Barley disease yield loss response curves

Take home messages

  • Losses in yield to a given disease epidemic vary with the levels of resistance in the varieties affected.
  • The GRDC funded project “Yield loss response curves for host resistance to leaf, crown and root diseases in wheat and barley” is gathering data on yield losses of varieties with different levels of resistance under epidemics of different severities for a range of pathogens of wheat and barley.
  • The barley foliar disease module of the project is being led by DAFQ, with experiments conducted nationally to explore the impact of leaf rust, net blotch (spot form and net form), scald and powdery mildew on current commercial barley varieties.
  • NVT disease resistance ratings categorize varieties into 9 resistance categories (R – VS). Data collected from the project will add precision to assignment of these resistance categories.
  • Growers and agronomists will be able to make better informed and more accurate decisions on disease management by implementing information gathered from the yield loss response curves project in their calculations.


Disease control measures and the economics of chemical intervention are often based on potential yield losses – a worst case scenario e.g. disease “X” can cause yield losses in excess of 50%. That may be true in a very susceptible variety under a prolonged and severe epidemic; so the decision to spray to protect yield is automatic.

However, are we always likely to incur a worst-case epidemic or is it going to be somewhere between nil and severe? Clearly it will be somewhere in-between. Furthermore, are we likely to suffer the same potential yield losses under a lesser epidemic? The answer is NO.

Similarly are the varieties we are growing all in the very susceptible (VS) category? If not, are we likely to suffer the same potential yield losses in varieties that have some resistance to disease ”X”? NO, again. Therefore, when growing a variety of intermediate resistance under an epidemic that is less than severe, what is the loss likely to be – 40%, 25% or 10%? We simply don’t know.

We need to be able to predict these losses under variable epidemics so that an agronomist or a grower can more accurately determine the risk from disease and the profitability of disease management strategies. Knowledge and predictability of the yield and quality losses that varieties of different resistance levels might incur under different epidemics, is crucial to these calculations.

Losses to diseases are a factor of:

  • Susceptibility of the variety
  • Severity of infection
  • Duration of the epidemic and
  • Environment

If we were able to quantify yield losses in terms of varietal resistance level and severity of the epidemic then we would have a much more accurate predictive tool to guide decisions on disease control measures. This is the basis of the GRDC funded Yield Loss Response Curves project.

Response curves relate a measure of productivity (yield), to a measure of disease (area under the disease progress curve) and are being constructed for a range of varieties of different resistance levels across a number of locations and years. The yield responses of varieties are measured by exposing them to different levels of disease severity under the same environment. Through the manipulation of disease severities, an epidemic continuum is established, allowing the yield response due to disease to be estimated for each variety across a wide spectrum of disease epidemics.

Data collection

The value of the yield response curves is directly correlated to the quality of the data used to generate them. Field experiments have been conducted annually since 2014 and will continue until 2018 to generate the raw data. A typical experiment consists of six varieties with different resistance levels as assigned by the NVT disease rating system, avoiding resistant to moderately resistant varieties. Varieties are then exposed to epidemics ranging from no disease to high disease. Typical epidemics are nil, very low, low, medium and high where epidemics are manipulated using different levels of initial inoculum, fungicide applications or combinations of both to give stepwise differences in grain yield for each variety.

As the crop and epidemics develop, a minimum of 5 disease assessments are taken on every plot and this information is expressed as area under the disease progress curve (AUDPC) for each treatment. AUDPC is a quantitative summary of disease severity over time i.e. the period over which assessments were taken.

Trials are taken through to maturity and harvested, yields recorded and grain quality parameters measured.


Yield losses

Yield data is generated with the prime objective of developing yield response curves for the type variety of specific resistance categories; however, it also serves to demonstrate the relative yield losses across resistance categories. The figures below summarise the yield responses of trials in 2014 to investigate barley leaf rust (Figure 1) and net form net blotch (Figure 2) and in 2015 to investigate spot form net blotch (Figure 3).

Figure 1. Comparative yields of 6 barley varieties with differing resistance categories under epidemics of leaf rust in 2014. This data corresponds to that presented in Figure 1 using response curves.
Figure 1. Comparative yields of 6 barley varieties with differing resistance categories under epidemics of leaf rust in 2014. This data corresponds to that presented in Figure 1 using response curves.

Barley leaf rust 2014

In this experiment the percentage loss in yield between the nil and high diseased plots and the resistance ratings of the varieties were Grout  (VS) 25.1%; Scope  (S) 12.3%; Compass  (S) 38.5%; La Trobe  (MSS) 9.4%; Mackay  (MRMS) 11.0% and Shepherd  (MR) 3.9%. The choice of varieties in this experiment was based on 2013 NVT resistance ratings where Compass was rated S. This trial demonstrated that Compass  was misclassified as susceptible (S) and was clearly VS to leaf rust. It also demonstrated that a little resistance can make a big difference to yield loss under some epidemics e.g. the yield loss in Scope (S) was about half that of the VS Grout  and less than one third of Compass.


Figure 2. Comparative yields of 6 barley varieties with differing resistance categories under epidemics of net form net blotch in 2014.
Figure 2. Comparative yields of 6 barley varieties with differing resistance categories under epidemics of net form net blotch in 2014.

Net form net blotch (NFNB) 2014

As seen in the leaf rust experiment, the VS variety – Charger  in this case - suffered the greatest yield loss (41.7%) under the high epidemic and still lost 29.3% under the very low epidemic (no inoculum applied) when compared to the nil disease treatment. This contrasted with the S variety Fairview  that lost 20.8% and 11.5% respectively and Compass (MRMS) which lost 14.9% and 5% respectively under the same levels of inoculum applied. Compass  has good adult plant resistance to the pathotype of NFNB used in this trial; but is susceptible in the vegetative stages. The losses recorded in Compass could have occurred from early infection of NFNB but are likely to have been exaggerated by late natural infection with leaf rust.

Figure 3. Comparative yields of 6 barley varieties with differing resistance categories under epidemics of spot form net blotch in 2015.
Figure 3. Comparative yields of 6 barley varieties with differing resistance categories under epidemics of spot form net blotch in 2015.

Spot form net blotch (SFNB)

The yield losses between the nil disease and high disease treatments in this experiment were 22.0% for Oxford, 20.4% for Shepherd, 18.2% for Scope, 15% for Commander, 10.1% for Flagship  and 15.4% for Compass. Again, the losses in Compass  (MRMS) may have been exaggerated by a late infection with leaf rust.

While the data presented here is only preliminary, it clearly demonstrates increasing losses with increasing susceptibilities. It also shows differences in yields from the different levels of inoculum applied; but the real value of this data is in generating yield response curves that provide additional information to the bar charts shown above.

Yield response curves

Yield response curves are derived by analysing yield and quality data using random regression techniques to estimate the yield potential (intercept) and yield response (slope) of each variety in the trial. Yield potential provides an estimate of the ability of a variety to yield in the absence of disease, while the yield response demonstrates the rate at which yield is lost per unit increase in disease pressure.

The yield response curves of six barley varieties subjected to a range of leaf rust epidemics in the 2014 experiment are presented in Figure 4. The resistance categories of varieties in the experiment ranged from moderately resistant (MR) to very susceptible (VS). The measure of leaf rust pressure was determined by combining the assessments conducted on the two leaves directly below the flag leaf using the AUDPC. Although described as “curves” the relationship between grain yield and disease was linear in this case.

Figure 4. Yield response curves for six barley varieties of different resistance categories under a continuum of leaf rust epidemics developed in 2014.

Figure 4. Yield response curves for six barley varieties of different resistance categories under a continuum of leaf rust epidemics developed in 2014.

What do the yield response curves tell us?

From Figure 4, it can be seen that the curve of each variety varies in length and slope. A comparison of how far the curves stretch along the x-axis provides a measure of the susceptibility of the varieties to leaf rust; the greater the AUDPC, the more susceptible the variety. In this case, the varieties Grout and Compass  exhibit the greatest amount of disease, closely followed by the variety Scope. This ranking of varieties is roughly consistent with the resistance categories assigned by NVT; however in 2014, Compass  was still rated as a susceptible variety (S) yet in terms of disease severity (AUDPC) and yield loss,  it clearly displays a very susceptible (VS) response (resistance rating updated in 2015). Conversely, the moderately resistant (MR) and moderately susceptible (MS) varieties Shepherd and Mackay show the least amount of disease.

Also demonstrated by Figure 4 is the difficulty in maintaining true nil disease treatments in the presence of leaf rust. Nil disease treatments were sprayed twice; however leaf area disease measurements still revealed loss of leaf area from the disease. This lack of a true nil disease control makes the estimation of the varieties yield potentials difficult, as extrapolation is required beyond the range of the data.

A comparison of the responses (slopes) of the varieties also proves useful in determining the rate of change at which yield is lost due to leaf rust for each variety. In this case, Figure 4 shows that the varieties Shepherd  and La Trobe  demonstrate near to equivalent responses to leaf rust, with flatter slopes than the other varieties. This indicates that with an equivalent unit increase in disease, these varieties lost less yield than the other varieties in the experiment. On the other hand, the variety Compass exhibits the largest slope, demonstrating that per unit increase in disease pressure, the yield lost by this variety is greatest.


The data presented in this paper come from 3 trials, each conducted at one location in one year and should be regarded as preliminary. Further trials are being undertaken and data from these will be used to validate and complement the above improving the confidence and precision of the resultant yield response curves.


The yield response curves project is a national initiative of GRDC coordinated by the Department of Agriculture and Food Western Australia. It is generating data on yield losses in cereals relative to varietal resistance and severity of disease epidemics. This data will be used to develop yield response curves for varieties within specific resistance categories under a continuum of disease pressures. 

These curves will be much more informative than a yield loss bar chart or an NVT resistance rating. They will provide a quantitative summary of yield loss in varieties or resistance categories over a range of disease severities. This will assist in choosing varieties and in planning disease management strategies. Through this project, growers and agronomists will have quantitatively based decision support to guide the adoption or disadoption of varieties and to more accurately formulate disease control budgets and application strategies to minimize the effects of diseases. Furthermore, the trials will help validate the NVT disease resistance ratings in terms of loss of yield and quality and not just disease severity.

In terms of applying the information presented in the response curves into disease control strategies, the curves in Figure 4 reveal that a grower could rely on the host resistance of Shepherd  alone to control leaf rust. The loss in yield of this variety would not warrant application of fungicide for leaf rust control. On the other hand, the curves demonstrate that Compass has the highest yield potential but also exhibits the greatest potential loss in yield under leaf rust epidemics. This information can help growers decide (a) if they want to risk growing a VS variety (b) the level of disease control required to match the yields of a more resistant, lower yielding variety and (c) the likely return on application of fungicide.


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 valued support.

Contact details

Greg Platz
Dept of Agriculture and Fisheries
604 Yangan Rd, Warwick QLD
Ph: 0408 733 055

Lisle Snyman
Dept of Agriculture and Fisheries
604 Yangan Rd, Warwick QLD
Ph: 0428 324 932

Clayton Forknall
Dept of Agriculture and Fisheries
13 Holberton St, Toowoomba QLD 4350
Ph: 07 4529 1218

Varieties displaying this symbol are protected under the Plant Breeders Rights Act 1994.

GRDC Project code: DAW00245