Relationship between disease severity and escape of Pseudoperonospora cubensis sporangia from a cucumber canopy during downy mildew epidemics

Authors


Abstract

Fundamental to the development of models to predict the spread of cucurbit downy mildew is the ability to determine the escape of Pseudoperonospora cubensis sporangia from infected fields. Aerial concentrations of sporangia, C (sporangia m−3), were monitored using Rotorod samplers deployed at 0·5 to 3·0 m above a naturally infected cucumber canopy in two sites in central and eastern North Carolina in 2011, where disease severity ranged from 1 to 40%. Standing crop of sporangia was assessed each morning at 07·00 h EDT and ranged from 320 to 16 170 sporangia m−2. Disease severity and height above the canopy significantly (< 0·0001) affected C with mean concentration (Cm) being high at moderate disease. Values of Cm decreased rapidly with canopy height and at a height of 2·0 m, Cm was only 7% of values measured at 0·5 m when disease was moderate. Daily total flux (FD) was dependent on disease severity and ranged from 5·9 to 2242·3 sporangia m−2. The fraction of available sporangia that escaped the canopy increased from 0·028 to 0·171 as average wind speed above the canopy for periods of high C increased from 1·7 to 3·6 m s−1. Variations of Cm and FD with increasing disease were well described (< 0·0001) by a log-normal model with 15% as the threshold above which Cm and FD decreased as disease severity increased. These results indicate that disease severity should be used to adjust sporangia escape in spore transport simulation models that are used to predict the risk of spread of cucurbit downy mildew.

Introduction

Cucurbit downy mildew, caused by Pseudoperonospora cubensis, is economically the most damaging disease of cucurbitaceous crops worldwide (Lebeda & Cohen, 2011). In the USA, cucumber, cantaloupe, muskmelon, pumpkin, winter and summer squashes, and watermelon are the primary commercial crops that are impacted by the disease (Ojiambo et al., 2010). Since the 1960s, the disease had been adequately controlled in cucumber by planting resistant cultivars (Holmes et al., 2006). However, there was a resurgence of the disease in 2004 which resulted in substantial yield losses for the cucumber crop in the eastern USA (Holmes et al., 2006) and host resistance alone is no longer sufficient to control the disease (Lebeda & Cohen, 2011). A change in the pathogen population structure as a result of the spread of virulent strains of P. cubensis from Asia has been suggested as the cause of the resurgence of the disease in Europe and the USA (Runge et al., 2011; Savory et al., 2011). The resurgence of cucurbit downy mildew now requires renewed and improved management options for effective disease control. Currently, disease control relies heavily on multiple applications of fungicides that can effectively protect yield when applied in a timely manner.

Like all obligate plant pathogens, P. cubensis can survive and reproduce on living plant tissue or foliage. Thus, the pathogen is thought to overwinter in frost-free regions of the USA in southern Florida where cucurbits are grown year-round and along the Gulf of Mexico on wild cucurbit species (Bains & Jhooty, 1976) and weed hosts. Although the pathogen can also overwinter via oospores (Cohen et al., 2011; Lebeda & Cohen, 2011), oospores have not been observed in vivo in the USA. Pathogen reproduction and dissemination is favoured by warm temperatures (20–25°C) and high humidity (>85% relative humidity, RH; Thomas, 1996) that prevail during the growing season in many parts of continental USA. The pathogen reproduces primarily through asexual production of sporangia and the primary mechanism by which it can be transported to disease-free fields beyond its overwintering range is through aerial dispersal of sporangia. The survival of sporangia in the atmosphere decreases rapidly with increasing exposure to solar radiation (Kanetis et al., 2010). However, sporangia can remain viable in the atmosphere for 1 to 2 days under overcast sky conditions (Holmes et al., 2004). Sporangia are deposited on plant surfaces by rain, wind and/or gravitational settling, and viable sporangia can germinate and enter host tissue via open stomata. Depending on the environmental conditions, a new crop of sporangia is produced between 4 and 12 days after infection (Lebeda & Cohen, 2011) and the cycle of aerial dispersal and infection can be repeated within a field until healthy leaf tissue is no longer available. Under favourable environmental conditions, the time from initial deposition of viable sporangia on the host to widespread defoliation of the canopy in the absence of fungicide sprays ranges from 2 to 3 weeks (Lebeda & Cohen, 2011).

Long distance aerial transport of P. cubensis sporangia from infected sources to disease-free cucurbit fields was first reported using anecdotal information of disease outbreaks along the Atlantic coast states in the USA (Nusbaum, 1944). A recent study suggests that P. cubensis sporangia can be transported aerially up to a distance of 1000 km from overwintering sources in southern Florida to northern states in continental USA (Ojiambo & Holmes, 2011). Thus, cucurbit fields throughout this region can be at risk to infection by P. cubensis sporangia transported from distant sources during periods of strong southerly winds (20 km h−1). To help manage this risk, an aerobiological disease forecasting system was developed as part of the Cucurbit Downy Mildew (CDM) ipmPIPE program to facilitate decision-making regarding the intensification of disease monitoring and the timing of initial fungicide applications in cucurbit fields in the eastern USA (Ojiambo et al., 2011).

Several factors (Aylor, 1986; Skelsey et al., 2009) determine the infection risk of plant pathogens, such as P. cubensis, whose infectious propagules are dispersed aerially: (i) number of spores available for transport; (ii) proportion of available spores that escape the canopy; (iii) dilution of spores by the wind and their removal from the air; (iv) survival of spores during transport; and (v) efficiency of spore deposition on susceptible host tissue. Currently, the CDM ipmPIPE forecasting system provides predictions of disease risk to fields throughout the eastern USA based on aerial transport and deposition of P. cubensis sporangia (Ojiambo et al., 2011). However, several factors that influence the risk of infection as a result of aerial transport of P. cubensis sporangia have not been quantified for incorporation into the forecasting system. For example, relationships between common field-based measures of disease severity and (i) the number of P. cubensis sporangia produced from lesions and thus, potentially available for dispersion, and (ii) the proportion of these sporangia that escape the canopy into the atmosphere above an infected field, are unknown. Establishing the aerial concentration and escape of sporangia from infected sources is key to successful modelling of inoculum dispersal. If escape of sporangia from an infected canopy is known, sporangia deposition on host plants in disease-free fields can be estimated using established particle dispersion models (Gifford, 1968).

In aerobiological studies involving plant diseases, sporangia production and escape have been investigated in fields where disease severity is assumed to be constant in time (e.g. Aylor & Taylor, 1983; Aylor et al., 2001, 2011; Andrade et al., 2009). In most of these studies, disease severity in the test plots is usually high with abundant sporulation and the primary objective is to relate sporangia escape to meteorological factors (e.g. Aylor & Taylor, 1983). Currently, the CDM ipmPIPE forecasting system determines the risk of infection by P. cubensis using new disease reports based on the assumption that all source fields of equal size have the same source strength irrespective of their levels of disease severity (Ojiambo et al., 2011). However, sporangia production for P. cubensis occurs in lesions that form on living plant foliage and individual lesions produce sporangia for only 1 to 2 days (Holmes et al., 2004). Consequently, as disease severity increases, sporangia production will increase up to a maximum and then decrease thereafter as the amount of healthy leaf area decreases. For obligate plant pathogens such as P. cubensis, necrosis and defoliation are usually widespread when disease severity exceeds 40% (James, 1974), which terminates the disease epidemic. Establishing the relationship between disease severity and sporangia production and escape would provide a basis to adjust these two variables based on disease severity in disease reports when new disease outbreaks are reported. Thus, the objectives of this study were to: (i) examine the temporal dynamics of P. cubensis sporangia production and escape from a cucumber canopy during the progress of cucurbit downy mildew epidemics; and (ii) establish the relationship between disease severity and aerial concentration and escape of sporangia from a cucumber canopy during downy mildew epidemics.

Materials and methods

Field sites, inoculum source and disease assessment

Field experiments were conducted at two research sites during the summer of 2011. The first site was located at the Central Crops Research Station in Johnston County near Clayton in central North Carolina. The second site was about 120 km away and was located at the Horticultural Crops Research Station in Sampson County near Clinton in eastern North Carolina. Seeds of cucumber cultivar Poinsett 76 were planted in experimental plots at Clayton and Clinton on 20 June and 5 July 2011, respectively. Poinsett 76 is moderately susceptible to downy mildew. Experimental plots at Clayton consisted of 11 rows that were 17 m long with 0·5 m spacing between rows. The experimental plot was at the northern end of a 1·5 ha cucurbit field that was used for downy mildew monitoring and cultivar evaluation trials. The surrounding area was generally flat (i.e. no hills or valleys) with no obstacle to wind except for a few trees on the east end of the larger cucurbit field plot. Experimental plots at Clinton consisted of 10 rows, each 20 m long with 0·6 m between rows. The experimental plot in Clinton was surrounded with 5·6 ha of cucumber and watermelon used for downy mildew monitoring, cultivar evaluation and fungicide trials. Except for a few shrubs separating different plots on the research station, the surrounding area consisted of low crops such as sweet potatoes and was flat with no obstacle to wind. Black plastic mulch with drip irrigation was used at the two sites. Within the experimental plots, holes in each row were punched with a bulb planter every 0·6 m and two seeds were planted per hole for a total of 28 to 33 plants in a row. When necessary, plots were reseeded 1 week after initial seeding to replace non-germinated seeds. Two weeks after initial seeding, plots were thinned to one plant per hole at the two sites. Beds covered in black mulch were raised about 0·11 m above ground and the top of the canopy of fully grown plants was 0·15 m above the plastic mulch.

Plants were monitored regularly for growth and once plants reached ‘tip-over’ or about 30 days old, 10 plants were randomly selected and tagged in each plot. The sixth, seventh and eighth leaf of every randomly chosen plant was tagged using zip ties to allow for consistent disease assessment throughout the experiment. Disease in the experimental plots and surrounding cucurbit plantings was initiated at the two sites by naturally windblown sporangia. Disease severity on all tagged leaves was visually assessed by estimating the percentage of leaf area with chlorotic and necrotic symptoms. The mean leaf area infected (%) across all the tagged leaves was used as an estimate of the disease severity of the plot. Disease progress was monitored on six occasions at each of the study sites starting at the initial disease symptom date (16 and 12 August for Clayton and Clinton, respectively) to the date at which disease severity reached approximately 40% (29 August and 1 September for Clayton and Clinton, respectively). Disease progress was also monitored in the adjacent sentinel plots and surrounding cucurbit plantings. During the experiment, the estimated disease severity in the experimental plots was similar (± 2%) to the disease severity in the sentinel plots and surrounding cucurbit plantings.

Plants were monitored regularly for growth and once plants reached ‘tip-over’ or about 30 days old, 10 plants were randomly selected and tagged in each plot. The sixth, seventh and eighth leaf of every randomly chosen plant was tagged using zip ties to allow for consistent disease assessment throughout the experiment. Disease in the experimental plots and surrounding cucurbit plantings was initiated at the two sites by naturally windblown sporangia. Disease severity on all tagged leaves was visually assessed by estimating the percentage of leaf area with chlorotic and necrotic symptoms. The mean leaf area infected (%) across all the tagged leaves was used as an estimate of the disease severity of the plot. Disease progress was monitored on six occasions at each of the study sites starting at the initial disease symptom date (16 and 12 August for Clayton and Clinton, respectively) to the date at which disease severity reached approximately 40% (29 August and 1 September for Clayton and Clinton, respectively). Disease progress was also monitored in the adjacent sentinel plots and surrounding cucurbit plantings. During the experiment, the estimated disease severity in the experimental plots was similar (± 2%) to the disease severity in the sentinel plots and surrounding cucurbit plantings.

Standing crop of sporangia

The number of sporangia produced in the infected plots was estimated in the morning (07·00 h EDT) on each disease assessment date as described by Aylor et al. (2001). Assessments of the standing crop of P. cubensis sporangia were made in 0·25 × 0·25 m2 sampling grids within the plots. Three sampling grids were selected randomly in each field plot on each day of sporangia assessment. All the lesions in the sampling grids were counted and recorded. For each sampling grid, three lesions were destructively sampled and placed separately in a tube containing 5 mL 12% CuSO4 solution to inhibit any sporangia germination (Arauz & Sutton, 1989). Each tube was vortexed for a few seconds to dislodge sporangia into the solution, and then leaf material was removed from the tube and discarded. The resultant spore suspension from each lesion was then sampled twice and a haemocytometer was used to estimate the number of sporangia per lesion. The standing crop of sporangia (sporangia m−2) was determined by multiplying the mean number of sporangia per lesion by the number of lesions per grid and then dividing the result by the area of the sampling grid. Standing crop of sporangia is both an indicator of sporangia production and a measure of the number of sporangia potentially available for aerial dispersal (Aylor et al., 2001).

Airborne concentration of sporangia

On each assessment day, the aerial concentration of P. cubensis sporangia, C (sporangia m−3), was monitored above the cucumber canopy using Rotorod spore samplers with retracting type heads (Model 82; Sampling Technologies, Inc.). Sporangia were collected on two plastic rods (2 × 2 × 32 mm) attached to the rotating arm (2400 rpm) of each Rotorod sampler. A thin layer of high vacuum silicon grease (Dow Corning Corp.) was applied on the sampling surfaces of the plastic rods to ensure adhesion of sporangia on the rods. The samplers were supported by three poles positioned in the centre of the plot in a triangular pattern. Four Rotorod samplers were mounted on each pole so that the centre of the collection rod was 0·5, 1·5, 2·0 and 3·0 m above the canopy. Rotorod samplers on a pole were activated simultaneously for a 10 min sampling period and then turned off for 20 min. Each pole was operated in succession such that a total of six consecutive collections of 10 min duration were obtained within each hour from 07·00 to 14·00 h EDT.

Plastic rods on each Rotorod sampler were carefully removed and replaced with clean ones between successive sampling periods. Sporangia captured on the greased surface were counted with the aid of a microscope at  × 100 magnification. Counts of sporangia were converted to C by accounting for the proportion of sample surface that was enumerated, sampling rate, duration of sampling period and the efficiency of the Rotorod sampler (Aylor, 1993).

Values of C for the 10 min sampling period from each height were averaged across the six consecutive collections within each sampling hour to obtain hourly concentrations of sporangia (Ch). Furthermore, Ch were averaged across the 8 measurement hours (07·00 to 14·00 h EDT) to generate a mean sporangia concentration (Cm) for each day and sampling height. The LOESS procedure of sas (v. 9.2; SAS Institute) was used to perform LOESS non-parametric regression analyses (Cleveland & Grosse, 1991) to model daily temporal trends for Ch at Clayton and Clinton for each level of disease severity. The smoothing parameter used in the LOESS regression analysis was chosen to minimize a bias-corrected Akaike information criterion (Hurvich et al., 1998). To qualitatively describe the relationship between disease severity and sporangia concentration at the sampling heights, three classes of disease severity were defined: low (0–5%), moderate (>5–20%) and high (>20–40%) based on a previous study on infection of cucurbits by P. cubensis (Neufeld & Ojiambo, 2012). Differences in Cm between three sampling heights (0·5, 1·5 and 2·0 m) and disease severity classes were examined using univariate analysis.

Meteorological measurements

Air temperature (°C) and relative humidity (RH,%) were monitored using Watchdog data loggers (Model 450, Spectrum Technologies Inc.) located in the centre of the triangulated poles within plots at Clayton and Clinton. Measurements were obtained every 5 min, averaged and recorded every hour. Wind speed data were obtained from the weather stations located at the research stations. Wind speed and wind direction were measured using a helicoid propeller and wind vane (Model 05103, R. M. Young) positioned 10 m above ground level. Measurements of wind speed and direction were obtained every 5 s, averaged and recorded every hour. Correlation analyses were performed using the CORR procedure of sas to determine if any association existed between Ch, air temperature and RH for each hour of the day using combined data from the two sites.

Wind speed (u, cm s−1) at 0·5, 1·5 and 2·0 m above the canopy was calculated using the wind speed measurements at 10 m height and the standard logarithmic wind profile equation (Lowry & Lowry, 1989):

display math(1)

where math formula is the friction velocity (cm s−1), k is the von Karman's constant (0·4), z is the height above the ground (cm), D is the zero-plane displacement height (cm) and z0 is the roughness length (cm). D and z0 are set to 0·7 h and 0·2 h, respectively, where h is the height of the cucumber canopy (0·3 m during the study).

Escape of sporangia from the canopy

Calculations of the number of sporangia that escaped the canopy were based on the assumption that the vertical diffusion of sporangia is determined by the vertical diffusivity of the air (Smith & Hay, 1961) and thus, the diffusivity of the sporangia (Ks) can be equated to the diffusivity of the momentum (Km). The friction velocity and the vertical rate of change of the average wind speed were used to calculate Ks using the following equation (Sutton, 1953):

display math(2)

where math formula, u and z are as defined above, and du/dz is the vertical wind speed gradient.

For each level of disease severity, the vertical flux of sporangia Fs (spores m−2 s−1) was calculated using the following equation (Aylor & Taylor, 1983):

display math(3)

where ∂C/∂z is the vertical gradient of sporangia concentration as determined from the Rotorod measurements at the sampling heights. Given that Fs values were calculated from concentrations measured just above and far from the leading edges of the uniformly infected cucumber fields, it can be assumed that they are representative of the flux at the height of the canopy. Calculation of Fs above the canopy using Eqn (3) ignores sedimentation under gravity that reduces Fs above the canopy by approximately 10% at a height of 1–2 m (Aylor & Taylor, 1983). Values of Fs obtained for each measurement period from 07·00 to 14·00 h EDT were integrated to generate the total daily flux (FD, sporangia m−2) for the day. The fraction of sporangia that escaped the canopy was then estimated by expressing FD as a proportion of the standing crop of sporangia.

Relationship between sporangia concentration, escape and disease severity

The relationship between Cm, FD and disease severity was first examined graphically for separate data collected at Clayton and Clinton. Preliminary analysis did not indicate significant differences in the shape parameters for plots of Cm, FD and disease severity from Clayton and Clinton sites. Thus, combined data from the two sites were used in the final analysis to establish the relationship between Cm, FD and disease severity. Based on the visual inspection of the plots, data were fitted to a three parameter log-normal model of the form:

display math(4)

where y is Cm or FD, x is disease severity, a is the vertical scale parameter, b is the log standard deviation and x0 is the log-mean. Data were fitted to the model in sas using the NLIN procedure. Goodness-of-fit of the model was evaluated based on the significance of parameter estimates, magnitude of asymptotic standard errors, and simple correlation between observed and predicted values of Cm and FD.

Results

Disease severity and standing crop of sporangia

Disease symptoms were first observed in the plot at Clinton on 12 August with a disease severity of 1·6% (Table 1). Initial symptoms of the disease at Clayton were observed on 16 August with a severity of 1·1%. At both sites, the disease increased rapidly and final disease assessment was recorded approximately 2 weeks after initial symptoms were observed in the cucumber plots. Disease severity in the plot at Clayton was 37·2% on 29 August and 35·1% at Clinton on 1 September (Table 1). Disease measurements were terminated thereafter due to severe necrosis and defoliation of the host plants.

Table 1. Standing crop of Pseudoperonospora cubensis, sporangia escape and meteorological variables for assessment periods at the field sites in North Carolina in 2011
SiteDateDisease severity (%)aStanding crop (spores m−2)bDaily total flux (spores m−2)cEscape fractiondWind speed (m s−1)eTemperature (°C)eRelative humidity (%)e
  1. a

    Disease severity was assessed visually as the percentage of leaf area infected on each date when sporangia were collected.

  2. b

    Standing crop of sporangia was estimated by counting number of lesions (x) in three 0·25 × 0·25 m2 sampling grids and the average number of sporangia washed from nine sampled lesions (y) as: (× y)/0·0625.

  3. c

    Daily total flux was obtained by integrating sporangia escape for measurement period from 07·00 to 14·00 h EDT.

  4. d

    The escape fraction was calculated by dividing the daily total flux by the standing crop.

  5. e

    Meteorological variables are averaged for the period when aerial sporangia concentrations were high (08·00 to 10·00 h EDT). Wind speed values are based on measurements at 10 m above the ground.

Clayton16 Aug1·1320·326·40·0821·727·957·8
17 Aug1·62300·4260·30·1132·521·671·0
24 Aug7·97370·3713·10·0971·927·882·1
25 Aug12·213129·52242·30·1713·613·656·5
26 Aug18·311090·11590·90·1432·932·758·9
29 Aug37·21789·7158·70·0892·233·049·7
Mean13·16000·1831·90·1162·526·162·7
Clinton12 Aug1·6531·15·90·0112·133·554·6
23 Aug9·45240·3804·80·1513·128·463·6
25 Aug16·116170·2594·80·0372·134·553·7
28 Aug22·65480·1121·10·0221·728·955·2
30 Aug27·95422·1119·90·0222·327·760·3
01 Sept35·12469·969·00·0281·530·638·5
Mean18·85885·6285·90·0462·230·554·3

The standing crop of sporangia ranged from 320 sporangia m−2 at Clayton to 16 170 sporangia m−2 at Clinton (Table 1). The standing crop of sporangia was dependent on the level of disease in the field and increased with increasing disease severity until a disease severity level of 12·2% and 16·1% at Clayton and Clinton, respectively, and then decreased thereafter. Yields of sporangia m−2 at disease severity levels >20% were 67 to 88% lower than the corresponding yields of sporangia m−2 at disease severity levels of 12·2% and 16·1%. Variations in the standing crop of sporangia were much more pronounced at moderate disease severity (5–20%) than at lower (0–5%) or higher (>30%) disease severity (Table 1).

Meteorological variables and dynamics of sporangia concentration

Hourly average air temperature, RH and wind speed from 07·00 to 14·00 h EDT followed a similar pattern at the Clayton and Clinton sites. Generally, air temperature increased while RH decreased with time of the day, except on 12 August at Clinton where temperature increased from 07·00 h EDT until 11·00 h EDT and decreased thereafter, while RH decreased until 11·00 h EDT and slightly increased thereafter (data not shown). During the study period, recorded temperatures ranged from 20 to 38°C at Clayton and 15 to 42°C at Clinton. Relative humidity ranged from 30 to 90% at Clayton and 25 to 92% at Clinton. Wind speed ranged from 0·04 to 4·65 m s−1 across the two sites during the study. Generally, wind speed remained constant or increased slightly with the hour of the day, except on 16 August at Clayton and 12 August at Clinton when wind speed fluctuated during the sampling period.

Aerial concentration of sporangia, Cm, was significantly (< 0·05) affected by the height above the canopy and level of disease severity (Table 2). The highest concentration of sporangia was recorded at 0·5 m above the canopy irrespective of disease severity at both Clayton (Fig. 1) and Clinton (Fig. 2). Across sites, values of Cm at 0·5 m above the canopy ranged from 15 sporangia m−3 at 1·1% disease severity to 21 122 sporangia m−3 at a disease severity of 18·3%. Values of Cm decreased rapidly with increasing height above the canopy and were lowest at 2·0 m, while no sporangia were captured at 3·0 m (Fig. 3). At the 0·5, 1·5 and 2·0 m heights above the canopy, Cm increased with increasing level of disease severity until about 16–18% disease severity and then decreased thereafter (Figs 1 & 2). The shapes of the vertical profiles of concentrations of sporangia and absolute values of Cm at the two sites were very similar, except at moderate levels of disease severity where absolute values of Cm were higher at Clayton than at Clinton particularly at 0·5 m above the canopy (Fig. 3).

Table 2. Concentration of Pseudoperonospora cubensis sporangia at three sampling heights above an infected cucumber canopy stratified by disease severitya
SiteHeight above canopy (m)Sporangia m−3
Low severity (0–5%)bModerate severity (>5–20%)bHigh severity (>20–40%)b
MeanSEMeanSEMeanSE
  1. a

    Sporangia concentrations are averages of spores m−3 of air sampled from 07·00 to 14·00 h EDT for each level of disease severity and height above the canopy of a cucumber crop.

  2. b

    SE denotes the standard error of the mean for a sample size of = 42 for low and high disease severity and = 84 for moderate disease severity at both locations.

Clayton0·5114·714·98711·91462·6748·4106·6
1·539·56·21233·7221·795·416·2
2·031·45·4630·1121·157·310·4
Clinton0·5102·810·53498·7577·2796·9142·9
1·590·49·2319·844·5119·719·1
2·056·38·4173·219·367·110·6
Figure 1.

Diurnal pattern of aerial concentration of Pseudoperonospora cubensis sporangia above a cucumber canopy during downy mildew epidemics at Clayton, North Carolina, USA. Curves are fitted by locally weighted regression to illustrate daily trends. Data are shown for four disease assessment periods that represent the range of disease severity levels observed during the study.

Figure 2.

Diurnal pattern of aerial concentration of Pseudoperonospora cubensis sporangia above a cucumber canopy during downy mildew epidemics at Clinton, North Carolina, USA. Curves are fitted by locally weighted regression to illustrate daily trends. Data are shown for four disease assessment periods that represent the range of disease severity levels observed during the study.

Figure 3.

Vertical profiles of Pseudoperonospora cubensis sporangia concentrations above infected cucumber plots, averaged from 07·00 to 14·00 h EDT for 25 and 26 August at the Clayton and Clinton sites, respectively. For moderate disease severity, the mean disease was 12·8% at both Clayton and Clinton; for high disease severity, the mean disease severity at Clayton and Clinton was 37·2% and 28·5%, respectively.

LOESS non-parametric regression indicated a marked daily trend in the concentration of sporangia with peak hourly concentrations between 08·00 and 10·00 h EDT (Figs 1 & 2). Sporangia concentrations averaged for 1-h periods (i.e. Ch) were not significantly (> 0·05) correlated with air temperature or relative humidity recorded in the same 1-h periods at both Clayton and Clinton. However, when measurements were grouped by level of disease severity, Ch were significantly (< 0·05) negatively correlated with RH (−0·97 < < −0·99) for the 09·00 and 10·00 h EDT periods when disease severity was high (Table 3).

Table 3. Pearson's coefficients for the correlation between Pseudoperonospora cubensis sporangia collected 0·5 m above a cucumber canopy (sporangia m−3) and relative humidity (%), stratified by level of disease severity and hour of day
Hour of dayCorrelations between sporangia concentration and relative humiditya
Low severity (0–5%)bModerate severity (>5–20%)bHigh severity (>20–40%)b
r >  F r F r F
  1. a

    P-values are for the test to determine if the correlation coefficient (r) is significantly different from zero.

  2. b

    Disease severity was assessed visually as the percentage of leaf area infected on each date when sporangia were collected.

07·00 0·9180·2559–0·1540·8048–0·4450·5550
08·000·9410·1731–0·5010·3900–0·8070·1932
09·000·9870·1029–0·4300·4698–0·9730·0267
10·000·5470·6314–0·0410·9484–0·9930·0073
11·000·9330·2332–0·2760·65300·2790·7206
12·000·7190·4888–0·4710·4234–0·1290·8713
13·00–0·7210·4877–0·2660·6651–0·4340·5664

Escape of sporangia from the canopy

Estimates of sporangia that escaped the canopy (Fs) were dependent on the disease severity and hour of day. Escape of sporangia during the day typically followed a diurnal pattern with maximum escape rates at 09·00 h EDT, irrespective of the level of disease severity (Figs 4 & 5). Fs increased with increasing disease severity until a maximum threshold and then decreased thereafter. Thus, Fs at Clayton was lowest at 1·1% disease severity with maximum of 7·98 sporangia m−2 s−1, increased with increasing severity until 18·3% disease severity where the maximum value was 926·4 sporangia m−2 s−1, and then decreased thereafter (Fig. 4); maximum sporangia escape at 12·2% and 37·2% disease severity was 824·1 and 69·2 sporangia m−2 s−1, respectively. Values of Fs followed a similar pattern at Clinton with sporangia m−2 s−1 being lowest at the lowest level of disease severity (maximum of 2·7 sporangia m−2 s−1) and highest at 9·4% leaf area infected (maximum of 378·4 sporangia m−2 s−1); sporangia escape decreased rapidly with increasing levels of disease until at the highest level of disease severity the maximum Fs was 17·6 sporangia m−2 s−1 (Fig. 5). Daily total flux, FD, followed a similar pattern to Fs with values being higher at moderate levels of disease (5–20%) than at low (<5%) or high (>20%) levels of disease (Table 1). FD ranged from 5·9 sporangia m−2 at Clinton to 2242·3 sporangia m−2 at Clayton. Extrapolation of the escape estimates to units commonly used in regional scale aerobiology models indicate that FD ranged from 0·09 to 8·07 × 1010 sporangia ha−1 at Clayton and from 0·02 to 2·9 × 1010 sporangia ha−1 at Clinton.

Figure 4.

Escape of Pseudoperonospora cubensis sporangia from a cucumber canopy at the Clayton field site during epidemics of downy mildew. On each assessment date, disease severity was assessed visually as the percentage of leaf area infected. Data are shown for four disease assessment periods that represent the range of disease severity levels observed during the study.

Figure 5.

Escape of Pseudoperonospora cubensis sporangia from a cucumber canopy at the Clinton field site during epidemics of cucurbit downy mildew. On each assessment date, disease severity was assessed visually as percentage of leaf area infected. Data are shown for four disease assessment periods that represent the range of disease severity observed during the study.

The proportion of the standing crop of sporangia that escaped the canopy (i.e. escape fraction) was variable and ranged from 0·022 to 0·171 (Table 1). There was no significant correlation between escape fraction and the meteorological variables averaged for the assessment period (07·00 to 14·00 h EDT). However, when these variables were averaged for the period of peak aerial concentrations of spores (08·00 to 10·00 h EDT), the proportion of the standing crop of sporangia that escaped the canopy increased with increasing wind speed (see Eqns (2), (3)), with less variation in sporangia escape at higher (≥2·5 m s−1) than lower (<2·5 m s−1) wind speeds (Table 1). Air temperature and relative humidity averaged for the same periods were not correlated with sporangial escape.

Relationship between sporangia concentration, escape and disease severity

Both Cm and FD increased with increasing disease severity until a given disease severity threshold and then decreased sharply thereafter with a thin tail as disease severity increased (Fig. 6). This relationship was well described by the log-normal model (< 0·0001). All three parameters in the model were significant (< 0·001) for both Cm and FD (Table 4) and plots of residual versus observed values did not reveal any systematic pattern in the residuals (data not shown). Simple correlations between observed and predicted values were high for both Cm (= 0·82) and FD (= 0·75). Parameters representing the log-mean and standard deviation in the model were similar for Cm and FD. The disease severity threshold above which there was decrease in both Cm and FD as disease severity increased was approximately 15%.

Table 4. Parameter estimates from non-linear regression analysis using the log-normal model to describe the relationship between mean Pseudoperonospora cubensis sporangia concentration (Cm), daily total sporangia escape (FD), and disease severity for combined data from two field sitesa
Response ParameterbEstimateAsymptotic SEcAsymptotic CILcAsymptotic CIUcF
  1. a

    Mean sporangia concentrations are averages of sporangia m−3 recorded from 07·00 to 14·00 h EDT at each level of disease severity at 0·5 m above the canopy of a cucumber crop.

  2. b

    Parameters a, b and x0 are the vertical scale parameter, log-standard deviation and log-mean, respectively.

  3. c

    SE is the standard error and CIL and CIU = lower and upper limits of the 95% confidence interval around the parameter estimates.

C m a 66994·539358·9545823·1088165·900·0001
b 0·390·070·240·550·0003
x 0 15·131·1112·6017·650·0001
F D a 24645·134824·8813732·4035557·930·0005
b 0·300·070·140·460·0009
x 0 13·870·9811·6516·080·0001
Figure 6.

Relationship between downy mildew severity and (a) mean aerial concentration of Pseudoperonospora cubensis sporangia, Cm, and (b) daily total escape of sporangia, FD, from naturally infected cucumber fields at Clayton and Clinton, North Carolina, USA. Solid circles (● or ○) are observed values. The curve shows the predicted Cm (a) and FD (b), obtained from fitting a log-normal model to combined data from the two sites.

Discussion

Establishing the escape of sporangia from source areas is critical to successful modelling of aerial transport of P. cubensis sporangia in the atmosphere and the spread of cucurbit downy mildew at local and continental scales. Downy mildew can be managed effectively with fungicide applications as long as infected fields are sprayed before initial symptoms or immediately after disease symptoms are observed. Consequently, knowledge of escape of sporangia of P. cubensis from infected cucurbit fields and wild inoculum sources is fundamental to the development of decision support systems for rational and effective management of the disease. Data on sporangia escape can also assist in assessing the risk from introduction and spread of the new and aggressive strains of P. cubensis (Runge et al., 2011; Savory et al., 2011).

Aerial concentration of sporangia showed a strong diurnal periodicity, irrespective of disease severity in the plots, with peak C occurring between 08·00 and 10·00 h EDT. These results are largely similar to those reported for other downy mildew pathogens (Aylor & Taylor, 1983; Aylor et al., 2001; Carisse & Philion, 2002). Release episodes of downy mildew sporangia have been associated with a decrease in RH, an increase in air temperature and evaporation of moisture from leaf surfaces (Populer, 1981; Sutton & Hildebrand, 1985; Carisse & Philion, 2002). In the present study, values of Ch were significantly negatively correlated with RH at 09·00 and 10·00 h EDT when disease severity was high. It has been shown that a decrease in RH causes hygroscopic twisting of sporangiophores as they dry, which in turn actively releases sporangia into the air (Lange et al., 1989). Unlike in a previous study conducted in Michigan (Granke & Hausbeck, 2011), a significant correlation between air temperature and Ch was not observed in the current study. This may be due to differences in the range of air temperatures during sporulation between the two studies. In the study by Granke & Hausbeck (2011), air temperatures ranged between 14 and 30°C, while in the present study air temperatures were much warmer and ranged from 23 to 40°C (except for one assessment date at Clinton). Sporulation of P. cubensis is optimum within a temperature range of 15–20°C (Thomas, 1996).

As expected for spores that are released from a ground level source (Csanady, 1973), aerial concentrations of P. cubensis sporangia were highest closer to the canopy and decreased with increasing height above the canopy. On average, the values of Cm at a height of 2·0 m were only 7% of values measured at 0·5 m above the canopy when disease severity was moderate or high. Based on the studies conducted on Venturia inaequalis, the rates of spore release and physical transport dilution of airborne spores by the wind are the main determinants of the vertical variation of the aerial concentration of spores released from a ground level source (Aylor, 1995). In the present study, calculated wind speeds ranged from 0·01 to 3·1 m s−1 at 0·5 to 2·0 m above the canopy and increased by more than 64% as the height above the canopy increased from 0·5 to 2·0 m. The pronounced decrease in Cm with height has also been reported for other oomycetes (Aylor & Taylor, 1983; Aylor et al., 2001). The shapes of the vertical profiles of Cm and corresponding values of Cm were similar at Clayton and Clinton, except at the lowest sampling height during moderate disease severity. These vertical profiles for P. cubensis are similar to those reported for the potato late blight pathogen Phytophthora infestans (Aylor et al., 2001). In addition, ratios of Cm at 1·5 m:2·0 m at Clayton (c. 1·62) and Clinton (c. 1·75) were not significantly different. Coupling these ratios with observations that downy mildew was not present outside the immediate area on the research stations, suggests that the experimental plots and surrounding cucurbit plantings were the main source of sporangia captured in this study.

Escape of sporangia varied with the hour of the day and disease severity in the plots. The results of the current study indicated that peak Fs occurred around 09·00 h EDT, which coincides with the peak period for Ch. Values of Fs reported in this study (maximum 926·4 sporangia m−2 s−1) were similar to those reported for P. infestans which ranged from 0 to 700 sporangia m−2 s−1 (Aylor et al., 2011). In addition to wind speed, canopy structure and the height of spore release inside the canopy are important factors that can influence C and Fs. However, the height of spore release has a relatively larger effect on C, and thus Fs, compared to the canopy structure (Aylor et al., 2001). The higher the inoculum source is located in the canopy and the stronger the wind velocity, then the greater the number of spores that will escape from the canopy. Fewer spores will escape the canopy at lower wind speeds as turbulence is less (i.e. low math formula) and spores are expected to settle out quickly in low turbulence (Aylor et al., 2001). Cucumber vines usually elongate along the ground surface and downy mildew progresses rapidly in the plots, retarding plant growth. As such, it is unlikely that changes in the cucumber canopy structure and height of spore release during the 2-week study period substantially influenced Fs.

In this study, the escape fraction of P. cubensis sporangia ranged from 0·022 to 0·171 of the standing crop. Aylor et al. (2001) suggested that the escape fraction for P. infestans sporangia from a potato canopy should initially be small at low disease severity and then increase when the foliage becomes more affected by disease and the canopy begins to disintegrate. However, the results of the current study indicate that the escape fraction of P. cubensis sporangia varied more with mean wind speed measured at periods of high C, than with disease severity. Thus, the escape fraction did not appear to be impacted by the disintegration of the sprawling, low canopy formed by the cucumber vines. The 10-m wind speeds recorded at the study sites (up to 3·6 m s−1) were sufficient to lift large numbers of sporangia from the cucumber canopy. A wind speed of 3 m s−1 can transport sporangia for 20 km in about 2 h. However, the spread of cucurbit downy mildew over such distances will depend on the viability of sporangia when they are deposited a host plant. Solar radiation is the most important physical variable that affects survival of P. cubensis sporangia in the atmosphere (Kanetis et al., 2010). Thus, the dispersal distance of P. cubensis sporangia will depend on inoculum source strength, sporangia escape fraction and survival of sporangia during transport. The calculations of Fs using Eqns (2), (3) are based on the assumption that there is sufficient fetch and that spore transport takes place during neutral atmospheric conditions (Sutton, 1953). However, strong solar heating and light winds can create unstable atmospheric conditions that could result in Fs values higher than those reported in this study.

In many aerobiological studies of plant pathogen dispersal, daily measurements of the standing crop of spores at the study site are used as the measure of inoculum source strength (e.g. Aylor & Taylor, 1983; Aylor et al., 2001, 2011; Andrade et al., 2009). Determination of the standing crop of spores is labour-intensive and takes a significant amount of time, and is thus impractical for regional forecasting programmes that rely on field surveys to delineate the geographic extent and strength of inoculum sources (Aylor et al., 2001). The results of the current study indicated that the quantity of sporangia that escapes an infected cucumber field varies greatly during the course of the disease epidemic but that there is a predictable relationship between the commonly used field measure of disease severity and Cm and FD. The effects of disease severity on Cm and FD were well described by a log-normal model and estimates of the log-mean and standard deviation were similar for Cm and FD. The results indicate an increase in Cm and FD as disease increases up to a threshold of 15% followed by a decrease in both variables with further increases in disease severity. This threshold relationship was also evident with the standing crop at the source which was used as a measure of source strength.

A substantive goal of this study was to generate data to serve as a foundation to improve the efficiency of the CDM ipmPIPE forecasting system (Ojiambo et al., 2011) as a decision support tool for the management of downy mildew. Specifically, the aim was to provide data that can be used to adjust aerial concentrations of sporangia and sporangia escape based on the reported level of disease severity. The CDM ipmPIPE forecasting system uses the FLEXPART (Stohl et al., 2005) particle dispersion model to simulate the long-range and mesoscale transport and deposition of P. cubensis sporangia. These simulations are coupled with weather data along the projected pathways of sporangia transport and deposition to predict the risk of disease outbreak. Regional scale aerobiology forecast models for plant pathogens are highly dependent on knowledge of the geographic distribution and intensity of spore escape from infected canopies (Isard et al., 2011). In the current implementation of the FLEXPART, the total number of sporangia released (>1) and the total mass of sporangia emitted (0·1–0·5 kg) assumes a linear relationship between disease severity and source strength. This study supports the contention that disease severity could be a useful variable in regional sporangia transport models such as the FLEXPART when estimating sporangia escape from a source field. The results indicate that the total number of sporangia released and total mass emitted should be adjusted based on disease severity at the source. For practical purposes, more studies will be required to validate the results reported in this study, particularly where the relationship between disease severity and aerial concentration of sporangia or sporangia escape is intended for use within a predictive framework of the FLEXPART. Such a validation will need to be conducted on a regional scale in several locations across the eastern USA for robust results. Clearly, the use of the biology of P. cubensis to parameterize key components within the FLEXPART should result in better representation of the simulations of sporangia transport and thus, significantly improve cucurbit downy mildew forecasts.

Acknowledgements

This research was supported by a grant from the United States Department of Agriculture, National Institute of Food and Agriculture Regional IPM Competitive Grants Program (Southern Region) Award No. 2010-41530-21134. The authors thank Mike Adams and Rob Kautz for assistance with the field experiments. The authors also wish to thank Ryan Boyles and Aaron Sims, North Carolina State University State Climate Office, for some useful discussion.

Ancillary