Introduction

Leptosphaeria maculans is a hemibiotrophic ascomycete fungus that causes blackleg disease (Phoma) in Brassica napus (canola, oilseed rape) crops. Leptosphaeria maculans survives between seasons on residue (“stubble”) of infected canola after harvest where sexual reproduction occurs between isolates of opposite mating type to form pseudothecia – a fruiting body that produces ascospores when fully matured (Salisbury et al., 1995). The release, attachment, and stomatal penetration of the primary ascospore inoculum on cotyledons and leaves initiate the first stage of infection resulting in leaf lesions (West et al., 2001). The fungus then has a biotrophic phase of growth from the lesion through the petiole to the stem (Hammond et al., 1985), developing necrotrophically to form crown cankers at the stem base which disrupt the flow of water and nutrients in the plant causing yield loss (Hall, 1992). Asexual conidiospores that develop in leaf lesions cause secondary infection to neighbouring plants, however, as the secondary infection contributes little to no effect on the severity of disease at the crown (Li et al., 2006), the lifecycle of blackleg is generally regarded as monocyclic, particularly in Australia, Canada, and Europe (Salam et al., 2007).

The severity of blackleg crown canker development in any given season is often associated with the epidemic severity in the previous season and the growth stage of the crop at infection. There is a positive correlation between blackleg crown canker severity at crop harvest and the number of ascospores released in the following cropping season (McGee, 1977; Petrie, 1995). This relationship was supported by a crop level investigation where the number of pseudothecia, and therefore ascospores, increased with increasing crown canker disease rating of infected plants (Lô-Pelzer et al., 2009). Subsequently, a high level of ascospores released early in the season results in severe crown canker (McGee & Emmett, 1977), hence the association of blackleg disease severity between growing seasons.

In addition, the onset of ascospore release typically coincides with early seedling growth—the critical seedling stage of the crop (West et al., 2001) which further contributes to severe blackleg disease epidemics (Hammond & Lewis, 1986; Marcroft et al., 2005). Existing agronomic interventions are targeted to reduce the density of ascospore inoculum and reduce the risk of infection at the early seedling stage. These include stubble management (burning/burying), crop rotation (> 2 years between canola crops), isolation of crops from infected stubble, choice of sowing date, use of cultivars with specific and/or partial resistances, and chemical treatments (Bondad et al., 2023). Reduced exposure to ascospores by burial, burning, isolation or long rotations, and early sowing effectively decreases the severity of blackleg disease epidemics (Aubertot et al., 2004; Guo et al., 2005; Marcroft et al., 2004; Turkington et al., 2000; Van de Wouw et al., 2021). Application of fungicides to seed or fertiliser at sowing and in crop during early seedling stages reduces the severity of infection (Khangura & Barbetti, 2002). However, understanding the mechanisms driving the development of blackleg inoculum – and the interplay between such development and the broader farming system – will be necessary if the disease is to be effectively controlled.

Temperature and moisture play a significant role in the development to maturation of pseudothecia. The different stages of pseudothecial maturation are based on the length and formation of the asci, – the saclike structures within a pseudothecium that contains the ascospores. Studies identify maturation rate to either the first observed mature pseudothecia (Pérès et al., 1999) or when 50% of pseudothecia mature (Toscano-Underwood et al. 2003). In controlled experiments in which L. maculans was exposed to continuous wetness, the maturation rate of the first pseudothecia was fastest at a constant temperature of 14 °C (14 days) and decreased with increasing temperature (21 to 34 days) (Pérès et al., 1999). In contrast, when 50% of the population was observed, development took approximately 23 days at 20 °C compared to 40 and 56 days at lower temperatures of 10 °C and 5 °C, respectively (Toscano-Underwood et al. 2003).

In the field, ascospore release generally occurs during periods of rainfall (West et al., 2001), but with fluctuations in temperature and rainfall distributions between seasons, locations and years, the onset of pseudothecial maturation and ascospore discharge pattern varies (Khangura et al., 2001; Salam et al., 2003). This has been shown in Canada (Guo & Fernando, 2005; McGee & Petrie, 1979), France (Aubertot et al., 2004), the United Kingdom (Huang et al., 2005) and Australia (Khangura et al., 2001). Weather-based simulation has been suggested to forecast the onset of pseudothecial maturation and ascospore release in all oilseed rape growing regions worldwide. The model called ‘Sporacle Ezy’ model (Salam et al., 2007) was developed from the Australian-focused model, ‘Blackleg Sporacle’ (Salam et al., 2003). This model was subsequently tested in Australia, Canada, France and the United Kingdom. The measurements for predictions are based on the appearance of the first mature cohort (group in a population) of pseudothecia or < 10% ascospores dispersed (Salam et al., 2007). Overall, these weather-based models performed reasonably well in most locations; however, model parameters required calibration to each new environment (e.g., lower limit for rainfall and number of favourable days for pseudothecial maturation). Pseudothecial development is generally assumed to start after harvest in the Sporacle model but observations that crown cankers generally appear at the start of the flowering period suggest that development of L. mcaulans pseudothecia begins prior to harvest (West et al., 2001).

Stubble management (“standing” stubble vs “lying” stubble) on the rate of pseudothecial maturation are not accounted for in current simulation models. Across a growing season, “standing” stubble releases 30% of the ascospores inoculum to that of “lying” stubble (McCredden et al., 2018), but little is known of the effects of stubble orientation on the maturation rate of L. maculans pseudothecia. That is, does the population in the “lying” stubble develop faster and release ascospores earlier than those in “standing” stubble?

In this study, we assess the dynamics of L. maculans pseudothecia maturation in populations on canola stubble from different environments, seasons and under different management practices to identify the mechanisms affecting pseudothecial maturation and ascospore release. An alternative modelling approach was derived to improve prediction of the onset of ascospore release based on the improved understanding of the interaction between pathogen development, environment and stubble management practices with comparison across different canola-growing regions globally.

Materials and methods

Stubble collection and sampling of L. maculans pseudothecia

Canola stubble was collected from commercial canola crops at Horsham, Victoria (-36.692723, 142.243028) in 2021 and 2022 and CSIRO Agricultural Research Station at Boorowa, NSW (-34.465928, 148.702374) in 2022. In 2021, stems of canola stubble in either a vertical “standing” position or a horizontal “lying” position in contact with the soil surface, were collected from early June. In 2022, “standing” stubble was collected at Horsham from early May and Boorowa from early April. At each location in both seasons, seven intact canola stems with roots attached were sampled weekly until the first week of June. Samples were air-dried at room temperature for at least 48 h to ensure that no further development of the inoculum occurred.

Ten samples of L. maculans pseudothecia were collected randomly over a 10 cm length from each piece of stubble (total 70 pseudothecia per stubble orientation at each timepoint) using forceps and were placed on a glass slide (25 × 75 × 1 mm). A drop each of distilled water and trypan blue (0.4% concentration) were added to the specimen before applying the cover slip (24 × 40 mm). Trypan blue was used to differentiate the dead tissue (blue colour) from the living tissues/cells with intact membrane (clear in colour). The asci were extruded from the pseudothecial fruiting bodies by the application of gentle pressure to the cover slip. The maturity class of each pseudothecium in the specimen was assessed under ZEISS® Axio Imager M2 at 40 × magnification. Immature, mature and empty pseudothecia were identified and recorded based on the description provided by (Toscano-Underwood et al., 2003). Pseudothecial development was assessed according to a 1–3 scale of pseudothecial maturity in which 1 = immature [stage classes A-C], 2 = mature [stage class D], and 3 = empty pseudothecia [stage class E].

Fungal development in lying and standing canola stubble

$$PMI=\frac{\sum {n}_{i}({st}_{i})}{N}$$
(1)

For each sample, a pseudothecial maturity index (PMI) was calculated as the weighted average of the observed stages using Equation 1. Where \({n}_{i}\) = number of pseudothecia in the developmental stage \(i\), \({st}_{i}\) is the value of the pseudothecial developmental stage (1–3), and N = total number of L. maculans pseudothecia assessed. Due to the similar physical resemblance of pycnidia and pseudothecia of other fungi to L. maculans pseudothecia on the stubble samples, the number of non-L. maculans fruiting bodies was recorded and deducted from the total number of samples collected. Pseudothecia containing L. maculans ascospores were easily recognised based on the spore appearance for each spore type (e.g., ascospores versus pycnidiospores) and pathogen species (e.g., L. maculans versus Alternaria brassicae). The changes in the number of mature [Stage 2] and empty [Stage 3] pseudothecia from a sample population as well as the PMI over time for each stubble type were fitted with a nonlinear regression model using the analysis of variance (ANOVA) model fit comparisons. The relationship between the rate of PMI increase in standing and lying stubble was assessed using the standard linear regression.

L. maculans pseudothecial maturation models

Generally, the development of L. maculans pseudothecia in canola stubble occurs when cool temperatures are combined with wet conditions (West et al., 1999). This theory is supported by controlled experiments (Pérès et al., 1999) and field observations (Salam et al., 2003). As wetness is essential for inoculum development in canola stubble (Toscano-Underwood et al., 2003), rainfall threshold values have been used for existing prediction models. The Blackleg Sporacle model accumulated 4 mm rain over a period of seven days to initiate development (Salam et al., 2003) which was later modified to a single rainfall event ≥ 1 mm in ‘Sporacle Ezy’ (Salam et al., 2007). This accepted modelling approach has been used as our control treatment.

The aim of our hydro-thermal time modelling approach was to accumulate the thermal time per favourable (wet) day determined by one of the two moisture conditions: i) rain ≥ 1 mm or ii) rain > soil evaporation. The first method is a simple and straightforward wetness indicator for pseudothecia to develop, while the second method requires the daily rainfall to exceed soil evaporation, describing when the soil surface, hence the stubble, is wet for an extended length of time (Unkovich, 2010). A moisture value of 1 indicates that the daily wetness requirement has been met, and the moisture value of 0 is when the day is “dry”, thus no pseudothecial development will occur. The temperature-driven development follows the experimental results of Pérès et al. (1999) where the observed thermal time for the first pseudothecium or the first cohort of pseudothecia to reach maturity under continuous wetness at 14 °C was 196 °C-day. This temperature value was specifically chosen due to the optimal value for pseudothecial maturation (Pérès et al., 1999).

Table 1 Parametric values used in the Sporacle Ezy model (Salam et al., 2007) and hydro-thermal time approach in predicting the onset of Leptosphaeria maculans pseudothecial maturation and ascospore release
Full size table

On a ‘wet’ day, pseudothecial maturation progresses based on the cardinal temperatures (Table 1). Where the controlled treatment used a constant maturation rate at temperatures between 6 °C and 22 °C, Hydro-thermal time used the hourly interpolation to estimate the potential maturation rate of L. maculans on infected canola stubble. Each day consisted of 24 interpolated temperature values based on the minimum temperature, maximum temperature, and daylength (Goudriaan & Van Laar, 1994). Using the hourly development rates, development was estimated for each hour that fell within the modified temperature range (≥ 5 °C and ≤ 22 °C) (Fig. 1). The minimum and maximum temperatures for pseudothecia maturation were derived from the observations of Toscano-Underwood et al. (2003) and Salam et al. (2003), respectively. The calculated rates for each temperature were then aggregated to determine the pseudothecial maturation per day. A daily cumulative sum of development represents the estimated date at which the first 50% cohort (group in a population) of pseudothecia reached maturity.

Fig. 1
figure 1

Illustration of interpolation method to estimate the rate per thermal unit values based on daylength, minimum (5 °C) and maximum (22 °C) temperatures adapted from Goudriaan and Van Laar (1994). The shaded area shows the cardinal temperature range for L. maculans pathogen development (5–22 °C)

Full size image

Model testing and evaluation

Predictions for the onset of pseudothecial maturation or ascospore release were performed and tested against the observed datasets in this study as well as published and unpublished data from Australia, Canada, France and the UK (Table 2). As it is intended that the simulations are a realistic representation for pathogen development on canola stubble, it is important to start the simulation for all sites and seasons at the time of harvest of the crop in the previous season. As these harvest dates were unavailable from published literature the canola crop module in APSIM (Robertson & Lilley, 2016) was used to provide approximate dates of harvest for each location and season in Australia. The cultivar ATR-Bonito was chosen to represent the crops generally sown in Western Australia.

The approximate harvest dates for spring canola varieties sown in Canada (Guo & Fernando, 2005) and winter varieties sown in autumn (Europe) were used in locations outside Australia (shown in Table 3) (Salam et al., 2007; West et al., 2001). Daily weather data (rainfall, maximum temperature, and minimum temperature) for the Australian sites were retrieved from SILO (Jeffrey et al., 2001) (https://www.longpaddock.qld.gov.au/silo/) and soil profiles from the APSoil database (https://www.apsim.info/Products/APSoil.aspx). The French, Canadian, UK and Poland weather data were provided by INRAe, a Canadian weather database (https://climate.weather.gc.ca/) and NASA POWER (https://power.larc.nasa.gov/), respectively.

Table 2 Additional observed datasets used to test the hydro-thermal time modelling approach in predicting the onset of pseudothecial maturation or ascospore release for species complex Leptosphaeria maculans
Full size table
Table 3 General harvest time of canola (oilseed rape) for each location and the start of simulation used in predicting the onset of pseudothecial maturation
Full size table

Model performance was assessed using the standard regression (observed versus predicted) and deviation statistics. In correlation-regression analysis, the statistical parameters applied were the coefficient of determination (R2), standard error of the regression, and the slope of the regression line. The deviation-based statistical parameters used were the Root mean-squared deviation (RMSD), RMSD-observations standard deviation ratio (RSR), Nash–Sutcliffe efficiency (NSE), and Percent Bias (PBIAS). The RMSD is a commonly used statistics which shows the difference (squared) of observed and simulated data whereas RSR is the normalised version of RMSD relative to the standard deviation of the observed data. A low RSR (0.0 ≤ RSR ≤ 0.70) equates to a low RMSD and thus shows a good model performance. The NSE indicates the magnitude of the residual variance relative to the observed data variance – a high value of NSE is generally viewed as satisfactory model performance (0.50 < NSE ≤ 1.0). The PBIAS determines the deviation tendency of predicted data from the observed: a positive PBIAS indicates model underestimation and a negative PBIAS shows model overestimation. Detailed information and equations used to calculate the recommended deviation-based statistics can be found in (Moriasi et al., 2007).

Results

Effect of stubble orientation on pseudothecial maturation rate

Weekly assessments of pseudothecial maturation were conducted to determine the effect of stubble management on the development of pseudothecial populations and timing of ascospore release in standing and lying stubble. At the first time of assessment in early June, approximately 40% of the sampled pseudothecia were mature on the lying stubble compared to only 9% on the standing stubble. ‘Empty’ pseudothecia started to appear when ≥ 30% of the pseudothecial population were mature, which was one week after the first date of assessment for lying stubble and three weeks after for the standing stubble. After these timepoints, ‘empty’ pseudothecia were present in all subsequent observations (Fig. 2).

The increase in the proportion of mature pseudothecia around July (35–56 days from first assessment) was similar for both stubble orientations. However, between August and December (63–196 days from first assessment) the proportion of mature and empty pseudothecia on the lying stubble fluctuated compared to the standing stubble on which the mature pseudothecia continued to increase and the fraction of empty pseudothecia was consistently around 30–40%.

In both standing and lying stubble, pseudothecia were not exhausted until June 2022, ~ 20 months after harvest.

Fig. 2
figure 2

Fractions of mature pseudothecia (●) and empty pseudothecia (○) at each assessment timepoint on “standing” stubble and “lying” stubble collected from a commercial crop grown in the 2020 season at Horsham, VIC. Stubble samples were collected between June 2021 and June 2022. The regression equations for the fractions of mature (solid lines) and empty (dashed lines) pseudothecia in lying (FML, FEL) and standing stubble (FMS, FES) were: FML = 0.32 + 5.29e−3x-3.63e−5x2 + 5.79e−8x3 (p < 0.001; R2 = 0.38), FEL = -0.081 + 9.44e−3x-4.50e−5x2 + 7.36e−8x3(p < 0.001;R2 = 0.73), FMS = 0.12 + 8.99e−3x-4.36e−5x2 + 5.37e−8x3 (p < 0.001;R2 = 0.73), and FES = 0.12 + 4.07e−3x-2.48e−5x2 + 5.25e−8x3 (p < 0.001;R.2 = 0.82)

Full size image

The PMI was calculated for each assessment timepoint on standing and lying stubble to determine the rate of maturation. The PMI for both stubble orientations follow a similar pattern for PMI but occurs at a faster rate on lying stubble (Fig. 3) suggesting that at a given time, more pseudothecia have reached maturity, and subsequently released ascospores, on the lying stubble which has been observed in Fig. 2.

Fig. 3
figure 3

Pseudothecial maturity index (PMI) over time on a one-year-old “standing” stubble and “lying” stubble collected at Horsham, VIC between 4th June 2021 to 3rd June 2022. The regression equation of the PMI was: PMI = 2.27 + 3.96x-2.06x2 + 1.67x3 + (-0.16*O), where O = 1 for standing stubble, and 0 for lying stubble (p < 0.001, R.2 = 0.88)

Full size image

A regression analysis was performed to generate the PMI increase of the species cohort on standing stubble relative to that on lying stubble (Fig. 4). For each increase of proportion value for mature pseudothecia in the lying stubble, there is a 0.91 (SE = 0.051) increment of the population in the standing stubble.

Fig. 4
figure 4

Linear regression for pseudothecia maturity index (PMI) increase on standing stubble relative to the PMI increase on lying stubble at Horsham in 2021 and 2022. Regression equation of the line was: y = 0.037 + 0.91x. The dots represent each timepoint

Full size image

Testing of the model

Pseudothecial maturation using hydro-thermal time modelling approach

Pseudothecial maturation in standing stubble collected at Horsham and Boorowa were simulated using the hydro-thermal time accumulation equal to 196 °C-day, a stubble management factor value of 0.91, and two moisture parameters: i) rain ≥ 1 mm and ii) rain > soil evaporation in designating a favourable day for development. The use of hydro-thermal time modelling approach demonstrated both temperature and moisture availability influenced development but the magnitude of the development rate for each wet day was driven by the diurnal cycle of temperatures in relation to the cardinal temperatures (Fig. 5). The observed dates used in this comparison were the dates when 50% of the population reached maturity. In the 2021 season at Horsham, the rainfall + thermal time underestimated the timing of pseudothecial maturation by 83 days, whereas rainfall + evaporation + thermal time underpredicted the date by only 8 days. On the contrary, simulation of pseudothecial maturation in the following season (2022) at Horsham showed that the latter parameter combinations delayed the development by 11 days compared to the 23 days underestimation of rain + thermal time only. Simulation results observed at Boorowa showed an overestimation of only 4 days using the rainfall + evaporation where the simple hydrothermal parameters predicted the date earlier by 17 days from the observed date.

Fig. 5
figure 5

Stepwise development of L. maculans pseudothecia on standing stubble at Horsham, VIC in 2021 and 2022 and at Boorowa, NSW in 2022 (start of simulation: 29th October 2020, 10th November 2021 in Horsham and 2.nd December 2021 in Boorowa) using the rainfall threshold ≥ 1 mm + target thermal time value of 196 °C-day (solid line), and rain > soil evaporation + target thermal time value of 196 °C-day (dashed line). The date when the development reaches 1 is the predicted onset of pseudothecial maturation. The observed dates for the two sites and years are represented by a dot (○)

Full size image

Rainfall threshold for hydro-thermal time calculation

Simulations using the hydro-thermal time approach with different moisture trigger was subsequently applied to the historical observed datasets in Australia to identify the well-suited wetness parameter describing the conditions of the stubble and represent realistic development (Fig. 6). Based on the correlation-regression statistics of observed versus predicted date of the onset of pseudothecial maturation from this study and historical datasets from Australia, hydro-thermal time with rain > soil evaporation as the moisture parameter more closely predicted the dates for the onset of pseudothecial maturation across different sites and seasons (R2 = 0.65; RMSD = 18 days) than rain ≥ 1 mm (R2 = 0.33; RMSD = 57 days). Both moisture indicators underestimated the timing of pseudothecial maturation and ascospore release (below the 1:1 line), however, the deviation between the observed and predicted dates with rain > soil evaporation was lower (2.46%) than rain ≥ 1 mm (29.71%). Between the two moisture parameters, using rain > soil evaporation for hydro-thermal time performed better than rain ≥ 1 mm.

Fig. 6
figure 6

A regression analysis of the observed versus predicted dates of the onset of pseudothecial maturation and ascospore release in 26 Australian datasets using the rain ≥ 1 mm or rain > soil evaporation as moisture parameters for calculating accumulated hydro-thermal time

Full size image

Hydro-thermal time and sporacle ezy model comparison

The combined performance ratings of the hydro-thermal time approach in Australia, Canada, France, and the UK (see Table 2) were compared with the predictions using the Sporacle Ezy model (Fig. 7). Based on the combined regression- and deviation-based (RMSD, RSR, NSE and PBIAS) statistics, the use of hydro-thermal time approach can accurately predict the onset of pseudothecial maturation as well as the environmentally calibrated Sporacle Ezy (Table 4) model. The predictions of pseudothecial maturation and ascospore release in each region with the hydro-thermal time approach using rain > soil evaporation had an RMSD value of 18 days in Australia, 12 days in Canada, 4 days in France, 7 days in the UK, and 16 days in Poland which were consistently lower than those for Sporacle Ezy (RMSD = 20 days, 13 days, 16 days, 20 days, and 22 days, respectively). Both models underestimated the timing of pseudothecial maturation and ascospore release (below the 1:1 line) but the deviation between the observed and predicted dates with the hydro-thermal time approach was lower (2.50%) than Sporacle Ezy (4.15%). Based on the calculated values for NSE and RSR, the performance rating of hydro-thermal time simulation using rain > soil evaporation was very good with an NSE value of 0.93 and an RSR value of 0.37, similar to that of the Sporacle Ezy with an NSE value of 0.87 and an RSR value of 0.37.

Fig. 7
figure 7

A regression analysis of the observed versus predicted dates of the onset of pseudothecial maturation and ascospore release in Australia (○), Canada (♦), France (●), the UK ( +) and Poland (∆) using the Sproracle Ezy and Hydro-thermal time modelling approach (rain > soil evaporation)

Full size image
Table 4 Performance evaluation of Sporacle Ezy and Hydro-thermal time modelling approach in predicting the onset of pseudothecial maturation and ascospore release using the regression- and deviation-based statistics
Full size table

Discussion

We found that crop management practices and environmental conditions were important factors driving the rate of pseudothecial maturation and ascospore release. Pseudothecial maturation was significantly delayed in standing stubble compared with lying stubble. In stubble of both orientations, empty pseudothecia were evident when the proportion of the mature pseudothecia reached approximately 30% indicating the onset of ascospore release. This quantum is similar to the results of Huang et al. (2007) in the UK, who found that ascospores were released when 30–50% pseudothecia reached maturation. With this observation, the potential delaying factor for the population in standing stubble relative to the conventional (lying stubble) was identified via correlation-regression analysis of the Pseudothecia Maturity Index (PMI) values for each stubble type. This value was subsequently used in modelling pseudothecial maturation in standing stubble collected at Horsham, VIC and Boorowa, NSW.

Our PMI calculations revealed that the fractions of mature to empty pseudothecia were high on the horizontally oriented stubble compared with the standing stubble. The standing stubble had a steady appearance of both mature and empty pseudothecia, with a rapid increase and then decrease of the two development stages at the beginning and after a year of observations. The difference in PMI over time between standing and lying is most likely due to greater soil contact increasing the moisture content of the straw compared to the standing stubble where only the base is in contact with the soil. The earlier and higher amount of mature and empty pseudothecia from the lying canola stubble than in the standing was expected, and it has been observed in the ascospores release study between the lying and standing stubble by (McCredden et al., 2018). The earlier study observed that the total number of ascospores released from the horizontally oriented straws during the growing season was 66% greater than that from the vertical ones. It is however worth noting that while standing stubble may be beneficial in terms of disease mitigation, it may be less beneficial with respect to other aspects, such as minimising soil water evaporation and moisture losses, and enabling decomposition of organic matter from the surface into the soil (Phelan et al., 2018).

Assessments on the population dynamics of L. maculans pseudothecia between the standing and lying stubble demonstrated that more pseudothecia mature earlier in the lying stubble presumably due to the greater contact of the stubble with soil which allows moisture to be available for a prolonged period than in the standing. With the significant response of pseudothecial maturation on moisture – as shown in this study – as well as temperature (Pérès et al., 1999; Toscano-Underwood et al., 2003), an alternative modelling approach was developed. Our new method calculates the cumulative thermal time target value of 196 °C-day on a favourable (‘wet’) day for pathogen development as depicted by the moisture parameter used: rain ≥ 1 mm or rain > soil evaporation. The first method (rain ≥ 1 mm, temperature) had fewer parameters required compared with the second method which required three parameters (rain, soil evaporation, temperature), although the former method can only presume that the stubble is wet for an indefinite period whereas the latter can identify the periods when stubble is not wet – a more accurate calculation of hydro-thermal time. These assumptions were supported by the simulations generated for the two study sites – Horsham and Boorowa (Fig. 5). When tested against independent Australian datasets (see Table 2), the second method (rain > soil evaporation) yielded more accurate predictions than the first method (rain ≥ 1 mm), therefore rain > soil evaporation was subsequently used for hydro-thermal time calculation and tested against international datasets which include sites in Canada, France, and the UK.

The overall performance ratings via correlation-regression and deviation-based statistics suggest that the use of hydro-thermal time approach can accurately predict the timing of pseudothecial maturation and ascospore release across different sites and seasons. The ratings were similar with those of the Sporacle Ezy model – the accepted modelling approach in predicting the onset of pseudothecial maturation and ascospore release (Salam et al., 2007). This highlights the biological modelling ability of the hydro-thermal time approach for pseudothecial maturation which is initiated at harvest of the previous crop and accounts for both temperature and moisture response of the pathogen. Despite having the same accuracy in predicting the timing of pseudothecial maturation and ascospore release in new environments and stubble management as with the Sporacle Ezy, the alternative modelling approach is more generically scalable (i.e., not calibrated with site-specific parameters, which can lead to over-fitting; (Harrison et al., 2019) and able to describe processes driven by crop and environmental signals. Process-based simulations, such as the hydro-thermal time approach, that capture the disease responses to the environment and crop management practices can provide significant contributions to the strategic planning of preventive measures for blackleg disease under changing climatic conditions (Bondad et al., 2023).