*Corresponding author Introduction
Finger millet (Elucine coracana L. Gaertn), that belongs to the family Poaceae (Gramineae), is the third most widely cultivated millet after pearl millet (Pennisetum glaucum) and foxtail millet (Setaria italica) in the semiarid tropical and subtropical regions of the world (Reddy et al., 2009). It was native and first domesticated in Ethiopian highlands and Western Uganda at least 5,000 years ago and was introduced to India, Sri Lanka, and China approximately 3,000 years ago (Upadhyaya et al., 2006). Currently, the crop is cultivated across 25 countries, semiarid regions, and tropical regions, up to an altitude of 2,300 m (Joshi et al., 2023). Ethiopia is one of the major producers of finger millet in addition to Uganda, India, Nepal, and China, and it is also native to the highlands of the country (Ayalew, 2015). In Ethiopia, the crop is mainly grown in the northern, northwestern, and western parts of the country, especially during the main rainy season (Adugna et al., 2011).
Finger millet is an important staple food crop in traditional low-input cereal-based farming systems in East and Central frica, where it is used as a subsistence and food security crop (Semahegn et al., 2021). Finger millet is potentially a climate-resilient and nutritious crop with high nutraceutical and antioxidant properties (Chandra et al., 2016). Very importantly, finger millet grain is gluten-free, rich in calcium, fiber, and iron, has excellent malting qualities, and has a low glycemic index, and because of these properties, the crop is a choice food for diabetics (Arjun et al., 2014; Gebre, 2019). The status of finger millet is now changing from a neglected and underutilized crop to an emerging high-potential crop for health food and functional food products with high value (Kandel et al., 2019). Thus, finger millet plays the key role in the livelihood of smallholder farmers in the semiarid areas of Africa and Asia.
Finger millet is one of the important food crops in Ethiopia, and it is grown as a staple food grain in parts of Ethiopia where drought takes its highest loss on crop production (Erenso et al., 2009). The major attributes of finger millet are its adaptability to adverse agro ecological conditions with minimal inputs, tolerant to moisture stress, produced on marginal land where other crops cannot perform and tolerant to acidic soil and termite (Gebre, 2019). Therefore, finger millets represent one of the critical plant genetic resources for the agriculture and food security of poor farmers that inhabit arid, infertile and marginal lands. As a result of increased drought and soil fertility degradation, a growing number of farmers are resorting to finger millet, and thus the area allocated for this crop has significantly increased over the last 10 years (Erenso et al., 2009). It is an important crop in parts of Gojjam, Gonder, Wollega, Illubabur, Gamo-Gofa, Eastern Hararghe, and Tigray. It also becomes an important crop in parts of the Ethiopian central rift valley including Arsi-Negelle, Shashemene and Siraro Woredas (Anchala et al., 2006), and has a great genetic potential in other semiarid areas of the country like the Somali region.
In Ethiopia, the grain is processed to make unleavened bread (locally referred to as enjera) and for malting to prepare local drinks such as a distilled spirit ‘Areki’ or local beers such as ‘tella’ and non-alcoholic drinks such as ‘karibu’ and ‘shamita’, while the straw is vital as a livestock feed and for thatching of houses (Gebreyohannes et al., 2021).
The global total area and production of finger millet are not known because the data from pear millet, finger millet, foxtail millet, and other minor millet types are reported together (FAOSTAT, 2020; Gebreyohannes et al. [2021] as cited in Gebreyohannes et al., 2024). In Ethiopia, finger millet is the sixth important crop after tef, wheat, maize, sorghum, and barley (Anteneh et al., 2019). It is produced on 366,301.45 hectares of land with 1,038,539.49 tons of production and a yield of 2.8 tons/ha (CSA, 2022). Amhara Regional State of Ethiopia is the first in terms of production area dedicated for the crop, which was around 243,448.48 ha, with potential production of 697,645.18 tons and productivity of 2.9 tons during the 2022 cropping season (CSA, 2022).
Despite the importance of finger millet for food security and livelihoods, its productivity is relatively low (2.8 t/ha) in Ethiopia compared with the potential yield of the crop (5.5 t/ha) achieved under experimental conditions (Kasule et al., 2023). The low average yield of finger millet is attributed to several production constraints: biotic stresses (fungal diseases, insect pests, weeds, and bird damage) and abiotic stresses (lodging, moisture stress in dry areas, poor soil fertility and salinity), little research emphasis given to the crop, non-adoption of improved technologies, and poor attitude toward the crop (Kinfe et al., 2017; Wekesa et al., 2019). However, worldwide, blast disease caused by the fungus Magnaporthe grisea (teleomorph: Magnaporthe grisea) is the major biotic constraint that affects finger millet production and productivity, leading to low grain quality and poor yields (Mgonja et al., 2007). The pathogen is a seed-borne disease that damages the foliage, neck, and finger at different stages of crop growth. It hibernates in seeds and infected crop debris and occurs during the rainy season (Manandhar et al., 2016). Before the milking growth period, diseases known as neck blast and finger blast cause a significant decrease in grain productivity. In severe years, losses in excess of 80% have been documented in East Africa (Mgonja et al., 2013). The disease is causing 10-80% yield losses in both Kenya and Uganda and even complete crop loss of finger millet in Kenya (Odeph et al., 2020; Takan et al., 2004). The estimated yield losses due to finger millet blasts in Ethiopia have been recorded as 42% (Lule et al., 2014). The finger and neck blast in finger millet may be jointly called a head blast to avoid complications arisen by the two different names (Ghimire et al., 2022).
Ghimire et al. (2022) report that in favorable conditions with lower temperatures and higher humidity levels (>70%), the crop is more vulnerable to leaf, neck, and finger blast disease during the growth season. According to Gashaw et al. (2014), blast disease is an important factor preventing finger millet production from expanding further in East Africa. However, the intensity of finger millet blast may vary depending on the production approach to growing across different agro-ecological zones in terms of both amount and nature of distribution due to differences in cultural, biological, and physical environmental factors (Allorent and Savary, 2005; Mwang'ombe et al., 2007). It is important to understand the various agro-ecological factors that affect disease intensity and how these related factors cause epidemics in order to develop environmentally friendly disease management strategies (Aytenfsu et al., 2019; Rusuku et al., 1997; Zewde et al., 2007).
Despite the importance of the disease and its potential to cause significant yield losses in finger millet, the distribution, significance, and variables influencing the development of head blat of finger millet in Ethiopia have not been extensively studied. For instance, surveys of finger millet head blast in various Ethiopian finger millet-growing regions showed that the disease is becoming the main factor limiting Ethiopia's output of finger millet (Mulualem and Melak, 2013; Gebreyohannes et al., 2021). However, using a logistic regression model, the relationship between disease intensity and biophysical factors was not determined. Determining the most significant factors and focusing efforts on developing strategies for controlling it might be aided by understanding how the many independent variables relate to the intensity of the disease. Therefore, the objectives of this study were to assess the distribution and importance of blast disease in various finger millet growing areas of northwestern Ethiopia and to determine the association of disease intensity with biophysical factors.
Materials and Methods
Description of survey area
A survey of finger millet head blasts was conducted in the Amhara Regional National State's four districts, which include South Achefer, North Mecha, and Bahir Dar Zuria from the west Gojjam administrative zone, and Dera, which is located in the south Gondar administrative zone, during the 2020 and 2021 main cropping seasons (Fig. 1). Districts were purposively selected based on finger millet production potentials and constrained by head blast. The latitudes and longitudes of the survey areas ranged from 11 19‵58‵‵ to 11 47‵58‵‵ North and 36 55‵37‵‵ to 37 34‵10‵‵ East. The surveyed districts are different in agro-ecological characteristics and weather conditions and are located at an altitude ranged from 1,900 to 2,050 meter above sea level (m.a.s.l.) Monthly average temperatures, total rainfall, and relative humidity of the districts in the cropping seasons were obtained from the web page (Table 1) which extended June to October.
Fig. 1.
Map showed the four districts in Ethiopia's Amhara region where the finger millet blast disease was surveyed during the 2020 and 2021 cropping seasons.
Table 1.
Agroecological characteristics and weather conditions of surveyed districts in Northwestern Ethiopia during the 2020 and 2021 main cropping seasons
| Zone | Districta | No. of field | Altitude (m.a.s.l.)b | Rainfall (mm) | Temperature (°C) | Relative humidity (%) | |
|---|---|---|---|---|---|---|---|
| Min. | Max. | ||||||
| West Gojjam | Bahir Dar Zuria | 64 | 1,950-2,050 | 288.5 | 11.7 | 24.9 | 76.6 |
| North Mecha | 64 | 1,930-2,050 | 312.1 | 12.1 | 25.5 | 82.3 | |
| South Achefer | 64 | 1,950-2,050 | 410.9 | 10.1 | 23.0 | 84.4 | |
| South Gondar | Dera | 64 | 1,900-1,950 | 234.1 | 15.7 | 25.3 | 68.7 |
a District: average rainfall, temperature, and relative humidity ranges of each district were obtained from meteorological stations the web page https://power.larc.nasa.gov/data-access-viewer/.
Sample unit and disease assessment
Based on the level of finger miller production, four farmer associations from each district were selected through the use of purposeful sampling. Eight fields per season were selected from each farmer association using a systematic random sampling method spaced 5 to 10 km apart along and main and accessible rural roads. Thus, a total of 256 fields were assessed in both years at the pre-flowering to grain filling growth phases during the 2020 and 2021 the main cropping seasons.
Through direct inspection, the disease incidence and severity of each sampled field were determined using a 1×1 m quadrat. The quadrat was used to sample plants by moving through the field in an “X” pattern and making a predeter-mined number of equally spaced paces, depending on the field's size. Five throws of the quadrat, each representing five replications per field, were made on sampling points. Disease incidence, as defined by James (1974), is the proportion of diseased plants to the total number of plants evaluated.
According to Babu (2011), the symptoms manifest as characteristic spindle-shaped spots with gray or whitish centers and brown or reddish-brown margins on the leaf laminae. The neck develops as an elongated black lesion that is typically one to two inches below the ear, and the finger appears brown (Fig. 2).
Fig. 2.
Finger millet blast disease symptoms in finger millet farmer field. (A) Elliptical or diamond shaped lesions on leaves with grey or whitish center; (B) finger millet blast symptom starts at the tip and proceeds toward the base of the finger; (C) neck blast symptoms develop as elongated black color lesion mostly 1-2 inches below the ear; (D) finger millet blast disease affect the field severely.
Five randomly selected plants from each quadrant were assessed for head blast severity, which is defined as the extent of plant tissue damage by Kashyap et al. (2022). The finger millet head blast disease infection was evaluated in the field by using blast disease severity measures for neck and finger blast; observations were made following finger formation during the panicle development stage until maturity was reached using standardized 1 to 5 scales. A score of 1 shows that fewer than 2% of fingers are affected, a score of 2 indicates that 2.1-10% of fingers are impacted, a score of 3 suggests that 10.1-25% of fingers are affected, a score of 4 indicates that 25.1-50% of fingers are affected, and a score of 5 indicates that more than 50% of fingers are affected by finger blast infection. According to the relative size of the neck lesions, a score of 1 indicates the absence of lesions or only very tiny pinhead-sized ones; a score of 2 indicates lesions between 0.1 and 2.0 cm; a score of 3 indicates lesions between 2.1 and 4.0 cm; a score of 4 denotes lesions between 4.1 and 6.0 cm; and a score of 5 indicates lesions larger than 6 cm on the neck for neck blast disease (Babu et al., 2013). Then, the severity grades were converted into a percentage severity index (PSI) for analysis using the formula of Wheeler (1969) as follows:
During the survey periods, biophysical parameters like altitudinal ranges, sowing date, variety type, fertilizer usage, weed management level and previous crop history were recorded for each sampled field to determine their relationship with the disease incidence and severity. Data on geographical information (latitude, longitude, and altitude) of each field was recorded using a Global Positioning System, and information on cultural practices like previous crop history, variety type, fertilizer usage, weed management level, and sowing date were obtained from growers through interviews.
Data analysis
Simple descriptive statistics was performed and descriptive parameters were used to describe the distribution and prevalence of finger millet head blast incidence and severity. Disease incidence and severity were classified into distinct groups of binomial qualitative data as described by Adila et al. (2021) and Belete et al. (2013). Class boundaries were selected so that classes contained approximately equal numbers by considering the mean value of the incidence and severity (Endalew et al., 2021). Thus, ≤45% and >45% were chosen for blast incidence and ≤45% and >45% for blast severity, yielding binary dependent variables. Categorized independent variables that were used in the analysis are presented (Table 2). Contingency tables of disease intensity and the independent variables were built to represent the bivariate distribution of fields according to two data classifications. An entry in a cell of a contingency table represents the frequency of fields falling into that cell (Table 2).
Table 2.
Independent variables by disease contingency table for logistic regression analysis of head blast disease of finger millet during 2020 and 2021 cropping season
| Independent variable | Variable class | Total number of fields | Number of fields in different blast disease categories | |||
|---|---|---|---|---|---|---|
| Incidence | Severity | |||||
| ≤45 | >45 | ≤45 | >45 | |||
| District | Bahir Dar Zuria | 64 | 34 | 30 | 37 | 27 |
| Dera | 64 | 44 | 20 | 48 | 16 | |
| North Mecha | 64 | 32 | 32 | 36 | 28 | |
| South Achefer | 64 | 23 | 41 | 25 | 39 | |
| Altitude | <1,950 m.a.s.l. | 96 | 57 | 39 | 60 | 36 |
| 1,900-2,050 m.a.s.l. | 160 | 76 | 84 | 86 | 74 | |
| Variety | Local Deke | 124 | 43 | 81 | 47 | 77 |
| Improved Necho | 132 | 90 | 42 | 99 | 33 | |
| Land preparation | <5 times ploughing | 69 | 20 | 49 | 25 | 17 |
| >5 times ploughing | 187 | 113 | 74 | 121 | 34 | |
| Sowing datea | Early June 28 | 173 | 55 | 118 | 73 | 100 |
| Late June 28 | 83 | 78 | 5 | 73 | 10 | |
| N-fertilizerb | <Recommended rate | 122 | 80 | 42 | 87 | 35 |
| >Recommended rate | 88 | 19 | 69 | 25 | 63 | |
| Recommended rate | 46 | 32 | 14 | 34 | 12 | |
| Weed managementc | Good | 166 | 90 | 76 | 94 | 72 |
| Intermediate | 66 | 39 | 27 | 45 | 21 | |
| Poor | 24 | 4 | 20 | 7 | 17 | |
| Previous cropd | Finger millet | 26 | 8 | 18 | 12 | 14 |
| Maize | 200 | 110 | 90 | 118 | 82 | |
| Potato | 5 | 3 | 2 | 2 | 3 | |
| Teff | 12 | 5 | 7 | 4 | 8 | |
| Wheat | 13 | 7 | 6 | 10 | 3 | |
| Season | 2020 | 128 | 59 | 69 | 43 | 85 |
| 2021 | 128 | 74 | 54 | 67 | 61 | |
The association of finger millet head blast with the biophysical variables were analyzed using a logistic regression model (Fininsa and Yuen, 2001; Yuen, 2006); using SAS procedure of GENMOD (Statistical Analytical Software, 2014). Logistic regression calculates the probability of a given binary outcome (response) as a function of the independent variables (Cheek et al., 1990). The model allows the significance of multiple independent variables affecting the response variables to be assessed. Yuen (2006) states that the logistic regression model assumes that if the probability of the outcome is represented by the symbol (P), then the logarithm of the odds of P (P / [1 - P]), which equals logit (P), is a linear function of the independent variables. In this case, the binary outcome was the probability that finger millet blast incidence exceeds 45% and severity exceeds 40% in a finger millet field. The GENMOD procedure gives parameter estimates and the standard error of the parameter estimates. Exponentiation the parameter estimate yields the odds ratio, which is interpreted here as the relative risks (Yuen, 2006).
The importance of the independent variables was evaluated in three ways. First, the association of an independent variable alone with disease incidence or severity was tested. This consists of testing the deviance reduction (DR) attributed to a variable when it was first entered into the model. Second, the association of an independent variable with disease incidence or severity was tested when entered last into the model with all other independent variables. Third, variables with high association to disease intensity when entered first and last into a model were added to a reduced multiple variable model (Shifa et al., 2011). A complete analysis of deviance table was generated for the final reduced multiple variable models, where DR was calculated for each variable as it was added to the reduced model (Shifa et al., 2011). The odds ratio was obtained by exponentiation of the parameter estimates for comparing the effect based on reference point (Fininsa and Yuen, 2001; Yuen, 2006).
The deviance (-2×log likelihood) was used to compare single and multiple variable models. The difference between the two models, known as a likelihood ratio test, was used to examine the importance of the variable and was tested against a chi-square value (Cheek et al., 1990) where the number of degrees of freedom in the chi-square value corresponded to the difference in df between the two models.
Results
Features of finger millet surveyed field
Aggregation of independent variables was used to analyze the survey of finger millet head blast disease into four distinct variable classes within a number of corresponding fields for each class as presented in Table 2. Of the finger millet fields, about 62.5% were found in the altitudinal range 1,950-2,050 m.a.s.l., while the remaining 37.5% were found below 1,950 m.a.s.l. Regarding the date of sowing, the majority (67.6%) of finger millet surveyed fields were sown in early June 28 in the main cropping season, while the remaining fields (32.4%) were sown in late June 28. In all districts assessed, the finger millet varieties Deke (48.1%) and Necho (51.6%) were examined. Out of the 256 finger millet fields examined for the survey, 72.3% were ploughed more than five times, while the remainder fields were ploughed lesser than five times for the purpose of growing finger millet in each district's two cropping seasons (Table 2). During the survey, 35.99% of the fields received the recommended rate (100 kg/ha) of N-fertilizer; however, 23.11% and 40.9% of the fields received N-fertilizer rates that differed compared to the preceding fields’ rates. The fields that were treated with proper weeding were approximately 64.8% of the total, with the other 35.2% exhibiting different levels of weed infestation. The majority of the finger millet fields in each district were assessed during the period of 2 years (Table 2). During the survey, white to grey-green lesions or spots on leaves were observed as blast disease symptoms caused by Magnaporthe grisea. The initial symptoms of these lesions had dark green borders, while the older lesions had an elliptical or spindle-shaped shape with whitish to grey centers and a red to brownish or necrotic border (Fig. 2A). Similar to this, the symptoms of a neck blast develop as an elongated, black-colored lesion that is typically located 1 or 2 inches below the ear and reveals symptoms indicative of a serious ear head infection (Fig. 2B). The symptoms that followed confirmed those reported to be associated with rice blast disease (Patro et al., 2020). The most commonly observed weed species during the surveyed periods were Elusine indica, Galinsoga parviflora, Eragrostis aspera, Cynodon dactylon, Setcirui pitmila, Bidens spp., and Ageratum conyzoides L. Digitaria spp.
Distribution and intensity of finger millet blast disease
During both cropping seasons, finger millet blast was prevalent and widely distributed every finger millet growing fields in each district. The incidence and severity of finger millet blast varied among the examined districts and study seasons (Table 3).
Table 3.
Incidence and severity (mean±SE) of finger millet head blast for different independent variables across districts of Northwestern Ethiopia, during 2020 and 2021 main cropping season
The districts with the greatest mean disease incidence were South Achefer (58.8%) and North Mecha (54.9%), whereas the district with the least incidence was Dera (30.8%) in 2020. Furthermore, in 2020, the districts of Bahir Dar Zuria (45.6%) and South Achefer (54.7%) recorded the highest mean severity of blast, whereas the district of Dera (30.1%) had the lowest. In terms of altitudinal variance, fields at an altitude of 1,950-2,050 m.a.s.l. showed the highest mean blast incidence (52.2%) and severity (47.5%), compared to fields found at an altitude <1,950 m.a.s.l., which had the lowest mean blast intensity (Table 3).
The local Deke variety showed the greatest mean disease incidence (55.1%) and severity (50.2%) of blast, whereas the improved Necho variety recorded the least mean disease incidence (37.5%) and severity (36.7%) in the 2020 main cropping season. In regard to ploughing, fields that were ploughed less than five times with an ox-plow showed greater disease incidence (55.0%) and severity (47.7%) than fields that were ploughed more than five times (Table 3). According to the survey results, late June 28 showed the lowest blast incidence (24.1%) and severity (29.7%) as compared to early June 28 sowing. Comparing poorly managed finger millet fields to well-managed fields, the former revealed the highest mean disease incidence (66.0%) and severity (54.7). Throughout the 2020 cropping season, there were no variations in the mean disease incidence and severity among the variable classes among previously planted crops (Table 3). Closest trends regarding disease incidence and severity were observed during the 2021 main cropping season (Table 3).
The t-test analysis showed that highly significant difference on variable altitude, variety, land preparation and sowing date (P<0.001) for finger millet blast incidence (Table 4).
Table 4.
Contingency t-test analysis of mean disease incidence and severity of finger millet head blast in Northwestern Ethiopia, during the 2020 and 2021 main cropping seasons
There is significant difference in finger millet head blast disease incidence between; 1,900-2,550 and <1,900 m.a.s.l altitude, Deke and Necho varieties, early (<June 28) and late (after June 28) sowing, and less than five and more than five ploughing (t=4.16, P=0.0001; t=7.6, P=0.0001; t=13.7, P<0.0001; and t=3.9, P=0.0002, respectively).
Similarly, the t-test analysis revealed that significant difference in finger millet head blast disease severity between early (<June 28) and late (after June 28) sowing date (t=10.62, P<0.0001), Deke and Necho varieties (t=7.66, P<0.0001), 1,900-2,550 and <1,900 m.a.s.l altitude (t=3.23, P<0.0014) and less than five and more than five ploughing land preparation (t=3.23, P<0.0015). While, the variable season showed non-significant difference with blast incidence and severity at P-values of 0.596 and 0.048, respectively (Table 4).
Association of finger millet blast intensity with biophysical factors
The association of all independent variables with finger millet blast incidence and severity are presented in Table 5. The majority of the independent variables, such as district, altitude, varieties, field preparation, sowing date, fertilizer, and previous crop, were the most important variables in their very highly significant (P<0.0001) association with blast disease incidence when entered first into the logistic regression model (Table 5). The variables weed management (χ2=6.6, 2 df) and year (χ2=5.23, 1 df) revealed significant (P≤0.05) association with incidence in the model when the variable entered into the first model. Furthermore, district, varieties, sowing date, fertilizer, and previous crop showed a highly significant (P<0.0001) association with blast disease incidence when entered into the last model. A few variables, including weed management (χ2=12.9, 2 df), land preparation (χ2=13.0, 1 df), and altitude (χ2=11.5, 1 df), maintained their high significant association with blast disease incidence despite entering into the last model (Table 5). However, the year lost significance. Similarly, more than half of the independent variables such as district (χ2=341.7 and 231.4, 3 df), altitude (χ2=80.1 and 14.7, 1 df), varieties (χ2=612.2 and 290.5, 1 df), land preparation (χ2=39.8 and 2 df), sowing date (χ2=412.1 and 376.5, 1 df), fertilizer (χ2=160.8 and 153, 2 df), and year (χ2=39.09 and 90.21, 1 df) had highly significant (P<0.0001) association with blast disease severity when entered first and last into the model, except land preparation (χ2=2.82, 2 df), which lost its highly significant when entered into the last model (Table 5). Additionally, weed management practice and previous crop were also found to be significantly (P≤0.05) associated with blast disease severity when entered first and last into the model (Table 5).
Table 5.
Logistic regression model for head blast intensity and their likelihood ratio test on independent variables, Northwestern Ethiopia, during the 2020 and 2021 main cropping season
The parameter estimates resulting from the reduced regression model, their standard errors and odd ratio are presented in Table 6. Significantly (P<0.0001), the probability of a high (>45%) head blat disease incidence was associated with the Deke local variety, less than five times plough practice, early sowing, and more than 100 kg/ha N-fertilizer. There were about 1.87, 1.13, 2.25, and 2.05 times greater probabilities of a high incidence of the disease for local Deke varieties, early sowing, fewer than five times plough practice, and N fertilizer in excess of 100 kg ha-1, respectively, compared to the reference group of each variable class (Table 6). Conversely, low (<45%) disease incidence had highly significantly (P≤0.0001 associated with Dera districts, improved Necho variety, lower recommended rate (<100 kg ha-1) N-fertilizer, potato as preceding crop, good to intermediate weeding practice (Table 6).
Table 6.
Analysis of deviance, natural logarithms of odd ratio and standard error of head blast incidence and likelihood ratio test on independent variables, Northwestern Ethiopia, during the 2020 and 2021 main cropping season
| Variablea | Variable class | Residual devianceb | df | LRT | Estimate loge (OR)c | SE | Odds ratiod | |
|---|---|---|---|---|---|---|---|---|
| DR | Pr>χ2 | |||||||
| Intercept | 5,115.4 | 17.78 | 0.0001 | -0.4724 | 0.10 | 0.62 | ||
| Cropping season | 2020 | 5,110.1 | 1 | 1.02 | 0.3114 | -0.0287 | 0.03 | 0.97 |
| 2021 | -0e | |||||||
| District | Bahir Dar Zuria | 4,344.0 | 3 | 83.80 | 0.0001 | -0.4549 | 0.05 | 0.63 |
| Dera | 308.50 | 0.0001 | -1.2154 | 0.07 | 0.30 | |||
| North Mecha | 1.24 | 0.2655 | -0.0469 | 0.04 | 0.95 | |||
| South Achefer | -0e | 10.0 | ||||||
| Altitude | 1,900-2,250 | 4,223.1 | 2 | 11.50 | 0.0007 | -0.1893 | 0.06 | 0.83 |
| <1,900 | 10.0 | |||||||
| Variety | Local Deke | 3,196.0 | 1 | 418.00 | 0.0001 | -0.6285 | 0.03 | 1.87 |
| Improved Necho | -0e | 000. | 10.0 | |||||
| Land preparation | <5 times | 3,055.3 | 1 | 13.00 | 0.0003 | -0.1185 | 0.03 | 1.13 |
| >5 times | -0e | 10.0 | ||||||
| Sowing date | <June 28 | 2,027.0 | 1 | 621.40 | 0.0001 | -0.8093 | 0.03 | 2.25 |
| After June 28 | -0e | 10.0 | ||||||
| N-fertilizer | <Recommended | 1,644.7 | 29.76 | 0.0001 | -0.2100 | 0.04 | 1.23 | |
| >Recommended | 294.44 | 0.0001 | -0.7184 | 0.04 | 2.05 | |||
| Recommended | -0e | 10.0 | ||||||
| Weed management | Good | 2,020.4 | 2 | 9.010 | 0.0027 | -0.1559 | 0.05 | 0.86 |
| Intermediate | 12.89 | 0.0003 | -0.2048 | 0.06 | 0.81 | |||
| Poor | -0e | 10.0 | ||||||
| Previous crop | Finger millet | 1,613.0 | 0.64 | 0.4224 | -0.0622 | 0.08 | 0.94 | |
| Maize | 4.54 | 0.0331 | -0.1423 | 0.07 | 0.87 | |||
| Potato | 22.98 | 0.0001 | -0.5497 | 0.10 | 0.58 | |||
| Teff | 0.15 | 0.6955 | -0.0395 | 0.10 | 1.04 | |||
| Wheat | -0e | 10.0 | ||||||
The parameter estimate, standard error and odds ratio resulted from reduced regression model are shown in Table 7. The probability of high blast severity (>45%) was very highly and significantly (P≤0.0001) associated with the Deke local variety, plough practice (less than five times), early sowing, and more than 100 kg ha-1 N-fertilizer (Table 7). The Deke local variety, the sowing dates (early June 28), the use of less than five plough practices, applying more than 100 kg/ha of N fertilizer, and Teff as the previous crop increase the probability of high disease incidence (>45%) by 1.7, 1.7, 1.6, 1.5, and 1.3 times, respectively, in comparison to the reference group of each variable class (Table 7). On the other hand, the Dera district, the improved Necho variety, the greater than five times plowing technique, the late sowing (after June 28), and the recommended (100 kg ha-1) N fertilizer were associated with the probability of a low (<45%) blast severity (Table 7).
Table 7.
Analysis of deviance, natural logarithms of odd ratio and standard error of head blast severity and likelihood ratio test on independent variables, Northwestern Ethiopia, during the 2020 and 2022 main cropping season
| Variablea | Variable class | Residual devianceb | DF | LRT | Estimate loge (OR)c | SE | Odds ratiod | |
|---|---|---|---|---|---|---|---|---|
| DR | Pr>χ2 | |||||||
| Intercept | 2,645.51 | 13.73 | 0.0001 | -0.4252 | 0.10 | 0.65 | ||
| Cropping season | 2020 | 2,606.43 | 71.24 | 0.0001 | -0.2309 | 0.03 | 0.79 | |
| 2021 | -0e | 10.0 | ||||||
| District | Bahir Dar Zuria | 2,264.78 | 3 | 21.99 | 0.0001 | -0.2243 | 0.05 | 0.80 |
| Dera | 171.21 | 0.0001 | -0.8775 | 0.07 | 0.42 | |||
| North Mecha | 0.40 | 0.5258 | -0.0257 | 0.04 | 0.97 | |||
| South Achefer | -0e | 10.0 | ||||||
| Altitude | 1,900-2,250 | 2,184.63 | 2 | 14.67 | 0.0001 | -0.2073 | 0.05 | 0.81 |
| <1,900 | -0e | 10.0 | ||||||
| Variety | Local Deke | 1,572.42 | 1 | 288.36 | 0.0001 | -0.5052 | 0.03 | 1.66 |
| Improved Necho | -0e | 10.0 | ||||||
| Land preparation | <5 times | 1,532.60 | 1 | 2.82 | 0.0931 | -0.0536 | 0.14 | 1.71 |
| >5 times | -0e | 10.0 | ||||||
| Sowing date | <June 28 | 1,156.50 | 1 | 230.26 | 0.0001 | -0.4745 | 0.03 | 1.61 |
| After June 28 | -0e | 10.0 | ||||||
| N-fertilizer | <Recommended | 0,988.33 | 2 | 6.39 | 0.0001 | -0.0944 | 0.04 | 1.10 |
| >Recommended | 114.31 | 0.0001 | -0.4327 | 0.04 | 1.54 | |||
| Recommended | -0e | 10.0 | ||||||
| Weed management | Good | 1,149.08 | 2 | 0.04 | 0.8430 | -0.0098 | 0.05 | 0.99 |
| Intermediate | 3.39 | 0.0657 | -0.1001 | 0.05 | 0.90 | |||
| Poor | -0e | 00.0 | 10.0 | |||||
| Previous crop | Finger millet | 0,976.42 | 0.11 | 0.7398 | -0.0250 | 0.08 | 1.03 | |
| Maize | 0.17 | 0.6810 | -0.0265 | 0.07 | 1.03 | |||
| Potato | 1.35 | 0.2447 | -0.1296 | 0.11 | 0.88 | |||
| Teff | 6.78 | 0.0092 | -0.2525 | 0.10 | 1.29 | |||
| Wheat | -0e | 00.0 | 10.0 | |||||
Discussion
Finger millet head blast (Magnaporthe grisea) was found widely distributed in all four surveyed districts of Northwestern Ethiopia. Despite variations in the disease's intensity across cropping seasons and districts, this finding confirmed that finger millet production is significantly threatened in the studied areas, and an earlier study by Gashaw et al. (2014) also reported that the disease is prevalent and a major production challenge for finger millet in western and northeastern Ethiopia.
The higher disease prevalence in the area of study might be due to environmental variables which encourage the development of the disease. In this regard, high relative humidity, temperatures, and rainfall were all associated with the incidence and severity of finger millet head blast diseases. The highest finger millet head blast incidence and severity were recorded in South Achefer district (57.8% and 53.6%, respectively), compared with Dera (34.2 and 37.4%). This could be attributed to the more favorable weather conditions in South Achefer district, as shown in Table 1, which included a temperature range of 10.1°C to 23.0°C, high rainfall (410 mm), and high relative humidity (84.4%) for the development of disease. Similar findings were also reported by Castejón-Muñoz et al. (2007), Castejón Muñoz (2008), and Koutroubas et al. (2009), who noted that relative humidity of 90.5%, rainfall of 93.2 mm, and average temperatures of 26.7°C were the ideal environmental conditions for blast incidence and severity. According to Prasad et al. (2015), warmer temperature and during the season is responsible for lower disease incidence and further progress. Therefore, during the period of time when blast disease develops, mild temperatures and increased rainfall along with high humidity should be the most significant components.
The influence of altitudinal variation on the pressure of finger millet blast disease was significant in 2020 and 2021 in different area. The disease may spread more slowly even when it shows symptoms early because of the low relative humidity (68.7%) and lower altitudinal range of the Dera district below 1,950 m.a.s.l (Table 1). However, the survey's findings revealed that finger millet blast was more prevalent and widely distributed in the South Achefer, Bahir Dar Zuria, and parts of the North Mecha district, within the altitudinal range of 1,950-2,050 m.a.s.l. Therefore, the mild temperature (23.0°C), high relative humidity (84.4%), and plenty of rainfall in South Achefer might have contributed to the establishment and development of finger millet head blast. However, Amayo et al. (2020) found that lower mean incidences appeared at altitudes below 1,000 and above 1,100 m.a.s.l. that the mean blast incidence was higher in altitudes between 1,000 and 1,100 m.a.s.l. The previous altitude range may have contributed to the infection and subsequent development of the disease due to the coexistence of high relative humidity, rainfall, and other factors. According to Bitew et al. (2021), altitude could have a substantial influence on the progress and spread of the disease when combined with favorable environmental conditions, although it cannot significantly affect the disease's development on itself.
The results of the survey showed that the farmers‵ local finger millet variety (Deke) had a higher incidence and severity of head blast disease than the improved variety (Necho). Farmers save seeds for their own local cultivars from one harvest season to the next or share seeds between themselves. This increases the spread of the disease inoculum in the district during study because there is a deficiency of prior knowledge regarding the presence of primary inoculums in the seeds. According to Ghosh et al. (2013), local cultivars in India exhibited higher disease incidence and more severe symptoms when compared to improved varieties.
The highest mean disease incidence and severity was seen in finger millet fields with fewer than five ploughing, compared to more than five ploughing. Crop debris building in the top soil layers might be the cause of this, as it raised concern about an increase in disease and may promote the overwintering and survival of numerous pathogens (Walters, 2009). Findings by Raveloson et al. (2018) suggest that primary inoculum might not be present in rice residues that are plowed under soon after harvest.
Another factor that was found to affect the incidence and severity of finger millet head blast in the study area was the sowing date. Thus, it was found that fields sown early in the cropping season had a high mean disease incidence of head blast. An extended period with high rainfall and relative humidity causes plant leaves to get wet, which favors pathogen infection. Researchers Kumar and Shukla (2012) reported similar results, indicating that finger millet plants that were sown early in the season had a significant incidence and severity of blasts.
As compared to the low recommended N-rate (<100 kg ha-1) and recommended N-rate (100 kg ha-1), the above-recommended N-rate (>100 kg ha-1) revealed high mean blast incidence and severity throughout all studied areas. This could be the result of using too much nitrogen fertilizer, which could cause the plant cells to develop succulently and get more susceptible to Magnaporthe grisea pathogen infection. The abundance of nitrogen fertilizer influences the buildup of ammonia within plant cells and provides ideal development conditions for Pyricularia oryzae (Ahmad et al., 2011). It has been found by Long et al. (2000) that high nitrogen applications applied early in the season, during the vegetative growth stage of rice plant development, increase disease incidence and the overall lesion area of rice blast. Likewise, findings were obtained by Kapoor and Sood (2000), who reported that, among other factors influencing the incidence and severity of blast, the rate of nitrogen fertilization has been found to influence the disease severity to a significant degree. On the other hand, Snoeijers et al. (2000) found that low nitrogen also made disease more likely to occur in weak plants that lacked adequate defenses against disease.
When compared to finger millet fields that are well managed, significantly weed-infested fields in all areas assessed showed a 66.0% increase in disease incidence and a 54.7% increase in disease severity. Fields that were not adequately well weeded showed significantly greater disease intensity. This may be due to the fact that weed infestation increases air humidity and decreases air circulation over the fields/crop canopy, which increases leaf wetness and increases the possibility of pathogen development and spore germination (Getahun et al., 2022; Hadiza et al., 2022). A large weed infestation may also limit crop vigor and succulent growth by competing for resources (Sahile et al., 2008). This could promote the growth of a number of weed species, including wild Eleusine spp., Digitaria spp., and Setaria verticillata. Likewise, in the field, the disease could be supported by those weed species. Pathogen isolates from weeds have been shown by Takan et al. (2004) to be able to infect finger millet and contribute in the early stages of disease development.
In 2021 main cropping seasons, the finger millet field with finger millet as pre-crops had greater mean head blast incidence (55.8%) and severity (55.1%) with than the finger millet field with other pre-crops. It is thought that fields containing finger millet crops from the previous year served as reservoir sources of the target pathogen's spore-loaded inoculum, which would account for the higher mean incidence and severity noted in those fields. Based on the findings of Raveloson et al. (2018), the presence of infected rice residues on the soil surface had a favorable impact on the beginning of blast in the field during the initial phases of the epidemics. These findings show that the causes of these outbreaks can be related to previously infected rice residues. Similar results were reported by David et al. (2012), who found that spores produced as the primary inoculum on the overwintering tissues cause early seedling infections by penetrating the leaf tissues and germinating on leaves.
The results of the survey data analysis using logistic regression analysis indicated that a number of independent variables, including districts, altitude, variety, sowing date and weed management, were associated with the incidence and severity of finger millet head blast, either alone or in combination, and were highly significant for the development of finger millet head blast epidemics. Furthermore, the model analyzed the relative importance of the independent variable and showed how these variables influenced the epidemic's increase or decrease as a function of the independent variables. Therefore, the study suggested that late planting dates, the use of resistant varieties, weed management, and other related farming practices should be given priority when designing management options in order to minimize the effect of head blast on finger millet productivity in the study areas and other locations with similar agro-ecological settings. Furthermore, comprehensive and consistent surveys should be conducted to generate a significant amount of information and to identify those variables that are most important for high disease intensity in order to develop sustainable integrated management strategies that are designed for all finger millet-growing environments, including surveyed districts.
PDF Links
PubReader
ePub Link
Full text via DOI
Download Citation
Print
