New Research Projects

Canola yield loss calculator: App

Prof F Meyer
Bureau for Food and Agricultural Policy (BFAP)

Background

During the 2018 canola harvest season, Canola yield loss trials were initiated by the Protein Research Foundation (PRF) in collaboration with SOILL in the Southern Cape and Swartland regions. Harvest losses between 152 kg/ha and 606 kg/ha were recorded, which implied R806/ha to R3 212/ha based on an average canola price of R5 300/tonne in the 2018 season. These were substantial losses and in order to help farmers curb and limit these losses (through harvester speed and settings) the harvest losses per farm or field had to be measured before and after changes to harvester speed and settings. The PRF and SOILL together with a few collaborating farmers developed a methodology by which farmers can determine their yield losses themselves (see Appendix A summarising the methodology).

In order to roll this out to more farmers, they approached the Oilseeds Advisory Committee (OAC) for assistance in appointing service providers to develop a user-friendly Mobile Application (App). The OAC appointed BFAP to develop an App Customisation for a first-time roll-out in the 2020 season with a dedicated group of farmers or app-users. This app is designed to feed georeferenced data into a database which can be used in various reporting outcomes and can also be shared with stakeholders effectively. The results of the 2020 harvest season are presented in the sections below.

Findings

The roll-out of the app was undertaken in collaboration with 11 agents from three agrochemical companies who visited 26 farms in 10 districts across the Western Cape. In total 33 fields were visited, and 50 yield loss measurements were recorded. Yield losses ranged between 0 and 350 kg/ha, implying that monetary losses up to R2 219.54/ha were incurred based on the average canola price of R6 341.53. Although these are still significant losses, they are lower than the ones recorded in the 2018 season, which ranged between 152 kg/ha and 606 kg/ha. It is also interesting to note that the 2020 season produced an all-time record harvest. According to the fifth production forecast of the Crop Estimates Committee (CEC) that was released on 18th December 2020, a total of 74 120 hectares of canola were planted and 163 256 tonnes were harvested, resulting in an average record yield of 2.2 tonnes/ha.

District analysis

Figure 2 below illustrates the average recorded harvest losses per district and production region (Southern Cape, Swartland and Overberg). The largest average losses per district (close to 150 kg/ha) were recorded in the Caledon and Napier districts, amounting to R950/ha losses. The worst performers in the Swartland region were the Durbanville and Moorreesburg districts with 82 kg/ha (R523/ha) and 87 kg/ha (R550/ha) yield losses respectively.

Cultivar analysis

A list of the newest cultivars was provided as part of the app form and Figure 3 presents the average harvest losses per cultivar. The gathered data suggests that the top performing cultivars include 45Y93 and 44Y87 Clearfield cultivars as well as the Quartz Conventional cultivar; these cultivars recorded 0 kg/ha, 10 kg/ha and 25 kg/ha (R158.54/ha) losses respectively. Given the limited number of observations per cultivar in the data collected (only one or two observations for some cultivars) these cannot be interpreted as conclusive results: repetitions over multiple seasons are required to identify consistently top-performing cultivars.

Equipment analysis

From Figure 4 and Table 1 it can be concluded that, for the points and farms sampled, Case harvesters registered the largest harvest losses with an average of 117kg/ha (R744.07/ha) while the New Holland and Claas harvesters recorded the lowest harvest losses, especially in the Malmesbury and Caledon districts. John Deere harvesters were most frequently used in the sample for this analyses.

Financial impacts

Figure 5 illustrates the effect that the 2020 average yield loss (78kg/ha), 2020 maximum yield loss (350kg/ha) and the maximum yield loss measured in 2018 (606kg/ha) can have on the canola gross margin (R/ha), relative to the baseline of no yield loss, per region. The baseline gross margin is based on an average realised canola price of R5 500/tonne. In the average yield loss scenario during the 2020 season, a farmer's gross margin was R467 per hectare lower than the baseline. For the historic maximum yield loss of 606kg/ha, Overberg farmers would have only made a R56 profit per hectare, and R1 026 in the Swartland.

Figure 6 depicts how much profit a farm would lose, given a total area canola planted (100ha, 200ha or 300ha) as a result of the various levels of yield losses. A farm cultivating 200ha of land, will lose R94 500, given the 2020 average yield loss of 78kg/ha, while the same farm, would lose just over R400 000, given the 2020 maximum yield loss of 606kg/ha.

Appendix A – Harvest Loss Methodology

The harvest losses were measured as follows:

• A standard size ice cream container was placed between the wheels of the harvester during the harvesting process (at the time of the study, the size of the ice cream container was measured at 279.2cm², see photos below).
• All leaves and sticks were carefully removed from the container after the losses were captured.

The harvest losses was calculated on a volumetric basis as follows:

The following harvest losses were measured (in weight): 0.5g, 1.0g and 2.0g (see photos 1,2 and 3 to see how these volumes look in a standard sized ice cream container).

Area of the ice cream container = length x width

Now, the relation of the size of the ice cream container to 1m² is calculated as follows:
Relation = 1 m² ÷ 0,04 m²
= 25

Therefore, the mass of seed collected in the container, needs to be multiplied by 25 to determine the seed collected per 1m².

The losses were thus calculated as follows:

0,5 g X 25 = 12,5 g/m². A hectare contains 10 000 m², therefore 12,5 g/m² = 125 kg/ha

1 g X 25 = 25 g/m² or 250 kg/ha

2 g X 25 = 50 g/m² or 500 kg/ha

At an average price of R5000/ton, a 2g collected in the container represents a loss of R2500/ha.

The following table (Table 1) can be utilised to translate harvest losses on a volumetric basis: Table 1: Measurement of harvest losses on a volumetric basis.

Use Table 1 as follows:

Calculate your harvester's CF relation in m² (correction factor = cut width divided by "kaf strook wydte" – width of row falling out of harvester after passing):

CF = cut width ÷ "kaf strook wydte"

If CF = 5 (see example marked in Table 1), the next steps are as follows:

Determine your harvest loss by collecting seed in the ice cream container and use the measurement silinder to measure the volume of seed collected. Then mulitply the volume of seed measured from the container with 25 to obtain the total volume of seed per 1m². If the losses collected were measured at 56 ml/m², that translates to 0.08t/ha or R424/ha at an average price of R5000/ton (see blue markings in Table 1).

Determine the drivers and impact of sunflower seed quality in South Africa

Prof F Meyer
Bureau for Food and Agricultural Policy (BFAP)

In 2014, the Bureau for Food and Agricultural Policy (BFAP) undertook a comprehensive sunflower value chain analysis for the Oilseed Advisory Committee (OAC). In the report, the importance of the oil content of the seed to ensure the economic sustainability of the industry was clearly articulated. Reference was also made to international markets, where price premiums are paid for sunflower seed with a higher oil content. Over the past season, the drastic decline in the oil content of sunflower seed has had a major adverse impact on the industry. Whereas the industry norm was typically the delivery of sunflower seed with an oil content of approximately 36% and higher, the oil content has dropped significantly, putting crushing margins and the overall competitiveness of the industry under pressure. In the light of this, BFAP has undertaken a study to provide a deeper understanding of the key drivers affecting sunflower seed quality, the impact of varying (lower) oil content on the industry, and potential interventions to achieve an upgraded state of the value chain.

In South Africa, yields and oil yield (t/ha) per hectare have been relatively stagnant over the past decade, managing to increase by an average annual growth of only 1% compared to maize yields, which have increased by an average annual growth of 2.5% over the same period. However, the oil content (%) of the seed simultaneously has decreased by an average annual growth of -0.6%. The oil content of sunflower seed produced in South Africa is also significantly lower than that produced in Argentina or the Balkan countries. Furthermore, cultivar trial data from international seed breeders shows that the same cultivar planted in South Africa yields a lower oil content than when planted under optimal conditions in Argentina. This research has clearly illustrated that seed yield and oil content are complex and quantitative traits that are not only controlled by many genes, but are also influenced to a great extent by environmental conditions, and both additive and non-additive genetic effects play an important role in the inheritance of seed yield and oil content.

Ample evidence exists that planting dates that are too early or too late have a negative effect on both the yield and oil content of sunflower seed, and the late plantings common in South Africa, where sunflower is mostly planted after the maize crop, may negatively influence the sunflower seed's oil production potential. The 2018 sunflower crop seems to be a typical illustration of this phenomenon, where exceptionally late plantings (in some cases as late as the first week in February) coincided with a sharp drop in the oil content of the delivered crop. In fact, all of the past five production season, except for 2017, had a late start, with the majority of plantings only commencing in late December and January in the major sunflower production areas. Sunflower seed production is the predominant cash crop in Argentina and the Balkan countries, whilst in South Africa it is maize. In South Africa, sunflower is often planted as a "catch crop" and preference is not given to the timing of production, such as optimal planting date, fertiliser applications, soil analysis or much of the required pest weed or disease programmes required for optimal production.

Thus, to achieve both high yields and a sufficiently high oil content as inherent in the genetic potential of the sunflower seed, the crop should be planted in areas with suitable climatic conditions (temperature and water stress) with optimal crop management (planting date, planting density, fertilisation). Currently, South African producers have a choice of 199 different sunflower seed varieties. Of these, 113 are classified as high oil hybrids with the potential of achieving higher than 40% oil content. All the 113 cultivars are GMO free, but some have speciality trials such as specific herbicide resistance (e.g. Clearfield or Clearfield Plus) or have specific fatty acid compositions (e.g. High Oleic).

From a seed breeding and plant sciences theoretical perspective, there should be a negative correlation between yield potential and oil content. However, based on a range of analyses undertaken on available data (the national cultivar trial database), this study could not find conclusive evidence that there was a negative correlation between yields and oil content. To some extent, these results could have been anticipated, given the wide range of external influences like the planting date and weather, which have a significant influence on both yields and oil content. Despite these inconclusive results, the fact remains that both higher yields (in the case of producers) and higher oil content (in the case of processors) are critical to boost the overall competitiveness of the industry. To this end, data from a recent pilot study undertaken by Syngenta and CEOCO was incorporated into this report. The pilot study, which involved the latest commercially sold high oil content variety, produced encouraging results, with sunflower seed being delivered to the crushing plant with an oil content well over 40% and average yields above 2 t/ha (in the 2020 season). During industry consultations, seed companies seemed to be confident that, with the correct cultivar selection, the average oil content of the South African sunflower crop could be improved significantly without compromising on yields in a meaningful way.

Consequently, this study explored the principles of an incentive system that rewards farmers for adopting high oil content seed that maximises oil yield per hectare, rather than just yields per hectare. This is not a new concept and is a sunflower contracting norm in international markets. To speed up the adoption of high oil content seed by producers, an incentive structure that stimulates uptake would be important to consider, especially if there might be a perception of a negative correlation between yields and oil content. Therefore, at the farm level, the incentive will translate into a gross margin trade-off between producing "high yield - lower oil content seed" and "low yield – high oil content seed". A summary of various international and domestic quality-based pricing models highlighted two options for the sunflower industry: a) Oil content-based price premium or oil content-based grading and b) Back-payment structure, by which the farmer gets to share in high quality-driven (high oil content) advantages at the crushing level.

The pilot study suggested the feasibility of a 1.5% premium, which would offer farmers a 0.24 t/ha "potential yield loss flexibility". In other words, this premium would be sufficient for farmers to achieve improved returns per hectare, up to a 0.24 t/ha yield loss (where the farmer would achieve returns equivalent to the average sunflower SAFEX price without yield loss). This study shows that the "potential yield loss flexibility" offered by such a price premium structure varies significantly with respect to yield, oil content and price premium level. The adoption of the high oil-yielding varieties would therefore also depend on the extent and likelihood of farmers consistently producing average oil content beyond 38%, which could be particularly difficult under the circumstances of shifting seasonal patterns. While the desire for crushers to procure higher oil content seed is well placed,¹ the uptake of new high oil-yielding seed varieties will also depend on the processors' appetite for and willingness to pay premiums that could compensate farmers for "potential yield sacrifices for high oil content seed". Again, this study has not found conclusive evidence of a negative correlation between yield and oil content.

Overall, it comes down to the farmers' willingness to adopt new high oil content seed and targeted agronomic practices to produce seed that consistently exceeds the threshold of 38%, preferably in excess of 40%, oil content, and the wherewithal of processors to pay premiums that will sustainably stimulate high levels of high oil-yielding seed production, while competing with crude oil imports. ²

Investigating the feasibility of testing the sunflower seed's oil content at silo level revealed that, firstly, the current sunflower seed-grading system does not explicitly define what constitutes a high or low oil content. Secondly, there is no standardised moisture content at which both yield and oil content are presented, i.e. representation at 9% moisture, which is mostly taken as the norm. Thirdly, although some silo depots have equipment to measure secondary quality characteristics (viz. NIR machines) and use it for measurements for grading purposes, this does not include the measuring of the oil content of sunflower seeds. In principle, oil content could be a measurement monitored at the silo level. However, since this currently is not a grading requirement and essentially a nice-to-have, it is consequently not measured, and machines are not necessarily calibrated for this. In addition, the roll-out of such a requirement will also require significant investment in order to make the equipment available at all silo depots.

The near-infrared reflectance (NIR) testing methodology could be the analysis of choice at both silo level and crushing plants due to the speed of the test, the ease of testing and its relative accuracy. However, not all silo depots currently have NIR machines available and, where they are available, a calibration for sunflower seed oil content would need to be purchased. Ring tests, however, as advocated by Agbiz Grain and SAGL, could go a long way to ensuring that oil content is measured accurately at all value chain nodes (silos, crushers, etc.), and some investment would be required to ensure that this type of testing equipment becomes more widely available.

Overall, one needs to also take cognisance of the fact that these relationships (pricing structures, profitability and farmer decisions) are by no means clearly defined cause-and-effect relationships when it comes to sunflower seed production in South Africa, as the causal relationships behind the performance of yield and oil content are not clear-cut, and the positioning of sunflower seed in the typical South African crop mix creates additional complexities.

Epidemiology and control of Phoma black stem and Alternaria leaf of sunflower in South Africa

Prof TAS Aveling
University of Pretoria

The final report is still awaited.

Nutritious snack multimix development for addressing under and overnutrition and promoting small, medium and micro-sized enterprises (SMME's) in low-income communities

Prof W Oldewage-Theron
University of the Free State

The primary goal of this study is twofold, firstly to address the the increasing need for an affordable, nutritive snack food by developing a nutrient-dense snack multimix from soy bean, peanut, seeds and baobab powder, secondly to address the double burden of disease (under- and overnutrition) (2020) in low-income households through the consumption of the developed snack foods and complementary nutrition education over a three-month period (2021). A secondary goal is to support the soy and oilseeds industry through SMME growth and development (2022).

The original plan was to implement the study in 2020 and complete it in 2022, but due to the Coronavirus pandemic and the subsequent travel ban, we could not implement the study as planned in 2020.

Progress to date:
We have started with product formulation and completed six theoretical snack food formulations, but no chemical analyses have been done as we will do this in South Africa during May 2021 instead of May 2020 as planned.

Proposed planting of soybean: Department of Agriculture, Escourt and Newcastle

Mr H Davies
Eden Social Development Foundation

Reporting is part of point 3.4.2 under Transformation Projects.

SAGL business plan: CAPEX

Ms W Louw
Southern African Grain Laboratory (SAGL)

The business plan submitted by the SAGL in respect of the establisment of a sustainability model for the SAGL as a laboratory to provide services to the agricultural industry was supported and the following funding was approved: