About Bioline  All Journals  Testimonials  Membership  News

African Crop Science Journal
African Crop Science Society
ISSN: 1021-9730 EISSN: 2072-6589
Vol. 6, Num. 1, 1998, pp. 9-18
African Crop Science Journal,Vol. 6. No. 1, pp. 9-18, 1998



Alemaya University of Agriculture, P.O. Box 138, Dire Dawa, Ethiopia
^1 CIAT, Kawanda A.R.I., P.O. Box 6247, Kampala, Uganda
^2 Department of Biological and Agric. Engineering. Georgia Stn., University of Georgia, Griffin, GA 30223

(Received 15 August, 1997; accepted 17 December, 1997)

Code Number:CS98002
Sizes of Files:
      Text: 23.7K
      Graphics: Line drawings and tables (gif) - 69.1K


Bean (Phaseolus vulgaris L.) is an important food and cash crop grown in diverse environmental settings in Ethiopia. Its production is very heterogenous in terms of ecology, cropping system and yield. This study analyses the agroclimatic resources of 18 representative bean growing sites in Ethiopia and assesses the potential yield and moisture deficit stress using DRYBEAN (DSSAT V.3) growth model. Annual rainfall of various locations ranged from 580 mm (Mekele) to 1995 mm (Gore). Seasonal rainfall varied from 270 mm (Babile) to 1650 mm (Gore). The length of the growing period is from only 80 days (Mekele) to more than 220 days (Jimma). Among the agroclimatic variables, annual rainfall, seasonal rainfall, length of the growing period and altitude (temperature) are important factors to cluster the bean growing regions into 3 major and 6 minor homogeneous groups in agronomic sense for strategic planning. Simulated yield potential varied from 1.6 t ha^-1 (Jijiga) to 3.3 t ha^-1 (Bako). Planting date has a significant impact on simulated yield in the sub-humid and semi-arid clusters. Yield losses for each day of delayed planting after the effective onset of rainfall reached up to 60 kg ha^-1 day^-1. Moisture deficit stress was found to be an important limiting factor in the semi-arid and moderately limiting in the sub-humid regions. The analysis established that bean improvement work should focus on the development of high-yielding, long-maturing genotypes for multiple cropping systems in the humid regions. In the semi-arid regions emphasis should be on the development of drought tolerant and early maturing cultivars which fit well with efficient soil and water conservation practices.

Key Words: Moisture deficit stress, Phaseolus vulgaris, resources, sowing date, yield potential


Le haricot (Phaeolus vulgaris L.) est une culture alimentaire et commerciale cultivee dans divers environments cadres en Ethiopie. Sa production est tres heterogene du point de vue ecologique, systeme de culture et le rendement. La presente etude analyse des resources agroclimatiques de 18 sites representatifs de production du haricot en Ethiopie et evalue le rendement potentiel et la tension de deficit hydrique en utilisant le modele de croissance "DRYBEAN (DSSAT V.3). La precipitation annuelle de diverses locations s'est rangee de 580 mm (Mekele) a 1005 mm (Gore) tandis que la precipitation saisonniere a varie de 270 mm (Babile) a 1650 mm (Gore). La longueur de la periode de croissance est seulement de 80 jours (Mekele) a plus de 220 jours (Jima). Parmi le variables agroclimatiques la precipitation annuelle, laprecipitation saisonniere, la longueur de la periode de croissance et l'altitude (temperature) sont de facteurs importants pour grouper les regions de production du haricot dans trois principaux groupes et six petits groupes homogenes dans le sens agronomique pour la plannification strategique. Le rendement potentiel simule variait de 1.6 tonnes/ha (Jijiga) a 3.3 tonnes/ha (Bako). La date de semis a un impact significatif sur le rendement simule dans les groupes sous-humides et semi-arides. Les pertes de rendement pour chaque jour de retard de semis apres l'approche effective de precipitatio atteignait 60 kg/hajour. La tension de deficit hydrique etait trouve etre un facteur limitatif important dans les regions semi-arides et moderement limitant dans les regions sous-humides. L'analyse a etablie que le travail d'amelioration du haricot devrait se concentrer sur le developpement des genotypes a haut rendement, a maturite tardive pour les systemes multiples de culture dans les regions humides. Dans les regions semi-arides, l'accent devrait etre place sur le developpement des varietes tolerantes a secheresse et des cultivars a maturite precoce qui conviennent bien avec les pratiques de conservation efficace de sol et d'eau.

Mots Cles: Tension de deficit hydrique, Phaseolus vulgaris, resources, date de semis, rendement potentiel


Bean (Phaseolus vulgaris) is an important cash crop and protein source of farmers in many parts of Ethiopia. It is primarily a crop of small-scale producers and generally few inputs are used (Wortmann and Allen, 1994). Bean is compatible with numerous other crops in mixed cropping systems and is grown in diverse agroecologies (Fig. 1). The major cropping systems include beans inter-cropped with maize, sorghum, root crops or bean grown in sole crop in the Rift Valley.

The national average yield of haricot bean is only 0.8-0.9 t ha^-1 under peasants farming condition (CSA, 1994). The low yields are attributed to various limitations including biotic constraints, low soil fertility, unpredictability of rainfall and poor management. In most bean growing regions rainfall is erratic in its distribution and the soil is often sandy and with low moisture holding capacity (Simane and Struik, 1993). Therefore, the key factor affecting the productivity are the length and distribution of moisture during the crop growing season. Development and impleme-ntation of plans for sustainable production require an understanding of the effect of the combined action of these factors. Although these were known, no serious attempts have been made to quantify climatic risks in Ethiopia.

Agroclimatic and crop growth models are useful tools for evaluating alternative management strategies and developing plans and policies for achieving maximum resource use efficiency. They are useful for investigation of long term weather risks.The definition of a dependable starting date and duration of the growing season, and the quantification of dry periods during the growing season represent major challenges in sustainable dryland farming areas (Stern et al., 1994).

Different growth models are used to: i) match crops, varieties and management options to specific weather, soil and farming situations; ii) assess sustainability of cropping systems in specific areas; iii) plan agricultural development on a regional scale; and iv) evaluate uncertainness and risk at field and farm levels. The goal of such systems is to improve the performance of decision makers while reducing the time and human resources required for analysing complex decisions and field experiments. Wafulu (1995) used the crop simulation model CMKEN, a locally adapted version of CERES-maize, to evaluate farmers' decisions with respect to management options and the inherent economic implications in Kenya. Thornton et al. (1995) also demonstrated the use of CERES-maize model to provide information concerning management options such as the timing and quantity of nitrogen fertilizer application and to quantify the weather related risks of maize production in Malawi. This paper presents a discussion on the application of agroclimatology and a dynamic crop growth model to define the agroecology of bean production in Ethiopia.

The main objectives of the study were to quantify the physical resource bases of haricot bean growing areas and cluster them for strategic planning, and to assess potential yields and possible cropping systems for sustainable production.


A total of 18 major bean growing locations distributed throughout Ethiopia were assessed for their agro-ecological characteristics (Table 1). Selections were made based upon completeness of required weather data.

Agroclimatic analysis of the locations to identify major weather constraints was done using INSTAT statistical packages (Statistical Service Centre, University of Reading, UK). In view of rainfall variability, all agronomic events, such as start of the season, were defined using the concept of dependable values, which is the minimum period that can be expected in three out of four years. The start of the season was defined as the first occasion when the available soil moisture content exceeds 40% of field capacity and 10 days running total rainfall exceeds 0.5 potential evapo-transpiration. The end of the growing season was defined to be when available soil water content drops below 10% of field capacity.

Correlation analysis was made to identify the most important climatic variables to bean production. The variables used are longitude, latitude, altitude, annual rainfall, modality of rainfall distribution, main season rainfall, length of growing season, maximum temperature and minimum temperature. Altitude, longitude, annual rainfall amount and distribution, seasonal rainfall and length of the growing period are used to classify bean growing regions into homogeneous ecoregions.

The dynamic crop growth DRYBEAN model (Hoogenboom et al., 1994), a process oriented model of DSSAT v3, was used to test management alternatives for different scenarios in bean growing regions. The model allows the quantitative determination of growth and yield of bean with crop growth simulated with a daily time step on the basis of physiological processes. The soil water balance component simulates surface run-off, evapotranspiration, drainage and water extracted by the plant. Four different varieties representing different growth habits and phenologies were used to estimate potential yields for each cluster group (Table 2). Based on the out come of the agroclimatic analysis and research recommendations, three planting dates (July 1, July 15 and August 1) were used to determine the effect of planting date. Simulations started two weeks before sowing. Six sites, one from each cluster, were selected to represent the different agroclimatic conditions. Selected sites were judged to be representative of the bean growing region in Ethiopia.

The importance of stress at a location is determined as the sum of the products of frequency of occurrence and severity of stress with severity grouped as < 0.10 (no stress), 0.1-0.3 (mild stress), > 0.3 (stressful environment). Stress severity (index) is calculated as the ratio of available soil moisture (extracted water) to total plant water requirement during a specified growth stage, where 0 represents no stress and 1.0 extreme moisture stress.

All data were subjected to ANOVA, using the MSTAT software (Michigan State University, 1991). LSD values were used to determine the significance levels.


The agroecology of bean growing locations in Ethiopia is diverse (Table 1). Altitude ranges from 1200 to 2212 m, while annual rainfall varies from 580 to 1950 mm. The length of the growing period is from as short as 80 to 220 days. Accordingly the seasonal rainfall during the growing period ranges from 120 to 1636 mm.

Annual rainfall, main season rainfall, length and end of growing period are negatively correlated with longitude (r= -0.73**, -0.81**, -0.78**, -0.63** and -0.48*, respectively) (Table 3). The modality of rainfall distribution is positively correlated with longitude (0.66*). Altitude is negatively correlated with only maximum and minimum temperatures (r=-0.63** and -0.86**, respectively). Annual rainfall was found to correlate positively with main season rainfall and length of the growing period (r=0.96** and 0.77**, respectively).

Based on the result of the correlation analysis latitude, altitude, annual rainfall and its modality, seasonal rainfall and length of the growing period were found to be the most important variables that affect distribution and productivity of beans. Using these variables, bean growing regions in Ethiopia are grouped into three major (humid, sub-humid and semi-arid) groups which are sub-divided based on temperature resource (Table 4).

The annual rainfall distribution of all the sites is highly variable. The pattern of rainfall distribution is mono-modal in all the regions with the exception of the eastern highlands due to the influences of Inter-Tropical Convergence Zone. The length of the season lasts for 120-220, 100-150 and less than 100 days for humid, sub-humid and semi-arid regions, respectively (Table 4). In the eastern highland (sub-humid high altitude) region the short (Belg) rainy season extends from April to June and receives about 25%, whereas the long rainy season (Meher) extends from July to October and receives about 45% of the annual rainfall.

A simulation of more than 12 years of historical weather data was run using DRYBEAN crop growth model at three different planting dates for the four varieties. Simulated grain yield shows significant differences among sites and planting dates, while the difference among varieties is not significant (Fig. 2). Mean yields over sites ranged from 1.6 t ha^-1 (Jijiga) to 3.3 t ha^-1 (Bako). The average long term yield potentials in decreasing order are 3.3 t ha^-1 in humid low altitude, 2.3 t ha^-1 in sub-humid low altitude, 2.2 t ha^-1 in humid high altitude, 2.2 t ha^-1 in sub-humid high altitude, 1.7 t ha^-1 in semi-arid high altitude and 1.6 t ha^-1 in semi-arid low altitude.

Date of planting was found critical in the sub-humid and semi-arid regions (Fig. 2). The analysis established that delayed planting after July 1, or July 15 resulted in yield losses for low rainfall northern, eastern and central Rift Valley regions. Yield losses are about 35, 42, 62 and 1.5 kg of grain yield per day delay in planting in Alemaya, Nazareth, Mekele and Jijiga, respectively. However, different planting dates between July and August gave similar simulated yields for both Bako and Nekemte due to their long length of the growing season.

Early varieties flower in a mean of 42 days, while the late ones flower in 54 days (Table 2). Number of days to physiological maturity is 76 days for early maturing and 90 days for late maturing varieties. Different planting dates do not have any effect on the phenology.

Analysis of variance to evaluate the importance of moisture deficit stress revealed that there was a significant difference among sites, cultivars and planting dates (Table 5, Table 6 and Table 7). Moisture stress deficit, using the average of the products of the frequency of occurrence and the severity of the stress was greatest during the seed filling stage (S3) in all cases. Among the sites, moisture deficit stress was most important in the semi-arid regions (Jijiga and Mekele) and sub-humid regions (Nazareth and Alemaya) (Table 5). However, in the humid regions bean production is not limited by moisture deficit stress. The indeterminate long maturing variety experienced the greatest stress during the grain filling stage? (Table 6). The latest planting date was most affected by moisture deficit stress during the grain filling stage (Table 7). Correlation coefficients for moisture deficit stress levels with yield are generally highest for seed filling stage (S3, r=-0.73**), intermediate during mid season ( S2, r=-0.59**) and non significant during early stage (S1, r=-0.28).


The bean growing ecosystems of Ethiopia are numerous and highly diverse. Their potential for production and their management requirements are determined by the interplay of many factors, including climate, soil type and a range of socioe-conomic and biological factors. Development of appropriate technologies requires a good understanding of constraints and different opportunities of the bean growing ecosystems. Such understanding is needed for the identification of problems, setting of research priorities and targeting of technology that is agroecosystem specific.

Among the climate variables studied in the bean growing region, precipitation, altitude (temperature) and soil type are found variable in the space dimension. Rainfall is the most important factor influencing the choice of crops and management practices in the semi-arid tropics. Other studies (Singh and Byerlee, 1990; Simane and Struik, 1993) have also clearly demonstrated that rainfall is the predominant factor influencing yield variability and thus the major climatic factor affecting crop production in dry-land agriculture. Other factors, temperature, evapotranspiration, wind speed and radiation are fairly stable and ranged within the optimum limits of haricot beans requirements (data not presented). The thermal characteristics are typical of the tropics. However, cloudiness in the western and low temperature in the eastern highlands are sometimes important yield limiting factors.

Amount and mode of rainfall distribution decreases along the west-east dimension. This agrees with the previous findings of Simane et al. (in press). Rainfall distribution in the eastern part, in the longitude zone between approximately 39.5 degrees E to 42.5 degrees E, is effectively bimodal as a consequence of the movements of the Inter-Tropical Convergence Zone. The western, southern and northern regions have mono modal rainfall pattern. The total length of the growing season ranged from 80 days in the north-east to 200 days in the west humid region. However, released varieties of bean need only 90 to 110 days and under-utilise the rainfall in the humid and sub-humid regions when grown in sole crop.

In Ethiopia, the choice bean cultivars and the cropping systems are not well designed using the inventory of the physical resources. For proper placement of cultivars and choice of associated management practices, clustering of bean growing regions into homogenous environmental groups based on natural resources is believed to be a stepping stone for sustainable strategic planning. Accordingly, the bean growing regions in Ethiopia are classified into three major regions viz: humid, sub-humid and semi-arid which are sub-divided based on temperature (altitude). This result disagrees with the previous classifications made by Wortmann and Allen, (1994), where they classified the bean areas into only southern, eastern (above and below 1500 m), central Rift Valley and western regions .

The humid (western) cluster is basically characterised by very high rainfall with monomodal distribution, long growing season and ecologically high production potential. However, in terms of rainfall use efficiency, growing beans in these environments as a sole crop may not be economically optimal. Rather inter-cropping of indeterminate bean types or relay cropping or double cropping using determinate types should be considered. Other negative effects, such as waterlogging and cloudiness should be considered to maximise resource use efficiency.

The sub-humid regions are more optimal in terms of rainfall use efficiency for bean production. In these environments terminal moisture stress is often a problem with late planting. The moisture availability period in the eastern region shows that two alternatives are possible: single cropping and double cropping. In the first case, the two rainy seasons are considered as one growing season. Therefore, late maturing crops/cultivars of sorghum and maize could be planted in April and harvested in November, but before frost occurs. In the second option, early maturing haricot beans could be grown in the small-rainy season. Crops with intermediate length of growing season (e.g., maize cv. Katumani, wheat or barley) may be grown during the long-rainy seasons.

In the semi-arid regions potential yields are low and variable due to moisture deficit stress. In the northern, central Rift Valley and much of the eastern region, development of terminal drought tolerance with determinate growth habit should be the primary focus of breeding programmes. Date of planting is critical in the dry sub-humid and semi-arid regions of the country and as a result there is always a yield loss associated with a delay in planting after the onset of rainfall.

Economic, political and social factors, however, have to be taken into account for detailed recommendations about how best to respond to the assessment in terms of land preparation, cultivars, cropping system, seeding and initial fertilisation rates, and time of seeding for the different agroecological regions.


This project was supported financially by CIAT/ East Africa, and by Alemaya University of Agriculture. We gratefully acknowledge Dr. Tekalign Mamo for reviewing the manuscript.


FAO. 1986. Early agrometeorological crop yield assessment. FAO Plant Production and protection paper-73.

Hoogenboom, G., Jeffrey, W.W., James, W.J. and Kenneth, J.B. 1994. BEANGRO: A Process-Oriented Dry Bean Model with a Versatile User Interface. Agronomy Journal 89:182-190.

Michigan State University 1991. User's Guide to MSTAT-C. Michigan State University, East Lansing.

Simane, B and Struik, P.C. 1993. Agroclimatic analysis: a tool for planning sustainable durum wheat (Triticum turgidum var. durum). Agriculture, Ecosystem and Environment 47: 36-47.

Singh, A.H. and Byerlee, D. 1990. Relative variability in wheat yields across countries and over time. Journal of Agricultural Economics 41:23-32.

Stern, R.D., Dennett, M.D. and Dale, I.C. 1982. Analysing daily rainfall measurements to give agronomically useful results. II. A modelling approach. Experimental Agriculture 18:237-253.

Thornton, P.K., Saka, A.K., Singh, U., Kumwenda, J.D.T., Brink, J.E. and Dents, J.B. 1995. Application of a maize crop simulation model in the central region of Malawi. Experimental Agriculture 31:213-226.

Wafalu, B.M. 1995.Application of crop simulation in agricultural extension and research in Kenya. Agricultural Systems 49:399-412.

Wortmann, C.W. and Allen, D.J. 1994. African bean production environments: their definitions, characteristics and constraints. Network on Bean Research in Africa, Occasional Paper Series No. 11, Dar es Salaam, Tanzania.

Copyright 1998, African Crop Science Society

The following images related to this document are available:

Line drawing images

[cs98002b.gif] [cs98002g.gif] [cs98002c.gif] [cs98002e.gif] [cs98002d.gif] [cs98002i.gif] [cs98002f.gif] [cs98002h.gif] [cs98002a.gif]
Home Faq Resources Email Bioline
© Bioline International, 1989 - 2024, Site last up-dated on 01-Sep-2022.
Site created and maintained by the Reference Center on Environmental Information, CRIA, Brazil
System hosted by the Google Cloud Platform, GCP, Brazil