Yield and Physiological Response of Common Bean in Three Semi-Commercial Farmers’ Fields in Limpopo Province

Common beans (Phaseolus vulgaris L) is one of the key grain legumes for human consumption. It is considered a cost-effective option for improving the diets of low-income consumers in developing countries who rely mainly on plant based protein. Common beans are rich in proteins and vital nutrients like vitamins, carbohydrates, fats and fibres (Garden-Robinson and McNeal, 2013). Its consumption in some African countries, can be as high as 60 kg per capita per year (Beebe et al., 2013) compared to a world average of 6.5 kg per capita per year (Joshi and Rao, 2017). In South Africa, the per capita consumption rate stood at 1.59 kg per year in 2016-17.
 
However, despite the importance of common bean in human nutrition, its production, in developing countries is far below consumption rates. As a result, common bean production has become commercial due to urbanisation and the globalisation of markets (Beebe et al., 2013). A DAFF (2015) report shows that in the period 2010 to 2014, common bean consumption rate in South Africa averaged around 150 000 tonnes per year, yet the production rate, is less than half of that requirement. To meet the nutritional needs of the populace, emerging farmers are getting more involved in the production.
 
Planting dates of common bean in South Africa are controlled by frost occurrence for those planted late in the season and by rains at harvesting for early plantings. In Limpopo, common bean is planted between February and March. The crop is normally grown under residual moisture plus irrigation. According to FAO (2016), the yield levels of common bean in Africa are still low (0.9 t/ha) when compared to the other regions e.g. Americas (1.0 t/ha) and Europe (2.5 t/ha). In addition, common bean yield in Africa is declining due to extension of the crop production into marginal areas with poorer soil fertility as well as into areas prone to droughts (Beebe et al., 2013).
 
Even though most emerging farmers in Limpopo province grow common bean under irrigation, the crop still experiences water stress on some farms due to poor irrigation management, the effects of which are reflected in the yield levels. Yield levels vary among famers due to several factors which extend to poor agronomic practices, seasonal water shortage and suboptimal resources. The need to reduce unproductive water losses and maintaining healthy, vigorously growing crops through optimized water use and agronomic management is paramount if common bean productivity is to be enhanced among these farmers. This study was established to understand the yield performance and water-related physiological response of irrigated common bean under three farmers’ field conditions in the Limpopo province.

Experimental sites
 
The study was carried out on semi-commercial farmers’  fields in Limpopo Province. Three farmers were identified in three locations: Dalmada, Masemola and Mokopane. All the three locations fall under BSh (Arid, steppe, hot arid) climate classification (Kottek et al., 2006). The average temperatures during the growing period (February to end of May) ranged from a minimum of 7.7°C observed on the 8^th of May to a maximum of 33.2°C observed on 15^th March. Very little to no rainfall was received throughout the growing period hence the crop depended mostly on irrigation. The relative weather conditions experienced around the study locations are shown in Fig 1.
 

 
The soils at Masemola and Mokopane were deep reddish brown light textured soils (loamy sand overlying sandy loam) while those at Dalmada were deep brownish sandy loam overlying sand clay loam. The soil pH at Masemola was 6.8; Mokopane 8.5 and Dalmada 6.9.
 
Management practices
 
All the three farms had access to adequate inputs from a local Seed and Agro-chemical supply company. The cultivar planted by all the farmers was OPS RS4. At Masemola, 4 hectares were planted between 26^th February and 5^th of March. Basal fertilizer was applied in the form of a compound fertilizer (13:10:7; NPK) at a rate of 200kg per hectare (banded in rows). Green sulphur was also applied as top dressing at a rate of 200kg/ha along with kelp fertilizer which contain trace minerals. All the three farmers had access to the required inputs as advised.
 
At Mokopane, two hectares were planted on the 20^th of February. The fertilizers applied at Mokopane were similar to those at Masemola. At Dalmada the planting was done on a one hectare plot on the 2^nd of March. Basal fertilizer was applied in the form of compound fertilizer (2:3:4; NPK) at a rate of 300kg/ha. Green sulphur and kelp were also applied as under Mokopane.
 
 
A planting density of 167 000 plants was targeted at all the three locations with inter-row spacing of 0.75m and in row spacing of 0.08m. Water application was through drip irrigation. Weeds were removed mechanically through hand hoeing.
 
Agronomic data
 
Biomass was collected when 50% of the plants had flowered. The biomass was collected by harvesting plants in a 0.5m row length. The plants were cut at the soil surface, oven dried at 65°C until constant weight. Samples were randomly collected at four different positions in each hectare. Grain yield was collected from 1m row length and as with biomass this was also randomly done at four different positions in each plot.
 
Leaf gas exchanges
 
Leaf gas exchanges were measured using an LCi-SD Photosynthesis System (BioScience, UK). The measurements were taken between 11h00 and 14h00 on clear sunny days. At Mokopane and Masemola the measurements were taken on four different dates i.e. 18, 32, 45 and 60 days after planting (DAP). At Dalmada measurements were only taken on 45 and 60 DAP. A fully grown mature leaf was targeted for gas exchange measurements. Instantaneous WUE (InsWUE) and Intrinsic WUE (IntrWUE) were calculated using the following equations:

 
  and       

 
Where A is photosynthetic rate, E is transpiration rate and gs is stomatal conductance.
 
Chlorophyll content and Leaf Area Index
 
Chlorophyll content was measured using two devices: 1) a portable chlorophyll content meter (CCM-200, Opti-Sciences, USA) and chlorophyll meter (Soil–Plant Analyses Development, SPAD-502 Plus, Konika Minolta). Normalised Difference Vegetation Index (NDVI) was measured using the Green Seeker Handheld Crop Sensor (Trimble, USA). Leaf area index (LAI) was measured using AccuPAR LP-80 Ceptometer (Decagon Devices, Inc.). These measurements were also taken four times over the season. All measurements were taken between 10h00 and 14h00.
 
Statistical analyses
 
All data were analysed using the SPSS 25.0 statistical package (SPSS, USA). Where significant differences were observed, means were compared using the Tukey’s honestly significant difference (HSD) test at 5% probability level.

Above ground dry biomass and grain yield
 
Above ground biomass and grain yield were both significantly influenced by location. The above ground biomass was higher at Mokopane (3 694 kg/ha) followed by Dalmada (2 800 kg/ha) and the least biomass was obtained at Masemola with 1 557 kg/ha (Fig 2). A similar trend was also observed with grain yield where the yield at Masemola was the lowest at 363 kg/ha. These results showed that despite the availability of resources to the three emerging farmers, the productivity levels at the three farms differed significantly. The grain yield at Masemola was less than 30% of the yield achieved by the other two farmers. This could be attributed to a combination of factors, mainly the chilling damage observed during pod filling as well as poor management practices, particularly irrigation management.
 

 
When tropical crops are subjected to low temperatures, photosynthetic activities related enzymes are affected leading to low growth rate or even cessation of growth (Charrier et al., 2015). The severity of the damage depends on the crop health status and the soil moisture (Niwas and Khichar, 2016). It was observed during data collection, even before the chilling damage, that the crop was experiencing some water stress. Weed management was also poor, probably due to lack of labour and the relatively larger area planted (4 ha) compared to the other farmers, compromising the crop nutrition status. The combination of a drying soil and the generally poor health status of the crop might therefore have contributed to extent of the chilling damage observed at Masemola. The generally poor crop management at Masemola was evidenced by the low biomass observed at flowering (Fig 2). Challenges observed at Masemola are similar to those reported for common bean farmers of Maharashtra in India (Jaybhay et al., 2018). However, the grain yield observed in this study particularly at Dalmada and Mokopane were within range of yields observed in the country. South Africa dry bean yields are between 1.5 – 3.0 tons/ha. The yields are nevertheless obtained through nitrogen fertiliser applications of upto 50kgN/ha.
 
Effect of location and DAP on NDVI, SPAD value and LAI
 
To understand the variation in biomass and grain yield observed between the locations, some critical physiological parameters such as the SPAD value, NDVI, LAI and gas exchanges were determined. NDVI is a measure of the crop’s vigour while the SPAD value measures the greenness of the plant and is closely related to chlorophyll content. NDVI showed no differences between Dalmada and Mokopane even though Mokopane showed a tendency to have relatively higher values. The SPAD value showed significant variation between the locations at 45 and 75 DAP. At 45 DAP, SPAD value was higher at Masemola compared to the other locations but 30 days later the SPAD value at Masemola had the lowest value. On the other hand, the chlorophyll content, did not show any variation between the locations at 45 and 60 DAP but only at 75DAP. At 75 DAP, Mokopane had twice as high chlorophyll content compared to those observed at the other two locations. The findings of this study showed interesting results in that NDVI values at Masemola were significantly lower compared to those at Mokopane while the reverse was true for the SPAD value where the values were higher at Masemola. Since both NDVI and the SPAD value are a reflection of the chlorophyll content (Indradewa et al., 2019), these contrasting results are very unusual and not expected. The measured chlorophyll content did not provide much information to explain the observed contrast in NDVI and the SPAD value since there were no significant differences between the locations until after the chilling damage at 70 DAP (Fig 3).
 

 
Leaf area index on the other hand varied significantly at 45, 60 and 75 DAP. Masemola had significantly lower LAI throughout when compared to the two locations. The highest LAI was observed at Dalmada at 45 DAP, but at the later stages of growth, the highest was recorded at Mokopane. The findings suggest that the plant growth was more vigorous at Mokopane which agrees with the higher biomass observed in Fig 2. Therefore, based on the observed grain yield and the biomass at these two locations, it seems that LAI and NDVI showed a better reflection of the yield compared to the SPAD value.
 
 
Bivariate correlation between NDVI and SPAD value showed that the two had positive relationships which were only significant at Masemola. The correlation analysis also showed that chlorophyll content was positively related to both but was again only significant at Masemola. In addition, it was observed that biomass and grain yield were negatively related to SPAD and chlorophyll content but showed a strong positive relationship with NDVI, E and gs (Table 2). On the other hand, both InsWUE and IntrWUE showed negative relationships with Ci, E and gs but were more strongly related to Ci, with R-values of -0.82 and -0.90 respectively (Table 3). InsWUE did not show any significant relationship with SPAD, NDVI, Chlorophyll and LAI while IntrWUE only showed a significant negative relationship with NDVI.
 

 

 
The effect of DAP on SPAD value, NDVI and LAI was only analysed for Mokopane and Masemola due to missing data at Dalmada. Masemola, differed significantly with other locations hence Fig 4 shows the comparison between Masemola and Mokopane. The SPAD value did not vary much with time from 18 DAP to 60 DAP at each location but still showed differences between the two locations (Fig 4). There was a significant reduction in the SPAD value at 75 DAP which was more abrupt at Masemola. The SPAD value at Masemola plummeted by almost 30 SPAD value units from 55 to 27 compared to a 10 unit drop observed at Mokopane. LAI index on the other generally increased with DAP at both locations until 60 DAP before reducing at 75 DAP. Unlike the SPAD value, LAI was higher at Mokopane compared to at Masemola. The LAI index at Mokopane ranged from 1.8 to as high as 4.8 compared to values of 0.6 to 2.0 observed at Masemola. NDVI showed a similar trend to that of LAI. NDVI values at Mokopane were consistently higher than those at Masemola. Fig 4 shows that NDVI increased with DAP, peaking at 45 DAP before decreasing again. The variation in NDVI of dry bean plants over the season could be described by the quadratic equation on Fig 4.
 

         
Seasonal variation in leaf gas exchanges with location and DAP
 
In addition to the three parameters discussed above, leaf gas exchanges were also measured (Table 1 and Fig 5). Firstly, the effect of location on leaf gas exchanges was assessed at different DAP (Table 1). There was an inclination for intercellular CO₂ concentration to be higher at 45 DAP when compared to at 60 DAP (Table 1). The values were also relatively higher at Mokopane compared to the other two locations. Dalmada had the highest photosynthesis rate (11.65 μmol m^-2 s^-1) at 45 DAP compared to other locations. However, at 60 DAP the highest photosynthesis rate was observed at Mokopane with 9.41 μmol m^-2 s^-1. Stomatal conuctance was higher at Mokopane at both 45 and 60 DAP but Masemola and Dalmada did not show any difference. Transpiration rate was also significantly higher at Mokopane compared to the other locations at 60 DAP. Water use efficiency (WUE) measured at leaf level also showed significant differences between the locations. Both instantaneous and intrinsic WUE were lower at Mokopane compared to Masemola and Dalmada which did not differ between themselves. 
 

 

 
Secondly, the effect of DAP on leaf gas exchanges were assessed (Fig 5). Intercellular CO₂ concentration was relatively low at Masemola compared to Mokopane. Also, A fluctuated more at Mokopane compared to at Masemola. Intercellular CO₂ concentration (Ci) at Mokopane ranged from 253 to 409 ppm while it ranged from 159 to 281 ppm at Masemola. Despite Masemola having relatively low Ci, it had higher photosynthesis rate at 18 and 32 DAP which however dramatically declined from 10.38 to 4.48 μmol m^-2 s^-1. Transpiration rate was also lower at Masemola and constantly decreased with time from 18 to 60 DAP while the transpiration rate at Mokopane initially decreased before increasing again at 60 DAP. Stomatal conductance (gs) at Masemola also declined with time as observed with E but the gs at Mokopane fluctuated with DAP.
 
The fluctuation of gs at Mokopane can be attributed to variations in soil moisture levels, probably due to some measurements being taken soon after irrigation events while others may have been taken several days after irrigation. Such variations in soil moisture can be mirrored in leaf water status, particularly in isohydric plants (Igarashi et al., 2015). Many studies have shown that gs responds to variations in soil moisture level (Serret et al., 2018, Munjonji et al., 2017). These previous findings suggest that at Masemola, the crop might have been subjected to continuous soil drying thus resulting in continuous decrease in gs. When roots are subjected to a drying soil they produce abscisic acid which in turn signal the closure of the stomata leading to reduction in gs, and subsequently E and A (Saradadevi et al., 2014, Osakabe et al., 2014). This is also supported by the observed data in Fig 5.
 
The lower rates of E, gs and Ci observed at Masemola resulted in higher instantaneous WUE (InsWUE) and intrinsic WUE (IntrWUE) (Fig 6). Initially, reduction in gs reduces E more than it reduces A resulting in higher water use efficiency at leaf level i.e. IntrWUE and InsWUE (Nobel, 2009). In Fig 6, it was observed that both IntrWUE and InsWUE were significantly higher and increased with time (i.e. with DAP) at Masemola compared to Mokopane despite gs at Masemola always decreasing. However, as the soil continued to dry out and crop water stress became more severe, both E and A were reduced drastically due to wilting of the leaves. The results showed that higher WUE at leaf level does not necessarily result in better yields but only proves that crops tend to use water more efficiently under limited water supply compared to when well-watered. Although this is not completely new finding in plant sciences, it however emphasises the need for deficit irrigation in drier areas where water supply is limited and hence the need to efficiently utilise the little water available.
 

In conclusion, this study reveals that irrigation water management is critical for emerging farmers if higher yields are to be attained. Whilst all the locations experienced similar chilling conditions, more damage was observed at Masemola due to poor irrigation management leading to significant losses in grain yield and income as well. The study also showed that NDVI is of great value to farmers and can be used to monitor crop condition throughout the season.

The work was supported by Risk and Vulnerability Science Centre at the University of Limpopo. The authors would also like to thank the farmers and research assistants involved in this study.