• The vegetation of the ca. 615 000 ha domain of the Garden Route Initiative was classified and mapped at a scale of 1:50 000 in order to develop surrogate biodiversity units to facilitate to development a conservation plan for the domain.
• The vegetation was mapped as untransformed units, as it was perceived to be before European settlement in the region. This proved to be a great challenge as vast areas have been altered to such an extent that only a few remnant patches of vegetation still remain in certain areas.
• We developed a classification system that follows a four-tier hierarchy. This facilitates analyses at three different scales, at biome, habitat type and vegetation unit level.
• Aquatic and terrestrial systems are recognized, with three aquatic biomes and six terrestrial biomes recognized in the area.
• Aquatic ecosystems cover more than 13 percent of the surface area, clearly indicating how prominent they are in the domain.
• A total of 39 habitat types are recognized; six of these are aquatic habitat types. Descriptions and a photograph are provided for each habitat type.
• At the vegetation unit level, 86 units are recognized, 15 in the aquatic ecosystems and 71 in the terrestrial ecosystems. Diagnostic descriptions are provided for each vegetation unit.

Limitations of the CBA map

• Although electronically possible these maps should not be used beyond a scale of 1:50 000. Zooming beyond this level will result in inaccuracies.
• Field mapping was done on hard copy satellite images printed at a scale of approximately 1: 30 000. We believe that field accuracy of the maps is in the order of 100 meter from mapped boundaries, but distortion in the satellite images may result in reduced accuracy in certain areas.
• Users should keep the original purpose of this map in mind, to develop a regional conservation plan. This vegetation map is not suitable for small-scale (< 1:50 000) studies or managerial plans.
• The vegetation units we recognize and their boundaries are not compatible with those of Mucina and Rutherford (2006), as their map is intended to function at a much larger scale (1: 1 000 000). We do agree conceptually at the higher order level (biome), but differ in many aspects at the lower levels of the classification.