Managing frost risk in warmer seasons may not be as easy as you think

  1. University of Queensland, Queensland Alliance for Agriculture and Food Innovation (QAAFI)
  2. CSIRO Agriculture Flagship
  3. Department of Agriculture, Food and Forestry Queensland  (DAFF)

Take home message

Growers need to consider carefully whether earlier sowing is justified in seasons where warmer temperatures are predicted.

Warmer temperatures may reduce the frequency of frost events but may also increase the rate of crop development bringing crops to the susceptible, post heading stages earlier.

Situation analysis of national frost impact indicates substantial losses in all regions averaging approximately 10% using current best practice.

In some regions, there are even greater losses in yield potential due to late sowing.

These results indicate that continued research into reducing frost risk remains a high priority despite increasing temperatures.

Analysis of frost impact on wheat is giving us our first nationwide assessment of the comparative impact of frost in different Australian cropping regions.  It is also providing important insights into how to manage frost risk in warmer cropping environments.

Warmer temperatures accelerate plant development causing crops to develop to the frost-susceptible, post-heading stages more rapidly.  So counter intuitively, planting earlier or even at the conventional date during warmer seasons may sometimes lead to increased frost risk.

Climate data from 1957-2013 was used to assess the frequency and severity of frost for each region of the Australian cropping belt.  Night time minimum temperatures have been observed to increase over much of the Australian cropping region during that period.  However, our analysis showed that frost risk and frost impact did not reduce over the whole cropping area during that time. 

Historic climate data from a grid database and for 60 locations representing each of the four major cropping regions of Australia was used to determine the frequency and severity of frost (Figure 1 top).  Crop simulation modelling using the Agricultural Production Simulator program (APSIM) was used to estimate crop yields.  Expert knowledge combined with data from frost trials was used to estimate crop losses.  Computer simulation allowed prediction of crop losses for all Australian cropping regions using damage information from a limited number of frost trial sites.  It also allowed simulation of potential yields using sowing dates optimised for yield in the hypothetical absence of frost risk, something that has not be achieved experimentally. 

             Figure 1.  Maps showing sites and regions for which climate data were analysed for the frequency and severity of frosts (top panel) and annual % change in yield loss due to frost from 1957 to 2013; negative values (yellow to red) represent areas where yield loss became worse over the recent decades (bottom panel). Estimations in the lower panel were for the cultivar Janz sown May 18th and are based on a ~5 x 5 km grid of climatic data.  Gridded climate data may not reflect local climatic conditions of particular paddocks within each grid as frost events are highly spatially variable.

Figure 1.  Maps showing sites and regions for which climate data were analysed for the frequency and severity of frosts (top panel) and annual % change in yield loss due to frost from 1957 to 2013; negative values (yellow to red) represent areas where yield loss became worse over the recent decades (bottom panel). Estimations in the lower panel were for the cultivar Janz sown May 18th and are based on a ~5 x 5 km grid of climatic data.  Gridded climate data may not reflect local climatic conditions of particular paddocks within each grid as frost events are highly spatially variable.

The study revealed that estimated current yield losses due to direct and indirect damage vary between regions and for crops of different maturities (Figure 2).  Frost damage averaged close to 10% nationally for all maturity types, for current sowing guidelines.  A similar result was observed for the northern region (about 10% yield reduction).  For direct damage estimates, simulations used sowing dates optimised for seasonal climate data applying current sowing guidelines to reduce frost risk. To determine the loss of yield potential due to late sowing currently required in many areas to reduce  frost risk, a wider range of sowing dates were considered for a hypothetical crop that would survive even the most severe of frosts.  In many areas this allowed earlier sowing dates leading to higher yield estimates.  When accounting for the current delay in sowing imposed by frost risk, estimated yield losses increased from approximately 10% to 20% nationally for early and mid-flowering genotypes and from approximately 11% to 18% for late flowering types (‘direct + indirect’ impact in Figure 2).  In the northern region, these losses were even greater, with estimates at 34, 38 and 23% for early, mid and late flowering cultivars, respectively (Figure 2). 

                         Figure 2.  Estimated yield losses (%) due to frost damage for crops sown at a fixed date corresponding to the best sowing date for local long-term yield achievement (‘direct’ frost damage); and crop losses due to both direct damage and relatively-late sowing of current best practice advocated to reduce frost risks (direct + indirect).

Figure 2.  Estimated yield losses (%) due to frost damage for crops sown at a fixed date corresponding to the best sowing date for local long-term yield achievement (‘direct’ frost damage); and crop losses due to both direct damage and relatively-late sowing of current best practice advocated to reduce frost risks (direct + indirect).

In some areas in each region, simulated frost impact has significantly increased between 1957 and 2013 (yellow, orange and tan areas, Figure 1, bottom panel). Estimated date of last frost has come later in some areas but earlier in others. However even in areas where last frost has come significantly earlier, increased temperatures have also increased the rate of development to frost-susceptible post-heading stages.  The modelling suggests that crop heading dates have been brought forward at a more rapid rate than the date of last frost, leading to an increased frost impact in some areas.

These trends over time may have implications for growers making planting decisions.  They indicate that sowing early to increase yield potential may not always be warranted in warmer seasons, even if the likelihood of frosts is expected to reduce.  By increasing the rate of crop development, warmer temperatures will cause the crop to develop more rapidly to the frost susceptible, post-heading stages, which may actually increase frost risk.

These results indicate that continued research to reduce frost risk remains a high priority despite increasing temperatures due to climate change.

Results from this Frost Situation Analysis will provide valuable insights allowing GRDC to better direct research resources.  They also provide valuable insights for managing frost risk now.

Acknowledgements

The research undertaken as part of this project is made possible by the significant contributions of growers through both trial cooperation and the support of the GRDC, the authors would like to thank them for their continued support.

The project is also supported by the Queensland Alliance for Agriculture and Food Innovation (a collaboration between UQ and DAFF) as well as by the University of Southern Queensland and the Department of Agriculture Western Australia.  The authors wish to thank Dr Ben Biddulph for his advice and for access to unpublished results from frost trials in WA.

Contact details

Dr Jack Christopher
University of Queensland, Queensland Alliance for Agriculture and Food Innovation (QAAFI)
Department of Agriculture, Food and Forestry, Leslie Research Centre
PO Box 2282, Toowoomba, Qld 4350
Ph: 07 4639 8813
Fx: 07 4639 8800
Email: j.christopher@uq.edu.au