SEARCH

SEARCH BY CITATION

Keywords:

  • bacteria;
  • disease management;
  • epidemiology;
  • pathogen dispersal

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Citrus canker (Xanthomonas citri subsp. citri (Xcc)) can cause yield loss and trade restrictions. The pathogen is dispersed in rain splash and spread is promoted by wind. The goal of this study was to gain some insight into the properties of short-distance splash dispersal of Xcc from ∼1·5 m-tall cankered grapefruit canopies in turbulent wind, common during rainstorms in Florida. Turbulent wind up to 19·9 m s−1 was tested in five experiments. Bacteria flux density (BFD, bacteria cm−2 min−1) was quantified at heights of 30, 70, 110, 130 and 180 cm above ground, and at four horizontal points (17, 51, 85 and 119 cm) at each height across the direction of the wind 1 m downwind. BFD varied among experiments, but the lowest BFDs were consistently detected at the greatest sample height. Despite differences between experiments, the relationship between log BFD and sample height was consistently described by a linear function (= 0·06–<0·0001, R2 = 0·75–>0·99). The BFD collected at the horizontal points across the wind path was variable. BFDs collected were sometimes significantly different, but no relationship was discernible. Stronger, turbulent wind resulted in greater BFD, with a linear function describing the relationship between log BFD and wind speed (= 0·2–0·02, R2 = 0·94–0·96). Multiple regression analysis demonstrated predictability of the proportion of total bacteria collected (= 141, < 0·0001, d.f. = 3, R2 = 0·53).


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Citrus canker is caused by the plant pathogenic bacterium Xanthomonas citri subsp. citri (Xcc). The pathogen infects species of cultivated citrus (Schubert et al., 2001), and is widespread in many tropical and subtropical citrus growing regions (Koizumi, 1985; Schubert et al., 2001). The disease is manifest as small, water soaked lesions that expand over a few days, becoming erumpent and necrotic. The symptoms can appear on leaves, shoots, branches and fruit, resulting in yield loss, and also, in the case of infected fruit, reduced marketability (Goto & Yaguchi, 1979; Schubert et al., 2001; Gottwald et al., 2009; Behlau et al., 2010).

Several studies have shown that rain splash in both still air and wind can result in dispersal of the pathogen from canker-infected plants (Serizawa, 1981; Pruvost et al., 2002; Bock et al., 2005, 2010a, 2011), and presence of Xcc in aerosol-sized droplets has been reported (Roberto et al., 2001). Local, within-orchard spread of citrus canker has been characterized in various spatiotemporal analyses (Danos et al., 1984; Gottwald et al., 1992a,b). But despite the implication of wind and rain in local epidemic development, the characteristics of local dispersal remain poorly defined. Furthermore, the importance of wind-blown splash removing inoculum from a canopy is the first stage in long distance dispersal events – there are reports of citrus canker developing >50 km from a known inoculum source following a hurricane, pointing to the combined effect of extreme wind and rain initially removing inoculum from a diseased canopy and dispersing it long distance (Irey et al., 2006; Gottwald & Irey, 2007).

Once removed from the inoculum source, various studies have characterized dispersal of fungal and bacterial pathogens in wind and/or rain splash over local, intermediate and long distances (McCartney & Fitt, 1985; Fitt et al., 1987, 1989; Aylor & Ferrandino, 1989; Aylor, 1990, 1999; Madden, 1992). Rain, combined with wind as the agent for pathogen dispersal, was first investigated by Faulwetter (1917) and it is established that wind enhances the distance that the small propagules of fungi or bacteria are dispersed dry, or in splash or aerosol droplets (Venette & Kennedy, 1975; Carnegie, 1980; Quinn et al., 1980; Fitt & Nijman, 1983; Brennan et al., 1985; Walklate, 1987; Aylor & Ferrandino, 1989; Aylor, 1990). Wind speed increases logarithmically with distance above the crop surface (Aylor & Ferrandino, 1989), so turbulent wind will enhance vertical mixing and encourage escape of splash and spray of locally dispersed droplets containing pathogen propagules from the canopy layer (Aylor, 1990) into the winds above the canopy that have greater velocity. Splash dispersal of Xcc occurs during rain showers in calm conditions in the canopy of a canker-infected citrus tree (Pruvost et al., 2002), but at wind velocities common in thunderstorms (and in tropical storms or hurricanes) the number of bacteria dispersed in splash beyond a canker-infected tree canopy increases (Bock et al., 2005, 2010b, 2011).

The heterogeneity in the quantity of Xcc dispersed over short distances from an infected citrus canopy in turbulent wind has not been fully characterized, and the functional relationship with height and wind speed remains to be confirmed (Bock et al., 2011). Whether the number of bacteria dispersed at a given height is uniform across the downwind plume has not been established, nor has the variability in quantity dispersed at each height. Diseased citrus trees are not uniform in structure, disease distribution, leaf or branch position, and these factors possibly contribute to uneven quantities of dispersed bacteria downwind. Improved understanding of the pattern and numbers of Xcc dispersed locally from canker-infected citrus canopies can be useful to help guide the design and implementation of disease management strategies, including use of wind breaks (Gottwald & Timmer, 1995; Behlau et al., 2008; Tamang et al., 2010).

The aim of these experiments was to simulate the turbulent wind in rainstorm conditions over short distances to: (i) describe the cross-sectional dispersal of Xcc in wind-driven splash immediately downwind from canker-infected grapefruit trees; (ii) ascertain whether the functional relationship immediately downwind with height was consistent; (iii) determine whether the dispersal pattern across the wind flow was consistent; and (iv) confirm previous observations on the relationship between wind speed and the number of bacteria dispersed.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Location, plant material, wind and rain generation

The experiments were done on four separate dates (15 December 2004, 22 and 23 March 2005 and twice on 18 May 2005) at a USDA containment field site in Broward Co., Florida. On 15 December 2004, three young, canker-infected grapefruit trees (cv. Ruby Red) in 38 L containers, as the inoculum source, were placed in a triangular formation with the third tree situated downwind and midway between the other two trees. On 22 and 23 March and 18 May 2005, only two trees acted as the inoculum source, and were placed next to each other, perpendicular to the wind path. Tree height was approximately 1·5 m, with a crown of approximately 0·8 × 0·8 m. Turbulent wind was generated by an axial fan and rain was simulated by overhead garden sprayers, as previously described (Bock et al., 2010a). Canker lesions initially release a surge of bacteria (Bock et al., 2005); to avoid this pulse of dispersal, plants were sprayed for 1·5 h to allow bacterial exudation from lesions to stabilize prior to conducting the experiment. Disease severity was assessed visually on the trees by counting lesions on a sample of 40 leaves on each tree. There were 15·8 lesions per leaf on the plants used on 15 December 2004 (SD = 14·9, 94·2% leaves infected), 7·7 lesions per leaf on the plants used on 22 and 23 March (SD = 6·5, 92·5% leaves infected), and 6·4 lesions per leaf on the plants used on 18 May 2005 (SD = 6·8, 90·7% leaves infected).

Splash was collected downwind using panel samplers (Parker et al., 2005). There were five panel samplers (each 30 × 12·5 cm, 375 cm2) on each of four vertical frames placed 1 m downwind, adjacent to each other across the wind path and facing the wind. Panel centre heights were 30, 70, 110, 130 and 180 cm above ground. Each panel was rinsed with 50 mL water after each sample run to ensure any remaining bacteria on the panel surface were collected. The sample was directed into a collection vessel at the base of the panel. Panel surfaces and collection vessels were thoroughly cleaned between samples. In all experiments, zero wind speed was obtained with the fan switched off. On 15 December 2004 the fan was operated at 1 and 3 m upwind from the canopy. In all other experiments the fan was operated at 1, 3 and 5 m upwind from the canopy. The experiment was repeated four times.

Samples were collected over a 10 min period, the volume measured, and plated out on KCB semiselective media (nutrient agar (NA) amended with kasugamycin (16 mg L−1), cephalexin (35 mg L−1) and chlorothalanil (12 mg L−1 tetrachloroisophthalonitrile)). In addition to the samples of splash, well water used for spray was also plated out to act as a control. The plates were incubated at 27°C for 5–7 days prior to counting colonies typical of reference cultures of Xcc, and the bacteria flux density (BFD, bacteria per square centimetre per minute; Paul et al., 2004) was calculated. Weather data (wind speed, temperature and rainfall) were recorded every 60 s using Davis Weather instruments (Weather Wizard III, Davis Instruments) as described by Bock et al. (2010a).

Data analysis

All analyses were performed in sas v9.2 (SAS Systems). Means and standard deviations were calculated for the weather variables for each sample period. Functional relationships were fitted to both untransformed and log transformed data. The log transformed data was used to compare the functional relationships between height of sample and BFD, and between wind speed and BFD between repeats of the experiments. A linear regression function was fitted to the transformed data. A negative exponential function was fitted to the untransformed data for height, and an exponential function for the effect of wind speed. The appropriateness of the linear function was ascertained using the coefficient of determination (R2) and significance of the model (based on F- and P-values). proc reg was used to fit all regression solutions. General linear modelling (GLM) with a dummy variable was used to compare the linear regression solutions from each of the experiments for the effects of height (height × date interactions) on log BFD and log volume, and to compare the regression solutions from each of the experiments for the effects of wind speed (wind speed × date interactions) using the ‘contrast’ statement in sas. No meaningful relationship was apparent between sample locations perpendicular to the direction of wind flow, and these data were analysed using GLM to compare the BFD collected at each horizontal location for each date with main effects of height, wind speed and panel position (and second order interactions) and means separation for panel position using Tukey’s HSD test (= 0·05). A non-orthogonal multiple regression analysis (proc reg) using forward selection combining data from the five experiments was performed to gain insight into the predictability of BFD based on the independent variables (height, horizontal position and wind speed) using the proportion BFD of the total collected ((sample BFD/maximum BFD collected in run) × 100). A test of co-linearity was made to check whether any of the independent variables were correlated.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Maximum mean wind speed measured at the canopy face ranged from 12·6 to 19·9 m s−1, depending on the experiment (Table 1). Wind speed when the fan was situated 5 m from the canopy ranged from 3·2 to 4·6 m s−1. With the fan switched off, ambient wind speed ranged from 0 to 1·5 m s−1. Depending on the experiment run and the applied wind speed, the downwind wind speed at the panel face ranged from 0 to 5·15 m s−1.

Table 1.   The maximum mean wind speeds applied to the canopy of canker-infected grapefruit trees to study the characteristics of dispersal of bacteria of Xanthomonas citri subsp. citri downwind at different heights and distances on four different dates in five repeats of the experiment
DateFan distance from canopy (m)Wind speed at canopy face (m s−1)Wind speed 1 m downwind (m s−1)
  1. aMeans are based on all heights and horizontal locations.

  2. bStandard deviations in parentheses.

15 December 200400·87a (0·49)b0·55 (0·37)
36·42 (0·19)0·85 (0·16)
113·95 (2·15)1·83 (0·38)
22 March 200500 (0)1·47 (0·63)
54·58 (0·36)4·39 (0·30)
37·55 (1·96)4·66 (1·03)
119·89 (0·44)5·15 (0·26)
23 March 200500·12 (0·19)0 (0)
53·36 (0·54)3·26 (1·05)
35·20 (1·35)4·24 (1·55)
115·51 (0·69)3·62 (0·070)
18 May 2005 (1)00 (0)0·16 (0·21)
53·22 (0·36)0·93 (0·25)
35·41 (1·47)1·18 (0·19)
114·05 (1·37)1·51 (0·33)
18 May 2005 (2)00 (0)0·08 (0·17)
53·21 (0·65)0·98 (0·17)
35·78 (0·85)1·26 (0·34)
112·57 (0·34)1·27 (0·25)

On all five occasions the experiment was repeated, there was variation in the BFD collected at each height and across the dispersal path (Fig. 1), although consistently fewer bacteria were dispersed above the height of the canopy (>150 cm) at the sampling distance of 1 m. BFD was greatest at the lowest heights, and a linear function described the relationship between log BFD and sample height (Fig. 2a) for all repeats of the experiment, although a negative exponential function fitted the untransformed data (Fig. 2b). Despite the same functional relationship, a pairwise comparison of the repeat experiments demonstrated significant differences among experiments in BFD collected (Table 2).

image

Figure 1.  The mean bacteria flux density (log BFD, bacteria cm−2 min−1) of dispersed inoculum of Xanthomonas citri subsp. citri collected at different heights and horizontal positions 1 m downwind from the canopy of infected citrus trees. Sampling periods were 10 min; surface fit based on a total of 20 sample points, each a mean of four wind speeds (except for 15 December 2004, which was based on three wind speeds).

Download figure to PowerPoint

image

Figure 2.  The relationship between sample height and bacteria flux density (BFD, bacteria per square centimetre per minute) of Xanthomonas citri subsp. citri collected by panel samplers downwind from canker-infected grapefruit canopies subject to simulated wind/rain events in five repeat experiments. (a) The linear function relationship for the log transformed BFD data, and (b) the negative exponential relationship for the untransformed BFD data. Panel samplers were positioned 1 m downwind of the plants, and wind speed was varied by placing the fan at different distances upwind of the plants. Parameters for the linear regression solutions for each date for the effect of height on log BFD: 15 December 2004, = 1·19, = −0·0036 (= 8·8, = 0·06, R2 = 0·75); 22 March 2005, = 1·63, = −0·0082 (= 46·5, = 0·007, R2 = 0·94); 23 March 2005, = 1·57, = −0·0078 (= 1467, = <0·0001, R2 = >0·99); 18 May 2005 (1) = 1·37, = −0·0067 (= 278106, = 0·0005, R2 = 0·99); 18 May 2005 (2), = 0·96, = −0·0049 (= 78·2, = 0·003, R2 = 0·96. F-distribution value that tests goodness of fit for the model, = probability the F-value is significant and R2 = coefficient of determination (proportion of variability accounted for by model).

Download figure to PowerPoint

Table 2.   Comparisona of dates for the relationship between bacteria flux density (log BFD, bacteria cm−2 min−1) of Xanthomonas citri subsp. citri collected and heightb. Samples were collected by vertical panels situated downwind from canker-infected canopies of grapefruit trees exposed to different wind speeds in five repeats of the experiment
Date comparisoncMean square F-valued P-valuee
  1. aGeneral linear modelling effects for BFD: height F-value = 286·4 (P < 0·0001); date F-value = 11·8 (P = 0·0002); height × date F-value = 5·8 (= 0·005).

  2. bHeights were 30, 70, 110, 130 and 180 cm, with the canopy inoculum source being approximately 1·5 m tall. Panel collectors were placed on four towers situated next to each other and perpendicular to the wind flow.

  3. cTests between experiment dates were made using the ‘contrast’ statement in SAS Proc GLM.

  4. d F: F-distribution value that tests the contrast.

  5. e P: probability the F-value is significant.

15 Dec 2004 vs. 22 March 20050·14115·80·001
15 Dec 2004 vs. 23 March 20050·11913·40·002
15 Dec 2004 vs. 18 May 2005 (1)0·0667·40·02
15 Dec 2004 vs. 18 May 2005 (2)0·0111·20·3
22 March 2005 vs. 23 March 20050·0010·10·8
22 March 2005 vs. 18 May 2005 (1)0·0141·60·2
22 March 2005 vs. 18 May 2005 (2)0·0738·20·01
23 March 2005 vs. 18 May 2005 (1)0·0080·90·4
23 March 2005 vs. 18 May 2005 (2)0·0586·50·02
18 May 2005 (1) vs. 18 May 2005 (2)0·0232·60·1

Bacteria flux density was greatest at highest wind speeds and a linear function described the relationship between the log BFD and wind speed (Fig. 3a) although an exponential function described the relationship between untransformed BFD and wind speed (Fig. 3b). These data consistently show that the rate bacteria in splash are dispersed over a short range (i.e. escape downwind from the canopy of the tree) increases rapidly at approximately 8 m s−1 in turbulent wind. A contrast of the repeat experiments showed no differences due to the effect of wind speed between any of these experiments (F-value for contrasts = 0–2·8, P-value = 0·2–1·0).

image

Figure 3.  The relationship between wind speed and bacteria flux density (BFD, bacteria per square centimetre per minute) of Xanthomonas citri subsp. citri collected by panel samplers downwind from canker-infected grapefruit canopies subject to simulated wind/rain events in five repeats of the experiment. (a) The linear function relationship for the log transformed BFD data, and (b) the exponential relationship for the untransformed BFD data. Panel samplers were placed at five separate heights 1 m downwind of the plants, and wind speed was varied by placing the fan at different distances upwind of the plants. Parameters for the linear regression solutions for each date for the effect of wind speed on BFD: 15 December 2004, = 0·26, = 0·078 (= 16·4, = 0·2, R2 = 0·94); 22 March 2005 = 0·15, = 0·079 (= 41·2, = 0·2, R2 = 0·95); 23 March 2005 = 0·28, = 0·078 (= 44·7, = 0·02, R2 = 0·96); 18 May 2005 (1) = 0·03, = 0·112 (= 48·7, = 0·02, R2 = 0·96); 18 May 2005 (2) = −0·06, = 0·095 (= 24·5, = 0·04, R2 = 0·93), where the linear function is represented as log y = a + bx (= intercept, = height parameter and y = the independent variable, log BFD), F: F-distribution value that tests goodness of fit for the model, P: probability the F-value is significant and R2: coefficient of determination (proportion of variability accounted for by model).

Download figure to PowerPoint

The GLM analysis of BFD collected in splash at different tower locations perpendicular to the wind flow were significantly different in all experiments except on 15 December 2004 and 23 March 2005 when there was no difference between tower locations for BFD collected, demonstrating that the quantity of bacteria dispersed from the canopy was often variable (Table 3).

Table 3.   Mean bacteria flux density (log BFD, bacteria cm−2 min−1) of Xanthomonas citri subsp. citri collected by four vertical towers of five panels each, situated 1 m downwind from canker-infected grapefruit canopies subject to different wind speeds during simulated wind/rain events in five experiments
DateHorizontal position (cm)a F-valueb P-valuec
175185119
  1. aThe horizontal position is the midpoint of each panel tower relative to the midpoint of the upwind inoculum source (across the direction of wind flow from left to right).

  2. b F: F-distribution value that tests goodness of fit for the model.

  3. c P: probability the F-value is significant.

  4. dNumbers with the same letter are not significantly different according to Tukey’s HSD test (P = 0.05).

15 Dec 20040·99ad0·76a0·68a0·73a1·40·3
22 March 20050·77ab0·88a0·79ab0·65b4·20·01
23 March 20050·77a0·85a0·78a0·62a1·30·3
18 May 2005 (1)0·77a0·75ab0·48c0·57bc8·00·0004
18 May 2005 (2)0·53a0·53a0·26b0·50a11·5<0·0001

Multiple regression analysis of percentage of maximum BFD collected (Table 4) confirmed predictability in the sample as a proportion of total bacteria dispersed at a given height and wind speed. The proportion of BFD collected was described by = intercept + b(wind speed) + c(panel height) − d(wind speed × panel height). The fit was highly significant (= 141, < 0·0001, d.f. = 3), with approximately half the variance accounted for (R2 = 0·53). This demonstrated that in these simulated studies the independent variables (and interactions) of wind speed and sample height were useful in predicting the proportion of the total BFD of Xcc collected downwind at a specific point on the axis of the sampling plane, with the β-value (1·37) indicating wind speed was the major factor influencing the prediction. The test for co-linearity confirmed none of the independent variables were correlated.

Table 4.   Non-orthogonal multiple regression analysisa for all experiment runs combinedb of the relationship between wind speed, height and horizontal position of sample in plume, and bacteria flux densityc (BFD, bacteria cm−2 min−1) of Xanthomonas citri subsp. citri collected by panels positioned downwind from canker-infected canopies of grapefruit trees exposed to rain splash and different wind speeds
Intercept (a)Parameters (β-valuesd) R2e Ff Pg
b c d
  1. aMultiple regression tested using forward selection, a(intercept) + b(wind speed) + c(panel height) − d(wind speed × panel height).

  2. bData from all runs in experiments on 15 December 2004, 22 and 23 March 2005 and 18 May 2005 were analysed together.

  3. c BFD as a percentage of the total collected in each experiment repeat ((sample BFD/maximum BFD collected in run) × 100).

  4. d β-Values (standardized estimates) indicates the relative contribution of each factor in the model to the prediction of the multiple regression.

  5. e R 2: coefficient of determination (proportion of variability accounted for by model).

  6. f F: F-distribution value that tests goodness of fit for the model.

  7. g P: probability the F-value is significant.

−8·724·72 (1·37)0·06 (0·14)−0·03 (−1·09)0·53141<0·0001

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

These results demonstrate that in turbulent wind the dispersal of Xcc downwind of a canker-infected canopy is not uniform over short distances (Fig. 1, Table 3). However, there was a consistently greater BFD collected at lower sample heights compared to samples taken higher above ground level. Turbulent wind (such as produced by a fan or in a storm) is known to play a role in dispersal of pathogens (Aylor, 1990), and a previous study showed that the BFD of Xcc declined with height and distance downwind from an inoculum source (Bock et al., 2011), but this is the first time that the pattern of dispersed Xcc has been described across the wind path immediately downwind of canker-infected citrus trees exposed to turbulent winds and rain simulating storm conditions, confirming that the BFD is variable.

The relationship between sample height and log BFD was described by a linear function, and the non-transformed data was described by a negative exponential function. This is in agreement with results described in a previous study (Bock et al., 2011). The functional relationship was consistent among experiments, suggesting that this is a reliable description of the vertical dispersal gradient under the conditions applied here, although differences between inoculum sources (tree structure, disease severity and distribution) will doubtless affect the pattern of wind blowing through the canopy, explaining the differences in the regression coefficients among experiments.

In calm conditions most of the spray and splash fell directly through the canopy with limited lateral spread (Pruvost et al., 2002), and splash in calm conditions is known to afford particularly short range transport (Brennan et al., 1985). Increased wind speed resulted in more Xcc dispersed in greater quantities of splash that was sprayed through the canopy instead of falling vertically to the ground. This was noted previously (Serizawa et al., 1969; Stall et al., 1980; Serizawa, 1981; Bock et al., 2005, 2010a), but in this study the relationship between log BFD and wind speed was described by a linear function (non-transformed BFD data had an exponential function), i.e. as wind speed increased, the rate of the increase in BFD was greater, resulting in more bacteria dispersed 1 m downwind when the canopy of a diseased citrus tree is impacted with turbulent wind. The trend for a rapidly increasing rate of splash dispersed bacteria at 1 m occurred at approximately 8 m s−1. At higher wind speeds a greater proportion of the splash collected at 1 m will have passed through the canopy and will be infused with Xcc. In addition, turbulent wind at higher wind speeds will cause more complete wetting of leaf surfaces and more effective mixing and removal of Xcc bacteria from the canopy, resulting in greater BFD dispersed, which will increase the potential for epidemic development both locally and over longer distances.

Work on infection of citrus with Xcc has shown disease severity increases at wind speeds of ∼8 m s−1, presumably due to water congestion of foliar tissues as inoculum is forced through stomata in between underlying mesophyll cells (Serizawa et al., 1969; Stall et al., 1980; Serizawa, 1981; Bock et al., 2010b), so the combined effects of wind speed on increasing dispersal and infection above this threshold of 8 m s−1 could have a dramatic effect on epidemic development in storms.

Xcc was collected at 1·8 m height (maximum canopy height was 1·5 m). Thus when the canopy face is exposed to higher wind speeds the impact of spray and splash potentially results in droplets containing bacteria being entrained in air above the canopy, where it has the chance to be dispersed further (Aylor & Ferrandino, 1989; Aylor, 1990, 1999), and ultimately cause more infection. However, once escaped from the canopy, the potential for dispersal to greater distances will depend on several factors including wind speed and droplet size (Venette & Kennedy, 1975; Carnegie, 1980; Quinn et al., 1980; Fitt & Nijman, 1983; Brennan et al., 1985; Walklate, 1987).

The multiple regression analysis demonstrated the predictability of the relationships using this system to generate wind and rain splash, previously shown in a similar study incorporating distance (Bock et al., 2011). The range of droplet sizes, and gradients of droplet deposition in relation to droplet size downwind of infected canopies has not been established, and tree canopy structure, disease distribution and wind flow in the canopy probably contribute to the amount of inoculum at any point downwind. These studies were conducted using a fan to generate wind so no attempts have been made to draw conclusions regarding the intermediate- or long-range distance dispersal of Xcc. Wind (and rain) in storms will doubtless be somewhat different, particularly as natural wind is a function of very large parcels of air associated with meteorological events. Nonetheless, the impact of turbulent wind on dispersal of Xcc over a short range in the immediate vicinity of a diseased canopy of a young tree typical of an immature orchard provides valid information on processes occurring over the short range.

The results here demonstrate variability in the quantity of bacteria collected, although the effects of height and wind speed on BFD were characterized by specific relationships. Knowledge of the ability of bacteria to enter rain splash that becomes entrained in wind suggests that orchard management approaches such as wind breaks can help minimize the risk of dispersal of Xcc and ultimately the incidence of infection both locally and over greater distances (Gottwald & Timmer, 1995; Behlau et al., 2008; Tamang et al., 2010). In areas particularly prone to wind/rain events these precautions should help reduce the damage caused by citrus canker.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This work was funded in part by the Florida Citrus Production and Research Advisory Council. Jose Renteria (USDA-APHIS-PPQ, Edinburg, TX) constructed the panel collection devices and Tara Zacharakis (Florida Atlantic University) helped provide assistance with performing some of the experiments.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Aylor DE, 1990. The role of intermittent wind in the dispersal of fungal pathogens. Annual Review of Phytopathology 28, 7392.
  • Aylor DE, 1999. Biophysical scaling and the passive dispersal of fungus spores: relationship to integrated pest management strategies. Agricultural and Forest Meteorology 97, 27592.
  • Aylor DE, Ferrandino FJ, 1989. Dispersion of spores released from an elevated line source within a wheat canopy. Boundary-Layer Meteorology 46, 25173.
  • Behlau F, Belasque J, Bergamin-Filho A, Graham J, Leite R, Gottwald TR, 2008. Copper sprays and windbreaks for control of citrus canker on young orange trees in southern Brazil. Crop Protection 27, 80713.
  • Behlau F, Belasque J, Graham J, Leite R, 2010. Effect of frequency of copper application on control of citrus canker and the yield of young bearing sweet orange trees. Crop Protection 29, 3005.
  • Bock CH, Parker PE, Gottwald TR, 2005. The effect of simulated wind-driven rain on duration and distance of dispersal of Xanthomonas axonopodis pv. citri from canker infected citrus trees. Plant Disease 89, 7180.
  • Bock CH, Graham JE, Gottwald TR, Cook AZ, Parker PE, 2010a. Wind speed effects on the quantity of Xanthomonas citri subsp. citri dispersed downwind from canopies of grapefruit trees infected with citrus canker. Plant Disease 94, 72536.
  • Bock CH, Graham JH, Gottwald TR, Cook AZ, Parker PE, 2010b. Wind speed and wind associated injury affect severity of citrus canker on Swingle citrumelo. European Journal of Plant Pathology 128, 2138.
  • Bock CH, Gottwald TR, Cook AZ, Parker PE, Graham JH, 2011. The effect of wind and rain on dispersal of bacteria of Xanthomonas citri subsp. citri at different heights downwind of canker-infected grapefruit trees. Journal of Plant Pathology 93, 66777.
  • Brennan RM, Fitt BDL, Taylor GS, Colhoun J, 1985. Dispersal of Septoria nodorum pycnidiospores by simulated rain and wind. Journal of Phytopathology 112, 2917.
  • Carnegie SF, 1980. Aerial dispersal of the potato gangrene pathogen, Phoma exigua var. foveata. Annals of Applied Biology 94, 16573.
  • Danos E, Berger RD, Stall RE, 1984. Temporal and spatial spread of citrus canker within groves. Phytopathology 74, 9048.
  • Faulwetter RC, 1917. Wind-blown rain, a factor in disease dissemination. Journal of Agricultural Research 10, 63948.
  • Fitt BDL, Nijman DJ, 1983. Quantitative studies on dispersal of Pseudocercosporella herpotrichoides spores from infected wheat straw by simulated rain. Netherlands Journal of Plant Pathology 89, 198202.
  • Fitt BDL, Gregory PH, Todd A, McCartney HA, MacDonald OC, 1987. Spore dispersal and plant disease gradients, a comparison between two empirical models. Journal of Phytopathology 118, 22742.
  • Fitt BDL, McCartney HA, Walklate PJ, 1989. The role of rain in dispersal of pathogen inoculum. Annual Review of Phytopathology 27, 24170.
  • Goto M, Yaguchi Y, 1979. Relationship between defoliation and disease severity in citrus canker (Xanthomonas citri). Annals of the Phytopathological Society of Japan 45, 68994.
  • Gottwald TR, Irey M, 2007. Post-hurricane analysis of citrus canker II: Predictive model estimation of disease spread and area potentially impacted by various eradication protocols following catastrophic weather events. Plant Health Progress, doi: 10.1094/PHP-2007-0405-01-RS.
  • Gottwald TR, Timmer LW, 1995. The efficacy of windbreaks in reducing the spread of citrus canker caused by Xanthomonas campestris pv. citri. Tropical Agriculture 72, 194201.
  • Gottwald TR, Graham JH, Engel DS, 1992a. Analysis of foci of asiatic citrus canker in a Florida citrus orchard. Plant Disease 76, 38996.
  • Gottwald TR, Reynolds KM, Campbell CL, Timmer LW, 1992b. Spatial and spatiotemporal autocorrelation analysis of citrus canker epidemics in citrus nurseries and groves in Argentina. Phytopathology 82, 84351.
  • Gottwald TR, Graham JH, Bock CH et al. , 2009. The epidemiological significance of post-packinghouse survival of Xanthomonas citri ssp. citri for dissemination of Asiatic citrus canker via infected fruit. Crop Protection 28, 50824.
  • Irey M, Gottwald TR, Graham JH, Riley TD, Carlton G, 2006. Post-hurricane analysis of citrus canker spread and progress towards the development of a predictive model to estimate disease spread due to catastrophic weather events. Plant Health Progress, doi: 10.1094/PHP-2006-0822-01-RS.
  • Koizumi M, 1985. Citrus canker: the world situation. In: Timmer LW, ed. Citrus Canker: An International Perspective. Lake Alfred, FL, USA: University of Florida, 27.
  • Madden LV, 1992. Rainfall and the dispersal of fungal spores. Advances in Plant Pathology 8, 3979.
  • McCartney HA, Fitt BDL, 1985. Construction of dispersal models. In: Gilligan CA, ed. Advances in Plant Pathology Volume 3: Mathematical Modelling of Crop Disease. London, UK: Academic Press, 10743.
  • Parker PE, Bock CH, Gottwald TR, 2005. Techniques to sample citrus canker bacteria (Xanthomonas axonopodis pv. citri) in wind-blown spray. Plant Disease 89, 132430.
  • Paul PA, El-Allaf SM, Lipps PE, Madden LV, 2004. Rain splash dispersal of Gibberellazeae within wheat canopies in Ohio. Phytopathology 94, 13429.
  • Pruvost O, Boher B, Brocherieux C, Nicole M, Chiroleu F, 2002. Survival of Xanthomonas axonopodis pv. citri in leaf lesions under tropical environmental conditions and simulated splash dispersal of inoculum. Phytopathology 92, 33646.
  • Quinn CE, Sells IA, Graham DC, 1980. Soft rot Erwinia bacteria in the atmospheric bacterial aerosol. Journal of Applied Microbiology 49, 17581.
  • Roberto SR, Gottwald TR, Graham JH, Riley T, 2001. Xanthomonas axonopodis pv. citri aerosol production in Miami, Florida. Summa Phytopathologica 27, 5660.
  • Schubert T, Shabbir J, Rizvi A et al. , 2001. Meeting the challenge of eradicating citrus canker in Florida – again. Plant Disease 85, 34056.
  • Serizawa S, 1981. Recent studies on the behaviour of the causal bacterium of the citrus canker. Proceedings of the International Society for Citriculture 1, 3957.
  • Serizawa S, Inoue K, Goto M, 1969. Studies on Citrus canker. I dispersal of the citrus canker organism. Bulletin of Schizuoka Prefecture Citrus Experiment Station 8, 815.
  • Stall RE, Miller JW, Marco GM, de Echenique BIC, 1980. Population dynamics of Xanthomonas citri causing cancrosis of citrus in Argentina. Proceeding of the Florida State Horticulture Society 93, 104.
  • Tamang B, Andreu MG, Rockwood DL, 2010. Microclimate patterns on the leeside of single-row tree windbreaks during different weather conditions in Florida: implications for improved crop production. Agroforestry Systems 79, 11122.
  • Venette JR, Kennedy BW, 1975. Naturally produced aerosol of Pseudomonas glycinea. Phytopathology 65, 7378.
  • Walklate PJ, 1987. A random-walk model for dispersion of heavy particles in turbulent air flow. Boundary-Layer Meteorology 39, 17590.