Ascochyta blight of field pea, caused by Didymella pinodes, Phoma medicaginis var. pinodella, Phoma koolunga and Didymella pisi, is controlled through manipulating sowing dates to avoid ascospores of D. pinodes, and by field selection and foliar fungicides. This study investigated the relationship between number of ascospores of D. pinodes at sowing and disease intensity at crop maturity. Field pea stubble infested with ascochyta blight from one site was exposed to ambient conditions at two sites, repeated in 2 years. Three batches of stubble with varying degrees of infection were exposed at one site, repeated in 3 years. Every 2 weeks, stubble samples were retrieved, wetted and placed in a wind tunnel and up to 2500 ascospores g−1 h−1 were released. Secondary inoculum, monitored using seedling field peas as trap plants in canopies arising from three sowing dates and external to field pea canopies, was greatest in early sown crops. A model was developed to calculate the effective number of ascospores using predictions from G1 blackspot manager (Salam et al., 2011b; Australasian Plant Pathology, 40, 621–31), distance from infested stubble (Salam et al., 2011a; Australasian Plant Pathology, 40, 640–7) and winter rainfall. Maximum disease intensity was predicted based on the calculated number of effective ascospores, soilborne inoculum and spring rainfall over two seasons. Predictions were validated in the third season with data from field trials and commercial crops. A threshold amount of ascospores of D. pinodes, 294 g−1 stubble h−1, was identified, above which disease did not increase. Below this threshold there was a linear relationship between ascospore number and maximum disease intensity.
Ascochyta blight is the most common disease of field pea in Australia (Davidson & Ramsey, 2000; Bretag et al., 2006), often causing 25% yield loss and sometimes up to 75% yield loss in individual crops (Bretag et al., 1995a; McDonald & Peck, 2009; McMurray et al., 2011). This disease, which has a worldwide distribution, is caused by the fungal pathogens Didymella pinodes, Phoma medicaginis var. pinodella and D. pisi. Phoma koolunga was recently identified as another component of the disease complex in South Australia (Davidson et al., 2009). Didymella pinodes is considered the major pathogen in this complex, and it produces airborne ascospores as primary and secondary inocula that are spread over long distances by wind (Bretag et al., 2006). Didiymella pisi is rarely associated with ascochyta blight in southern Australia. Inoculum of the two Phoma species consists of rain-splashed conidia (Punithalingham & Gibson, 1976; Davidson et al., 2009), although the perfect stage of P. medicaginis var. pinodella has been reported in laboratory conditions (Bowen et al., 1997).
Limited genetic resistance to ascochyta blight has been identified in field pea germplasm (Bretag et al., 2006) and disease control depends on agronomic practices such as delayed sowing and strategic application of foliar fungicides to minimize infection. However, foliar fungicides are not always cost-effective in the low rainfall environments of southern Australia where field pea yield is often less than 2 t ha−1, while in the high and medium rainfall environments the requirement for economic benefit restricts fungicide use to no more than two applications per crop (McMurray et al., 2011).
The major control strategy for ascochyta blight of field pea in Australia until recently has been to delay sowing crops until 4–6 weeks after the first autumn rains, to minimize infection from airborne ascospores of D. pinodes (Bretag et al., 2000; Davidson & Ramsey, 2000). However, delayed sowing in southern Australia often leads to yield loss as a result of heat and moisture stress during the flowering and grain-filling stages (McDonald & Peck, 2009; McMurray et al., 2011). In response to recent weather patterns comprising less rainfall and shorter growing seasons, optimum sowing dates were revised to 3 weeks after the first autumn rains in medium and medium–high rainfall regions, and to within 1 week in low rainfall regions (McMurray et al., 2011).
Daily rainfall and temperature influence the timing of ascospore release from infested field pea stubble, leading to seasonal and regional variation in ascochyta blight risk. Consequently, a forecasting system for predicting release of ascospores of D. pinodes, G1 blackspot manager, was developed to identify sowing dates that minimized the risk of ascochyta blight without delaying sowing longer than necessary. Growers were advised to delay sowing until at least 50% of the ascospores had been released and fallen on bare soil (Salam et al., 2011b).
G1 blackspot manager predicts the fraction of the ascospores available in a given season that are released on individual days, but the total number of ascospores available as inoculum in that season can vary widely. Hence the rule of sowing after 50% of ascospores have been released relates to a wide range of potential inoculum amounts. In addition, the absolute number of ascospores is influenced by proximity to infested stubble (Salam et al., 2011a) and it is likely that the number of ascospores also varies with the severity of disease on the stubble. The effect of the amount of ascospores available as inoculum on disease development has not been established.
The ability of ascospores to infect field pea is affected by humidity during the infection process (Schoeny et al., 2007) which, in turn, will affect disease. In addition, disease severity at the end of the growing season is influenced by secondary inoculum cycles of conidia and ascospores within the crop. Conidia, which develop from pycnidia within ascochyta blight lesions, are dispersed over short distances by rain (Schoeny et al., 2008) but, under Australian conditions, these are considered of minor importance (Bretag et al., 2006). However, pseudothecia of D. pinodes develop on infected, senescent leaves and on infected stems during crop maturation, releasing ascospores which spread quickly throughout the crop. Subsequent rainfall events promote infection by these ascospores, further increasing disease (Roger & Tivoli, 1996; Bretag et al., 2006). Consequently, there appears to be no direct relationship between the number of ascospores released from infested stubble and disease severity at the end of the growing season, although G1 blackspot manager identifies disease risk from the pattern of ascospore release from stubble.
Other sources of inoculum can also be important in the early establishment of ascochyta blight in field pea crops, e.g. soilborne inoculum which is able to survive for several years in the absence of the host (Wallen et al., 1967; Wallen & Jeun, 1968; Bretag et al., 2006; Davidson et al., 2011). While the pathogens are commonly detected on seed, seedborne inoculum is not considered a source of inoculum for ascochyta blight epidemics (Bretag et al., 1995b; Moussart et al., 1998).
The aims of this study were to: (i) establish the relationship between disease development and numbers of ascospores of D. pinodes released, the secondary inoculum of all the causal pathogens produced within the crop canopy, and rainfall; and (ii) develop a disease predictor, for ascochyta blight of field pea and validate it with independent field data.
Materials and methods
Ascochyta blight in field pea disease management trials
Ascochyta blight was assessed in naturally infected field pea disease management trials described in detail by McMurray et al. (2011) in three regions which differed in annual rainfall and length of growing season. These trials were conducted in 2007, 2008 and 2009 in part to assess fungicide efficacy and the effect of sowing date on disease and yield. The experimental sites were located in three areas of South Australia as follows: (i) a medium–high rainfall region (mean annual rainfall 464 mm) represented by Kingsford Research Station (34·5°S, 138·8°E), 50 km north of Adelaide, in 2007 and 2009, and the nearby Turretfield Research Station (34·6°S, 138·8°E) in 2008; (ii) medium rainfall (mean annual rainfall 429 mm) and short growing season represented by Hart, 140 km north of Adelaide (33·8°S, 138·4°E), in 2007, 2008 and 2009; and (iii) a low rainfall (mean annual rainfall 325 mm) and short growing season represented by Minnipa (32·9°S, 135·2°E), c. 600 km northwest of Adelaide, in 2007, 2008 and 2009. The trials at the medium–high rainfall sites and the medium rainfall site each had three times of sowing, with the first sowing date as soon as practicable after the first autumn rains (‘early-sown’) and subsequent sowing dates 3 weeks (‘medium-sown’) and 6 weeks (‘late-sown’) later. The trials at the low rainfall site consisted of the two earlier times of sowing, as the latest sowing date was impractical for agronomic reasons.
In order to estimate the amount of soilborne inoculum present at each site, 500 g soil were collected per site, prior to sowing, as described in Davidson et al. (2011), and subjected to DNA tests for D. pinodes plus P. medicaginis var. pinodella and P. koolunga as described in Davidson et al. (2011). The relationship between quantity of soilborne DNA and disease intensity on field pea plants at the end of winter (designated DiseaseAugSoil) was inferred from data in this previous study using Eqn (1):
Disease intensity was measured by sampling six plants of field pea cv. Kaspa collected at random from untreated buffer plots, established at each sowing date, every 2 weeks for the trials at Hart, Kingsford and Turretfield, and every 4 weeks for the trials at Minnipa, from seedling stage in June to crop maturity in October. The plants were assessed for the number of internodes on the main stem with 100% area diseased (termed ‘girdled internodes’). Internodes with partial infection were assessed for proportion of surface area diseased and the fraction was added to the total number of girdled internodes. Data for the six plants assessed per sowing treatment were averaged. Microscopic examination of representative lesions on stems and leaves of each plant was conducted to determine the presence or absence of pseudothecia containing ascospores of D. pinodes, and pycnidia containing conidia of D. pinodes, P. medicaginis var. pinodella and P. koolunga.
Daily weather data (rainfall, maximum and minimum temperature) for the three sites for each season were accessed through SILO Patched Point (SILO, 2010). The Rosedale data point represented Kingsford and Turretfield (being 7 km east of Kingsford and 2 km southwest of Turretfield), the Blyth data point represented Hart (11 km south of the site) and the Minnipa data point was on-site. Rainfall at each site × sowing date × year was summed from sowing date to end of July (designated RainWinter) and summed for August and September (designated RainAS).
Disease intensity (number of girdled internodes) at the end of August (i.e. end of vegetative phase) was regressed against RainWinter and the number of additional internodes that became girdled during crop maturation (i.e. September and October) was regressed against RainAS, using GenStat 14. These relationships were accommodated in the maximum disease estimator described below.
Estimation of airborne primary inoculum from infested stubble
The number of ascospores of D. pinodes released from infested field pea stubble was assessed in different seasons and in different regions using stubble with varying degrees of infection. Each year from 2006 to 2009, naturally infected field pea crops and research trials with ascochyta blight were selected following random monitoring. Disease was assessed on standing crops prior to harvesting as described above for 20 plants selected at random, and stubble was collected from these crops immediately after harvest (Table 1). In December 2006 and November 2007 stubble was collected from Kingsford. In December 2007 and November 2008 and 2009 three lots of field pea stubble, each with different degrees of disease intensity (Table 1), were collected from commercial field pea crops within a 10-km radius of Hart. Disease intensity on the Hart stubble was categorized as low, moderate or high based on disease observations within each season: four, seven and 15 girdled internodes, respectively, on stubble collected in December 2007, and four, eight and 13 girdled internodes, respectively, on stubble collected in November 2008. The stubble collected from the Hart region in November 2009 followed a severe ascochyta blight epidemic and the low disease category comprised nine girdled internodes, the moderate category 12 girdled internodes and the high category 19 girdled internodes. Stubble was stored in dry conditions for 1–7 weeks until processed. Segments (c. 12 cm long) of the stems with ascochyta blight lesions were placed into nylon mesh bags (20 × 20 cm with pore size 1 mm2, 20 pieces per bag). Ascochyta blight lesions completely covered more than seven girdled internodes of the stem pieces. The stubble was incubated on the soil surface at either Kingsford or Hart, whichever was closest to the stubble collection site (Table 1). Steel mesh (7·5- × 7·5-cm grid size) was laid over the nylon mesh bags to prevent disturbance by wind and animals were excluded with a 1-m-high steel mesh enclosure. In 2007 and 2008 the stubble collected from Kingsford was also incubated at Waite Campus, Urrbrae (34·9°S, 138·6°E). An automatic weather station (AWS) (Measurement Engineering Australia) was placed at each site to record daily rainfall and daily average temperature. Missing data were substituted by data from the nearest Bureau of Meteorology site (http://www.bom.gov.au/climate/data).
Table 1. Release of ascospores of Didymella pinodes from field pea stubble after incubation in the field
Every 2 weeks a bag of stubble representing each category of disease intensity was collected from each incubation site (Table 1). The bags of stubble were sent to the Northam Laboratory of the Department of Agriculture and Food Western Australia (DAFWA), where the number of mature ascospores was assessed. Each stubble sample was wetted for 5 min and placed in a wind tunnel for 1 h, such that D. pinodes ascospores released from the stubble were captured on sticky tape (1 × 30 mm) mounted on rotor rods. The number of ascospores on the sticky tape was counted using a light microscope at × 400 magnification (Galloway & MacLeod, 2003; Salam et al., 2011b) and data presented as ascospores g−1 stubble h−1. The number of ascospores captured every 2 weeks from the different batches of stubble incubated at Hart was compared within a season using correlation analysis in genstat 14.
Monitoring of secondary inoculum
The timing of release and relative amounts of secondary inoculum (conidia and ascospores) of ascochyta blight pathogens were monitored for field pea canopies with different sowing dates and regressed against rainfall and disease intensity in the field pea canopy. The relative number of spores (airborne, soilborne and/or splash-dispersed combined) at Kingsford and Turretfield Research Stations in the canopy for each sowing date over three seasons (2007 to 2009) was monitored indirectly by counting the number of lesions on trap plants of field pea seedlings as described by Roger & Tivoli (1996) and Schoeny et al. (2007). Trap plants consisted of a tray containing 3-week-old field pea cv. Parafield seedlings, 12 per tray, that were raised in the greenhouse before being placed in the field. Trap plant placement is detailed in Table 2. All trap plants within the field pea canopies were placed in control plots which received no fungicide applications. After 7 days of exposure the trap plants were returned to the greenhouse and placed in plastic trays covered with lids. Water was added to the trays to a depth of 2 cm to provide high humidity and the temperature was maintained between 18 and 25°C. After 4 days the trap plants were placed on an open bench so that humidity decreased, the plants were incubated for a further 3 days and the total number of lesions on each plant in the tray was counted. The mean number of lesions per plant was calculated for each tray. Lesions on trap plants external to the field pea canopy from May to the end of July were assumed to be caused by primary inoculum from infested stubble and from soilborne inoculum. Lesions on trap plants within the field pea canopy were assumed to be caused by both primary inoculum and secondary inoculum produced in the lesions on the infected plants. The relative amounts of secondary inoculum from each canopy from May to the end of July were estimated by subtracting the number of lesions on trap plants outside the trials from the number of lesions on traps within the trials. After July, when airborne primary inoculum was depleted, the relative amounts of secondary inoculum per canopy were estimated as the total number of lesions on the traps inside the pea canopies; soilborne inoculum was presumed to be equivalent across the three times of sowing in each trial.
Table 2. Placement of field pea seedling trap plantsa from 2007 to 2009
Position of trap
Trial sowing date
Date first trap placed in field
Date last trap placed in field
Trays of field pea seedlings, 12 seedlings per tray, were grown in the greenhouse until 3 weeks old then exposed to field conditions for 1 week, after which they were returned to the greenhouse to allow ascochyta blight lesions to develop.
All trials were within 1·5 km of ascochyta-blight-infested field pea stubble.
No sowing date is presented because the trap plants were external to the field pea canopy.
The disease management trials (McMurray et al., 2011) included three sowing dates; trays were placed in the canopy arising from the first sowing date in 2007 and all three sowing dates in 2008 and 2009.
10 m from infested field pea stubble and 300 m from field pea trialb
The average number of lesions on the weekly trap plants adjacent to infested pea stubble in 2007 and 2008 was compared with the number of ascospores released from infested pea stubble over the same period, using Spearman's rank correlation in GenStat 14. The significance of the correlation coefficient was determined in this analysis using Student's t-distribution with n − 2 degrees of freedom at 5% probability.
Multiple linear regression in GenStat 14 was used to analyse the relationship between (i) the total weekly rainfall (mm) and the average number of lesions on the weekly trap plants in canopies of the disease management trials from 2007 to 2009, and (ii) disease intensity in the disease management trials as number of girdled internodes (averaged over the six plants) and the average number of lesions on the weekly trap plants. Data across years were tested for homogeneity before pooling. Lesion numbers and disease were square-root-transformed for the regression analyses to standardize the residuals.
Maximum disease estimator
The experiments described above, as well as predictions from G1 blackspot manager (Salam et al., 2011b) and data on distance from infested stubble (Salam et al., 2011a), were used to develop a three-step model that estimated (i) the number of ascospores available as primary inoculum, designated ‘effective number of ascospores’, (ii) maximum disease intensity at the end of winter using the relationship between the calculated effective number of ascospores and rainfall recorded in the disease management trials described above, and (iii) the final maximum disease intensity at the end of the growing season as the sum of the disease at the end of winter plus disease following rainfall during crop maturation (August and September), using data collected from the disease management trials in 2007 and 2008 described above. This model was named the ‘maximum disease estimator’.
The maximum disease estimator started with either 5000 or 10 000 ascospores per region [SporesInit] based on the total number of ascospores captured from the stubble at Hart and Kingsford each year and the observation that numbers larger than 10 000 saturated the model. SporesInit numbers were reduced according to the following ratios: (a) percentage ascospores remaining on infested stubble at sowing for each site × sowing date as predicted by G1 blackspot manager (Salam et al., 2011b) [%Spores]; (b) distance from known infested field pea stubble around trial sites [Distance] (Salam et al., 2011a); (c) relationship between cumulative winter rainfall [RainWinter] from sowing to end of winter (August) and disease intensity as observed in the disease management trials. A rainfall lower limit was set for no disease and an upper threshold of rainfall was set for maximum disease. The effective number of ascospores was calculated using Eqn (2):
The effective number of ascospores calculated in Eqn (2) was regressed against the observed disease intensity at end of winter (August) minus the amount of disease attributed to soilborne inoculum (Eqn (1)) in all trials and sowing dates to generate the relationship in Eqn (3). One data point was omitted from the regression because the observed disease at this point was much less than the observed disease for similar numbers of ascospores in other year × site combinations. The amount of disease from soilborne inoculum (DiseaseAugSoil) was added to Maximum Disease (Aug) in Eqn (3):
A maximum disease intensity was set for the end of winter based on observations in the disease management trials. Stepwise correlation analysis was performed in Microsoft excel 2007 between the effective number of ascospores calculated in Eqn (1) and the observed disease intensity (number of girdled internodes) at the end of August to identify when the number of girdled internodes reached a maximum, after which any additional ascospores had minimal influence on disease. The maximum value for this parameter was included in the final disease model.
Finally, the relationship between rainfall in August and September [RainAS] and disease intensity during crop maturation was added to the disease calculated in Eqn (4):
Model parameters are shown in Table 3. Final maximum disease intensity predicted for all the sowing dates and sites in the 2007 and 2008 trials was plotted against the effective numbers of ascospores calculated in Eqn (2), and the final maximum disease intensity predicted for the sowing dates × sites for the medium–high and medium rainfall regions in 2007 and 2008 was linearly regressed [PredictedRegression] against the percentage of ascospores present at sowing calculated from G1 blackspot manager (Salam et al., 2011b).
Table 3. Parameters of maximum disease estimator model
Regression parameter for ‘Disease at end of August’ and ‘Effective numbers of ascospores’.
Regression constant for ‘Disease at end of August’ and ‘Effective numbers of ascospores’.
Regression parameter for ‘Increase in Disease after Winter’ and ‘Cumulative Rainfall August and September’.
Regression constant for ‘Increase in Disease after Winter’ and ‘Cumulative Rainfall August and September’.
Survey of commercial field pea crops and validation of maximum disease estimator
Data from the 2009 field trials described above and from a survey of commercial field pea crops in 2009 described below were used to validate the maximum disease estimator. Each year from 2007 to 2009, all field pea crops within a 10-km radius of Hart were identified and mapped. Approximate sowing dates were calculated in winter from the mean number of internodes on 20 plants selected arbitrarily in the crops. This information was used to group crops into sowing categories similar to the sowing dates in the field trials described above: early (late April to early May), medium (mid- to late May) and late (early June onward). No data were available on crop rotation or soilborne inoculum. Crops representative of each sowing group were selected for assessment of ascochyta blight in late September or October. Selection within each sowing group was based on proximity to infested field pea stubble, such that crops on or adjacent to, within 500 m of, or more than 500 m from infested stubble were represented. Twenty plants were selected in a W transect across the field (Davidson et al., 2001), one every 50 paces. Plants were assessed for growth stage (vegetative, flowering, early pods, mature pods), total number of internodes, and number of internodes girdled by ascochyta blight. The effect of sowing period on disease intensity was analysed in 2007 using two-sample t-tests in GenStat 14; no data were collected for proximity to field pea stubble in 2007. In 2008 and 2009 the effect of sowing period and proximity to field pea stubble on disease intensity was analysed by unbalanced analysis of variance in GenStat 14 using crops as replicates.
To validate the maximum disease estimate, the observed disease intensity in the 2009 field trials and in the 2009 commercial crops was compared with maximum disease intensity predicted by the maximum disease estimator using Lin's concordance correlation coefficient. This analysis assesses the linear relationship between the two measurements and the degree to which the pairs fall on the 45° line through the origin (Lin, 1989). Where multiple observations had the same maximum disease prediction, only the maximum observation was included in the correlation, because only maximum disease observations were used to generate the relationship in Eqn (3).
Linear regression, in GenStat 14, was performed between observed disease intensity at the end of the season and the percentage of ascospores present at sowing calculated by G1 blackspot manager (Salam et al., 2011b). The regression slope was compared with the slope for PredictedRegression described above, using t-tests at α = 0·05.
Ascochyta blight in field pea disease management trials
No disease was observed when cumulative rainfall from sowing (April–May) until the end of July was less than 50 mm. Disease intensity at the end of winter (August) reached a maximum of 11 girdled internodes when the cumulative rainfall, from sowing until the end of July, was more than 100 mm. There was a significant linear regression (r2 = 0·621, P <0·001) for cumulative rainfall over this period between 50 and 100 mm rainfall and disease intensity at the end of August (Fig. 1). This result was used to set lower (<50 mm) and upper (>100 mm) limits for winter rainfall categories in the maximum disease estimator (Table 3). There was also a significant regression (r2 = 0·597, P <0·001) between the increase in disease in spring (September and October) and the rainfall during August and September.
The combined quantity of DNA of the pathogens (D. pinodes, P. medicaginis var. pinodella and P. koolunga) prior to sowing the field pea trials was zero in the Minnipa trials, 153–171 pg g−1 soil in the 2007 trials, 1024–1510 pg g−1 soil in 2008 and 23–490 pg g−1 soil in the 2009 trials at Hart, Kingsford and Turretfield. The disease intensity at the end of August attributed to soilborne inoculum was estimated using Eqn (1) to be less than 0·37 internodes per plant in all trials.
Estimation of airborne primary inoculum from infested stubble
The total number of ascospores released from infested stubble and captured in the wind tunnel in one season was between 3962 and 9986 at Kingsford and Waite Campus, respectively (Table 1). In 2007 and 2008, ascospores were first detected from stubble retrieved in January and detection ceased in June at Waite Campus in both years and in late July or August at Kingsford in 2007 and 2008, respectively (Fig. S1). The patterns of spore release at Kingsford and Waite Campus were significantly correlated in both seasons (2007, r =0·74, P <0·001; 2008, r =0·59, P <0·01). There was a peak in ascospores for stubble samples in April–May in both years, associated with rainfall above 25 mm in a 2-week period. An additional peak in February 2007 from stubble incubated at Kingsford in 2007 coincided with 40·2 mm of rainfall in the 2-week incubation period. At Waite Campus, over the same period, rainfall was less than 14 mm, and few ascospores (0–661 g−1 h−1) were recorded at this site in the summer months of both years. Another variation occurred in the 2 weeks preceding and following 2 April 2007, when more spores were captured from stubble incubated at Kingsford than at Waite Campus; this was associated with four additional rain days at the former site. Ascospore release from May to July during 2008 peaked slightly later at Kingsford than at Waite Campus; Kingsford had fewer rain days, and/or less rainfall, than Waite Campus in all but one of the 2-week incubation periods (Fig. S2).
The total number of ascospores captured in the wind tunnel in each season at Hart was similar to that of Kingsford and Waite Campus for stubble with low and moderate disease intensity (as described in Table 1), between 4205 and 11 830, except in 2008 when the stubble with moderate disease released 53 320 ascospores over the season. The stubble with most disease released up to 159 059 ascospores in one season (Table 1). At Hart, ascospores were trapped from stubble samples collected from January until mid-September, mid-July and late June in 2008, 2009 and 2010, respectively, and the maximum numbers were recorded during the months of April and May (Fig. S3), coinciding with rainfall of more than 20 mm in a 2-week period (Fig. S4).
The numbers of ascospores released from the three lots of field pea stubble incubated each year at Hart were significantly (P <0·05) correlated within each season (Table 4), except for the stubble from plants with high disease intensity compared to stubble with moderate disease intensity collected in 2009. Although release patterns were similar for each stubble sample, the disease intensity on the stubble at the time of harvest affected the number of ascospores detected. In 2008 and 2009 more ascospores were released from the stubble with high disease intensity than from the stubble samples with low or moderate disease intensity. The largest number released from a stubble sample in 2008 and 2009 was 26 205 and 91 695 ascospore g−1 h−1, respectively. However, in 2010 more ascospores (22 866 ascospores g−1 h−1) were obtained from the stubble with moderate disease intensity (12 girdled internodes) than from stubble with high disease intensity (19 girdled internodes). The stubble incubated in 2010 was the most severely diseased of all the material used, but the largest number of spores captured in the wind tunnel occurred in 2009 from stubble with disease comprising 13 girdled internodes.
Table 4. Correlation between numbers of ascospores of Didymella pinodes captured in a wind tunnel each fortnight from field pea stubble and disease severity on the stubble incubated in the field at Hart, South Australia for 2008, 2009 and 2010
15, 13 and 19 nodes represent high disease severity; 7, 8 and 12 nodes represent moderate disease severity; 4, 4 and 9 nodes represent low disease severity.
15 nodes vs 7 nodes
15 nodes vs 4 nodes
7 nodes vs 4 nodes
13 nodes vs 8 nodes
13 nodes vs 4 nodes
8 nodes vs 4 nodes
19 nodes vs 12 nodes
19 nodes vs 9 nodes
12 nodes vs 9 nodes
Monitoring of secondary inoculum
The average number of ascochyta blight lesions on the field pea seedlings in trays placed near infested field pea stubble or outside but close to field pea trials at the beginning of the growing season (May) varied from 1281 per seedling in 2007 to below 30 per seedling in 2008 (Fig. 2). In June and July in both years, the average number of lesions was below 40 per seedling and remained at this low level for the rest of the season. Conversely, mean numbers of lesions on trap plants placed within the field pea canopy increased as the season progressed, reaching a maximum in September each year, with peaks of 177, 805 and 341 lesions per seedling in 2007, 2008 and 2009, respectively (Fig. 2). In 2008, one earlier peak of 548 lesions per seedling occurred in August and in 2009 there were four earlier peaks, ranging from 133 to 211 lesions per seedling, from July to August. Associated with these peaks, rainfall was greater in June and July 2009 than in 2007 and 2008. There was a total of 130 mm from the start of the growing season to early July 2009, while only 95 mm fell by early August 2008 and a similar amount by September 2007. Pseudothecia containing ascospores of D. pinodes were detected on field pea plants in the trials during these peak periods, as were pycnidia containing conidia of D. pinodes, P. koolunga and P. medicaginis var. pinodella. Conidia were also observed outside these peak periods (data not shown).
The correlation between numbers of ascospores trapped from infested field pea stubble and lesions on trap plants in 2008 was significant (r =0·86, P =0·006) for traps placed next to infested field pea stubble and significant (r =0·71, P =0·053) for traps placed next to the field pea disease management trial. Numbers of ascospores and numbers of lesions on trap plants showed similar patterns in 2007 (Fig. 3).
Within a season, earlier sowing resulted in more lesions on trap plants, presumed to arise from secondary inoculum (Fig. 2). In 2008, trap plants placed in the first sowing date treatment had a total of 2396 lesions for the season, those in the second sowing date treatment had a total of 1238 lesions and those in the third sowing date treatment, 295 lesions. In 2009 the total number of lesions on trap plants placed in the first, second and third sowing date treatments was 2117, 877 and 606, respectively.
The number of lesions (square root) on the trap plants within the field pea canopy, analysed using all data from 2007 to 2009, increased (P <0·001, r2 = 0·4) with increasing disease intensity (number of girdled internodes) in the trial and with weekly rainfall (P <0·002, r2 = 0·4). The interaction between rainfall and plant disease was not significant. Few lesions were observed on trap plants when trial plants had an average of less than one girdled internode, although where rainfall of 33·4 mm was recorded large numbers of lesions developed on trap plants placed in the canopy, even though the average disease intensity on the plants in the trial was less than two girdled internodes. Numerous lesions were typically observed on trap plants when total weekly rainfall was 10 mm or more and also when rainfall was less (6·2 mm) if the trial plants had an average of four or more girdled internodes (Fig. 4).
Maximum disease estimator
There was a significant linear relationship between the effective number of ascospores and number of girdled internodes at the end of winter (August) (Eqn (2)); and the stepwise correlation analysis identified a maximum disease intensity at the end of August of 11 girdled internodes at 294 g−1 h−1 effective ascospores, after which disease did not increase.
Disease increase after winter was correlated with rainfall in August and September (Fig. 1). Maximum disease intensity at the end of the growing season was estimated using the model for each sowing date in each disease management trial and plotted against the calculated effective number of ascospores (Fig. 5). Disease intensity increased linearly (disease at end of growing season [observed] = 1·567 + 0·0386 × effective number of ascospores; r2 = 0·802, P <0·001) to a maximum of 20 girdled internodes when effective numbers of ascospores were 294 g−1 h−1, after which disease did not increase. The value of 0·37 girdled internodes at the end of August attributed to soilborne inoculum was calculated to increase up to five girdled internodes by the end of the season because of the additional effect of spring rainfall. The predicted maximum disease intensity at the end of the season had a linear relationship with percentage of ascospores at sowing for the combined data from Hart, Kingsford and Turretfield sites in 2007 and 2008 (r2 = 0·687, P <0·001) (Fig. 6). At these sites 50% of ascospores remaining at sowing resulted in a predicted disease intensity of 10 or fewer girdled internodes. At the low rainfall site, Minnipa, the estimated disease in 2007 and 2008 remained at zero girdled internodes when 75% of ascospores remained at sowing, because low rainfall reduced effective ascospores to less than 8% of the potential number.
Survey of commercial field pea crops and validation of maximum disease estimator
In 2007, 2008 and 2009, respectively, 52, 45 and 41 commercial field pea crops were mapped within a 10-km radius of Hart. Sowing dates were evenly spread in all three seasons; 32·6% were in the early-sown category, 35·6% were in the medium-sown category and 32·7% were late-sown. The majority of the crops were in close proximity to infested field pea stubble from the previous season; 49·4% were either adjacent to or planted into field pea stubble, 29·9% were no more than 500 m from field pea stubble, and only 20·6% of crops were more than 500 m from field pea stubble. All crops were affected by ascochyta blight and disease was assessed in 18, 15 and 22 crops in 2007, 2008 and 2009, respectively.
In each consecutive year of the study, disease significantly increased with earlier sowing (P ≤0·01) (Table 5). The intensity of ascochyta blight ranged from 0·4 to 14·8 (average 5·4) girdled internodes in 2007, and from 1·8 to 12·7 girdled internodes (average 5·9) in 2008. In 2009 the minimum disease intensity was 8·3 and the maximum was 20·2 girdled internodes (average 13·2). In 2008 proximity to infested stubble significantly (P <0·01) increased ascochyta blight at each sowing period. Disease was lowest in crops sown in the medium or late period which were not adjacent to infested stubble (Table 5b). In 2009 disease was lowest in crops sown in the late period not adjacent to infested field pea stubble (Table 5c).
Table 5. Mean severity of ascochyta blight (number of girdled internodes) on 20 plants per crop sown in three sowing periods, either on or adjacent to, or not adjacent to infested field pea stubble in a 10-km radius of Hart, South Australia in 2007
Early (early May)
Mid (mid–late May)
Late (early June onward)
Entries for 2007 are means of observations; entries for 2008 and 2009 are predictions from a regression model generated by unbalanced analysis of variance.
Within each year, numbers followed by the same letter are not significantly different at P <0·001.
Maximum least significant difference = 1·00; average least significant difference = 0·93; minimum least significant difference = 0·82
Adjacent or on field pea stubble
Not adjacent to field pea stubble
Maximum least significant difference = 1·08; average least significant difference = 1·02; minimum least significant difference = 0·93.
There was a significant correlation (r =0·8436, P <0·001) and concordance (Cb = 0·9339) between observed maximum disease intensity and predicted maximum disease intensity for the 2009 field trials and the survey of commercial field pea crops in the same year. There was also a significant linear relationship (r2 = 0·294, P <0·001) between observed disease at the end of the season in 2009 and percentage of ascospores present at sowing predicted by G1 blackspot manager (Fig. 6). There was no significant difference between the slopes of the linear regressions between percentage of ascospores present at sowing and observed or predicted disease.
A threshold number of ascospores of D. pinodes above which severity of ascochyta blight on field pea did not increase was identified. Below this threshold there was a linear relationship between numbers of ascospores and maximum disease intensity. This finding verified the assumption made in the G1 blackspot manager model (Salam et al., 2011b) that the timing of peak release, identified through the percentage of the total ascospores available at a given time, and sowing date were the primary factors influencing ascochyta blight disease in field pea in southern Australia. At the peak the numbers of ascospores released from the stubble exceeded the threshold for maximum disease intensity and crops should be sown after this event to limit development of epidemics. Not only were early sown canopies exposed to more primary inoculum, they also produced ascospores before the later sown canopies and in greater numbers. Winter and spring rainfall were positively correlated with disease intensity, and disease was not observed when winter rainfall was less that 50 mm, an important finding for low rainfall regions.
The effective number of ascospores calculated by the maximum disease estimator varied with the initial numbers on the stubble, the percentage of spores remaining at sowing (according to G1 blackspot manager), rainfall and distance from infested stubble. Disease intensity at the end of winter, when primary inoculum was no longer present, reached a maximum of 11 internodes girdled, irrespective of the effective number of ascospores. This was supported by the data used for validation, where an estimated 3500 effective spores did not result in more disease than 294 effective spores, whereas between 0 and 294 effective spores, there was a linear relationship with maximum disease. A similar relationship between ascospores discharged from infested stubble and disease has been reported for blackleg (phoma stem canker) of canola (oilseed rape) (Wherrett et al., 2004). The results from the current study confirm that timing of ascospore release is more important than the absolute numbers of spores. Crop growth stage and rainfall at the time of maximum ascospore release are likely to be determining factors in the establishment of disease on the crops during autumn and winter. Schoeny et al. (2003, 2007) also related ascochyta blight severity in field pea to rainfall during disease onset and crop growth.
The maximum disease estimator revealed that when 50% or less of ascospores were present at sowing, the disease intensity at crop maturity was 10 internodes or less (c. 30% of the entire stem) girdled with ascochyta blight. This is in agreement with the 20–40% of stem infected calculated by Salam et al. (2011b) from Western Australian data. Salam et al. (2011c) estimated the yield loss associated with this disease score to be 20% or less and similar figures were observed in the disease management trials in South Australia described by McMurray et al. (2011). In the low rainfall region of Minnipa, a higher percentage of spores can be present at sowing without increasing the risk of disease. This information is especially important for farmers in these dry regions where early sowing is essential to allow maximum yield potential.
The total number of ascospores released from infested field pea stubble per season varied with site and season but the factors that influence seasonal variation are not known. The total number was not directly related to disease on the stubble incubated in these experiments. It was anticipated that most ascospores would be obtained from the 2009 stubble incubated in 2010, on which disease intensity was greatest, but this did not occur. It is possible that when numerous crops of ascospores are released within the canopy as secondary inoculum, as occurred in 2009, the pathogen in the subsequent stubble is depleted and produces fewer ascospores as primary inoculum in the following autumn. Field pea residues are known to develop successive crops of mature pseudothecia, initially prolific, exhausting the inoculum by 50 weeks (Carter, 1963; McDonald & Peck, 2009).
Ascospore numbers did vary with disease intensity on stubble within a season, showing that the more diseased the crop the greater the potential to release ascospores from the subsequent stubble. Irrespective of the amount of disease, the pattern of release remained the same within a site and season. In 2010 more ascospores were obtained from the stubble with moderate disease than from the most diseased stubble, although this apparent anomaly may have been a result of experimental error.
Although airborne ascospores of D. pinodes are considered the primary inoculum during establishment of field pea crops, soilborne inoculum has also been associated with disease (Davidson et al., 2011), but distinguishing between the different sources of inoculum is difficult. Using the relationship identified by Davidson et al. (2011) to estimate the effect of soilborne inoculum in these trials indicated that a minor component of the inoculum in these trials came from the soil, and the majority from airborne ascospores. Nevertheless, when soilborne inoculum is more abundant than measured here, combined with high rainfall it has the potential to cause significant disease. As data on soilborne inoculum were not available in the survey of commercial crops, this aspect could not be considered in the validation process, which may have contributed to the variability of the results.
Lesions on trap plants attributed to secondary inoculum increased during August or later as the crop matured, depending on rainfall and disease in the crop. This secondary inoculum results in an increase in ascochyta blight epidemics in late winter and spring (Bretag, 1991; Roger & Tivoli, 1996). Examination of the trap plant data identified the compounding effect of sowing date on secondary inoculum. In the current study, the number of spores produced in-crop was greatly increased by early sowing where conditions were conducive for disease, and ascospores were produced much earlier in early-sown crops than in crops sown later, leading to more ascochyta blight in spring. Late-sown crops are exposed to relatively few ascospores from the previous year's stubble and, subsequently, produce fewer in-crop spores, leading to little disease. These results support the practice of later sowing to control ascochyta blight in field pea, as was demonstrated in the field trials. Roger & Tivoli (1996) found that the number of conidia produced in-crop increased with disease. In the current study the greatest peaks of secondary inoculum and increase in disease occurred when pseudothecia containing ascospores were readily detected on the diseased plants, confirming that ascospores as secondary inoculum have an important role in an epidemic (Bretag, 1991). As ascospores are produced on senescent plant material, early intervention to control disease to prevent premature senescence at the base of the plant could be a strategy to minimize or limit the development of epidemics (Roger & Tivoli, 1996).
Production of secondary inoculum in ascochyta-blight-affected field pea crops was also strongly linked to quantity and timing of rainfall, and the progressively earlier detection of lesions on trap plants placed within field pea canopies from 2007 to 2009 was linked to higher rainfall during the later growing seasons. Given that there was no interaction between disease intensity and rainfall in terms of the number of lesions detected on trap plants, rainfall was deemed the most suitable parameter for use in the predictive model to calculate the increase in disease during spring.
The survey of commercial field pea crops in the Hart district provided support for the previous observations on the influence of sowing date and distance from infested pea stubble on disease, the latter identified by Salam et al. (2011a). Many growers appear to have ignored basic agronomic disease management strategies of distance from stubble and/or delayed sowing. This may be because of constraints in field selection on the property and the yield risk associated with short dry seasons when sowing is delayed. In these circumstances, G1 blackspot manager allows growers to identify the disease risk linked to their agronomic decisions.
In conclusion, there was a threshold number of ascospores of D. pinodes from infested field pea stubble below which this primary inoculum was directly related to disease intensity at the end of winter. Disease did not increase above this threshold. The primary inoculum was also a determining factor in the amount of secondary inoculum produced in the crop, with rainfall leading to increased disease during spring. The influence of rainfall meant that the management of this disease was especially important in medium–high and medium rainfall areas. Research to identify reasons for the failure of industry to implement current recommendations for field selection and distance from infested stubble is warranted to improve the adoption of integrated disease management strategies aimed at minimizing exposure to inoculum.
The South Australian Grains Industry Trust and the Grains Research Development Corporation provided financial support for this study. Technical assistance was provided by Marzena Krysinska-Kaczmarek and Michelle Russ of SARDI, Urrbrae; Pip Payne and Tessa Humphreys of Department of Agriculture and Food Western Australia. Larn McMurray and Stuart Sherrif of SARDI, Clare, and Peter Hooper, Hart Field Site Group, assisted with the surveys of commercial field pea crops in the Hart district and collection of stubble pouches. The SARDI New Variety Agronomy group led by Larn McMurray at Clare and William Shoobridge of SARDI at Minnipa conducted field pea disease management trials.