South African Carbon Sinks Atlas

How the Model Works

The national carbon stock consists of a set of linked and interacting sub-stocks (called ‘pools’) which change over time: slowly in the case of soil carbon, moderately quickly in the case of woody biomass, and rapidly in the case of herbaceous and litter carbon. The carbon flows between the pools, and between the land and the atmosphere, land an ocean, and land and human systems are called fluxes.

Figure A: Components of a generalised terrestrial carbon cycle, with box sizes (representing stocks) and arrows (representing fluxes) roughly indicative of their relative size in South Africa, where NEE = Net Ecosystem Exchange; NEP = Net Ecosystem Productivity; NBP = Net Biome Productivity; GPP = Gross Primary Production; Ra = Autotrophic Respiration; Rh Heterotrophic Respiration; Re = Ecosystem Respiration; Rfire = Fire Emissions.

Carbon Pools and Fluxes

Carbon is the basis of all living organisms, and in the case of vegetation, carbon molecules typically constitute between 40% and 50% of the oven dry biomass. This carbon is referred to as terrestrial ecosystem organic carbon (which for simplicity will be refer to as “organic carbon”). It excludes the fossilized carbon stocks found in coal and oil reserves. Photosynthesis results in plants absorbing CO2 from the atmosphere and converting it into vegetation. Some of this vegetation is eaten by animals, some is lost to respiration, and some eventually becomes trapped in the soil as Soil Organic Carbon (SOC). Over time the amount of organic carbon found at any specific terrestrial location tends to reach a relatively constant equilibrium in regard to the quantity of carbon found in the different carbon pools. A disturbance to any one of the carbon pools tends to have ripple effects through all the carbon pools. For instance loss of vegetation (e.g. through clearing natural vegetation to form a crop field) will result in loss of SOC. The key carbon pools and fluxes are summarised in Figure A.

The terrestrial ecosystem organic carbon stock can be divided into a number of carbon pools (see Figure A). A common approach, as used within the National Terrestrial Carbon Sinks Assessment 2020 (NTCSA 2020), is to separate the organic carbon into carbon contained in vegetation and carbon found as soil organic carbon (SOC). The vegetation carbon can be separated into the carbon found in trees (including tree crops such as commercial plantain forestry, fruit trees and grape vines) and carbon stocks from herbaceous vegetation (grass, forbs, annual food crops). Further is useful to divide vegetation biomass (and its related carbon stocks) between the above ground biomass that can be seen and observed and the below ground biomass (roots) that is not visible and harder to assess. Finally, there is carbon in the dead plant material found on top of the soil and referred to as litter. Carbon found in the animal fauna is not considered due to it being insignificant compared to the other carbon pools.

Carbon pools analysed per land unit

  • All carbon pools are calculated based on individual land units (LU) of 1 km by 1 km. The total carbon within South Africa, or within any specific sub-region of South Africa such as a biome or province, is calculated by summing all the land units within that biome or province.

  • Three land cover products, NLC1990, NLC 2014 and NLC 2018 were used in calculating total terrestrial carbon stocks. These were run using two different soil organic carbon estimates.

Assessing Land Use Change

  • NLC products are either at 30m resolution (1990, 2014) or 20m (2018 and subsequently) and are used to determine the area of each land cover within each LU. This data was used to identify land cover changes that may have impacted on the carbon stock of the vegetation. A reduced set of 17 land cover classes (out of up to fifty) were identified that were shared between all three NLC products used.

Hypothetical change in a parcel of land from the reference period until 2018 as recorded by national land cover data. During the reference period there is only natural vegetation, by 1990 some of the land is cleared to form agricultural fields (grey). By 2014 the agricultural fields have expanded and some land is taken up by settlement. By 2018 both settlement and agriculture have expanded and the amount of natural vegetation is greatly reduced.

Calculating Total Ecoystem Organic Carbon

TEOC = SOC + (AGB[woody] +BGB[woody] + DW + AGB[herb] +BGB[herb] + AGL)*CF

Carbon Stocks Model

Total ecosystem organic carbon (TEOC) is calculated as the sum of a number of individual organic pools.

TEOC = Total Ecosystem Organic Carbon, SOC = Soil Organic Carbon, AGB = Above Ground Biomass, BGB = Below Ground Biomass, AGL = Above Ground Litter (which will include Dead Wood).

Tree biomass and carbon

Tree biomass was estimated based on satellite based Radar data, lower lever Lidar data and ground verification data. A model of tree biomass for the entire country was developed for the year 2014 and this is used as the tree biomass data. Carbon was assumed to be 42% of the entire tree biomass. Root biomass was assumed to be a set ratio of above ground biomass, with this ratio changing along a rainfall gradient.

Soil organic carbon (SOC)

The ISRIC 250m resolution soil carbon map was used to give a reference soil carbon. By reference we mean the SOC values that would be expected in un-disturbed natural vegetation at that specific location. The National Land Cover maps (1990, 2014 and 2018) were used to understand the proportion of land that had been converted from natural vegetation to another land cover. If land had been converted from natural land to an alternate land cover, then conversion factors were applied to the ISRIC reference soil carbon to estimate how much SOC would have been lost. These conversion factors differed based on land use, climate and biome and were derived from published and unpublished field data for South Africa. In situations where no South African data was available to estimate loss of SOC as a result of land use, IPCC default values were used.

Herbaceous biomass and carbon

Herbaceous biomass was assumed to be a function of mean annual rainfall and the same mean value is used for all time periods. Below ground (root) biomass is assumed to be a fixed proportion of above ground biomass. Where crop fields have replaced natural vegetation, herbaceous biomass was based on district level mean crop yields per crop varieties, with crop biomass determined by the length of the growing season and mean mass of crop at harvest. Carbon was assumed to be 42% of biomass.

Litter Biomass

There are exceptionally few studies on the amount of litter biomass in South Africa and how this varies over space. Mean values per biome were based on available literature and a single value was used for the entire biome of each of South Africa’s nine Biomes. In addition litter was assumed to include dead wood, which was assumed as 2% of standing woody biomass in areas of communal land tenure (with their high fuelwood demand) and 10% standing woody biomass in all other areas.