A National Review of Land Degradation in South Africa: The Influence of Biophysical and Socio-Economic Factors¶
M. Timm Hoffman; Simon Todd (2000). A National Review of Land Degradation in South Africa: The Influence of Biophysical and Socio-Economic Factors. Journal of Southern African Studies, Vol. 26, No. 4 (Special Issue: African Environments: Past and Present). Taylor & Francis.
| Type | journal article |
| Scope | Wider apartheid literature |
| Method | mixed |
| Period studied | c.1987–1998 |
| Regions | South Africa (national, 367 magisterial districts), former homelands (communal areas), Eastern Cape (Ciskei, Transkei), KwaZulu-Natal, Northern Province, Orange River districts |
| Themes | agriculture land, bantustans removals influx, inequality poverty unemployment, methods data sources, transition post apartheid |
| Citation key | hoffman2000 |
Argument¶
The authors present the first national, district-level audit of soil and veld degradation in South Africa. It covers commercial and communal (former homeland) areas and rests on the consensus perceptions of agricultural extension officers and conservation technicians gathered at 34 workshops in 1997-98. Degradation is perceived as far worse in communal areas, especially on the steep eastern and northern escarpment, than in commercial districts. Stepwise regressions on 31 district variables link degradation to both biophysical factors (slope, temperature) and socio-economic ones (rural share, unemployment, dependants per household). The authors argue that land tenure is a surrogate for poverty and susceptible environments, not a cause in itself, and that poverty combined with fragile land produces the worst degradation.
Evidence and method¶
34 participatory workshops (1997-98) with 453 mostly Department of Agriculture extension officers and resource conservation technicians, scoring degradation by land use type for each of 367 magisterial districts. These were combined with GIS-derived biophysical and climate data, Development Bank of Southern Africa 1995 provincial statistical reviews (demography, labour, production), 1991 census data and nine case studies. The final regression analysis used 348 districts.
Place in the debates¶
- Causes of land degradation: tenure/overgrazing and population pressure versus poverty and environmental susceptibility (see Agricultural policy, productivity, farm size and land reform). Rejects tenure as a cause. Argues that communal tenure per se is not responsible, and that poverty, unemployment, dependency and steep, hot, erodible environments better explain degradation.
- Desertification debate (Acocks's 'expanding Karoo') and whether degradation is inexorably increasing (see Agricultural policy, productivity, farm size and land reform). Agrees with Roux that the eastern Karoo margin is no longer the main focus. Rejects the claim that veld degradation is rising everywhere, noting that it has been reversed in some commercial areas through state intervention and education.
- Validity of perception-based versus objectively measured degradation (see Data reliability and econometric method). Defends the use of expert perceptions as reasonably balanced and partly corroborated by other studies. Concedes that no comprehensive reliability analysis was done and that the data are subjective.
- Apartheid-era neglect of communal areas in research and state support (see Agricultural policy, productivity, farm size and land reform). Argues that past work was local or sectoral and neglected homeland areas, producing incomplete maps and discriminatory solutions.
Findings¶
-
Past degradation research ignored a national, quantitative, multidisciplinary view. (p. 3)
No national, multidisciplinary synthesis based on quantitative measures exists.
-
The audit rests on a large participatory expert exercise. (p. 5)
In total, 34 workshops were held, each lasting five to seven hours and attended by a total of 453 people.
-
Commercial/communal status alone explained over half the variance in degradation, acting as a proxy for other variables. (p. 11)
Initial analysis of the data showed that the proportion of commercial area in a district ('% commercial') explained more than 50 percent of the variance in the degradation indices.
-
Veld in the communal areas was perceived as roughly twice as degraded as in commercial areas. (p. 12)
The grazing lands of the communal areas were generally perceived to be about twice as degraded as those of the commercial areas (Table 2).
-
The soil degradation model combines socio-economic and biophysical predictors and explains just over half the variance. (p. 12)
In the full data set of 348 magisterial districts, eight of the variables used in the stepwise regression analysis accounted for more than 52 percent of the variance in the soil degradation index
-
Veld degradation was poorly predicted, with biophysical factors dominant and unemployment the only socio-economic predictor. (p. 14)
The set of variables used was less successful in predicting veld degradation index values; less than a quarter of the variance was explained in the regression model
-
The reliability of the perception-based data has not been formally tested. (p. 14)
A comprehensive analysis of the reliability or potential inaccuracy of the workshop results has not been undertaken.
-
Participants perceived that veld degradation has been reversed in some commercial areas through state intervention. (p. 15)
The perceptions of workshop participants were that veld degradation, in the commercial areas at least, has been successfully reversed in some places
-
Degraded districts are steep and hot, which reduces plant cover. (p. 16)
Magisterial districts with higher degradation levels are dominated by steeply sloping soils and occur in generally hotter climates where plant cover is likely to be lower.
-
Communal tenure is not the cause of degradation. (p. 17)
Nowhere in our analysis do we suggest that communal tenure per se is the cause of land degradation in the communal areas of South Africa.
-
Tenure is a proxy for poverty, unemployment and dependency, given historical and political reasons. (p. 17)
land tenure is a surrogate variable for a suite of much more useful indicators of degradation such as poverty, high levels of unemployment and a high dependency ratio.
Evidence ledger¶
Sources used¶
| Source | Kind | Period | How used |
|---|---|---|---|
| Land degradation audit workshops (34 participatory workshops, 1997-98) with agricultural extension officers and resource conservation technicians, covering 367 magisterial districts | author survey or fieldwork | 1997-1998 (perceptions of change over previous ten years) | Generates the consensus soil, veld and combined degradation indices and land use data per district, which are the dependent variables and mapped findings. |
| Hoffman et al., 'Land degradation in South Africa' (unpublished report, NBI and PLAAS, 1999) | commission or official report | 1996-1999 | Underlying commissioned audit from which the paper draws its results, methodology and reliability checks. |
| Nine case studies in select commercial and communal magisterial districts, with consultation of farmers, villagers, administrators, traditional authorities and NGO personnel | interviews oral history | 1997-1998 | Used qualitatively to corroborate the workshop results on perceived degradation. |
| Provincial statistical macro-economic reviews (nine provinces), 1995 | official statistical series | 1995 (growth variables 1981-1991) | Source of the demographic, labour/employment and economic production variables used as regressors. |
| 1991 population census (population density and rural share, Central Statistical Services) | population census | 1991 | Provides population density and the '% rural' classification used as demographic predictors. |
| GIS data sets of the Computing Centre for Water Research; Schulze et al., South African Atlas of Agrohydrology and Climatology (1997) | maps spatial data | 1997 publication | Supplies the biophysical and climatic variables (altitude, slope, rainfall, temperature, etc.) averaged per district. |
| Environmental Potential Atlas for South Africa (ENPAT) database | maps spatial data | 1997 | Source for the runoff and erodibility variables. |
| 1995/96 livestock stocking density data (LSU/HA) | official statistical series | 1995/96 | Stocking density variable included in the regression for the combined index. |
| Versveld, Le Maitre and Chapman (1998), Alien Invading Plants and Water Resources in South Africa, WRC TT99/98 | secondary literature compilation | 1998 | Alien plant distribution map used to validate workshop results on alien plant distributions. |
| Hill Kaplan Scott (1985), Transkei Soil Conservation Programme: Reconnaissance Survey Report | commission or official report | 1985 | Independent soil erosion assessment correlated with workshop data as a reliability check. |
| Roux (1990) veld condition survey of farmers, extension officers and range scientists | author survey or fieldwork | 1990 | Prior survey showing extension officers' veld scores lie between farmers' and range scientists', supporting rater reliability. |
New data¶
Consensus-based soil, veld and combined degradation indices plus land use shares and change data for all 367 magisterial districts from 34 workshops (453 participants), merged with 31 biophysical, land use, demographic, labour and economic variables for 348 districts used in regressions.
Inference¶
Quantification: Statistical analysis.
Perception-based district-level scoring from expert workshops, mapped by GIS and compared between commercial and communal land tenure categories (Table 2), followed by stepwise multiple regression (cross-section of 348 districts) of degradation indices on biophysical and socio-economic variables; causal claims explicitly avoided.
Key claims¶
| Claim | Basis | Strength |
|---|---|---|
| Veld degradation in communal grazing lands is about twice that in commercial areas, though many commercial areas suffer bush encroachment and alien plant invasions. (p. 753) | Workshop-derived veld degradation index by tenure category (Table 2). | Partial data |
| Eight variables explain over 52 percent of the variance in soil degradation, combining socio-economic (dependants, rural share, unemployment, settlement) and biophysical factors (slope, temperature, erodibility). (p. 753) | Stepwise regression Model 1 on 348 districts using DBSA, census and GIS variables. | Systematic data |
| Veld degradation is only weakly predicted by the variables (under a quarter of variance), with biophysical factors dominant and unemployment the only socio-economic predictor. (p. 755) | Stepwise regression Model 2 (Table 3). | Systematic data |
| Poverty (unemployment, dependency) in environmentally susceptible settings, rather than land tenure per se, is associated with greatest degradation; tenure is a surrogate for these. (p. 758) | Regression results in which '% commercial' was a surrogate and was excluded; correlation of poverty indicators with communal areas; no tenure analysis. | Partial data |
| Perceptions of extension officers are reasonably reliable and if anything understate the commercial-communal gap. (p. 756) | Correlations with Versveld et al. and Hill Kaplan Scott data, nine case studies, Roux's survey, and the authors' subjective assessment of rater bias. | Illustrative or anecdotal |
| Veld degradation has been reversed in some commercial areas, contradicting the claim that rates are increasing everywhere, and the eastern Karoo is no longer the main focus of desertification concern. (p. 756) | Workshop participants' perceptions and index maps; comparison with Roux and Acocks. | Partial data |
The author's caveats¶
Data are perceptions rather than objective measurement; no comprehensive reliability analysis was undertaken, and rater bias is acknowledged. Districts are very large (mean 3328 km2). The rural-population variable is a poor proxy for rural density and the 'settlement' category is too general. Regression does not imply causation, and '% commercial' was excluded as a surrogate variable. Changes in area affected were not recorded.
Gaps in the evidence¶
The argument lacks independent measured indicators (remote sensing, erosion plots, vegetation surveys, soil loss) to validate the perception-based indices and to track change over time. Finer-scale socio-economic data (rural population density, household income, settlement type, tenure arrangements within communal areas) at sub-district level are needed to test the poverty-degradation link; some were obtainable from census and satellite imagery even then. A single cross-section with few time points cannot address causation or the historical (apartheid land policy) origins of the spatial pattern.
Replicability: Data described but not public.