Question NW578 to the Minister of Agriculture, Land Reform and Rural Development

Reply:

a) There was an elaborate process that culminated in a narrow criterion, determining whether the property was vacant or underutilised (partially vacant). The process involved a desktop spatial analysis; physical verification; desktop agricultural potential analysis; verification of land restitution claims; and final verification utilising local departmental knowledge. Each one of these steps contain detailed activities that can be outlined should that be considered necessary. The focus, for purposes of this question, will be restricted to the desktop spatial analysis.

The desktop spatial analysis included the identification and sourcing of data sets; definition of the terms vacant and partially vacant (underutilised); defining the parameters to identify vacant and partially vacant agricultural state land; mapping agricultural state land; and performing a desktop Geographic Information System (GIS) analysis on the data.

Identification and Sourcing of Data

The datasets listed below were identified and obtained from the identified data custodians:

• Cadastral data from Agriculture, Land Reform and Rural Development;
• Deeds Registration data from Agriculture, Land Reform and Rural Development;
• Land parcels identified as vacant by the Department of Public Works and Infrastructure;
• National Land Cover for the year 2017 from Agriculture, Land Reform and Rural Development;
• South African Protected Areas Database for the year 2018 from Environment, Forestry and Fisheries;
• Indigenous Forests Inventory from Environment, Forestry and Fisheries;
• Land Capability (Terrain Capability) from Agriculture, Land Reform and Rural Development; and
• Traditional community boundaries from Traditional Affairs.

Definition of Terms and Parameters for determining Vacant Land

• The term vacant refers to a piece of land that is not occupied or not being used for any specific purpose whilst the term partially vacant refers to a property when only parts of it have been classified as vacant.
• The 2017 Land Cover dataset, which was used, is a national land cover classification generated using multispectral satellite imagery with a multi-temporal analysis technique, produced by National Geospatial Information within Agriculture, Land Reform and Rural Development.
• The table below shows the selected land cover classes from the national land cover data which were utilised to identify agricultural land as vacant or not vacant.

Selected Land Cover Classes, 2017

No	Land Cover Class	Classification
1	Natural wooded	Vacant
2	Shrubland	Vacant
3	Grassland	Vacant
4	Waterbodies	Not vacant (depending on coverage)
5	Wetlands	Not vacant (depending on coverage)
6	Barren land	Vacant
7	Cultivated	Not vacant
8	Built-up	Not vacant
9	Mines and quarry	Not vacant
10	Planted Forest	Not vacant

• Indigenous forest dataset was used to classify properties located in indigenous forest as not vacant.
• The Protected Areas Database was used to classify properties located within these areas as not vacant.
• Although terrain capability is one of three building blocks (soil, climate and terrain) for land capability, only terrain capability was used in this analysis. Terrain capability includes aspects such as slope gradient, slope direction, slope shape, ground roughness, altitude and streams (drainage network). Land with a terrain capability value between “Very Low and Low to Moderate” was subtracted from vacant and partially vacant land.
• Land that is located within traditional community boundaries was regarded as not vacant.
• State land acquired in terms of the Pro-active Land Acquisition Strategy (PLAS) and by the Commission on Restitution of Land Rights, was regarded as not available for redistribution.
• Land that is claimed in terms of the Restitution of Land Rights Act was regarded as not available for redistribution, depending on the phase of the claim and options selected by the claimants.

Criteria Used

• In view of what has been explained above, the ultimate criteria that was determined is that:
  • • • • The land cover class is either grassland, barren land, shrubland or natural wooded;
        • the property is located on suitable terrain capability;
        • the property is less than 10% within an indigenous forest;
        • the property is less than 10% within a protected area;
        • the property is not located in an area which is likely being used for communal grazing;
        • the property is not part of Pro-active Land Acquisition Strategy (PLAS) land or State land acquired by the Commission on Restitution of Land Rights for settlement of claims;
        • the land parcel is not un-registered;
        • the aggregated vacant or partially vacant land available per property is bigger than 50 hectares; and
        • The property is not claimed in terms of the Restitution of Land Rights Act.

b) The final list was a product of collaboration amongst various business units of the Department in the manner outlined in the table below:

Desktop Spatial Analysis:	Chief Directorate: Property Management and Advisory Services
Desktop Agricultural Potential Analysis:	Directorate: Land Use and Soil Management
Verification of Land Restitution Claims:	Commission on Restitution of Land Rights
Final Verification utilising local departmental knowledge	Provincial Shared Services Centres

c) No personal communication was sent out to farmers, however there were media briefings that were generally accessible to the public.