Select cities to compare

Production

In 2013, the Western Cape Department of Agriculture started an extensive project to map all agricultural areas using an aerial survey. This mapping exercise combined an aerial survey with on-the-ground validation and resulted in one of the most detailed agricultural land use maps in South Africa. The mapping itself is readily available through an online tool (Cape Farm Mapper). In 2017/2018 the information was updated through another aerial survey and groundwork. The mapping is done during different agricultural seasons, to account for different land use that might happen between summer and winter seasons. The primary data generated by this survey are the total hectares planted within the region, subdivided by crop type. However, in order to convert this to total production, average yields were provided by the Western Cape Department of Agriculture. While these numbers are provided by expert estimates and not based on exact measurements, they enable the calculation of local production to be based on regional averages (based on the Western Cape province), rather than on national yields.

Compared with the 2013 production, the 2017/2018 local production figures decreased significantly. However, this period was heavily impacted by a drought which impacted farming decisions (including the decision not to plant that season). Lacking from the production data is the milk production, as well as fish catch (both have outstanding queries for data).

Data Quality Indicators

Imports & Exports

NOTE: this descriptions applies both to imports and exports.

Data come from the Freight Demand Model (FDM), developed by GAIN. This dataset covers freight movements throughout South Africa, at magisterial district level. The methodology behind this dataset includes the collection of freight data from specific transport modes where records are available (such as freight by rail or container ship), and modeling based on demand and supply industry data to complement these datasets. Within Cape Town, there are a number of different magisterial districts. However, their boundaries do not perfectly overlap with the municipal boundaries of the city. This boundary mismatch is limited to the more rural parts of the metropolitan municipality, and likely not of great impact to the freight movements. However, the spatial data quality rating is set to 2 for this reason. Flows are broken down by commodity, and include both an origin and destination. There are a total of 72 different commodities available within the dataset, out of which 21 were identified as food flows. These were regrouped to match the classification system. For each commodity in the freight dataset there is a breakdown available that defines which particular items are included (using a product breakdown based on Harmonized System classifications). However, the actual data are only available at the top level. There are two categories that are primarily impacted by this limitation. Firstly, there is a ‘Processed foods’ category. This group includes foods that have been processed within the food industry, such as muesli or frozen potato chips. Unlike other food reporting datasets (such as the FAO consumption data), food flows in this dataset are not aggregated based on the raw ingredients but instead reported on in their final product form. Consequently, the ‘Processed foods’ category is one of the most important categories (making up almost 50% of imports and over 90% of exports) but the exact composition remains a black box for the purpose of this study. The total quantity of processed foods that are exported, are larger than the total imports of processed foods. This means that food crops (such as wheat or apples) are processed into other products (such as bread or juice), and then exported. This fits Cape Town’s profile as a food processing hub, but it makes it impossible to trace different food groups through the entire food system. This is a limitation of the datasets used. Another challenge is the category labeled ‘Other agriculture’. This category includes freight flows mostly associated with agricultural activities themselves, rather than food flows. For instance, oats (mostly consumed by livestock) and live plants (used for planting) are included in this section. However, this category also includes any ‘Live animals’ that are imported and exported. In the case of live animals, these directly become part of the food flows of the city if they are imported to be slaughtered in one of the city’s abattoirs.

ASSUMPTION: The large imports of ‘Other Agriculture’ are dominated by imports of live animals, which are subsequently slaughtered in the city and enter the food system. For this reason, this category is included in the food flow analysis, even though it will include some imports and exports of agricultural products not part of the city’s food flows.

Data Quality Indicators

Consumption

City-specific food consumption data are not readily published in South Africa. Battersby and colleagues (2014) went through an in-depth exercise to review possible data sources and to test different methods of obtaining city-specific food consumption data for Cape Town. They used three different sources - all of which publish data on a national level: the Statistics South Africa General Household Survey, the Food and Agricultural Organisation (FAO) Country data report, and Department of Agriculture, Forestry and Fisheries (DAFF) Agricultural Abstract data. Their review concluded that the only way to obtain consumption data in kilograms per person for all food groups was to combine these different data sources, as none of them provided a complete dataset in the right units.

Almost a decade later, data availability on national food consumption has improved. FAO data reports now provide South African per-capita consumption figures for all of the food groups under study. This means that it is no longer a requirement to convert monetary or percentage-based dietary data info weight. However, it was still not possible to find sources that publish specific consumption data for Cape Town.

Battersby and colleagues interviewed and consulted food system specialists to understand how to best estimate Cape Town specific consumption based on national figures. It was suggested that food spend in Cape Town would be about one third higher than the average national spending. They subsequently increased the consumption figures by one third. They compared their numbers with three other datasets, and found that these numbers aligned with other work that looked at different aspects of the food system but that could validate their approach.

We used the FAOSTAT data, and multiplied the per-capita figures by Cape Town's population. This might mean that the figures are an undercount, as Cape Town has previously shown to have a consumption that is one third higher than the rest of the country. However, instead of multiplying the figures, we have indicated this as a lower reliability and geographical correlation indicator. Due to the height of Cape Town's consumption in relation to other cities in the study, and looking at the other flows within Cape Town's food system, we decided against increasing the consumption figures as was done in the aforementioned study (because it would further increase the already existing "gap" that exists between the flows entering and leaving food supply).

Data Quality Indicators

Food loss and waste

There are a number of possible sources for food waste and loss. However, all of these sources have shortcomings. The following sources were considered:

FAO Food Loss and Waste Database.

For 2019 data, the source is The African Postharvest Losses Information System (APHLIS). The 2019 data do not include losses for fruit. These losses are only available for 2021, which cite L Haywood (2021), Increasing reliable, scientific data and information on food losses and waste in South Africa. FAOSTAT was therefore discarded as a direct source, and the original sources were reviewed instead.

Global food loss and waste estimates show increasing nutritional and environmental pressures

This recently published study attempts to provide a global database on food waste and loss, providing country-level figures on food loss by stage. However, data for South Africa is not based on national insights but instead grouped using the same percentages as all other countries in the 'Sub-Saharan Africa' region. This already makes the data less reliable, as the economic profile of South Africa quite strongly differs from the average Sub-Saharan African country. Furthermore, the underlying sources for this region (Table S8) are principally the FAO Food Loss and Waste Database, and "Assumptions based on values for South & South-East Asia". For these reasons, this publication was discarded as a primary source for South Africa.

African Postharvest Losses Information System (APHLIS)

This project provides postharvest losses of cereal grains in sub-Saharan Africa. These postharvest loss profiles are derived from "peer-reviewed literature, and contextual factors provided by local experts". For South Africa, data are ranked as "incomplete" and the cited literature is generally not South Africa-specific. Furthermore, data are only available for four cereals. For these reasons, this source was discarded as a primary source for South Africa.

Increasing reliable, scientific data and information on food losses and waste in South Africa (Oelofse et al, 2021)

This technical paper looks specifically at how to improve the reliability of food loss and waste data in the South African context. Unlike any of the previous publications, this has a South African focus and it seeks to improve datasets derived from FAOSTAT's Food Balance Sheets by adjusting it for the local conditions which often deviate from the average Sub-Saharan country.

Due to the localized approach, food waste and loss factors from this paper were considered most appropriate. This publication calculates national figures, so a large assumption was made that these figures remain similar for Cape Town. The paper provides estimated losses (in tonnages) for South Africa for five stages:

• Food production
• Post-harvest handling and storage
• Processing and packaging
• Distribution
• Consumption

Losses were available for seven broad food groups. By calculating the percentages for each stage, it was possible to apply these same factors to the different Cape Town food flows. Losses to food produced in Cape Town was calculated by taking the losses of all categories except for Consumption (because Food Consumption data is available separately, and not all local production is necessarily consumed locally). Losses to imports were calculated based on the loss factors of Distribution. However, this entails an unverified assumption that no imports are processed inside the city. In fact, some of the imports are processed and packaged inside Cape Town. This loss figure is therefore expected to be undercounted.

Lastly, there were a number of food groups that did not fit within the seven categories available in this document. Below follows an overview of which waste or loss rates were used instead.

PRODUCTION Mushrooms were based on "Roots and tubers" Treenuts were based on "Oilseed and pulses"

IMPORTS/EXPORTS Eggs, Processed Food, and Beverages were based on the average figures from all categories. Other Agriculture is based on Meat (because this category involves "Live Animals" expected to be slaughtered, see data description for Imports/Exports)

CONSUMPTION Alcoholic Beverages, Stimulants, Spices, Vegetable Oils, and Sugar & Sweeteners were based on "Roots and tubers" which had one of the lowest rates, and we expect low rates (due to long shelf life) Treenuts were based on "Oilseed and pulses" Animal fats and Offal were based on "Meat" Aquatic Products, Other were based on "Fish and seafood" Eggs and Miscellaneous was based on the average rate

Data Quality Indicators