost is estimated to cost the Australian grains industry over $300 M every year. The GRDC National Frost Initiative conducts RD&E to manage the impact of frost and maximise grower profit. The initiative has three components;

  1. Genetics – develop more frost-tolerant wheat and barley germplasm and rank current wheat and barley varieties for frost susceptibility;
  2. Management – develop best practise crop canopy, stubble, nutrition and agronomic management strategies to minimise the effects of frost, and search for innovative products that may minimise the impact of frost; and
  3. Environment – predict the occurrence, severity and impact of frost events on crop yields and frost events at the farm scale to enable better risk management.

Widespread damage in northern region in 2017 growing season

The 2017 season will be remembered for the widespread and frequent frost events across the northern GRDC region.

For a local perspective of the damage to canola, chickpeas and cereals in the northern region as of September 2017 see:

Resources available to help growers deal with frost affected crops, GRDC media release

The 2017 damage in the northern grains region followed severe damage in the western region and parts of the southern region in 2016 and 2017 and widespread stem frost in the southern region in 2014. Any regional overview of frost impact will underplay the damage experienced in individual paddocks and for individual grain businesses.

Unfortunately frost was not the only climate concern for grain growers in the northern region in 2017. Although Queensland and northern NSW had a wet October, most of the region experienced rainfall in the lowest decile for the six months April to September. The second half of September was extremely hot. On the 23rd of September, NSW recorded the hottest day since records were kept in 1911 (BoM 2017).

Across the grain growing regions of northern NSW and southern Queensland the mean maximum temperature for the week ending 28th of September was 8 to 12o C above average. Experienced agronomists will point to difficulties in separating the impact of frost from drought and heat on final wheat yield.

Cold temperatures do some damage each year

The severe frost damage experienced in 2017 is a low frequency but high consequence event. It is likely that there are more frequent, but less damaging losses in most years. According to GRDC & WA DIPIRD (2017) cold damage can occur when wheat plants are exposed to temperature less 5°C which can cause spikelet damage if this occurs during pollen development (Z39 -45). From 0°C to -2°C moisture is drawn from the leaves resulting in desiccation damage. The greatest damage is freezing damage which might be expected at 0°C; however 0°C is the melting point of ice not the freezing point of water. Freezing usually occurs at temperatures below -2°C and the damage is caused by ice crystals physically rupturing cell walls and membranes.

Ten days after a frost event, bleached leaves, stems, heads and reproductive tissue might be evident (GRDC & WA DIPIRD 2017).

An indirect cost comes from strategies to avoid frost

In addition to the direct damage from frost, there is an indirect effect from conservative sowing time /flowering time strategies to avoid frost. This is captured in the statement in the 1970s by the pioneer of frost research at Tamworth, Dr Bill Single, that the fear of frost does more damage than frost itself. Local advisers and growers will have their own views on whether this is still the case but some recent modelling research indicates that this indirect cost is greatest in the northern region.

Figure 1 depicts the agro-ecological zones in Australia and was reproduced from Mushtaq et al. 2017 – see paper for details. Economic benefits (AUD ha–1) of various levels of post head emergence frost (PHEF) tolerance both direct (first bar) and direct plus indirect (second bar).

Figure 1. (reproduced from Mushtaq et al. 2017 – see paper for details). Economic benefits (AUD ha–1) of various levels of post head emergence frost (PHEF) tolerance both direct (first bar) and direct plus indirect (second bar).

Under the assumptions of the simulation modelling (described in detail in Mushtaq et al. 2017 and Zheng et al. 2015) the greatest benefit of frost tolerance (because of the greatest current damage) is found in the northern grains region and parts of the western region. The northern region also has the greatest indirect impact (shown as difference between each pair of bars). Not only is thinking about the trifecta of frost, heat and water stress important when diagnosing and attributing damage to frost in a year like 2017, this modelling shows that it is essential when managing frost. The importance of radiation, water, frost and heat in identifying the ideal flowering time is supported by extensive modelling for sites between Dubbo and Eyre Peninsula by Flohr et al. 2017. Similar results have been shown for canola (Lilley et al. 2017).

The need for a risk management approach

The interaction of frost with heat and water stress is a good reason for a risk management approach to frost. A further reason is the acknowledgement that frost is not an issue that can be simply solved or removed from grain farms. Rather it is a something that has to be lived with and managed. Frost risk differs for each grain grower; not only does each paddock have a unique physical exposure to frost, each grain business has its’ own financial exposure and the people behind the business have different risk appetites. Under these circumstances, being prescriptive is dangerous.

Frost management decisions at different times

There are some common features of frost management across grain farms. For example, decisions are different depending on the time of the year. A pre-season planning might be held with an adviser in say January or February, this contrasts to decisions in the immediate lead up to sowing or responses to frost within a growing season. Although there are overlaps, separating the timing and types of decisions provides a useful framework and is consistent with the format of the GRDC Frost Tips and Tactics.

Not only do the decisions occur at different times of the year, there are differences in the type of decisions. Running a farm involves many day to day operational decisions but these are influenced by longer term strategic decisions which set the overall direction of the farm and a series of tactical decisions made each season. One way to distinguish between tactical and strategic decisions is that tactical decisions respond to the state of the system such as stored soil water, time of season break, and potentially a seasonal climate forecast. This framework is used in GRDC business management fact sheets (Making effective business decisions, June 2013, Simple and effective business planning, May 2014).  Unlike larger corporations, in a grain farming business the same person is usually making the strategic and tactical decisions while carrying out many of the day to day operations.

The time lines or planning horizons of operational, tactical and strategic decisions can be matched to weather forecasts for operational decisions, seasonal climate forecasts for tactical decisions and strategic decisions using longer term climate records, including how these are shifting with climate change. In the following section we have matched weather and climate information to the timing and type of decision. This is based on discussion with farmers and advisers and the purpose is to sort weather and climate information by decisions and complement the information in the GRDC NFI Tips and Tactics fact sheet.

1. Strategic pre-season planning

Planning enterprise mix across the farm such as crops vs livestock vs hay. Crop choice for different paddocks. Decisions about leasing extra land and/or purchasing and selling land.

Information currently available: Many experienced farmers are aware of the frostier parts of their farms, some have data loggers.

The spatial climate information can be supplemented with historical climate data that is analysed the BoM data such as Australian CliMate and Yield Prophet and FlowerPower in WA.

Emerging resources: The NFI is funding research on the fine scale mapping of frost across paddocks using loggers and remote sensing. This will optimise the use of this equipment.

Ongoing challenge: As the cost of loggers and imagery becomes cheaper and more available, the spatial coverage will greatly improve. However there is an ongoing challenge is to link 2 to 3 years of fine scale records with the 50 years of Stevenson screen data from the Bureau of Meteorology.

One of the impacts of climate change is that it makes historical records less reliable for future risk assessments. There are some concerning shifts in frost likelihood that make it difficult to know how to use past data.

2. Tactical adjustments at sowing time

Decisions include area of dry sowing, refining choice of crop and variety and changing input levels. Choosing varieties. Input levels. Making plans for extra hay production.

Information currently available: Some farmers tend to use Australian CliMate, Yield Prophet or FlowerPower when there is a sowing opportunity outside of the normal sowing window

There has been a history of using SOI based forecasts for frost likelihood. In general when there is a forecast for an increased chance of El Nino, expect more frost.

Emerging resources: We can expect there will be ongoing improvements in the decision aids such as CliMate, Flowerpower and YieldProphet. They will be aided by improved phenology predictions.

Climate forecasts of the likelihood for frost at this time of the year currently only have marginal skill. However there is increasing attention to forecasting of extremes.

Ongoing challenge: Seasonal forecasting remains a relatively low signal to noise and many growers will require very large shifts in forecasts of extremes to change decisions.

We are hampered by relatively rudimentary understanding of the exact relationship between minimum temperature and wheat yield.

3. Responding to frost forecast within the season

Management options are greatly restricted once the crop is sown, however a frost warning for the coming week can be useful for herbicide decisions and as a prompt to check for damage. For some enterprises, a warning can be used to plan for hay and grazing options. A forecast for the coming months might influence nitrogen topdressing decisions and possibly forward selling. Some farmers might graze a crop to slow development.

Information currently available: The BoM issues short term frost warning

In terms of seasonal outlook, the state of climate drivers such as ENSO become clearer as the season progresses.

Emerging resources: The accuracy of 1-7 day forecasts is continuing to improve.

The experimental multi-week forecasts from the Bureau of Meteorology will provide some information on the likelihood of lower than expected minimum temperatures.

Ongoing challenges: Although forecasts within season will be more accurate than pre-season, the question remains as to whether they will be good enough to change decisions. It is also challenging knowing how to link uncertain forecasts to uncertain damage functions.

4. Responding to a frost

To cut for hay or graze or leave for recovery

Information currently available: The main task after a frost is to rapidly assess the damage.

This can be aided by accessing temperature data from the Bureau of Meteorology and other networks of temperature loggers from departments of primary industry and NRM bodies.

Emerging resources: NFI has guidance on where to place loggers and is researching innovative methods of rapid frost assessment. There is also excellent material on identifying frost damage.

Simulation modelling like YieldProphet along with other spreadsheet based decision support system (DSS) can help with the decisions to cut for hay or graze

Ongoing challenges: If a frost occurs relatively early in the season, the decision to cut for hay, graze or leave for recovery is still made difficult by uncertainty in estimating the potential damage and the potential for recovery.

5. Post season evaluation

Severe frost is a relatively low frequency but high consequence event. It is important to place the season in context and avoid the natural human response of either over reacting or under reacting to a major event.

Information currently available: Although the BoM network of stations is relatively coarse, the access to archive maps for individual nights is excellent. In some regions this network is enhanced with local data.

Emerging resources: Improvements in networks of loggers and links to remote sensing will improve the assessment of temperature.

Ongoing Challenges: Placing a single year in context will always be difficult in a variable and changing climate. There is an abundance of psychology evidence that as human’s we will always struggle to distinguish between decisions that are wise/unwise and those that are lucky/unlucky.

Analysis of frosts in the region in 2017

Table 1. Minimum temperatures at a range of locations in GRDC northern grains region for 19, 20, 28 and 29 August 2017









































Table 1 shows the minimum temperatures for widespread and damaging frost events on the 20th and 29th of August. Overnight temperatures at ground level or the top of a wheat canopy can be up to 5°C lower than those measured in a Stevenson screen. The offset used in the DSS Wheatman was that head height was 2.2 degrees colder than the Stevenson screen, but differences of up to 10°C have been recorded.

Figure 2 is a map of Australia showing the mean sea level pressure on 19, 20, 28 and 29 August 2017. Source Bureau of Meteorology. Blue arrows have been added and show southerly flow of air.Figure 2 is a map of Australia showing the mean sea level pressure on 19, 20, 28 and 29 August 2017. Source Bureau of Meteorology. Blue arrows have been added and show southerly flow of air.

Figure 2. Weather maps showing the mean sea level pressure on 19, 20, 28 and 29 August 2017. Source Bureau of Meteorology. Blue arrows have been added and show southerly flow of air.

Being located near the centre of a high pressure system provides the stable, descending, dry air required for the clear, calm, night conducive to a rapid temperature fall at dusk and a radiation frost. This raises the question as to why a radiation frost doesn’t occur each time there is a high pressure system. Part of the explanation comes from the synoptic pattern of the previous day providing a southerly flow of air (shown as blue arrows). A fuller understanding of the process comes from examining more levels of the atmosphere than the ground level.

Improving the understanding of atmospheric dynamics behind frost events

A team at CSIRO Hobart, led by James Risbey, set out to better understand the synoptic weather events leading up to and during frosts. An understanding of the synoptic drivers provides a basis for confidence and testing of seasonal forecasting, especially if there are broad-scale patterns in the atmospheric dynamics across the southern hemisphere. In discussion with industry, eight high quality climate stations were selected across the Australian grains belt Merredin, Katanning, Kyancutta, Snowtown, Nhill, Wagga Wagga, Gunnedah, and Miles. The historical record from 1955 to 2014 was used to identify mild (T min < 2), moderate (T min < 0), and severe (T min < -2) frost events in the three month period between 15th August and 15th November.

The left hand panel of Figure 3 shows the backtracking of air in the days prior to medium frosts at Miles. This highlights that even northern sites require the inflow of air from a long way south. Not only is the air cold, because it is descending from the middle of the troposphere, it is very dry. Such extreme southerly origin air trajectories are not associated with most high pressure systems in the region and occur mostly in association with the developing blocking high. The blocking highs develop rapidly and persist and are efficient at drawing in and entrapping the cold, dry air. Similar patterns are shown for other locations in the western and southern regions.

Figure 3A is a picture of Australia showing back track of air parcels for spring frost events at Miles Qld (1955 to 2014). The numbers refer to the days prior to the eventA.

Figure 3B is a map of Australia which shows backtrack of air parcel for 29th August 2017.


Figure 3. A: Back track of air parcels for spring frost events at Miles Qld (1955 to 2014). The numbers refer to the days prior to the event. B: Backtrack of air parcel for 29th August 2017.

Both the frost events of the 20th and 29th of August involved a front coming through, then a high developing with cold, dry descent trajectories and very cold air mass. These are consistent with the general pattern derived from the 1955 to 2014 data. The late August event was an especially persistent and strong pattern.

The atmosphere flows from west to east sets up what is called a zonal flow which follows the latitudinal lines. The contrast is meridional flow along longitudinal lines. Widespread frosts require meridional flow, that is the interaction of the front and high pressure systems enhance the southern transport of cold dry air. The answer to why each high pressure system is not accompanied by a frost lies in the fact that most high pressure systems are relatively shallow circulation features and do not have the deeper vertical organisation required to entrain very cold, dry air from higher latitudes. The developing blocking high system associated with frost has the appropriate vertical structure to provide cold, dry entrainment. Blocking highs are much rarer at a given location, and thus frost is a relatively rare event.

Research in this project has also shown the importance of synoptic patterns that set up meridional flow to bring hot northern air for spring heat events. The encouraging aspect of the research is that these patterns are not only show consistent synoptic patterns over cropping regions, there is also a strong pattern in the broad-scale southern hemisphere circulation.

Concluding remarks

There is much more climate research on heat and drought than spring frost events. This is not surprising given the enormous interest in heat events due to the direct impact on human health and safety, infrastructure, bushfires and demand for electricity. Likewise drought has always attracted research due to the widespread impact on agriculture but also ecology and increasingly, on urban water supplies. In contrast, spring frosts are only really of concern to grain farmers and horticulturists. This not only applies to the climatology, as there is better understanding on the impact of heat and drought on wheat than post head emergence frost.

Although freezing temperatures are an issue for wheat growth in many parts of the world, most of the literature on spring frosts or post head emergence frost damage comes from Australia and South America. The GRDC National Frost Initiative is guided by growers and agronomists and will continue to provide updates on findings as it builds on previous research, much of it from the northern region.


Australian Bureau of Meteorology 2017.  Special Climate Statement 62—exceptional September heat in eastern Australia October 2017

DPIRD (WA) and GRDC 2017 Frost frequently asked questions.

Flohr, B. M., J. R. Hunt, J. A. Kirkegaard and J. R. Evans (2017). "Water and temperature stress define the optimal flowering period for wheat in south-eastern Australia." Field Crops Research 209: 108-119.

GRDC 2016  Managing Frost Risk Tips and Tactics. GRDC Grownotes February 2016

Lilley, J. M., B. M. Flohr, J. Whish and J. A. Kirkegaard (2017). "Optimal flowering periods for canola in eastern Australia." Proceedings of the 18th Australian Society of Agronomy Conference, Ballarat, VIC.

Mushtaq, S., D.-A. An-Vo, M. Christopher, B. Zheng, K. Chenu, S. C. Chapman, J. T. Christopher, R. C. Stone, T. M. Frederiks and G. M. Alam (2017). "Economic assessment of wheat breeding options for potential improved levels of post head-emergence frost tolerance." Field Crops Research 213: 75-88.

Zheng, B., S. C. Chapman, J. T. Christopher, T. M. Frederiks and K. Chenu (2015). "Frost trends and their estimated impact on yield in the Australian wheatbelt." Journal of Experimental Botany 66(12): 3611-3623.

Climate analysis tools


The research undertaken as part of the project “Assessing forecast and management options for mitigating extreme temperature impacts on grains CMA00002” is made possible by the significant contributions of growers through the support of the GRDC. The authors would like to thank them for their continued support. The guidance and encouragement of John Shepherd is especially acknowledged. PTH and DST time is also supported through a project funded by GRDC and the Commonwealth Government’s R&D for Profit Program “Forewarned is forearmed - equipping farmers and agricultural value chains to proactively manage the impacts of climate extremes.”

Contact details

Peter Hayman
SARDI Climate Applications, GPO Box 397, Adelaide SA 5001
Ph: 08 8303 9729
Email: Peter.Hayman@sa.gov.au

Dane Thomas
SARDI Climate Applications, GPO Box 397, Adelaide SA 5001
Ph: 08 8303 9729

James Risbey
Ph: 03 6232 5086
Email: James.Risbey@csiro.au

GRDC Project Code: CMA00002,