Association of Chocolate Spot and Faba Bean Rust Epidemics with Climate Change Resilient Cultural Practices in Bale Highlands , Ethiopia

Field survey was conducted in south eastern Ethiopia to determine the disease intensity of chocolate spot and rust and to investigate the association of disease intensity (incidence and severity) with environmental factors and crop cultural practices. A total of 280 faba bean fields were surveyed in eight districts, and type of cropping system, weed management practices, crop growth stage, previous crop in the field, and sowing date were recorded. The associations of disease intensity with independent variables were evaluated using logistic regression model. Mean disease incidence of chocolate spot varied from 70.9 to 93.2% in most fields while percentage severity index (PSI) ranged from 10.5 to 47.1%. In a reduced multiple variable model, chocolate PSI ≤ 30% showed high probability of association with mixed cropping system, good weed management practices, late planting, and when faba bean was rotated with vegetables and cereals. The mean disease incidence of faba bean rust was varied from 23.6 to 78.2%, while the mean PSI of rust was varied from 4.8 to 37.9%. In Gasera, Dinsho, and Agarfa districts, poor weedmanagement practices, fields planted in the month of July, and when previous crop was legume had a high probability of association to (>20) rust PSI in a multiple variable reduced model. In contrast, soil types, fertilizer applied, and fungicides sprayed were not associated with disease intensity. The present study has identified cropping system, planting date, previous crop, district, and weed management practices as important variables that influence faba bean chocolate spot and rust epidemics in diversified fields. Therefore, proper weeding management practices, late planting, crop rotation habit, and other related farm practices should be carried out to reduce chocolate spot and rust impact until resistant faba bean genotypes are developed and distributed to the area.


Introduction
Faba bean (Vicia faba L.) is an important pulse crop produced all over the world for humans as a source of protein and carbohydrate and for animal feeds [1].It is also an excellent crop for fixing atmospheric nitrogen and as green manure [2,3].In addition, faba bean is used as a source of cash crop to farmers and foreign currency in Ethiopia [4].Faba bean is the sixth most important grain legume with 4.3 million tons of annual production worldwide.Next to China, Ethiopia is the second largest producer of faba bean [5,6] with annual production of 0.88 million tones and nearly covers 0.43 million hectares of lands counting about 31.2% of total pulse production of the country [7].
In spite of its various importance, the productivity of faba bean in Ethiopia remains far below the crop's potential (3.8t/ha), which has been constrained by several biotic and abiotic factors [8][9][10][11].Climate change (variability), edaphic factor, water stress, and colds are among major abiotic constraints of faba bean production.Moreover, climate variability due to increased temperature and erratic precipitation over time has increased susceptibility of faba bean which also favoured disease development [12].As a result of climate change different fungal pathogen evolved, seriously affecting the production of faba bean [13,14].Thus, climate change and associated factors of disease scenarios demand changes in crop and disease management strategies.
More than seventeen pathogens have been reported so far on faba bean from different parts of Ethiopia [11].Of them the most important yield limiting diseases are chocolate spot (Botrytis fabae), rust (Uromyces viciae-fabae), and black rot (Fusarium solani).Particularly, chocolate spot and faba bean Advances in Agriculture rust are the most serious and economically damaging diseases that affect the foliage, limit the photosynthetic activity, and reduce the productivity of faba bean, globally [15][16][17].Chocolate spot is a major limiting factor in the main faba bean growing regions of Ethiopia resulting with yield losses ranging from 34.1 to 61.2% [14,18].Faba bean rust is also a severe disease which causes about 70% of faba bean yield losses across the world in general [10] and in the Middle East, North Africa, Ethiopia, and parts of Australia in particular [16,19,20].In Ethiopian context there is rust devastating disease following chocolate spot, which causes 2 to 15% and 14-21% seed yield loss, respectively, in lower and intermediate altitudes [12].Together, rust and chocolate spot may cause total crop failure under severe epidemic conditions, resulting with more than 50% yield losses [16,21].
Sahile et al. [22] reported that different production practices such as planting date, weed management, cropping system, and previous crop can influence disease occurrence, epidemic development, and the damage caused to the crops.The disease severity can also be influenced by certain farming systems, field location, and/or season.Management options for chocolate spot and rust disease in Ethiopia include use of resistant cultivars, chemical control (Mancozeb), and late planting, but these options were neither widely disseminated nor adopted by end users (farmers) [8].
Information concerning the distribution and importance of chocolate spot and rust has not been systematically analyzed in major faba bean growing areas of the country; especially in the south eastern parts the information is totally lacking.Accordingly, as the diseases become a recurrent problem in the major faba bean production areas of the country, it is imperative to develop an effective integrated chocolate spot and rust management package, which is sustainable in agricultural yield development, knowledge of the relationships of the disease with different variables [23].Thus, the information obtained from survey work is necessary for planning management strategies; the present study aimed at investigating (i) the incidence and severity of chocolate spot and rust diseases and (ii) association of chocolate spot and rust with farmers cultural practices in the major faba bean growing areas of Bale Highlands, south eastern Ethiopia.

Materials and Methods
. .Survey Areas.The survey was conducted in eight districts of Bale highlands (i.e., Agarfa, Dinsho, Gasera, Goba, Gololcha, Goro, Ginir, and Sinana) during the main cropping season of 2017.The selection of the districts was based on faba bean production and productivity across south eastern Ethiopia as well as their accessibility to road.Mainly the districts were differed in altitude, weather, and soil characteristics (Table 1).The rainfall data of the surveyed districts were obtained from meteorological stations.A total of 280 faba bean fields were inspected to assess the magnitude of chocolate spot and rust disease.The time of survey was chosen so as to coincide with the time of flowering, podding, and late podding stages of faba bean, when the diseases reached its maximum [10,22].The districts surveyed were shown in Figure 1.
. .Sampling and Disease Assessments.Fields were sampled randomly at the interval of 5-10 km along roads.Depending on the topography and the relative importance of faba bean cultivation within each district the distances between fields were determined.In each sampled field, four quadrants (2m×2m) 10m away from each other were sampled by making diagonal moves with four throws in order to represent four replications per field.Following each throw, the number of healthy and infected (chocolate spot and rust) leaves of 12 plants within the area were recorded.A subsample of twelve leaves per plant and four leaves per canopy layers (upper, middle, and lower) were selected from the main stem, yielding a total of 144 leaves per field.Mean of canopy layers was determined per plant and then averaged per field for data analysis.Chocolate spot and rust incidence were rated as mean percentage of diseased plant leaves within the quadrant.Severity was rated on leaves from the same 12 representative plants in each quadrat, using a 1-9 scale [24], in which 1 indicates no visible symptom and 9 represents disease covering more than 80% of the foliar tissue.Disease severity scores were converted into percentage severity index (PSI) for the analysis [25,26].

=
Sum of numerical ratings × 100 No. of plants scored × Maximum score on scale (1) During the survey, cultural practices like soil type (vertisol, loam, and clay), cropping system (mixed or sole), weed management practices, type of fungicide applied, previous crop, planting date, and crop growth stages of every sampled field were recorded.
. .Data Analyses.Disease incidence and PSI were classified into distinct groups of binomial qualitative data.Thus, ≤75 and >75% were chosen for chocolate spot incidence yielding a binary variable and class boundaries of ≤30 and >30% were chosen for chocolate spot PSI data (Table 2).Class boundaries of ≤50% and >50% for rust incidence and ≤20 and >20% were chosen for rust PSI data.
Logistic regression analysis was used to determine the association of faba bean rust and chocolate spot incidence and PSI with independent variables (climate change resilient cultural practices) using SAS Procedure GENMOD [11,22,[27][28][29].Logit link function was used to determine the effects of climate change resilient cultural practices to the response variable.The importance of independent variables was evaluated three times for their effect on the response variable.At first, all the independent variables were tested with faba bean rust and chocolate spot incidence and PSI in a single variable model.Second, the association of an independent variable with disease incidence or PSI was tested when entered last into the model with all other independent variables.Third, those independent variables which showed  high significant association to the disease incidence and PSI were added in to a reduced multiple variable model.Furthermore, parameter estimates and standard error were analyzed in both single and multiple models.Odd ratio was obtained by exponentiating the parameter estimates for comparing variable classes based on reference point.Analysis of deviance and likelihood ratio statistics were used to compare single and multiple variable models.The difference between likelihood ratio tests was used to examine the importance of variables and tested against a Chi-square value.Data analysis and interpretation were done as described by Yuen [30].

Results
. .Characteristics Features of the Surveyed Fields.The altitude of surveyed fields ranged from 1805 in Goro district to 2600 m.a.s.l. in Goba district (Table 1).Aggregation of independent variables were used to analyze the survey of faba bean chocolate spot and rust into distinct variable classes within a number of corresponding fields for each class as presented in Table 2. Out of the total surveyed 280 faba bean fields, 81.43% of them were sole cropped, while 18.57% had been intercropped with wheat, barley, maize, and oat.Fields having poor weed management showed the highest incidence and severity of chocolate spot and rust compared to other respective variable classes.The weeds commonly observed during the survey were species of Trifolium, Rumex, grasses, and Solanum nigrum.Faba bean fields were at three growth stages during the survey, with 5.4% at flowering, 37.8 % at podding, and 56.8% at the late podding stages.About 37% of faba bean fields were sown in the month of July, while 19.6% and 43.2% were sown in the months of August and September, respectively (Table 2).
The highest chocolate spot mean disease incidence of 90.4% and PSI of 37.5% were recorded from fields planted in the month of July, while the lowest incidence (85.2%) and PSI (23.4%) were recorded in September (Table 3).On the other hand, the highest mean incidence (91.5%) and PSI (37.9%) were recorded from fields highly infested with weeds when compared with field managed in a good way.In some fields, faba bean was planted along with other crops such maize, barley, wheat, oats, potato, garlic, and onion.Consequently, a relatively lowest (87.5%) mean incidence of chocolate spot was recorded from such mixed fields, contrary to faba bean planted as sole cropping system with the highest mean incidence of 90.5%.The highest (89%) mean incidence of chocolate spot was recorded from fields planted with legume crop in the preceding year.The highest (36.2%) mean PSI was observed at late podding stage and the lowest (30.9%) at flowering stage.
The association of all independent variables with chocolate spot incidence and PSI is presented in Table 4.The independent variables such as district, planting date, previous crop, cropping system, and weed management practices were highly associated with chocolate spot incidence and PSI when entered into the logistic regression model as a single variable.However, when all variables entered last into the regression model, soil type, altitude, and growth stage showed significant association with chocolate spot incidence in addition to other variables.Among the independent variables, district, previous crop, altitude, growth stage, weed management, planting date, and cropping system were the most significantly important variables that are associated with chocolate spot incidence as a result of the reduced multiple variable model indicated.
The analysis of deviance for the variables added one by one to the reduced model showed the importance of each variable and variable class.The parameter estimates resulting from the reduced regression model and their standard error are given in Table 5.
High chocolate spot incidence has a high probability of association to Gasera district, when previous crop was legume, late podding stage, poor weed management practice, July sowing, and sole cropping system.In Gasera district, there was about twenty-six times greater probability of chocolate spot incidence observed that exceeds 75%.Again, among the independent variables, such as district, sowing date, previous crop, cropping system, and weed management were found as the most important variables in their association with chocolate spot PSI when entered first and last into the model with other variables (Table 4).A chocolate spot PSI of ≤30% showed high probability of association with cereals and vegetable as previous crop, faba bean planted late (September), mixed cropping system, good weed management practice, and low altitude (Table 6).

. . Rust Disease Intensity and Its Association with Climate
Change Resilient Cultural Practices.The highest mean incidence of faba bean rust was recorded in Gasera (78.2%),Goba (56.6%), and Ginir (56.5%), whereas the lowest was recorded in the districts of Goro (23.6%) (Table 3).The highest mean rust severity was recorded in Gasera (37.9%) and Dinsho (24.5%).The highest mean incidence (53.5%) and severity (26.4%) of rust were observed on fields sown in July while the minimum mean incidence (39.4%) and severity (11.2%) were in September.The highest (56.8%) mean incidence of faba bean rust was recorded from fields found at high altitude (>2500m a.s.l.) and the lowest (45.3%) was recorded from fields found at low altitude (<2200).
The highest mean incidence of faba bean rust was recorded from fields highly infested with weed (63.2%) than fields with less weed population.The highest mean incidence of rust (50.2%) was recorded from fields planted with legume crop from the previous year.Rust severity was found to increase with the advancement in growth stage, in which maximum mean rust severity was recorded at late podding stage (36.7%) and on fields sown faba bean alone (25.3%) compared to their respective other variable classes (Table 3).The independent variables varied in their association with rust incidence and PSI (Table 7).Of them, district, planting date, previous crop, growth stage, and weed management showed significant associations (P < 0.01) with rust incidence.These independent variables were tested in a reduced multiple-variable model.The parameter estimates resulted from the reduced regression model and their standard errors were presented in Table 8.A rust incidence of ≤50% showed a high probability of association with good weed management, late sowing (September), and rotation of faba bean with cereals and vegetables.Conversely, a high rust incidence was associated with poor weed management practices, early sowing, flowering stage, and continuous monocropping of faba bean (Table 8).However, the fertilizer, fungicides, soil types, and altitude were not significantly associated with the rust incidence.
Five variables, district, planting date, previous crop, altitude, growth stage, and weed management practices were associated with rust PSI when entered as single variable into the model (Table 7).The significance of district, previous crop, growth stage, weed management, and planting date were tested by adding these variables into a multiple reduced model (Table 9).Analysis of deviance for the variables, parameter estimates, and their standard errors is given in Table 9. Gasera, Dinsho, Agarfa, and Goba districts, fields sown in the month of July along with poor weed     management, had a high probability of association to high (>20%) rust PSI.There were about 150 times greater probabilities that rust PSI would exceed (>20%) in Gasera district than Sinana district and 19 times higher in the fields planted during the month of July compared to September.

Discussion
Faba bean chocolate spot and rust were found to be widely spread across all the studied districts and the incidence and severity were varied among altitude ranges, cropping systems, and other cultural practices.Furthermore, the higher total annual rainfall (878 mm-1170 mm) as well as the temperatures (5-32 ∘ C) of the area (Table 1 and Figure 2) has created favorable environmental condition for both diseases.When temperatures are mild (15-22 ∘ C) and relative humidity is high, chocolate spot can kill the plant [31], while rust is favored by warm temperatures and cloudy weather conditions [20].Therefore, chocolate spot and rust seriously damage faba bean and their spread which were strongly influenced by climatic conditions [32].Evident from the finding of this study is that intercropping faba bean with cereals (barley, wheat, maize, and oats) and vegetables (potato, garlic, and onion) appears to be an effective strategy to control chocolate spot and rust.These also concur with the results of previous researchers who found that intercropping reduced disease severity relative to monocrops in 79% of studies involving fungal pathogens, 72% of studies of virus diseases, 100% of bacterial studies, and 37% of nematode studies [33].The advantages of diversification may accrue to smallholders in the developing tropics, where intercropping is still widely practiced [34].Due to the high tillering capacity of cereal crops as compared to legume, intercropping cereals and vegetables with faba bean resulted with modifying the microclimate in favour of disease epidemic and reducing effective spore population by trapping.Besides, cereals could act as a physical barrier by intercepting spores carried by wind from neighbouring infected faba bean plants thereby reducing the amount of effective inoculum available to infect new tissue [35].Similar effects have also been reported in pea-cereal intercrops intercepting Mycosphaerella pinodes spores [36].Sahile et al. [9] have found chocolate spot reduction in faba bean-maize and faba bean-barley intercrops and not in a faba bean-pea intercrop.However, Dillon Weston [37] stated that faba bean intercropped with pea, barley, and oats was almost chocolate spot-free.Schoeny et al. [36] considered that reduction in host density was responsible to a large extent for reductions of M. pinodes dispersal in pea-barley intercrops, with a significant contribution of non-host plant by providing a physical barrier.Throughout the surveyed areas it was observed that plants were infected by chocolate spot and rust in poorly weeded fields.Thus, high weed population not only severely reduces the yield of the crop but also favours the development of disease epidemics (Agegnehu and Fessehaie 2006).This might be in fields with dense weed populations; there was competition for space, moisture, and soil nutrients as a result of which the faba bean plants were less vigorous and prone to the disease.Similarly, white rot of garlic [38] and root rot of faba bean [39] have been reported that high weed density in non-weeded fields increased disease severity.Besides, the presence of a high weed population in a field increases the humidity within the crop canopy (microclimate) which is more favorable for B. fabae infection and the development of chocolate spot disease epidemics [22].

Advances in Agriculture
Maximum chocolate spot and rust incidence were observed from fields sown legume crop in the previous year particularly in Gasera and Agarfa districts.The farmers of surveyed area produce faba bean crop continuously and the preceding crop residual appears to be a source of inoculum.This continuous growing of faba bean leads to accumulation of plant haulms in the field; as a result, there was an increase in the inoculum level, as well as rapid amplification of the disease.In cooler production regions, the uredospores are important means of survival between cropping seasons for faba bean rust (Shifa et al. 2011).Secondary spread is by means of uredospores, which can readily germinate on plant surfaces under humid conditions and dispersed by wind [20].In support of this, Zewude et al. (2007) reported that continued growing of garlic leads to accumulation of the white rot sclerotia in the soil that increases the occurrence of white rot.The importance of crop rotation is also affirmed by Belete et al. [39] who stated that faba bean cultivation is declining steadily due to root rot, as farmers substitute this crop with other legumes, mainly grass pea (Lathyrus sativus).
Planting date was also found as another factor influencing the incidence and severity of faba bean chocolate spot and rust in the study area.Accordingly, high mean disease incidence of chocolate spot was recorded from fields sown in July (early season).This particular period of time is characterized by high rainfall and humidity that favour leaf wetness which caused infection on plant leaf, due to the extended infection period of the pathogen in the air.Early infection produces typically dark-brown spots that do not pose a serious threat to the crop.However, when environmental conditions become more favourable for the development of the disease (that is, mild temperatures and high humidity) aggressive lesions may appear.In this case, a rapidly expanding necrosis takes place which may eventually lead to defoliation and death of the whole plant [31].Hence, in agreement with the literatures more severe epidemic occurred at Gasera, Agarfa, and Dinsho districts by the milder temperature favouring epidemic development.Similar results were also found by Belete et al. [39] that high root rot incidence occurred in the faba bean plants sown in early season.Likewise, Shifa et al. [1] showed high faba bean rust severity in fields sown between June and July, a time considered as an early sowing period in Hararghe highlands.Outside the Ethiopia context, Hawthorne et al. [40] have indicated that early sowing faba bean may result in a bulky crop, which creates an environment most conducive to rust disease in Australia.Therefore later sowing reduced the disease risk in faba bean fields.

Conclusions
This study indicated that chocolate spot and rust diseases were widely distributed and major challenging diseases in faba bean growing districts of Bale highlands.The survey data analyzed using logistic regression model identified climate change resilient cultural practices that are associated with rust and chocolate spot epidemics either singly or in combination.The present study has identified that intercropping of faba beans with cereals and vegetables reduced the impact of chocolate spot and rust development on faba bean.In addition, the study has identified the importance of sowing dates, previous crop, and weed management practices as climate change resilient cultural management components in chocolate spot and rust management strategies.Therefore, proper weeding practices, sowing at appropriate time, crop rotation with non-legume, mixed cropping system, and other related cultural practices should be carried out to reduce rust and chocolate spot impact on faba bean production and productivity until ecofriendly integrated management strategies are developed for the faba bean growing districts of Bale highlands.

Figure 1 :
Figure 1: Map of Ethiopia and the Bale highlands showing the surveyed districts.

Figure 2 :
Figure 2: Monthly rainfall in bars and monthly minimum and maximum temperature (curves) during main cropping season (July-November) in 2017 in eight surveyed districts in Bale highlands.

Table 1 :
Characteristic features of surveyed faba bean fields in eight districts of Bale highlands, south eastern Ethiopia.

Table 2 :
Independent variable by disease contingency table for logistic regression analysis of chocolate spot and rust during 2017 main cropping season in eight districts (n=280) of Bale highlands, south eastern Ethiopia.
a Percentage severity index.b Faba bean intercropped with cereals (wheat, mai, and oats) and vegetables (potato, garlic, and onion).

Table 3 :
Mean incidence and percentage severity index (PSI) of faba bean chocolate spot and rust for different independent variables during 2017 main cropping season in Bale highlands, south eastern Ethiopia.
a Percentage severity index.b Standard deviation.

Table 4 :
Independent variables used in logistic regression modeling of chocolate spot incidence and percentage severity index (PSI) and likelihood ratio test (LRT) for ten variables entered first and last into a model.
DF, degrees of freedom; LRT, likelihood ratio test; DR, deviance reduction; Pr, probability of 2 value exceeding the deviance reduction; VEF, variable entered first into model; VEL, variable entered last into model.

Table 5 :
Analysis of deviance, natural logarithms of odds ratio, and standard error of added variables in a reduced model analyzing chocolate spot incidence.
DF, degrees of freedom; LRT, likelihood ratio test; DR, deviance reduction; Pr, probability of 2 value exceeding the deviance reduction.* , reference group.a Variables are added into the model in order of presentation in table.b Unexplained variations after fitting the model.c Estimates are from the model with all independent variables added.d Standard error of the estimate.e Exponentiating the estimates.

Table 6 :
Analysis of deviance, natural logarithms of odds ratio, and standard error of added variables in a reduced model analyzing chocolate spot percentage severity index (PSI).
DF, degrees of freedom; LRT, likelihood ratio test; DR, deviance reduction; Pr, probability of 2 value exceeding the deviance reduction; * , reference group.a Variables are added into the model in order of presentation in table.b Unexplained variations after fitting the model.c Estimates are from the model with all independent variables added.d Standard error of the estimate.e Exponentiating the estimates.

Table 7 :
Independent variables used in logistic regression modeling of faba bean rust incidence and percentage severity index (PSI) and likelihood ratio test (LRT) for ten variables entered first and last into a model.
DF, degrees of freedom; LRT, likelihood ratio test; DR, deviance reduction; Pr, probability of 2 value exceeding the deviance reduction; VEF, variable entered first into model; VEL, variable entered last into model.

Table 8 :
Analysis of deviance, natural logarithms of odds ratio, and standard error of added variables in a reduced model analyzing faba bean rust incidence.
DF, degrees of freedom; LRT, likelihood ratio test; DR, deviance reduction; Pr, probability of 2 value exceeding the deviance reduction; * , reference group.a Variables are added into the model in order of presentation in table.b Unexplained variations after fitting the model.c Estimates are from the model with all independent variables added.d Standard error of the estimate.e Exponentiating the estimates.

Table 9 :
Analysis of deviance, natural logarithms of odds ratio, and standard error of added variables in a reduced model analyzing faba bean rust percentage severity index (PSI).
DF, degrees of freedom; LRT, likelihood ratio test; DR, deviance reduction; Pr, probability of 2 value exceeding the deviance reduction; * , reference group.a Variables are added into the model in order of presentation in table.b Unexplained variations after fitting the model.c Estimates are from the model with all independent variables added.d Standard error of the estimate.e Exponentiating the estimates.