Predicting the impact of environmental factors on citrus canker through multiple regression.

Climatic conditions play a significant role in the development of citrus canker caused by Xanthomonas citri pv. citri (Xcc). Citrus canker is regarded as one of the major threats being faced by citrus industry in citrus growing countries of the world. Climatic factors exert significant impacts on growth stage, host susceptibility, succulence, vigor, survival, multiplication rate, pathogen dispersion, spore penetration rate, and spore germination. Predicting the impacts of climatic factors on these traits could aid in the development of effective management strategies against the disease. This study predicted the impacts of environmental variables, i.e., temperature, relative humidity, rainfall, and wind speed the development of citrus canker through multiple regression. These environmental variables were correlated with the development of canker on thirty (30) citrus varieties during 2017 to 2020. Significant positive correlations were noted among environment variables and disease development modeled through multiple regression model (Y = +24.02 + 0.5585 X1 + 0.2997 X2 + 0.3534 X3 + 3.590 X4 + 1.639 X5). Goodness of fit of the model was signified by coefficient determination value (97.5%). Results revealed the optimum values of environmental variables, i.e., maximum temperature (37°C), minimum temperature (27°C), relative humidity (55%), rainfall (4.7-7.1 mm) and wind speed (8 Km/h), which were conducive for the development of citrus canker. Current study would help researchers in designing better management strategies against citrus canker disease under changing climatic conditions in the future.

Introduction

Citrus is one of the major fruit crops grown in >140 countries situated in tropical and subtropical regions of the world [1-7]. It is a popular fruit cultivated globally due to its easier availability, popularity, commercial importance, and a major source of vitamin C. It has a significant contribution towards human nutrition. The Citrus orchards are attacked by Xanthomonas citri pv. citri (Xcc), which is a noxious pathogen causing citrus canker disease [8-10]. Xanthomonas citri is a straight, rod shaped, mono-flagellum, gram-negative bacterium with a wide host range [11]. Citrus canker is one of the most notorious citrus diseases in Florida, USA, leading towards mass eradication of plants in the entire state [12]. The disease is still spreading even after spending an estimated 12 million US$ annually on its management [13-15]. The Xcc infestation initially seen as a small lesion on the leaves, which is expanded in later stages and ultimately become necrotic surrounded by water-soaked, oily margins. Chlorotic rings of yellow color appear on the surfaces of fruits, leaves and stems as disease progresses and these rings eventually develop into a crater-like appearance [6, 16]. The Asiatic citrus leaf-miner (Phyllocnists citrella) significantly accelerates the severity of citrus canker through feeding and transferring bacterium to new leaf growth [17].

The Environmental factors play a crucial role in the dispersion and multiplication of Xcc. The bacterium multiplies rapidly in the lesions and transfer to other plants under ample moisture availability [18]. Furthermore high rainfall and wind speed accelerate the dispersion rate of the disease [19]. Extreme weather conditions such as tropical storms and hurricanes result in long-distance dissemination of the pathogen and increase disease infestation [20-23]. Environmental factors significantly affect host susceptibility, succulence, vigor, survival, and multiplication rate of the pathogen, pathogen dispersion, and spore penetration and germination rates. Earlier studies indicated that temperature between 30–38°C with high relative humidity play a crucial role in the disease development [24, 25].

A couple of similar studies [26, 27] reported that temperature, wind velocity, relative humidity and rainfall are the main factors responsible for the spread and development of citrus canker disease. Strong wind driven rain splashes provided conducive environment for the dispersal of citrus canker [28-30]. Similarly, hot summer and mild winter with some alternations of low and high temperature make the pathogen more aggressive [31]. The Xcc multiplies in the lesions of leaves, stems, and fruits, and disperse to new healthier host plants by rain splashes. During the presence of ample moisture on lesions, the bacteria ooze out and disperse to infect new plant growth. Rainstorms during monsoon increase the epidemics of citrus canker in the presence of active source of inoculum [32, 33].

Climate anomalies are resulting in unstable conditions for plants’ growth since past years, which ultimately exert immense pressure on food production. The studies relating to epidemiological factors helps in management decisions by determining the conducive environmental conditions for the development of diseases such as citrus canker [34, 35]. Hence, development and utilization of disease predictive model with relation to environmental factors is probably the most effective way to manage the prevalence of citrus canker disease. However, unfortunately no such work has been previously done under the climatic conditions of Pakistan. Correlation of environmental variables in the prevalence and perpetuation of Xcc under native climatic conditions of the country has been computed and interpreted in this study. The results would provide novel perceptive to manage citrus canker disease under changing environmental conditions.

Materials and methods

Data collection

This three-year study collected citrus canker disease incidence data from the experimental area of Department of Plant Pathology, University of Agriculture Faisalabad, Pakistan. Data relating to environmental variables including maximum and minimum temperatures (°C), wind speed (Km/h), rainfall (mm) and relative humidity (%) were obtained from meteorological station located at Agronomy research area, University of Agriculture Faisalabad, Pakistan. The relationship of these environmental variables with the disease development was predicted by regression and correlation analysis and a predictive model was developed based on the obtained results.

Regression analysis

Regression analysis was used to determine the relationship between environmental variables and disease development/incidence [36, 37]. Two different regression models, i.e., simple, and multiple regression models were used in the study. The mathematical equations for these models are presented as Eq 1 and Eq 2.

Here, Y acts as response variable in case of disease, while X works as explanatory variable. The β0 denotes as intercept and β1 is the slope.

For multiple linear regression models, more than one explanatory or predictor variables (X) are included as compared to the simple linear regression analysis.

Here, x characterizes the compilation of predictors x1, x2,… xi in the model, and β1, β2,…βi act for the corresponding regression coefficients and ∈ is the random error or interruption in the experiment [38, 39].

Characterization of environmental factors conducive for citrus canker

All environmental data as well as disease incidence (%), and alterations in the environmental factors and disease incidence were analyzed by least significant difference test (LSD at P<0.05) [40]. The effect of environmental factors on citrus canker disease was modeled by correlation. Mean square error (MSE) Mallows Cp and R2 were used a criteria for selecting the best models [41, 42].

Goodness of fit of the model

The correlation was used for determining the goodness of fit of the model [43-45]. The varieties/cultivars in which >50% of the environmental variables exerted significant effect were plotted and most conducive environmental factors for disease development were determined. The manipulation of these factors on disease infestation was tested by drawing a comparison between observed and predicted disease incidence values by multiple regression models [45, 46]. Furthermore, disease predictive model depending on environmental conditions was developed which has significant influence on citrus canker disease development.

Statistical analysis

The data consisted of an average of three replicates and differences among treatments were estimated by one-way analysis of variance (ANOVA). The means were compared using least significant difference post-hoc test (P < 0.05) where ANOVA indicated significant differences. All statistical computations were made on SPSS 20.0 [47]. Microsoft excel 2016 was used to calculate the standard errors of the meas. Graphical presentation was completed on Origin Pro 9.0 (OriginPro, Northampton, USA). The minimal dataset of the study has been uploaded as S1 Dataset.

Results

Development and evaluation of citrus canker predictive model

Multiple regression equation of citrus canker predictive model for two years was Y = +24.02 + 0.5585 X1 + 0.2997 X2 + 0.3534 X3 + 3.590 X4 + 1.639 X5. Here (Y = disease incidence X1 = maximum temperature, X2 = minimum temperature, X3 = relative humidity, X4 = rainfall and X5 = wind speed). The R2 value expressed that model was statistically fitted well for environmental variables. Some data points deviated from the reference line according to the normal probability (Fig 1), while most values were scattered equally around the residual line in case of residual vs. fit model which showed better fit (Fig 2). Few data points were little far from the line of reference i.e., near to zero; -3.5 to + 4 primarily exhibited as an error in the regression model. Model was designed according to [36].

Fig 1

Plot of normal probability for the citrus canker disease predictive model based on two years data.

Fig 2

Residual versus fit plot for the regression model of citrus canker disease.

Assessment of disease predictive model by comparing dependent variables with regression coefficient through physical theory

Analysis of variance of regression articulated that maximum and minimum temperature, relative humidity, rainfall, and wind speed significantly contributed towards disease development. The R2 value of 97.5% expressed that model was statistically suitable under given environmental conditions. Variable’s coefficients of regression model for citrus canker are given in Table 1.

Table 1

Regression model’s coefficients of variables for citrus canker disease.

Parameter	Coefficient	Stander Error	t-Stat	P-value
Intercept	-24.02	2.56	-9.40	0.000*
Maximum temperature (°C)	0.5585	0.0570	9.80	0.000*
Minimum temperature (°C)	0.2997	0.0302	9.92	0.000*
Relative humidity (%)	0.3534	0.0495	7.14	0.000*
Rainfall (mm)	3.590	0.319	11.26	0.000*
Wind speed (Km/h)	1.639	0.147	11.14	0.000*

* = Significant at P< 0.05.

Estimation of model for predicted and observed values

For assessing the reliability of model, value differences of observed and predicted data points were estimated. Among observed values, fourteen data points were beyond reference line (standard error = 1.81517) and created an error in experiment. According to graphs, maximum prediction (464 out of 480 values) values have differences (less than 5) were consolidated between 95% confidence interval (C.I) and 95% predictive interval (P.I) which showed that there was a good fit between predictive and observed values (Fig 3).

Fig 3

A fitted line plot for citrus canker disease with observed and predicted data points at 95% confidence and predictive intervals.

Correlation of environmental variables with the development of citrus canker disease on various citrus varieties during 2017–18 and 2018–19

Maximum and minimum temperature, relative humidity, rainfall and wind speed had significant positive correlation (P≤ 0.05) with citrus canker incidence during both years on thirty varieties (Blood red (Citrus sinensis cv. blood red), Malta (Citrus reticulata cv. malta), Mayer lemon (Citrus limonia cv. mayer lemon), China lemon (Citrus limonia cv. china lemon), Feutral’s early (Citrus reticulate cv. feutral’s early), Sweet lemon (Citrus limettioides), Jaffa (Citrus sinensis cv. jaffa), Succari (Citrus sinensis cv. succari), Mungal singh (Citrus reticulata cv. mungal singh), Grapefruit (Citrus paradise), Tangerine (Citrus reticulata cv. tangerine), Musambi (Citrus sinensis), Pine apple (Citrus sinensis cv. pine apple), Valentia late (Citrus sinensis cv. valentia late), Kinnow (Citrus reticulata cv. kinnow), Key lime (Citrus aurantifolia), Trifoliate orange (Citrus poncirus), Pomelo (Citrus grandis), Orange (Citrus sinensis), Bitter orange (Citrus aurantium), Rough lemon (Citrus jambhiri), Lemon (Citrus limon), Citron (Citrus medica), Citrus paradise cv. foster, Citrus paradise cv. duncan, Citrus paradise cv. shamber, Citrus sinensis cv. washington navel, Citrus sinensis cv. rubby red, GAL GAL (Citrus pseudolimon), Persian lime (Citrus latifolia) (Tables 2 and 3).

Table 2

Pearson correlation matrix of various environmental factors with citrus canker disease on different citrus varieties during 2017–18.

Sr #	Variety Name	Max. T (°C)	Min. T (°C)	RH (%)	RF (mm)	WS (Km/h)
1	Citrus sinensis cv. blood red	0.768**	0.906**	0.945**	0.453**	0.812**
2	Citrus reticulata cv. Malta	0.758**	0.899**	0.949**	0.457**	0.821**
3	Citrus limonia cv. mayer lemon	0.713**	0.853**	0.971**	0.577**	0.781**
4	Citrus limonia cv. china lemon	0.765**	0.906**	0.945**	0.466**	0.813**
5	Citrus reticulata cv. feutral’s early	0.824**	0.942**	0.921**	0.477**	0.756**
6	Citrus limettioides (Sweet lime)	0.806**	0.920**	0.938**	0.468**	0.790**
7	Citrus sinensis cv. Jaffa	0.639**	0.774**	0.889**	0.484 **	0.706**
8	Citrus aurantifolia swingle (Kagzi lime)	0.559**	0.700**	0.930**	0.494 **	0.855**
9	Citrus reticulata cv. mungal singh	0.730**	0.818**	0.860**	0.451**	0.667**
10	Citrus reticulata cv. Kinnow	0.784**	0.928**	0.892**	0.493**	0.756**
11	Citrus reticulata cv. Tangerine	0.696**	0.790**	0.810**	0.297Ns	0.535**
12	Citrus sinensis cv. Succari	0.710**	0.865**	0.977**	0.515**	0.819**
13	Citrus sinensis cv. pine apple	0.687**	0.785**	0.807**	0.484**	0.621**
14	Citrus sinensis cv. valentia late	0.705**	0.858**	0.978**	0.546**	0.827**
15	Citrus paradise	0.781**	0.915**	0.934**	0.456**	0.809**
16	Citrus aurantifolia (Key lime)	0.446**	0.597**	0.865**	0.529**	0.833**
17	Citrus poncirus (Trifoliate orange)	0.661**	0.772**	0.896**	0.567**	0.823**
18	Citrus grandis (Pomelo)	0.558**	0.679**	0.895**	0.390Ns	0.740**
19	Citrus sinensis	0.761**	0.899**	0.952**	0.461**	0.817**
20	Citrus aurantium (Bitter Orange)	0.594*	0.662**	0.693*	0.301Ns	0.605**
21	Citrus jambhiri (Rough lemon)	0.757**	0.899**	0.952**	0.468**	0.814**
22	Citrus medica (Citron)	0.502**	0.584**	0.717**	0.352Ns	0.434**
23	Citrus japonica (Kumquat)	0.585**	0.682**	0.710**	0.565**	0.678**
24	Citrus latifolia (Persian lime)	0.430*	0.562**	0.817**	0.457**	0.730**
25	Citrus pseudolimon (GAL GAL)	0.661**	0.672**	0.619**	0.250Ns	0.518**
26	Citrus sinensis cv. ruby red	0.584**	0.735**	0.971**	0.510**	0.818**
27	Citrus sinensis cv. washington navel	0.435**	0.509**	0.620**	0.256Ns	0.514**
28	Citrus paradise cv. Shamber	0.792**	0.920**	0.941**	0.448**	0.786**
29	Citrus paradise cv. Duncan	0.761**	0.899**	0.952**	0.461**	0.817**
30	Citrus paradise cv. Foster	0.763**	0.902**	0.949**	0.448**	0.813**

The values are Pearson’s correlation coefficient

Ns = non-significant

* = significant (P <0.05)

** = Highly significant.

Table 3

Correlation of environmental factors with canker disease on different varieties of citrus for 2018–19.

Sr#	Variety Name	Max. T (°C)	Min. T (°C)	RH (%)	RF (mm)	WS (Km/h)
1	Citrus sinensis cv. blood red	0.619**	0.760**	0.945**	0.922**	0.796**
2	Citrus reticulata cv. Malta	0.608**	0.75**	0.949**	0.922**	0.799**
3	Citrus limonia cv. mayer lemon	0.509**	0.664**	0.949**	0.851**	0.805**
4	Citrus limonia cv. china lemon	0.561**	0.709**	0.964**	0.894**	0.821**
5	Citrus reticulata cv. feutral’s early	0.692**	0.814**	0.915**	0.921**	0.800**
6	Citrus limettioides (Sweet lime)	0.699**	0.802**	0.909**	0.886**	0.769**
7	Citrus sinensis cv. Jaffa	0.544**	0.658**	0.900**	0.832**	0.757**
8	Citrus aurantifolia swingle (Kagzi lime)	0.228**	0.421**	0.842**	0.804**	0.830**
9	Citrus reticulata cv. mungal singh	0.555**	0.653**	0.787**	0.786**	0.724**
10	Citrus reticulata cv. Kinnow	0.623**	0.772**	0.943**	0.937**	0.813**
11	Citrus reticulata cv. Tangerine	0.551**	0.676**	0.858**	0.880**	0.807**
12	Citrus sinensis cv. Succari	0.564**	0.654**	0.610**	0.572**	0.410**
13	Citrus sinensis cv. pine apple	0.684**	0.795**	0.922**	0.923**	0.783**
14	Citrus sinensis cv. valentia late	0.578**	0.734**	0.957**	0.935**	0.877**
15	Citrus paradise	0.655**	0.793**	0.930**	0.924**	0.791**
16	Citrus aurantifolia (Key lime)	0.378Ns	0.516**	0.839**	0.859**	0.831**
17	Citrus poncirus (Trifoliate orange)	0.747**	0.867**	0.827**	0.924**	0.728**
18	Citrus grandis (Pomelo)	0.649**	0.775**	0.935**	0.941**	0.844**
19	Citrus sinensis	0.687**	0.827**	0.898**	0.859**	0.789**
20	Citrus aurantium (Bitter Orange)	0.633**	0.747**	0.924**	0.880**	0.828**
21	Citrus jambhiri (Rough lemon)	0.613**	0.761**	0.945**	0.942**	0.808**
22	Citrus medica (Citron)	0.701**	0.788**	0.891**	0.835**	0.732**
23	Citrus japonica (Kumquat)	0.383Ns	0.531**	0.868**	0.863**	0.683**
24	Citrus latifolia (Persian lime)	0.703**	0.816**	0.899**	0.921**	0.827**
25	Citrus pseudolimon (GAL GAL)	0.369Ns	0.491**	0.741**	0.678**	0.666**
26	Citrus sinensis cv. ruby red	0.451**	0.613**	0.808**	0.835**	0.822**
27	Citrus sinensis cv. washington navel	0.553**	0.586**	0.649**	0.699**	0.635**
28	Citrus paradise cv. Shamber	0.678**	0.798**	0.910**	0.876**	0.771**
29	Citrus paradise cv. Duncan	0.631**	0.771**	0.943**	0.929**	0.815**
30	Citrus paradise cv. Foster	0.628**	0.765**	0.945**	0.924**	0.808**

The values are Pearson’s correlation coefficient

Ns = non-significant

* = significant (P <0.05)

** = Highly significant.

Characterization of environmental factors conducive for the development of citrus canker disease on five varieties during 2017–18 and 2018–19

Five citrus varieties, i.e., Jaffa, Kagzi lime, Mayer lemon, Succari and Grapefruit were used for the determination of environmental factors conducive for the development of citrus canker. All environmental variables had positive significant correlation with citrus canker on all varieties during both years. The highest disease incidence (up to 55%) was recorded for Grapefruit under 37.2°C maximum temperature, while low disease incidence (7%) was observed on Jaffa under 37°C (Fig 4). Minimum disease incidence of 7% and less than 8% was noticed on Jaffa under 27.9°C minimum temperature (Fig 5). Jaffa showed 6.5% disease incidence during both years under 79.8% relative humidity as compared to disease incidence on Kagzi lime (18.9%), Mayer lemon (25.8%), Succari (40.5%) and Grapefruit (55.7%) (Fig 6). The similar variety Grapefruit showed 55.7% disease incidence 7.3 mm and 4.9 mm rainfall during 1st and 2nd year, respectively. Jaffa expressed relatively low disease incidence (<7%) (Fig 7). Disease incidence of less than 10% was recorded on Jaffa under 8 km/h wind speed during 2017 and 2018 (Fig 8). It was noticed that disease incidence increased from 8.5 to 55.7 and 8 to 55.4% on Grapefruit with rain splashes increasing, during both 2017–18 and 2018–19, respectively and same pattern was observed on all other varieties.

Fig 4

Relationship of maximum temperature with the development of citrus canker disease with during 2017–18 (A) and 2018–19 (B).

Fig 5

Relationship of minimum temperature with the development of citrus canker disease during 2017–18 (A) and 2018–19 (B).

Fig 6

Relationship of relative humidity with the development of citrus canker disease during 2017–18 (A) and 2018–19 (B).

Fig 7

Relationship of rainfall with the development of citrus canker disease during 2017–18 (A) and 2018–19 (B).

Fig 8

Relationship of wind speed with the development of citrus canker disease during 2017–18 (A) and 2018–19 (B).

Discussion

Host susceptibility, pathogen virulence and favorable environmental conditions are necessary for disease development. Environmental factors like temperature, rainfall, wind speed, and relative humidity are crucial elements for different diseases [48, 49]. Sudden fluctuations in these environmental conditions can favor the development of diseases [50-52]. These climatic conditions play a significant role in resistance/susceptibility of plants against pathogens. They can also change the growth pattern, production, dissemination, infection, survival of pathogen and the interaction between the host and causal agent [53-55].

All environmental factors expressed significant positive correlation with all tested varieties/cultivars in the present study. The highest disease incidence was observed under 20–28°C and 30–38°C minimum and maximum temperature, 47–74% relative humidity, 8 km/h wind speed and 4 mm rainfall during both years. Predictive model based upon two years data was developed, Y = +24.02 + 0.5585 X1 + 0.2997 X2 + 0.3534 X3 + 3.590 X4 + 1.639 X5 (Y = disease incidence X1 = maximum temperature, X2 = minimum temperature, X3 = relative humidity, X4 = rainfall and X5 = wind speed). The R2 value of 97.5% expressed that that model is statistically appropriate and revealed that all these prescribed factors contributed positively towards the development of citrus canker. Results of present study are supported by several earlier studies [17, 20, 56, 57] reporting that temperature, high relative humidity and strong wind with rain splash play a crucial role in the development of citrus canker diseases. Numerous studies concluded that mild temperature, humidity and wind driven rain exhibited major effect on disease development [32, 33, 58].

Environmental conditions prevailing after the contact of pathogen with the plants can help greatly in the development of disease. Temperature greatly influenced the initiation and development of plant diseases. Various pathogens complete their life cycle and multiplied much rapidly during the favorable temperature [59, 60]. Different plant diseases and pathogens prefer different lower and higher temperatures. Some bacterial pathogens like Pseudomonas grow faster in the presence of low temperature, while others like Xanthomonas and Ralstonia grow much faster under high temperature. Temperature is also responsible in favoring and inhibiting the expression of certain genes, rapid production of pathogenesis related proteins involved in disease resistance and susceptibility by affecting the genetic machinery of host cells [61, 62].

In contemporary studies, temperature (maximum and minimum) relative humidity, strong wind with heavy rainfall gave significant correlation with citrus canker disease on maximum varieties. A strong interaction between disease development and increase in relative humidity was seen, which has been witnessed in earlier studies [63, 64]. Pathogen can spread up-to 50 km/h in the high wind speed and by reducing wind speed in orchards can reduced the dispersal of Xcc [65, 66]. Rainfall also expressed highly positive correlation with the development of disease. Incidence of canker disease was significantly increased with increasing in rainfall [35]. These results were supported by some recent studies [21, 67] reporting that temperature with other environmental factors had a major role in disease development. Disease index of citrus canker was highest under increased temperature and relative humidity during the month of July followed by month of August and September [68, 69]. High precipitation with sharp wind also contributed in the multiplication of bacteria. The incidence of citrus canker was greater when the rainy season started in the month of September [70, 71].

The winds also take part in the prevalence of diseases by spreading the pathogens, increasing the number of lesions and somehow by accelerating the drying of wet surfaces of the plants. It also facilitates bacteria in releasing spores and transferred form diseased portions to healthy ones. Results of the present study are also supported by an early study [72] indicating that wind becomes more drastic and lethal when it is accompanied by heavy rain, especially for citrus canker. These wind-blown rain splashes caused injuries on the surface of plants which help a number of bacteria and other pathogens to get entry into the plants [42]. Rainstorm during monsoon also increases the epidemics of various bacterial diseases in the presence of active source of inoculum and spread through strong winds [73].

Conclusion

All the environmental variables, i.e., maximum, and minimum temperature, relative humidity, rainfall, and wind speed had significant positive correlation with citrus canker development on all varieties. Due to sudden fluctuations in the weather conditions, continuous monitoring of environmental variables is necessary for accurate prediction of citrus canker and its management. Installation of weather stations in major citrus growing areas would be helpful in risk assessment and forecasting systems in a specific area. Based on data collected from different areas, a Decian Support System can be developed for precise management of disease.

(XLSX)

Click here for additional data file.