Three participatory on-farm trials were conducted during the 2021 cropping season in three locations in the Kori and Bagruwa chiefdoms in the Moyamba district. The trial sites were located at Bai Largor (N 08˚16.528', W 012˚08.942') in the Kori chiefdom and Bassah (N 07˚54.057', W 012˚27.212') and Njala Kanima (N 07˚54.941', W 012˚28.253') in the Bagruwa chiefdom.

The experimental design was a randomized complete block design in three replications. The treatments were sole groundnut, sole cassava and cassava-groundnut intercropping. The area for each treatment plot was 300 m² (30 m × 10 m).

The land was brushed and the plant debris was cleared. The sole cassava and groundnut were planted on flat land at a plant spacing of 1 m × 0.75 m and 0.3 m × 0.2 m respectively. To reduce competition between the cassava and the groundnut in the cassava-groundnut intercrop, the spatial arrangement of the cassava was manipulated without compromising the plant densities used in the sole cropping systems, the plant spacing for the cassava in the intercropping system was 1.5 m × 0.5 m. The cassava and groundnut were planted simultaneously on the same land. Routine management practices were carried out throughout the trial period. The crop residue of the groundnut was returned to the cassava-groundnut intercrop plots after harvest.

Yield and yield components of groundnut

At full maturity (3 months after planting), the groundnut was harvested from a sample plot (200 m²) in each replication to determine:

Above groundnut biomass yield (kg∙ha⁻¹): The above-ground biomass cassava was calculated as follows:

Abovegroundbiomass = Weightofbiomassfromsampleplot ( kg ) × 10000   m 2 200   m 2 × 1   ha (1)

Average number of filled pods per plant: Twenty plants were randomly selected from the harvested plants and the number of filled pods for the twenty plants was counted. The average number of filled pods per plant was determined by dividing the total number of filled pods by twenty.

Fresh pod yield (kg∙ha⁻¹): The weight of total filled pods harvested from the 200 m² sample plot was determined and used to calculate the fresh pod yield per hectare as follows:

Freshpodyield = Weightoffilldpodsharvestedfromsampleplot ( kg ) × 10000   m 2 200   m 2 × 1   ha (2)

Yield and yield components of cassava

At full maturity (12 months after planting), the cassava was harvested from a sample plot (200 m²) in each replication to determine:

Above cassava biomass yield (kg∙ha⁻¹): The above-ground biomass cassava was calculated as follows:

Above-groundbiomass = Weightofbiomassfromsampleplot ( kg ) × 10000   m 2 200   m 2 × 1   ha (3)

Average number of roots per plant: Twenty plants were randomly selected from the harvested plants and the number of filled pods for the twenty plants was counted. The average number of roots per plant was determined by dividing the total number of roots by twenty.

Fresh root yield (kg∙ha⁻¹): The weight total of the fresh roots harvested from the 200 m² sample plot was determined and used to calculate the fresh pod yield per hectare as follows:

Freshpodyield = Weightoffilldpodsharvestedfromsampleplot ( kg ) × 10000   m 2 200   m 2 × 1   ha (4)

Land use efficiency

The land equivalent ratio (LER) was used as a criterion for measuring the land-use efficiency of intercropping cassava with groundnut when compared with the sole crop of cassava and groundnut. The LER was calculated as follows:

LER = Yieldofintercropgroundnut Yieldofsolegroundnut + Yieldofintercopcassava Yieldofsolecassava (5)

Profitability analysis

Enterprise budgeting technique was used to determine the profitability of the sole cassava, sole groundnut and cassava groundnut intercropping systems.

Totalrevenue (SLL∙ha⁻¹): The total revenue was estimated by multiplying the marketable yield of the harvested crops for each treatment plot by the unit price of the marketable yield at the farm gate.

Totalcost of production (SLL∙ha⁻¹): The total cost of production for each treatment plot is comprised of the sum of the cost of inputs and labour. The cost of input was determined by multiplying the number of planting materials planted per plot by the unit cost of the planting materials. The cost of labour was determined by multiplying the number of man-days for each activity by the unit cost per man-day.

Grossprofit (SLL∙ha⁻¹): The gross profit was determined by subtracting the total cost of production from the total revenue.

Benefit-costratio (BCR): The BCR of the various treatments was calculated as the ratio of total revenue over the total cost of production.

Sensitivity analysis

A sensitivity analysis was conducted on the BCR for six scenarios of yield loss in cassava and groundnut (Table 1) which may likely occur because of the negative effects of adverse weather events.

Soil samples were collected from the on-farm trial sites to quantify changes in soil organic carbon stock due to the treatment effect. The change in soil organic carbon stock at the harvest of the groundnut and cassava was monitored against a baseline (after land preparation) of the on-farm trials for the four treatments. The treatments were sole groundnut, sole cassava and cassava-groundnut intercrop. A minimum of three sampling strata (relatively homogenous strata) was identified. Within each homogeneous unit (stratum), at least five soil cores were collected at a depth of 20 cm to form a composite sample using an auger. The composite samples represented the total area of the strata. The composite soil samples were then placed in well-labelled sample bags and transported to the Analytical Services Laboratory of the Njala Agricultural Research Centre (NARC) laboratory for the determination of total carbon content. The soil bulk density was collected using a bulk density kit and separately placed into tightly fitted containers to prevent moisture loss.

The soil samples were transferred into soil sample containers with labels matching those from the field for the determination of soil bulk density and total carbon content.

Soil Bulk Density: The soil bulk density samples collected from the field were immediately weighed to determine the initial weight. They were then oven-dried to constant weight and weighed again. The results obtained were used to calculate the bulk density in g∙cm⁻³ [11].

Total Carbon: The soil samples collected were air-dried for three days, crushed and sieved to pass through the <2 mm mesh. Samples of 0.5 g of air-dried (2 mm sieved) soil were placed into a 500 mL conical flask. Potassium dichromate and sulphuric acid were added to digest the organic matter in the samples. Then, 50 mL of deionized water was added and allowed to cool before titrating with ferrous ammonium sulphate [12].

Soil Organic Carbon stock: The soil organic carbon stork was calculated as follows:

SOCstock = d × BD × SOC × CFst

where:

d = depth of horizon/depth class [m].

BD = bulk density [kg∙m⁻³].

SOC stock = soil organic carbon stock (kg∙m⁻²).

SOC = soil organic carbon constant of <2 mm sieve soil.

CFst = correction factor for stoniness ((1 − % stones)/100), including subtraction of gravel and stones.

The SOC stock in kg∙m² was converted to kg∙ha⁻¹ using the following formula;

SOC stock (kg∙ha⁻²) = SOC value (kg∙m⁻²) × 10,000.

The net change in soil organic carbon was determined by the following formula:

Δ SOC = SOCt 2 − SOCt 1

where:

ΔSOC = annual change in soil organic carbon stock (kg∙ha⁻¹).

SOCt₁ = Soil organic carbon stock before the on-farm trial (kg∙ha⁻¹).

SOCt₂ = Soil organic carbon stock at the end of the on-farm trial (kg∙ha⁻¹).

Analysis of variance was conducted to determine the effects of different treatments using a mixed linear model MIXED procedure of SAS [13]. In the mixed model analysis, the treatments were used as fixed factors and “replicate” within on-farm trials were considered as random factors. The effects of the various treatments were compared by computing least square means and standard errors of difference (SED) using the pdiff option in the MIXED procedure. The significance of the difference was evaluated at a 0.05 level of probability.

Groundnut

Intercropping groundnut with cassava had a significant influence (p < 0.05) on the aboveground biomass, the average number of filled pods per plant and the fresh pod yield of groundnut in all the locations (Table 2). In each location, the highest aboveground biomass, average number of filled pods per plant and fresh pod yield of groundnut were recorded in the sole groundnut cropping system. In comparison to the sole crop, the average aboveground biomass, the average number of filled pods per plant and fresh pod yield of the intercropped groundnut across the locations significantly decreased by 33%, 15% and 31% respectively. Previous studies have reported a decrease in groundnut yield when intercropped with cassava [14]. The decrease in the number of pods per plant

could be attributed to the decrease in biomass. A significant (p = 0.03) and positive correlation (r = 0.51) were found between the above-ground biomass and the number of pods per plant. The positive correlation indicates that a decrease in the aboveground biomass could lead to a decrease in the number of pods per plant. The low pod yield of groundnut in both the sole and intercrop groundnut could also be attributed to high incidence rosette diseases and the low amount of phosphorus (54.79 to 79.42 mg∙kg⁻¹) and nitrogen (0.21 to 0.31 g∙kg⁻¹) at the trial sites.

Cassava

Intercropping cassava with groundnut had no significant influence (p > 0.05) on the aboveground biomass, the average number of roots per plant and the fresh root yield of cassava in all the locations (Table 3). However, the sole plots of cassava had the highest aboveground biomass, number of roots per plant and fresh root yield in all the locations. A similar result of no significant difference in the aboveground biomass was found between the treatments of pure cassava and the relative treatments of cassava-groundnut intercrop [15]. The no significant influence on the above groundnut biomass of cassava may be due to the different growth habits of the two crops while cassava has erect growth and groundnut is low growing. Besides, the cassava plants in the intercropping systems were mulched with groundnut residues. This may have contributed to an increase in the total nitrogen in the soil after the decomposition of the groundnut residues thus enhancing the vegetative growth of the cassava plants in the intercropping system to some extent. The previous research findings have shown that groundnut

SED: Standard Error of Difference at 0.05 level of probability.

residues have the potential to contribute to total nitrogen stocks if retained in the field [10]. Even though intercropping had no significant influence on the aboveground biomass of cassava, the aboveground biomass for the sole cassava was higher than the intercropped cassava. Average across locations, intercropping cassava with groundnut decreased the aboveground biomass, the average number of roots per plant and fresh root yield of cassava by 17%, 11% and 17% respectively. The decrease could be associated with the competition for plant growth resources during the root initiation stage of the cassava i.e. during the first three months after planting. At this stage, the groundnut plants were still in the field and not harvested. Besides, the fresh root yield was significantly and positively correlated with the aboveground biomass (p = 0.0515, r = 0.47) and the number of roots per plant (p < 0.0001, r = 0.82). This result shows that higher values of aboveground biomass and number of roots per plant of the sole cassava resultantly led to higher values of fresh root yield obtained from the sole cassava when compared to the intercropped cassava.

Land use efficiency

The land equivalent ratio (LER) was favourable (LER > 1) for the cassava-groundnut intercropping systems in all the locations (Table 4). This result indicates a benefit of using intercropping compared to sole cropping in terms of land use efficiency. Previous studies on cassava groundnut intercropping have

reported similar results of favourable land use efficiency advantage of intercropping over sole cropping in cassava-groundnut-based intercropping systems [16] and [17].

Profitability analysis

Table 5 presents the profitability analysis of the sole cassava, sole groundnut and cassava-groundnut intercropping system for the different locations.

Totalrevenue: At each location, the cassava groundnut intercrop generated higher revenues when compared to the sole groundnut and sole cassava. In the cassava-groundnut intercrop, cassava accounted for 75% of total revenue even though it had a lower farm gate price (SLL 1300 kg⁻¹) when compared to the groundnut (SLL 6000 kg⁻¹). This shows that the kilograms of cassava sold were more than the groundnut. The revenue generated from the intercrop cassava was also higher than the sole cassava in each location even though the total root yield for the sole cassava was higher than the intercrop cassava. For the profitability analysis, the farmers separated the total yield of the cassava into tradable and non-tradable storage roots based on the root size. The non-tradable roots were not included in the profitability analysis. Average across locations, the tradable root size accounted for 96.4% and 79.9% of the total yield for the intercrop cassava and sole cassava respectively. The high percentage of tradable cassava roots was obtained from the intercropped cassava the cassava roots were relatively bigger than cassava roots of sole cassava. This could be attributed to the enhancement of soil fertility resulting from the decomposition of the groundnut residues that were mulched around the cassava plants of the intercropping system. Results from previous have shown that intercropping cassava with groundnut had a positive effect on the storage root size of cassava [18].

Totalcost of production: The total cost of production varies among the sole groundnut, sole cassava and cassava-groundnut intercrop. The highest cost of

Note: Fresh cassava roots were sold at a farm gate price of SLL 1300 per kg; Fresh groundnut pods were sold at a farm gate price of SLL 6000 per kg, 1 bundle of cassava stems (50 stems per bundle 1.4 m long); 1 bag of unshelled groundnut was bought at SLL 300,000 per bag, Labour was rated at SLL 15,000 per man day.

production was obtained from the cassava-groundnut intercropping system, which could be attributed to the combined cost of planting both the cassava and groundnut. The total cost of production for the intercropping system was 16.2% and 19.5% higher than the sole groundnut and sole cassava respectively. The cost of labour for land preparation, planting, weeding and harvesting of groundnut was rated at SLL 15,000 per man-day whilst harvesting of cassava was rated at SLL 20,000 per man-day. The total cost of labour differed for the different operations for each of the cropping systems and accounted for 96.4%, 92.3% and 90.7% of the total cost of production for the sole cassava, sole groundnut and cassava groundnut intercrop respectively.

Gross profit and benefit-cost ratio: The cassava-groundnut intercropping treatment had the highest gross profit and benefit-cost ratio even though it had the highest total cost of production. Even though intercropping cassava with groundnut incurred the highest cost of production per hectare the average benefit-cost ratio across the locations for the cassava groundnut intercrop increased by 123% and 13% when compared to the sole groundnut and sole cassava. This increase could be attributed to the higher revenue generated from the total sale of cassava roots and groundnut pods. However, the results of previous studies revealed that the cassava-groundnut intercropping systems were more profitable than the sole crops of cassava and groundnut [14]. The gross profit and benefit-cost ratios were also favourable for the sole cassava. On the contrary, the gross profit and benefit-cost ratios were however not favourable for the sole groundnut. This could be attributed to the lower revenue obtained from the sole groundnut plots resulting from the low yield of the groundnut which may have been caused by the high incidence of the rosette disease and low amount of phosphorus (54.79 to 79.42 mg∙kg⁻¹) and nitrogen (0.21 to 0.31 g∙kg⁻¹) at the trial sites.

Sensitivity Analysis

Figure 1 presents a sensitivity analysis of the effect of different scenarios of yield loss in cassava and groundnut on BCR for the sole cassava, sole groundnut and cassava-groundnut intercrop. In the event of adverse weather conditions (seasonal droughts or heavy rainfall which may cause erosion or flooding), which may cause a 40% yield loss in both the cassava and groundnut, the benefit-cost ratio would only be favourable for the cassava groundnut intercropping system. The BCR for the sole cassava could be favourable in the event of a 20% yield loss in cassava whilst the BCR for the sole groundnut was not favourable for none of the yield loss scenarios. This result confirms that the cassava-groundnut intercropping system is a climate-resilient crop production practice for increasing crop productivity within fragile cropping systems that are pruned to the adverse weather events and the insurgence of pests and diseases associated with the negative effects of climate change.

Table 6 presents the results of the analysis of the SOC stock for the uncropped adjacent site, sole groundnut, sole cassava and cassava groundnut treatments for each trial site location. In comparison to the baseline, the changes in SOC stock for the uncropped adjacent site were favourable at each location. On average

across the locations, SOC for the uncropped adjacent site increased by 18.4% and 31.2% at the harvest of the groundnut and cassava respectively. At each location, the changes in the SOC stock for the cassava-groundnut intercrop were only favourable at the harvest of the cassava. This could be attributed to the decomposition of the groundnut residues that were mulched with the intercrop cassava. Average across locations, the SOC stock for the cassava-groundnut intercrop at harvest of the cassava increased by 3.4% when compared to the baseline. Previous research studies have shown that groundnut residues have the potential of increasing SOC if retained in the field, conversely, subsequent removal of the groundnut residues could lead to a sizable loss of SOC from the field [10]. On the contrary, the changes in the SOC stock for the sole cassava and sole groundnut were not favourable at both the harvest of the groundnut and cassava, which shows that conventional tillage practice without residue management, could lead to the loss of SOC in the field [19].