Pre sow assessment of crown rot risk in the northern region

GRDC codes:

DAS00137 – National improved molecular diagnostics for disease management
DAN00143 – Northern NSW integrated disease management

Authors

Steven Simpfendorfer, NSW DPI Tamworth & Alan McKay, SARDI, Adelaide

Introduction

PreDicta B is a DNA based soil test which detects levels of a range of cereal pathogens that is commercially available to growers through the South Australian Research and Development Institute (SARDI). The main pathogens of interest in the northern grains region detected by PreDicta B are Fusarium spp. (crown rot), Bipolaris sorokiniana (common root rot), Pythium (damping off) and both Pratylenchus thornei and P. neglectus (root lesion nematodes, RLNs). Over recent years PreDicta B has been shown to be a reliable method for assessing RLN populations but is perceived by industry to be less reliable in assessing levels of crown rot risk in the northern region. This is potentially due to the crown rot fungus being stubble-borne while PreDicta B is a soil based test. Consequently, there may be sampling issues which need to be resolved to improve the reliability of detecting crown rot inoculum levels prior to sowing.

The following paper reports on collaborative research conducted by NSW DPI and SARDI across central/northern NSW from 2010-2013 to determine the accuracy of PreDicta B to predict the risk of crown rot infection prior to sowing and progress in improving the reliability of this technique within the region.

Survey 2010-2012

NSW DPI conducted a winter cereal pathogen survey of 248 paddocks annually across 12 agronomy districts in central and northern NSW from 2010-2012. A one hectare area was established in each focus paddock and 20 small cores were collected in a grid across the trial area targeting the previous winter cereal crop rows to a depth of 0-30 cm prior to sowing in each year. The soil samples were sent to SARDI for PreDicta B analysis. Winter cereal stubble if evident was collected from each focus paddock when taking soil samples then trimmed, surface sterilised and plated on laboratory media for the recovery of Fusarium spp. Fifty crowns, collected from different plants, were plated from each paddock to provide and incidence of crown rot (Fusarium) infection. Over the three survey years this allowed 307 comparisons of Fusarium DNA levels assessed at sowing using PreDicta B with the actual incidence of infection that developed by harvest as determined from laboratory plating (Table 1).

Table 1. Relationship between levels of Fusarium detected using PreDicta B at sowing and incidence of crown rot infection based on laboratory plating after harvest from 307 paddocks (2010-2012) z

PreDicta B	Plating 25%> High	Plating 11-24% Medium	Plating 3-10% Low	Plating 0-2% Nil
>2.0 High	28	8	8	1
1.4-2.0 Medium	17	13	6	1
0.6-1.4 Low	12	15	9	10
<0.6 BDL	30	27	65	57

PreDicta B risk categories have been established for crown rot based on log DNA levels at sowing (Table 1). Similarly, previous NSW DPI research has established what constitutes a low (3-10%), medium (11-24%) and high (>25%) level of crown rot infection based on laboratory plating (Table 1).

In 107 paddocks (35%) PreDicta B at sowing predicted the exact level of infection that developed in the crop when measured after harvest (darker grey shading Table 1). For example, soil DNA levels indicated a high risk of crown rot development in 28 paddocks and over 25% of plants were infected with Fusarium at harvest in these paddocks.

However, these categories are fairly tight. The predicted risk of crown rot development at sowing using PreDicta B was within one category of the actual level of infection measured after harvest in 121 paddocks (39%, light grey shading). For example, soil DNA levels indicated a medium risk of crown rot development in 17 paddocks (i.e. should have been between 11-24% infection) but greater than 25% infection actually occurred making this a high level of crown rot infection. In many of these situations these actual levels of infection that developed were only just outside the predicted levels at sowing. Hence, taking this approach PreDicta B correctly predicted the risk of crown rot development at sowing within one category of what developed in 74% (228) of the paddocks over the three years.

In 10 paddocks (3%) PreDicta B overestimated the risk of infection compared to that which actually developed. For example, soil DNA levels indicated a high risk of crown rot development in 8 paddocks but only low levels of infection (3-10%) were measured at harvest. This could potentially relate to inter-row sowing of the wheat crop between the previous cereal stubble rows which has been shown in previous research to significantly reduce the incidence of crown rot infection.

The bigger concern was that in 69 paddocks (22%) PreDicta B underestimated the risk of crown rot compared to the levels which actually developed.  For example, soil DNA levels indicated crown rot levels were below the detection limit (BDL) in 30 paddocks but greater than 25% of plants were infected with Fusarium at harvest. This is considered a ‘failure to warn’ with the question being why?

Detection issue?

A potential cause of the failure to warn in 22% of paddocks could be the inability of the current PreDicta B tests to actually detect the species of Fusarium causing crown rot across the region. Currently there are three separate tests within PreDicta B that detect common species causing crown rot across Australia. There are two tests which detect variations in F. pseudograminearum (Fp) populations and a third test which detects both F. culmorum (Fc) and F. graminearum (Fg) but cannot differentiate between these two species.

A total of 180 Fusarium isolates (1 or 2 isolates per paddock) were collected from crown rot infected winter cereal crops throughout central and northern NSW in late 2012 and early 2013. The isolates were sent to SARDI who extracted DNA and tested them against the three current PreDicta B Fusarium tests. The DNA was also sent to CSIRO laboratories in Canberra where each isolate was sequenced to confirm the species identification. A total of 84.5% of the isolates were identified by PreDicta B to be Fp which were all confirmed by sequencing to also be Fp. That is, there were no variants of Fp identified by sequencing which are not being detected by the current PreDicta B tests. A further 9.5% of isolates were identified by PreDicta B to be Fc or Fg. Sequencing determined that these isolates actually consisted of 5.6% Fc and 3.9% Fg. However, there were no variants of Fc or Fg which were not detected by the current PreDicta B test, it simply cannot differentiate between these two species. The remaining 6.0% of isolates were identified by sequencing to be a Fusarium sp. chlamydosporum complex which is not detected in the current PreDicta B tests. Further work is required to confirm the distribution and importance of these isolates and incorporate their detection into the PreDicta B tests if warranted.

Sampling issues?

Fifty crowns and 50 first above ground nodes were cut from primary tillers and plated separately from stubble collected out of each paddock. This allowed the relative survival of Fusarium below ground (crowns) and above ground (1^st nodes) to be determined at each site. In around 5-10% of paddocks where PreDicta B underestimated the crown rot risk there was much lower survival of Fusarium in the crown tissue relative to the 1^st node. For example at site WE9A in the Wellington district in 2011 there was 8% Fusarium recovery from the crowns but 54% from the 1^st node above ground. PreDicta B is a soil based test so with the collection of cores targeted at the previous winter cereal rows it can provide a good measure of Fusarium levels in the crowns below ground but is restricted in its ability to detect levels in above ground stubble. This could potentially be an issue with sampling especially following wet summers which would reduce survival in the crowns relative to above ground residues.

In the remaining 5-10% of sites where PreDicta B underestimated the crown rot risk more traditional plating of winter cereal stubble from the site would also have failed to warn of the risk as generally there was no stubble evident to plate. This may be related to hosting of Fusarium inoculum on grass weeds that were not adequately sampled or cultivation/harrowing/mulching of the paddock after collecting soil cores which more evenly distributed inoculum across the paddock and into the main infection zones for the crown rot fungus. The impact of these practices on distributing crown rot inoculum requires further research.

Can we improve PreDicta B assessment of crown rot risk?

One of the big advantages of PreDicta B is its ability to assess the relative risk of a range of soil-borne cereal pathogens within the one sample. In the northern region PreDicta B has been primarily used to assess RLN populations. Consequently, a composite sample taken from the top 30 cm of soil has usually been used in the northern region based on this being the recommended sampling depth with traditional manual nematode counts within the region. In other regions a shallower sampling depth (0-10 cm or 0-15 cm) is recommended with PreDicta B. In 2013 we aimed to determine if this deeper sampling depth in the northern region, to account for RLN populations deeper in the soil profile, is potentially compromising the accuracy of PreDicta B to measure levels of other soil-borne pathogens including Fusarium.

In 2013 each of the six ranges in 11 NSW DPI pathology trials, 11 cereal NVT sites and 2 grain and graze trial sites were cored using PreDicta B. A separate 0-15 cm (40 cores) and 0-30 cm (20 cores) bulked soil sample was collected from each range at each of the 24 field sites spread from central NSW up into southern Qld. All cores were targeted at the previous winter cereal rows if evident. Previous winter cereal crop stubble was also collected across each separate range at coring if present and used to spike set soil samples. Twenty-five lowest nodes (1 cm segments around node) were cut from the corresponding stubble sample and added to half of samples collected at each depth. All samples were then sent to SARDI for PreDicta B analysis. After harvest stubble will be pulled from three check wheat varieties at each site and these plots will be re-cored for RLN numbers using PreDicta B. This information will be used to validate and calibrate if required the sampling strategy and resulting risk categories across the northern region. Unfortunately this information is not currently available and will also need to be repeated over a few seasons to fully refine risk categories and sampling strategies.

What have we found so far?

Pratylenchus thornei (Pt) populations did vary with sampling depth across the 24 sites (Figure 1a). Points above the 1:1 diagonal line indicate higher Pt populations in the 0-30 cm sampling then in the 0-15 cm sampling. Conversely, points below the diagonal line represent sites with higher Pt populations in the 0-15 cm then in the 0-30 cm sampling. Interestingly, there were five sites above the line and four sites below the line which varied from the 1:1 line by 0.3 Pt/g or greater. However, this relatively minor variation in Pt numbers with sampling depth only resulted in a slight shift in risk category at one site. At the Coolah NVT site the 0-30 cm samples indicated a low risk of Pt with 1.8 Pt/g soil but the 0-15 cm sampling averaged 2.1 Pt/g soil which just pushed the site into a medium risk category (2.1 to 15.0 Pt/g soil). Even though the Bullarah NVT site averaged 1.3 Pt/g soil in the 0-15 cm sampling and only 0.2 Pt/g soil in the 0-30 cm sampling both depths would still have resulted in this site being classified in the low risk category (0.1 to 2.0 Pt/g soil).

Figure 1. Populations of Pratylenchus thornei (a) and Bipolaris sorokiniana (b) detected using PreDicta B at 24 sites in 2013 at two samplings depths (0-15 cm vs 0-30 cm). Diagonal lines represent a 1:1 relationship. 

Bipolaris levels are expressed on a log scale which flattens out variation in numbers with sampling depth (Figure 1b). However, levels do not appear to vary greatly between a 0-15 cm versus a 0-30 cm sampling depth. Generally there were more sites with higher levels in the 0-15 cm compared to the 0-30 cm indicating that Bipolaris is more concentrated in the surface which is being diluted with a deeper sampling.

Similarly, Fusarium DNA is expressed on a log scale. There was only one site (Bullarah NVT) where a higher level in the 0-30 cm sample would have classified the crown rot risk as medium while the 0‑15 cm sampling indicated a low risk (Figure 2). However, there were four sites (Tulloona, Gilgandra, Westmar and Narrabri) where greater values in the 0-15 cm samples indicate a higher crown rot risk level than in the 0-30 cm samples. As with Bipolaris, this indicates that Fusarium is more concentrated in the surface which is being diluted with a deeper 0-30 cm sampling depth.

Figure 2. Populations of Fusarium detected using PreDicta B at 24 sites in 2013 at two samplings depths (0-15 cm vs 0-30 cm). Samples spiked with stubble fragments excluded from comparison.

Addition of stubble to soil samples

Previous cereal stubble was only present at 7 of the 24 sites in 2013 to allow addition of stubble fragments to soil samples. Bithramere was the only site where the addition of stubble did not increase the predicted crown rot risk level. Even though the log Fusarium DNA/g increased from 2.5 to 4.3 with the addition of stubble, both values represented a high risk of crown rot development at the Bithramere site in 2013. The addition of stubble at the remaining six sites increased the crown rot risk level from low to high at two sites (Coonamble and Westmar), low to medium at two sites (Gilgandra and North Star) and medium to high at two sites (Bullarah and Tamworth).

Laboratory plating of harvest samples will determine if the addition of stubble to soil samples improves the accuracy of PreDicta B to determine crown rot risk prior to sowing. Adding stubble is likely to increase the overestimation of crown rot risk while reducing the likelihood of underestimation or ‘failure to warn’. This is probably a preferred situation for growers and advisors. The addition of stubble will also reduce sampling issues following wetter summers which can result in greater survival of Fusarium in above ground residues then in the crowns as occurred in some of the previous survey paddocks between 2010-2012. Research to refine the sampling strategy is continuing.

Conclusions

RLNs being soil-borne appear to be more flexible with sampling technique to obtain an accurate risk level prior to sowing. However, the crown rot fungus is stubble-borne so detection is more sensitive to the sampling technique used to collect the soil samples. Punching 3-6 cores between the previous crop rows in a paddock, as with a soil nutrition test, may give a reasonable estimate of RLN levels but is likely to provide a poor indication of the crown rot risk. Recent collaborative research in the northern region between SARDI and NSW DPI has demonstrated that use of a smaller diameter soil core (e.g. Accucore) to collect 15-30 cores (depending on sampling depth) targeted at the previous cereal row if evident provides a good measure of both RLN and crown rot risk along with a range of other pathogens. This number of cores collected spatially across the paddock is required to account for the potential variability in the distribution of crown rot inoculum. Important change to sampling; it is now recommended up to 15 short pieces of cereal stubble from previous cereal crops and/or grass weeds be added to PreDicta B soil samples to enhance detection of the Fusarium spp. that cause crown rot.  Where stubble is present, add one piece per sampling location. Each piece should be selected from the base of separate crowns; discard stubble above the first node. 

Soil cores – collect up to three cores from 15 different locations within the target area; take cores from the previous cereal rows and retain any stubble collected by the core.  The number of cores per location will vary depending on core diameter and sampling depth. Maximum sample weight should not exceed 500g. Sampling depth (0-15 cm or 0-30 cm) does not appear to greatly impact on detection of the various pathogen levels in the northern region when the collection of cores is targeted at the previous cereal rows. However, the actual sampling depth needs to be recorded on the sample bag when collected as it is used to refine reporting of results to adjust for pathogens which are more concentrated at the soil surface. Significant stubble disturbance (harrowing, cultivation, mulching etc.) increases the risk of crown rot development if the stubble is infected with Fusarium. Collection of soil samples prior to stubble disturbance is likely to underestimate the crown rot risk.

If you are not willing to follow the recommended PreDicta B sampling strategies then DO NOT assess disease risk levels prior to sowing. 

Acknowledgments

This project was co-funded by NSW DPI, SARDI and GRDC under the national improved molecular diagnostics for disease management project (DAS00137) and a previous project DAN00143. Assistance provided by Robyn Shapland, Finn Fensbo, Amy Alston, Liz Farrell, Kay Warren and Karen Cassin (NSW DPI); Herdina, Russell Burns, Aidan Thomson, Ina Dumitrescu, Danuta Pounsett, Irena Dadej, Daniele Giblot-Ducray (SARDI); and Diane Hartley (CSIRO) is greatly appreciated.

Contact details

Dr Steven Simpfendorfer
NSW DPI
Mb: 0439 581 6720439 581 672
Email: steven.simpfendorfer@dpi.nsw.gov.au

Dr Alan McKay
SARDI
Ph: 08 8303 937508 8303 9375
Email: alan.mckay@sa.gov.au