NationalEcosystemClassification

Working Document

INTRODUCTION In this document we describe the overall approach to developing ecosystem classification systems and maps for the terrestrial, marine, inland aquatic and estuarine realms of South Africa. We discuss the array of uses and applications of the maps and the process that supportssupport their maintenance and updatesupdate. The document then presents the latest facts and figures for the ecosystem types within each realm. South Africa’s wide range of bioclimatic, ocean¬ographic, geological and topographical settings have resulted not only in high species diversity and endemism (),(ref?), but and also high ecosystem diversity and endemism across all realms. There is a wide variety of terrestrial biomes and aquatic ecoregions in South Africa, its surrounding seas and sub-Antarctic territory; ranging from the unique Fynbos biome to the exten¬sive savannas and grasslands of the eastern interior, and from the Subtropical Indian Ocean through the Warm Temperate Agulhas Shelf to the cold upwelling-influenced shelf of the Southern Benguela. Situated 1 700 km south of the country, the Prince Edward Islands (PEI) and their surrounding seas add a cold, sub-Antarctic set of ecosytem types to South Africa’s territory. The wide variety of inland aquatic and estuarine bioregions in South Africa also reflects this bioclimatic, geological and oceanographic heterogeneity . Classification and mapping of ecosystem types underpins a variety of applications relating to the management of natural resources, biodiversity planning, land‐use planning, marine spatial planning, allocation of environmental flows, and environmental assessment and monitoring in all realms. Ecosystem types are increasingly being used as units of assessment in national and international processes linked to biodiversity conservation and ecological infrastructure management. For example, reporting on international processes linked to the Convention on Biological Diversity (CBD) and Sustainable Development Goals (SDGs) is not possible without national-scale ecosystem classification systems and maps. (mention RLE here?) South Africa has a long history of mapping and classifying terrestrial vegetation units, with the first maps emerging in the 1930s (Mucina and Rutherford 2006). However, these maps were focused on rangeland potential. In the early 2000s, a growing need for a map of terrestrial ecosystem types that classified and delineated the biological diversity of natural ecosystems drove the development of the first national vegetation classification system and highly detailed map. This map altered both our foundational model of the biophysical environment and the operational landscape for implementers and policy-makers in the terrestrial biodiversity sector. Custodians of biodiversity finally had foundational ecosystem units that were appropriately mapped and classified for use in biodiversity planning, assessment, and monitoring. While the history of work in other realms does not reach as far back as that for terrestrial vegetation work, there have been major advances in the classification systems and maps of ecosystem types in the inland aquatic, estuarine and marine realms over the last two decades. Classification systems and maps for inland wetlands, estuaries, rivers and marine ecosystem types have been used along with terrestrial information in biodiversity planning and assessment since at least the early 2000s . A key message that emerged from the National Biodiversity Assessment (NBA) 2011 (Driver et al. 2012) was the need to further develop and strengthen a somewhat disparate set of national ecosystem classifications into a more comparable and aligned system for classifying national ecosystem types across all realms (marine, estuarine, inland aquatic and terrestrial). It was also clear that a convening organisation would be necessary to realise this goal, and the NBA 2011 highlighted the potential role of the South African National Biodiversity Institute (SANBI) in catalysing, facilitating and co‐ordinating this process, with the CSIR providing support in the estuarine and fresh water aquatic domains. Between 2015 and 2019, substantial resources were invested by SANBI and the CSIR in the further development of ecosystem classification systems and maps in all realms during South Africa’s third National Biodiversity Assessment process, the NBA 2018 (Skowno et al., 2019). Where opportunities arose, additional resources were leveraged through strategic engagement with the Department of Science and Innovation which were willing to prioritise research on ecosystem classification in research strategies and funding calls (such as the DSI’s Marine and Antarctic Research strategy and the NRF’s ACEP call). While the classification systems for realms are not completely unified, the systems were deliberately are sufficiently integrated and aligned for cross-realm assessments (with futher effort needed in this regard (Table 1). ).

PART A: APPROACH AND GENERAL PRINCIPLES OF THE THE SOUTH AFRICAN CLASSIFICATION OF SYSTEMECOSYSTEM TYPESS OF REALMS
Approach and Guiding Principles
The classification systems of the four realms followed several general principles. In all maps and classification systems: i) AnAn The original state (or baseline) of an ecosystem type is defined according to its pre-colonial extent and ecological condition . This baseline year is generally agreed to be approximately 1750 AD. Many landscapes in South Africa still retain remnants of the abiotic and biotic components from this historical period. Therefore, it is possible to map some ecosystem types that may potentially represent the historical state. However, in practice, many ecosystem types are defined in their current state using on more recent datasets due to the lack of suitable historical data. Further refinements are made to better represent the historical extent and ecological condition functionality? of ecosystems as additional new historical data emerges.
ii) Anthropogenically changed transformed landscapes or seascapes (post 1750AD) are defined as … and not classified as an ecosystem type. These areas are treated as not representative of the baseline ecological condition of the landscape and are thus mapped and classified through separate processes (e.g. land cover and land use classification maps). iii) Classifications are based on a nested hierarchical approach such that each realm has multiple levels, and lower levels in the hierarchy accumulate up to a higher level in a many -to- one relationship. Categories at a specific level of these hierarchies represent the ecosystem types of a realm, and may differ between realm classification trees? /from one to another, depending on the spatial representability and expert-agreed opinion of representation. iv) Ideally, eEach ecosystem type defined within a classification system must have the following elements associated with it to qualify as a valid type : a. a name b. a delineation (spatially explicit) c. a full description of the abiotic, biotic and/or functional characteristics d. must be assigned to a corresponding higher level within the nested hierarchy e. must be accompanied by suitable reference material in the form of published documents or accepted datasets f. must cover the extent of the full realm.
v) All classification systems should provide are practical units for applied processes within the context of the realm. Although finer scale hierarchical levels are allowed within each realm, a practical landscape or seascape scale hierarchical level must be identified and maintained at all times. With suitable scales of application, e.g. what is represented at the finer scale – or is that logical? vi) The units used to represent an ecosystem type in an NBA have been developed with the best most recently available data and isareareis underpinned by expert review. Each realm has a committee of regional and national experts who guide the strategic and technical objectives of the classification and mapping. These experts reviewed the initial classifications and maps, and should continue to review any proposed editions .

Key concepts and definitions. An Ecosystem type The South African Biodiversity Act defines an ecosystem as a “…dynamic complex of animal, plant and micro-organism communities and their non-living environment interacting as a functional unit”. This definition can be sensibly applied at a range of spatial scales but in South Africa it has been applied to a local scale, such as the vegetation types described in Mucina and Rutherford (2006) which that have an effective map scale of 1:100 000 – 1:250 000. The International Union for Conservation of Nature (IUCN) definition of ecosystem types for ecosystem red listing (following Keith et al. 2013) is “Ecosystem types are the units of assessment, and are… complexes of organisms and their associated physical environment within a specified area. They have four essential elements: a biotic complex, an abiotic environment, the interactions within and between them, and a physical space in which these operate. Many related concepts that have evolved at different scales (subnational to national) may be treated as ecosystem types…” (IUCN 2016, page 7). This definition is well suited to South Africa , where the Red List of Ecosystems (RLE) has formed a key element of national National bBiodiversity aAssessmentsbiodiversity assessmentsNBAs since 2004. Note: An ecosystem type is not equivalent to a habitat type or biotope. The concept of a habitat is closely related, but not identical to that of ecosystems. A habitat is usually defined as space that is physically and biologically suitable for the habitation of a particular species, or in the case of a biotope (similar to habitat type but more commonly used in English-speaking countries), a suite of species . Therefore, unlike an ecosystem type, a habitat type usually cannot be defined without a target species in mind (Bogaart et al. 2019). A Classification system (Typology) The national classification systems (referred to as “typologies” in some international literature) applied to South African ecosystems are hierarchical in nature. This is seen as a positive attribute that allows for use across a wider range of processes and scales than a single level classification. For example, the lowest levels in the classification hierarchy for the national vegetation map (i.e. types and subtypes) are used in RLE processes while the higher levels in the hierarchy (i.e. biomes and bioregions) are used in global change modelling (Guo et al. 2017).
Biome
The definition of a biome is still being discussed and debated globally in vegetation science (Mucina 2019). However, a working definition is vital for many practical applications because the biome concept is widely regarded as the largest-scale unit in the classification systems of most countries. Mucina and Rutherford (2006) define a biome as “ …a biotic community finding its expression at large geographic scales, shaped by climatic factors, and perhaps better characterized by physiognomy and functional aspects, rather than by species or life‐form composition. Biomes are frequently used as tools to provide large‐scale (regional to global) backgrounds in a range of ecological and biogeographical studies.” Biomes broadly correspond with climatic regions, although other environmental controls, such as biogeography, are sometimes important. They are defined in terms of all living organisms and of their interaction with the environment (not only with the dominant vegetation or habitat in which species reside). There is no single authoritative list of biomes for South Africa or globally. While some biomes are recognized by all terrestrial authors (e.g. tropical rainforest, taiga), many different biomes are proposed for less well-defined ecosystems, especially those on ecotones, such as savannas and woodlands. In the South African classification system for terrestrial, marine, freshwater inland aquatic and estuarine systemsy(rivers and inland wetlands) and estuary, the biome is the acknowledged as potentially the highest level in the classification system but it is not currently explicitly included in all realms as a classification level. Nevertheless, estuarine and Inland Aquatic ecosystem types fall within the same bioclimatic zones defined by the terrestrial map. Marine biomes could be classified and delineated as more data emerges.

Inland wetlands were grouped into bioregions (some subsumed) which relate to the biomes. While Although the concept of a biome is conceptually and geographically identical for the terrestrial, freshwater inland aquatic and estuarine classification systems, Tthe concepts used toprocess of defining define and delineateiondelineatingon biomes biogeographical regions (similar to biomes) in the estuarine, coastal and marine realm is still under development and may deviate slightly.given that they are driven by ever changing sea currents and climatic integration and more ‘ephemeral’ in nature. Broad regions subjected to similar ocean processes with similar species assemblages were identified and broadly grouped together as a biogeographical region.

Reference period for the map
In regions where substantial changes to the vegetation cover have occurred due to human activities (including agriculture and urban development), modelled reconstruction of natural vegetation is applied (see Neuhaus 1984, as cited in Mucina and Rutherford 2006). While Although the approach is necessary for RLE type assessments, it is acknowledged that, in regions where knowledge of patterns and processes to inform the reconstruction models is lacking, this approach will introduce uncertainty into the map and the subsequent assessments (see below). Ecoregions In North American classifications, ecoregions have been defined as large-scale ecosystems or groups of ecosystems driven by microclimate, species cover, topography, geology and hydrology (Mucina 2019). However, the IUCN considers ecoregions to be “large areas containing many ecosystems, which may or may not have a strong functional relationship with one another” (Bogaart et al. 2019)., page?). Ecoregions can be nested within a hierarchical structure and grouped into higher order biogeographic regions. Ecoregions in the South African classification systems resemble the IUCN definition more closely because they are considered to be constituted byit is a group of ecosystem types that share similar larger scale biogeographic processes, and these units nest within the biomes within the classification hierarchy. Land use land cover (LULC ) Land use and land cover (LULC) is defined by Bogaart et al. (2019) as “the observed physical and biological cover of the earth’s surface and includes natural vegetation and abiotic (non-living) surfaces. At its most basic level, it comprises all of the individual features that cover the area within a country. For the purposes of land cover statistics, the relevant country area includes only land and inland waters. The area of coastal waters is excluded.” In South Africa, land cover classes are not used to define represent as the baseline of ecosystem types in the historical classification system.for a number of reasons: it reflects anthropogenic transformation of land rather than the original abiotic, biotic and functional processes of terrestrial or aquatic realms. These LULC classes are therefore rather incorporated in a used to represent the separate classification system that classifies the current ecological condition of landscapes within South Africa. Uncertainty and quantification of error? Or confidence? in classification and mapping
Ecosystem classification systems and their spatial representations (maps) are seen as simplified models of complex natural patterns, with numerous sources and types of errors and uncertainties (Mucina and Rutherford 2006). • Positional accuracy and alignment of boundaries at the transition between contrasting ecosystem types is largely driven by scale of data capture, methods used and the quality of the spatial informant layers (e.g. aerial photography and satellite imagery). Where the transition between structurally similar ecosystems types is gradual, placement of the boundary line is somewhat subjective – leading to uncertainty in the final products that is largely independent of scale. Where the transitions between ecosystem types are very subtle the margin for uncertainty can be higher than for ecosystem types that have more abrupt transitions (Bogaart et al. 2019) e.g. the transition between mesic thicket and forest versus the transition between grassland and forest. Classification accuracy can be compromised by errors in zonal data used in top-down classification and by errors in the community / species data used in bottom-up classification. For example, an inaccurate digital elevation model will lead to errors in the altitudinal classification of high-altitude grassland types, and inaccurate species identifications could lead to errors in community assignment.

In addition, some features may meet the criteria of more than one ecosystem type in a realm or even between realms, as in reality biodiversity is distributed across features are a continuum and not as discrete features. A, as more data becomes available some features may need reclassification based on physical or biotic data.

None of the ecosystem classification systems and maps in South Africa include explicit accuracy assessments or considerations of uncertainty; this should be addressed in future development of the national ecosystem classification system (NECS). The marine ecosystem types were categorised into five categories of certainty based on the amount of data available per type.

An integrated approach to mapping and classificatio n Consistent progress has been made in developing and refining national-scale ecosystem classification systems and maps in the terrestrial, inland aquatic (rivers and inland wetlands), estuarine, and marine realms of South Africa (Figure 1). A complete description per realm can be found in Appendix 1.

Figure 1. Overview of South Africa’s national ecosystem classification system across the different realms. The coast is considered a composite realm consisting of marine, terrestrial, estuarine, river and inland wetland ecosystem types (A). The terrestrial realm overlaps slightly with the estuarine realm where the estuarine function zone consists of a recognised vegetation type (B). The estuarine realm and the marine realm overlap around estuary mouths (C). Internationally, estuaries are also considered a type of wetland (D ).

The development and refinement of ecosystem classification systems and maps have typically progressed independently within each realm. over the years and different NBAs. For many applications this does not present a significant problem (e.g. for offshore marine protected areas design), but for other applications (e.g. cross-realm systematic biodiversity assessment or integrated conservation planning, especially in the coastal zone) it is important to have an integratedcohesive map with no overlaps or gaps between realms. In the past, planners and scientists have integrated the maps on an ad hoc basisthemselves using a variety of approaches and rule sets for weighting the different maps. However, previous disparities between classification and mapping in the different realms were greatly reduced during the refinement for the NBA 2018. Representatives from each realm-specific committee co-created the mapping and classification of ecosystem types along transition areas between respective realms such that an integrated map of ecosystem types across our realms has now been achieved (Figure 2). To facilitate future alignment in these cross-realm transition zones, a National Ecosystem Classification Committee (NECC) has been established and will meet when necessary to ensure regular interaction among the ecosystem classification committees and mapping teams. For example, estuary mouths require careful alignment with coastal and marine maps, the upper reaches of estuaries often require careful alignment with rivers, inland wetlands, and the vegetation map currently has the outer boundary of the Eestuarine Functional Zone (EFZ ) ecosystem included as a subfield to prevent confusion).

Figure 2. Ecosystem map integration and alignment efforts (b has been extracted from Harris et al. 2019)

Alignment with international processes The South African ecosystem types and classification systems are aligned with the principles outlined by international processes that utilise ecosystems as foundational units, such as the IUCN ecosystem red listing and Ecosystem Accounting (ref) which utilise ecosystems as foundational units. Under these processes, ecosystems types should be guided by five basic criteria. Table 2 outlines these sixfive criteria and provides an indication of which criteria areis met within the South African classification of ecosystem types across all four realmssystem . Table 2. Alignment of the South African classification of ecosystem types across all four realms system with criteria defined by the IUCN (Bogaart et al. 2019) International CriteriaCriteriaum. Ecosystem types must be: South African classification of ecosystem types Description and alignment of South African ecosystem classification systems
i Based on ecological principles that should be underpinned by scientific data and expert consultation;  All levels of the classification in all realms are underpinned by ecological and biophysical data. Each realm has a committee of national and regional experts who review the classification and mapping. ii Map-ableSpatially explicit so that the area and extent of units can be spatially delineated;  All types defined in the classification system have been identified and delineated in space. iii Classification units must be mapped comprehensively across the target land or seascape;  The terrestrial, marine, estuarine and freshwater and aquatic ecosystem maps delineate the full extent of the South African territory. iv Ecosystem types must be mutually exclusive conceptually and in space;  The terrestrial and aquatic , marine, estuarine and freshwater ecosystem maps are mutually exclusive delineate the full extent of the South African territory. *Previously, separate benthic and pelagic ecosystem maps may challenge this criterion. v The size and scale of units need to be practical for implementation;  All units from across the four realms have been mapped at the landscape or seascape-scale and have already been used in regional and national implementation processes. vi The classification should be aligned to other international processes at a finer or coarser scale within matching hierarchies (Bogaart et al 2019).   International parameters are still being defined. Thus far, higher levels within the classification hierarchy can be matched to corresponding levels within other international systems such as the IUCN.

The IUCN (Bogaart et al. 2019) recommends considering the following elements as key inputs for the definition of an ecosystem: a) Terrestrial ecosystems: climate, mean annual temperature and seasonality, total annual precipitation and seasonality, total annual evapotranspiration, topography and geomorphology, lithology and soil structure and chemistry, phenology, vegetation growth form, life history, leaf type and phenology, adaptations to oxygen stress, animal feeding preferences, direct or indirect anthropogenic impacts. b) Freshwater Inland aquatic ecosystems and wetlands: geomorphology, stream order, fluvial zone, sediment size, channel pattern, bed form, hydrology, water chemistry, fish, macroinvertebrates, and vegetation; and for lakes: origin of the feature, stratification of the water column, trophic status, salinity, permanency, morphology, landscape position, dominant vegetation type . c) Marine: horizontal zonation, vertical layering, water nutrients and transparency, currents, substrate characteristics, pelagic biota, benthic biota, and anthropogenic influences. d) Estuarine: biogeographical region, extent, shape (geomorphology), mean annual runoff, marine connectivity, sediment processes, mixing processes, habitat (e.g. mangroves, saltmarshes) and biological assemblages. In general, the approach of the South African classification of ecosystem types aligns with the six criteria outlined for a classification system by the IUCN across all four realms. However, the key elements used to define ecosystem types within the classification system by the IUCN aligns more strongly with the terrestrial and estuarine realms and less strongly with the marine and freshwater realms. With the exception of the use of vertical layering as a defining feature of ecosystem types in the marine realm , all other variables have been included in the considerations for defining ecosystem types within the classification systems. Therefore, the South African classification system CcSs of Ecosystems?is aligned with current approaches to defining ecosystem typesecosystems globally .

PART B: APPLICATIONS OF THE SOUTH AFRICAN CLASSIFICATION OF SYSTEM OF ECOSYSTEM TYPESS? IN SOUTH AFRICA National ecosystem types are a powerful unit of reference for a range of purposes linked to research, policy development, management, regulation and decision making. Ecosystem types have been used extensively over the past few decades in assessments of biodiversity. and resource allocation. These foundational resources have played a key role in the identification of priority areas in conservation planning, and the declaration of new protected areas. A summary of established and emerging applications follows.

Applications Key applications informed by the South African classification of ecosystem types ecosystem classification systems and associated ecosystem maps are shown in Figure 1 and described in Table 3. A central motivation for the development of a national ecosystem classification system and map is that it underpins our ability to identify and list threatened or protected ecosystems in terms of the National Biodiversity Act. An initial red list of terrestrial ecosystems was published in the Government Gazette in 2011 ; and further development of the national ecosystem classification system is an essential prerequisite if South Africa is to move forward with Red List Evaluations (RLEv?) in marine and coastal, inland aquatic and estuarine realms. In 2016, a joint publication by SANBI and the World Conservation Monitoring Centre (WCMC) entitled “Mapping Biodiversity Priorities ” succinctly described and illustrated the use of ecosystem type maps in systematic biodiversity prioritisation and assessment processes (Figure 4).

  Table 3. Application and uses of ecosystem classification and mapping systems as illustrated in Fig 1. Process Description Red listing ecosystems (referred to as threat status of ecosystems in South Africa) This is the ecosystem equivalent of red lists for species and provides the basis for listing of threatened ecosystems in terms of the Biodiversity Act. Listed threatened ecosystems are referred to in the EIA regulations, thus directly influencing land‐use decisions, and must be taken into account in municipal Spatial Development Frameworks (SDFs). Assessing protection levels of ecosystems This is fundamental to ensuring that the protected area network includes a representative set of ecosystem types, using a number of thresholds, and directly informs protected area expansion, including national and provincial protected areas expansion strategies. Monitoring and reporting Monitoring and reporting on status and trends in biodiversity at the ecosystem level, for example in the National Biodiversity Assessments (NBAs) (Driver et al. 2012). Without nationally consistent, agreed ecosystem types or units, it is extremely difficult to assess trends and report meaningfully on these over time. Ecosystem types provide the basis for headline indicators for ecosystems. Different hierarchical levels in an ecosystem classification system (e.g. biomes, marine ecoregions) can be used to summarise and report on status and trends at a range of spatial scales for different purposes or audiences. Together with species red-list assessments (RLAs), the ecosystem-level assessment in the NBA informs the core of the reporting to the Convention on Biological Diversity (CBD).
Systematic biodiversity and conservation plans Including ecosystems types in spatial biodiversity and conservation plans that identify geographic priority areas for managing and conserving biodiversity, such as Critical Biodiversity Areas (CBAs) and Freshwater Ecosystem Priority Areas (FEPAs). These feed directly into land‐use planning, marine spatial planning and environmental authorisations, and are essential and powerful tools for mainstreaming biodiversity in a range of socio‐economic sectors. Ecosystem Services Providing ecosystem units that link to ecosystem service flows, both potential and actual (e.g. grazing potential, flood regulation, nutrient sequestration, and erosion control). This can help with formulating models, hypotheses and assumptions relating to the provision of ecosystem services and the link between biodiversity and ecosystem services. Climate change modelling Providing ecosystem units that can be used to analyse climate change distribution shifts, as has been done nationally for biomes and in more detail for fynbos . Biodiversity offsets Implementing biodiversity offsets, in which unavoidable loss of biodiversity as a result of certain types of development is offset by securing broadly equivalent biodiversity elsewhere. Clearly defined ecosystem types provide the scientific foundation for determining when offsets are required, calculating offset ratios and identifying offset receiving areas. Ecosystem management Ecosystem types provide useful management units for agricultural and conservation related activities such as livestock/wildlife management, fire management etc. Ecological Infrastructure Providing ecosystem units that link to ecological infrastructure, the assets that underpin many ecosystem service flows. This can help us to map ecological infrastructure and include it on maps of biodiversity priority areas, and thus to influence planning and decision making in a range of sectors. Natural capital accounting Providing the basis for national ecosystem accounting that can be done for ecosystem assets (stocks) and/or ecosystem services (flows). Ecosystem accounting relies in the first instance on having a clearly defined set of ecosystem units or ecosystem assets, which are accounted for in physical terms. Bogaart et al. (2019), page ?) state that “A classification describing the ecosystem types and a map showing their occurrences in the ecosystem accounting area are essential components of ecosystem accounting as it allows tracking changes in ecosystem assets over time.” Environmental flow allocations (water quantity and quality) Aquatic ecosystem types characterise physical and biotic processes and can thus be used as proxies for predicting sensitivity to anthropogenic pressures such as flow reduction and increased nutrient loading in environmental flow assessment in data-poor environments. Aquatic ecosystem typing is one of the fundamental datasets for extrapolating freshwater flow requirements across a region in low confidence assessments (van Niekerk et al. 2015). Fisheries resource management (aquatic systems) Ecosystem types provides guidance towards fisheries resource management, e.g. sensitivity towards pressures, guidance on harvesting limits, including gear and effort control restrictions. Other research Ecosystem types and higher classification levels are useful stratification layers for sampling processes. Education Higher classification levels such as biomes provide useful context in the curricula of primary, secondary and tertiary education.

Figure 4. Extract from Mapping Biodiversity Priorities (SANBI and UNEP-WCMC 2016) illustrating the how ecosystem typologies and maps are used in planning and assessment at a national scale.

PART C: A DESCRIPTION OF THE SOUTH AFRICAN CLASSIFICATION OF ECOSYSTEM TYPESSYSTEMS FOR ECOSYSTEM [updated December 2019] Note: the names, descriptions, and statistics of the classification systems for the terrestrial, freshwater, inland aquatic, estuarine and marine realms follow the most recent updates to these datasets i.e. end of August 2019 . The need for a South African Classification System for Ecosystem Types (SACSET ) the classification of ecosystem types has largely been driven by concerns over changes in extent and ecological condition of landscapes and seascapes. Even the earliest attempts at classifying the terrestrial landscape in 1929 by Pole-Evans (Gunn 1977 ) was driven by concerns over potential degrading effects of land use practices in the country. The need for the classification and map to be guided by empirical data and expert review became the standard practice across realms and feedback between users and custodians has driven the iterative development and improvement in our mapping and classification systems over time (Table A7).

The A hierarchical structure Fine-scale applications, such as Environmental Impact Assessments (EIAs), and coarse-scale global reporting processes, such as Ecosystem Accounting, have driven a nested hierarchical structure across all classification systems of ecosystem types (Table 4). While there is variation at mid and lower levels of the hierarchy across the realms, all realms have an operational level that is used for applied processes and this level is regarded as the ecosystem type. Furthermore, climatic and biogeographical factors are common to environments at a coarse scale, therefore, congruence at higher levels of the system are inevitable. This is especially true for the biome and/or biogeographical level, the highest levels in the classification hierarchy that are present across all realms. The current classifications should allow for flexibility within the system additional levels can be included at upper and lower levels in the hierarchy. However, existing levels should remain stable over time. Ecosystem types may move position between these levels should new data indicate that the scale at which they have been defined justifies an upgrade or downgrade. However, these decisions should be guided by a defined set of protocols and managed by a committee of informed experts to prevent spurious and radical changes (see Part D).

Table 4. Classification hierarchy structure in each realm. Text highlighted in bold represents the level that is used for many applied processes in South Africa. Levels that correspond across realms are highlighted in matching colours. *Currently inactive due to insufficient data. Numbers followed by a (+) indicate the number of ecosystems from the Prince Edward Islands marine and terrestrial classifications . Terrestrial Marine River Inland wetlands Estuarine Hierarchy level No. units Hierarchy level No. units Hierarchy level No. units Hierarchy level No. units Hierarchy level No. units Realm 1 Realm 1 Realm 1 Realm 1 Realm 1

1. Biome 9 0. * Biome 0. Biome 9 0. Biome 3 0. Biome 9 1. CS_L1 SYSTEM (Marine, Estuarine or Inland aquatic) 0. EFZ 1
2. Bioregion 41 1. Ecoregion 6 1. L1_Ecoregion 30 2. CS_L2 Bioregion Regional setting (37 vegetation bioregions are currently used as a surrogate data to represent this layer. Further research is required to determine unique Landscape units.) 37 1. Biogeography zone 4 2. Bathyregion 15 2. L2 Region 139? 3.CS_L3 Landscape setting (four landform units were used to predict HGM units, including plains, benches, slopes and valleys. Not used in the NBA 2018 anymore. 4 2. Estuary and micro estuary type 9 3. Substratum 59
3. Vegetation type (1:3000-250000) 459 +6 5. Ecosystem type (<1:3000-250000) 150 +29 5. River Type (1:50000) 222 4. CS_L4A HGM unit (1:2000-50000) 4 or 6 3. Ecosystem Type (1:2000-50000?) 22
4. Vegetation subtype 23 3. Flow x? * L4B HGM 4. Estuary name
   332 4. Geozone x? * L4C HGM * L5 Hydrological Regime * L6 Descriptors (geology, natural vs artificial, substrate, vegetation cover etc)

Current classifications and maps of ecosystem types Most current ecosystem classifications and maps were updated concurrently with the NBA 2018 process. Consequently, the ecosystems described for each realm in Table 5 has an assessment for the ecosystem threat status, protection level and several other metrics described in the NBA (accessed here: http://hdl.handle.net/20.500.12143/6370). The Ollis et al. (2013) Inland Aquatic classification system pre-dated the NBA 2018 and the South African classification of Inland Aquatic ecosystem types largely follows this classification. However, this classification is undergoing minor modifications and hence the classification system used for the NBA 2018 was slightly adapted from Ollis et al. (2013) and is fully described in Van Deventer et al. (2018a).

Terrestrial realm

South Africa’s terrestrial realm can be categorised into nine biomes and 459 ecosystem types, approx¬imately 80% of which are endemic. The moist, winter-rainfall region in the southwest of the country is home to the unique Fynbos biome. Adjacent to this lies the Succulent Karoo biome, an arid winter-rainfall biome with the highest diversity of succulent plants in the world. The Nama-Karoo biome covers the arid, summer-rainfall, western interior. The Savanna biome (the largest biome in southern Africa) dominates the northern and eastern summer-rainfall regions of South Africa. The Grassland biome occurs mostly on the cooler, high-lying central plateau and has high levels of endemism. The Albany Thicket biome occurs in the eastern and southern Cape and contains a unique combination of plant forms with an Eocene origin and unique evolutionary history. The Forest biome (with Warm Temperate and Subtropical types) is the smallest biome and is characterised by patches distrib¬uted across the winter and summer rainfall areas of the country. The Indian Ocean Coastal Belt biome represents the southernmost extent of the wet tropical seaboard of East Africa. The Desert biome occupies a small portion of the extreme northwest of the country, forming the southernmost extent of the Namib Desert. South Africa’s sub-Antarctic territory (cross-realm) consists of Prince Edward Island, Marion Island and surrounding seas (collectively known as the Prince Edward Islands, PEIs), and is situated 1 700 km southeast of the mainland. These tiny islands and surrounding seas have a very different biodiversity profile from that of the mainland. The islands are volcanic in origin and experience a cold temperate or polar climate with a strong oceanic influence; with five terrestrial ecosystem types described.

Terrestrial ecosystems (indicates ecosystems in the sub-Antarctic territory) Lineage: National scale vegetation maps compiled by Acocks in the 1950’s and 60’s (Acocks 1988) advanced terrestrial ecosystem classification and mapping for the purposes of rangeland management. Modern maps can trace some concepts and units back to the pioneering work of Acocks - though significant advances have been made through new thinking based on biodiversity concepts, new data and remote sensing products. The protocols for updating and improving the national vegetation classification and map are well established and the product is well accepted and utilised (Dayaram et al. 2017). Current version: National Vegetation Map of South Africa 2018 (Dayaram et al. 2019 is based on the Vegetation of South Africa, Lesotho and Swaziland, compiled and edited by Mucina and Rutherford in 2006. The map includes South Africa’s sub-Antarctic islands of Marion and Prince Edward which follows the same classification system as the mainland ecosystems. Number of units: The classification system recognises 459 unique vegetation types are recognised for the mainland and five for the terrestrial sub-Antarctic territory. This level of the classification system is used in ecosystem assessments and red-listing. However, one mainland ecosystem type is excluded from assessments as it occurs exclusively in Lesotho. Hierarchical levels: Four hierarchical levels are recognised in the classification system Biomes (9 + 2), Bioregions (41), Vegetation Types (459 + 5*) and SubTypes (23)

Challenges and limitations: Vegetation types are mapped at difference scales across biomes (e.g. 1:3000 for vegetation types along the coast, to 1:50000 for some vegetation types in the Nama Karoo Biome). This disparity in scale should be considered when the data and classification is used. Current assessment: The national vegetation map of South Africa (NVM 2018) has been updated to: a. include an improved forest map; b. remove inland wetlands and estuarine lagoon ecosystems – that are better represented in the dedicated National Wetland Map version 5 (NWM5; Van Deventer et al., 2018a;b) and Classification System for Aquatic and other inland wetland systems (Ollis et al. 2013; 2015); c. include additional thicket ecosystem types based on Vlok et al. (2002); d. include edits to maps of various ecosystem types; and e. align with a new coastal ecosystems classification system and map (Harris et al. 2019).
Previous assessment: For the 2011 National List of Threatened Terrestrial Ecosystems- a Red List of Ecosystems using the South African system (Botts et al. in review) - the 2006 vegetation map was supplemented with the national forest types recognised by the Department of Agriculture, Forestry and Fisheries (Von Maltitz et al. 2003; Berliner 2005; Berliner 2008) and priority areas identified in provincial systematic biodiversity plan for three provinces . Future steps: Improvements and updates to the vegetation map are ongoing. Future plans are to maintain integration along the marine/estuarine/terrestrial interface. Further refinements in the Forest Biome are currently being explored, as well as possible contributions from existing vegetation maps in the North-West and Free State provinces.

 

Marine realm

South Africa’s marine realm includes the Atlantic, Indian and Southern Oceans with the contrasting cold Benguela upwelling system and the warm, fast-flowing Agulhas current system. This diverse oceanographic setting, combined with complex geology and topog¬raphy, drives exceptional marine biodiversity and a wide array of ecoregions and ecosystem types. Three shelf ecoregions are recognised; the Cool Temperate Southern Benguela, the Warm Temperate Agulhas and the Subtropical Natal-Delagoa. The deep ocean beyond the shelf edge includes two ecoregions: The Southeast Atlantic and the South¬west Indian. The Southern Benguela comprises two sub-regions, the Namaqua and Cape regions that separate at Donkin Bay (north of St Helena Bay) on the west coast. In addition, the Natal-Delagoa ecore¬gion includes the Delagoa, KwaZulu-Natal Bight and KwaZulu-Natal–Pondoland regions, which have distinct biodiversity patterns. These ecoregions and sub-regions include 150 marine ecosystem types that include several functional ecosystem groups: Sandy Shores, Rocky and Mixed Shores, Islands, Bays, Kelp Forests, Soft Shallow Shelf, Shallow Reef and Rocky Shelf, Deep Soft Shelf, Deep Rocky Shelf, Slope, Plateau and Abyss. South Africa’s sub-Antarctic territory (cross-realm) consists of Prince Edward Island, Marion Island and surrounding seas (collectively known as the Prince Edward Islands, PEIs), and is situated 1 700 km southeast of the mainland. These tiny islands and surrounding seas have a very different biodiversity profile from that of the mainland and its oceans. The islands are volcanic in origin and experience a cold temperate or polar climate with a strong oceanic influence. There are 29 marine ecosystem types covering the shore, the territorial waters and Exclusive Economic Zone, and these range from temperate ecoregions in the north to polar ecoregions in the south. As part of the Southern Ocean, our sub-Antarctic marine ecosystems contribute to a globally important carbon sink and play an integral role in climate regulation.

Marine ecosystems (*indicates ecosystems in the sub-Antarctic territory) Lineage: The first national scale marine ecosystem classification system and map, produced by Lombard et al. (2005), identified 34 broad marine biozones based on advice from several regional expert workshops to define ecoregions and identify key depth divisions. Sink et al. (2012) followed this with a more detailed marine benthic and coastal habitat map combined with pelagic biozones from Lagabrielle (2009) for the NBA 2011. The sub‐Antarctic ecosystems in South Africa’s marine territory linked to the PEI had not been included in the marine and coastal ecosystem classification at that stage, although Lombard et al. (2007) collated and mapped both species and ecosystem features in work to design the Prince Edward Island Marine Protected Area. In the latest update to the marine ecosystem map (2018) existing datasets were revisited and spatially comprehensive map of ecosystem types was produced for both the mainland and Sub-Antarctic territory, and was based on a combination of benthic and pelagic biota, and abiotic processes and features.

Factors used to classify ecosystem types included: • biogeography (ecoregions, such as Namaqua, Agulhas, Delagoa). Including large scale patterns linked to sea temperature, productivity and chlorophyll • Depth (recognised as the greatest driver of pattern, with many variables including temperature, light, oxygen and wave energy changing with depth), • substrate and grain size (e.g. rocky, sandy, muddy, gravel, mixed), • wave exposure (sheltered, exposed or very exposed), • pelagic patterns and features such as frequency of eddies, and distribution of temperature and chlorophyll fronts • riverine input and turbidity • The classification includes the taxa and processes in both the benthic and the pelagic stratum of the ocean. Current version: Marine Ecosystem Map 2018 developed for the National Biodiversity Assessment 2018 (Sink et al. 2019) Key changes in the 2018 classification and map of marine ecosystem types include: • amalgamation of offshore benthic and pelagic habitat types into a single ecosystem type; • improved fine-scale shore mapping with alignment and integration between estuarine, marine and coastal vegetation types in the coastal zone (Harris et al. 2019b) • the inclusion of kelp forests (LV. Dunga, University of Cape Town MSc thesis in prep.), bays and stromatolites (Perissinotto et al. 2014) • the introduction of finer-scale depth strata across shelves (inner, mid and outer shelf) and on the slope (upper, mid and lower slope); • improved bathymetric data and refinement of boundaries between bathomes (depth zones); • refined biogeographic information including the introduction of a new KwaZulu-Natal Bight sub-region and an extension of the Cape ecoregion; • recognition of ecosystem types that are mosaics of repeating patterns of both consolidated and unconsolidated habitats on the shelf • consideration of fluvial fans, plumes and river-influenced ecosystem types in the offshore environment; and • consideration of substantial new datasets to inform the subtidal classification of ecosystems.

Hierarchical levels: The highest level in the classification is the Ecoregions (6 + 4*) followed by Bathometric regions (15), Seabed regions (59), with the ecosystem types (150 + 29*) at the lowest level.

Number of units: 150 marine ecosystem types. The Prince Edward Islands marine ecosystems added an additional 29 ecosystem types. Challenges and limitations: Efforts to spatially delineate and classify the marine environment are very challenging. This is due to the inaccessible marine environment, technical difficulties of sampling in the ocean where depth, pressure, strong currents and darkness demand additional innovation and the high costs involved. Future steps: Updates to the Marine Ecosystem Map will be ongoing.

Inland aquatic realm

South Africa is among the most water-scarce coun¬tries per capita in the world, and has a high temporal and spatial variability of rainfall. This results in highly variable runoff and river flow regimes, and a relative scarcity, but surprisingly rich variety, of inland wetlands. The diversity of river and inland wetland ecosystem types (together comprising the inland aquatic realm) is underpinned by the strongly contrasting bioclimatic zones – the arid western interior (summer rainfall), the mesic eastern grassy biomes (summer rainfall), the arid western coastal regions (winter rainfall) and the mesic winter-rainfall southwestern Cape. The South African Inventory of Inland Aquatic Ecosystems (SAIIAE) is a collection of datasets which spatially depicts the extent of river and inland wetland ecosystem types, as well as the representation of other water features, artificial systems, pressures and protection of these systems (Van Deventer et al. 2018b).

Rivers Lineage: The first national map of river ecosystem types was developed for the NSBA 2004 (Nel et al., 2004). A second map, taking a slightly different approach, was developed as part of the National Freshwater Ecosystem Priority Areas project (NFEPA) and published in the Atlas of Freshwater Ecosystem Priority Areas for South Africa (Nel et al. 2011). Current version: Hierarchical levels: River ecosystem types were delineated based on three factors: • 31 freshwater ecoregions (e.g. the Highveld which has flat plains and gentle meandering rivers, the Eastern Coastal Belt which has steeply incised rivers and confined valleys) • Three flow variability classes (i.e. perennial, seasonal or ephemeral) • Four slope categories or longitudinal zones (mountain streams, upper foothills, lower foothills and lowland rivers)

The 31 ecoregions are considered the highest level in the hierarchy, the ecoregions combined with the flow variability represent the second level and the third level is the combination of ecoregion, flow and slope. Number of units: These three factors were combined to identify 223 river ecosystem types that represent the riverine biodiversity of the country. Sections of different rivers that fall within the same river ecosystem type are likely to share broadly similar ecological characteristics and functioning. Challenges and limitations: Representation of the rivers spatially is a challenging aspect of this realm. In the current system the rivers are mapped as polylines with stream order at two scales (1:500 000 and 1: 50 000). River area is not considered explicitly and riverine ecosystems are linked to the polylines through buffers. Future steps: There is considerable scope to revise and refine the river ecoregions and the River Ecosystem Classification Committee has made this task a priority for 2018 /2019. River ecosystem types need to be ground-truthed, and may need to be refined. Biological survey data should be improved and included in this process.

Inland wetlands

Lineage: The lineage of the NWM5 is described in the SAIIAE report (Van Deventer et al., 2018b). The historic NWMs resulted from a combination of data modelled from remote sensing and heads-up digitising. NWM5 improved on these previous versions through removing overestimation (commission) errors and improving the representation (reducing omission errors) through integrating fine-scale datasets, and digitising wetlands in 13 focus areas (Van Deventer et al., 2018a;b; in review). Current version: National Wetland Map version 5 (NWM5)

Hierarchical levels: See Table 4. At the first level of the Classification System for Wetlands and other Aquatic Ecosystems in South Africa (Ollis et al. 2013; 2015), inland aquatic systems are distinguished from marine and estuarine systems; The second level distinguish the regional setting. While the proper representation of these areas still need to be developed, 37 vegetation bioregions (Dayaram et al., 2019) were used as surrogates for the NBA 2018 to describe the first part of the inland wetland ecosystem types; the third level distinguish the landscape setting. Landforms (benches, valley floors, slopes and plains) were originally included to predict the hydrogeomorphic (HGM) units as ecosystem types in the NFEPA project, and used in the NBA 2011. This approach was no longer needed for NWM5, used in the NBA 2018, since the inland wetland ecosystem types were improved directly at Level 4A of the Classification System for Wetlands and other Aquatic Ecosystems in South Africa; At the fourth level’s first sub-level, HGM units describe the second part of the inland wetland ecosystem types. Owing to uncertainties in accuracy of representing four (channelled valley-bottom; unchannelled valley-bottom; depression and flats) of the six units (seeps and floodplains) at a desktop level, these four units were amalgamated into valley-bottom and depression wetlands, respectively, for the NBA 2018, while the NWM5 depicts all six classes. The 5th and 6th level, the hydrological regime and other descriptors have only been partly populated in NWM5, however, were not used for the NBA 2018.

Number of units: The 37 vegetation biomes were combined with the four amalgamated HGM units resulting in the spatial representation of 135 inland wetland ecosystem types. Eight limnetic depressions were also identified and assessed separately for their ETS and EPL. Challenges and limitations: The representation of the regional setting needs to be researched and developed, while continuous improvement to the HGM units is required. It is estimated that inland wetlands are highly underrepresented and the accuracy of HGM units are low for nearly 70% of the extent of the country. Consequently, updates to these two levels will change the extent and number of inland wetland ecosystem types for future NBAs. Furthermore, since the National Wetland Monitoring Programme (NWMP) is not yet operational, the ecological condition of inland wetlands had to be modelled, limiting the confidence of the assessment. Future steps: The improved representation of the inland wetland ecosystem types, through improved representation of the Regional Setting at Level 2 and HGM units at Level 4A. The depth of the eight limnetic depressions needs to be confirmed by the Department of Water and Sanitation. Modelling the ecological condition of inland wetlands require a process of systematic, in-field verification to determine the accuracy and confidence.

Estuarine realm

South Africa has 290 estuaries and 42 micro-estuaries, which have been classified into 22 estuarine ecosystem types and three micro-system types. This represents a high diversity of estuary types stemming from diverse climatic, oceanographic and geological drivers. The comparatively small, wave-dominated South African estuaries generally have restricted inlets, with more than 75% closing for varying periods when a sandbar forms across the mouth. Four bioregions apply to South African estuaries: the Cool Temperate (Orange to Ratel), the Warm Temperate (Heuningnes to Mendwana), the Subtropical (Mbashe to St Lucia) and the Tropical (uMgobezeleni to Kosi).

Estuarine ecosystems Lineage: Prior to 2011, spatial data for most of South Africa’s estuaries was limited to point data along the coastline (Turpie 2004). A major advance in the NBA 2011 (Van Niekerk and Turpie 2012) was to map the area of the estuarine functional zone for each of the country’s 291 estuaries, including the open water area as well as its associated floodplain. The NSBA 2004 used 13 estuary groups based on Whitfield’s 1992 South African Estuarine Classification System. In the NBA 2011, estuaries were classified into 46 estuarine ecosystem types.

The most recent iteration is a brand new classification system based on a combination of estuary size, marine connectivity, geomorphology, climate biogeography. Micro-estuaries have also been mapped and classified and areas influence by rivers and other off-shore process have also now been mapped. The estuary ecosystem types were based on four factors: • estuary size, which is related to a range of estuarine characteristics such as tidal flows and diversity of habitat types within the estuary, • marine connectivity i.e. whether the estuary is permanently open to the sea or whether it closes from time to time, • prevailing current, salinity structure, which refers to whether the estuary is marine dominated, freshwater • geomorphology, influences runoff from the catchment as well as the volume of the estuary, type of freshwater input from the catchment, which can be predominantly clear, turbid, or black water (from tannin rich, nutrient poor rivers). • Climate (with a focus on freshwater input); and • Biogeography biogeographical zonation that divide the South African coast into four major zones; the Cool Temperate (Orange to Ratel), the Warm Temperate (Heuningnes to Mendwana), the Subtropical (Mbhashe to St Lucia) and the Tropical (uMgobezeleni to Kosi), the latter being a new addition to the estuarine biogeographical provinces. Current version: National Estuary Map 2018 (van Niekerk et al in press) Hierarchical levels: Four hierarchical levels are recognised in the classification system Biogeographical zones (4), Estuary Types (9), Estuary Ecosystems (22) and Estuary Name (332)

Number of units: 22 types 332 features Challenges and limitations: While the classification focused on developing a data-driven classification system, key datasets were found to be lacking that could facilitate a more detailed assessment. Critical parameters that would have allowed for a more nested or hierarchal classification include information on seasonal salinity regimes, mouth states, water clarity, estuary topography and bathymetry, and sediment structure. Accurate records/observations of the duration of the closed mouth condition is especially critical for future classification updates. In addition, very little information was available on the invertebrate community for most of South Africa estuaries. Overall, data on biological responses are more than 30 years old (e.g. national bird counts date from the 1980s) and national fish surveys date from the 1990s. While new data are being collected on some of the larger systems, very little information is currently being gathered on the numerous smaller estuaries, especially those in remote areas along the coast. Without a major investment in baseline information from numerous poorly studied South African estuaries, it will be difficult to refine and possibly expand this new classification system in the future. Future steps: Future steps include the ongoing refinement of estuarine vegetation maps by Janine Adams and team (Nelson Mandela University) within the EFZ, initiating national surveys of key abiotic features (e.g. sediment structure and bathymetry) and biotic assemblages (e.g. invertebrate, fish and birds), investigating remote sensing as a method for estimating seasonal open water area changes, and LiDAR surveys to improve topography data.

 

Coast

For the NBA 2018, an ecologically determined, cross-realm coast was defined and used, which spans the terrestrial, estuarine and marine realms. The South African coast is microtidal (<2-m tide range) and mostly high energy, with generally exposed to very exposed conditions from the Tropical/Subtropical northeast coast to the Cold Temperate west coast. It comprises a wide range of coastal vegetation types (from forests to arid shrublands), dunes, cliffs, beaches, rocky and mixed shores, estuaries, mangroves, kelp forests, reefs, bays, and river-influenced shelf regions that extend as far offshore as the shelf edge in some places. With this heterogeneity comes exceptionally high coastal biodi¬versity and high levels of endemism, especially among dune plants, beach fauna and other invertebrate taxa. There are 189 ecosystem types that are considered coastal: 25 estuarine, 79 terrestrial and 85 marine, all of which are fundamentally influenced by both the land and sea. The coastal zone is a secondary sub-classification of features that have been delineated in the terrestrial, marine and estuarine maps. Hence the coastal map draws on existing polygons affected by coastal processes from the respective realm maps.

Coastal ecosystems Lineage: Until the NBA 2018, the coast was not recognised as a cross-realm ecological entity. The seams of the terrestrial, estuarine and marine maps of ecosystem types did not align along the shore, precluding any coast-specific biodiversity planning and assessment. Effectively, the shores were the only ecosystem types that were considered coastal, and were included as part of the marine classification system. However, a new coastal ecosystem classification system and map were respectively defined and delineated for the NBA 2018 (Harris et al. 2019). It includes a fine-scale seashore (backshores and shores) zone that seamlessly stitches the realm maps together, and that includes all terrestrial, estuarine and marine ecosystem types influenced by both terrestrial and marine ecological processes in an ecologically defined coastal zone (see Harris et al 2019 for more information). Current version: Integrated Coastal Map 2018 Hierarchical levels: The coast inherits the hierarchies of the respective realms from which each ecosystem type is drawn, with shared levels for which all realms have information being those included in the coast hierarchy: i.e., realm, biogeography (bioregion/ecoregion/biogeography zone), and ecosystem type. There is also an intermediate classification level below realm, called sub-realm based on the proximity of ecosystem types to the shore. The coastal classifiers are thus: 1. Realm (3: terrestrial, estuarine, marine), 2. Sub Realm (5: semi coastal (terrestrial), coastal (terrestrial), estuarine (EFZ and estuarine shore), shore (marine), inner shelf and river influenced (marine), 3. ecosystem type (same as ecosystem types in their respective realms), 4. Coastal subtype (a further classification of ecosystem types exists, currently for terrestrial seashore ecosystems only, e.g. barcanoid dunes)

Number of units: Terrestrial (79), Marine (85), Estuarine (25). Challenges and limitations: An ongoing challenge will be to ensure alignment of between the marine, estuarine and terrestrial maps as they move forward with future iterations. Organisational systems (e.g. committee leads) have been set up to ensure that any changes along the interface between these maps occurs as a collaborative effort. Limitations to the current map include: (1) coastal/semi-coastal vegetation types are included in entirety based on proximity to the shore, with no current evidence of or sub-typing based on specific dependence or influence from marine processes, with the result that there are some gaps in the terrestrial coastal zone even immediately adjacent to the shore; (2) there is no national classification system for mixed shores nor field sampling on which this and an assessment of biogeography can be based; (3) exposure for rocky shores is based on expert opinion, and modelling studies have highlighted some inaccuracies; (4) there are some questions regarding sandy beach biogeography, with the current data not supporting the same number of ecoregions as the other shore ecosystem types that in turn proves challenging for constructing classification hierarchies; (5) the extent of estuarine plumes and fans needs further work to improve mapping of river-influenced ecosystem types. Future steps: Address the five limitations listed above.

Sub-Antarctic territory

South Africa’s sub-Antarctic territory (cross-realm) consists of Prince Edward Island, Marion Island and surrounding seas (collectively known as the Prince Edward Islands, PEIs), and is situated 1 700 km southeast of the mainland. These tiny islands and surrounding seas have a very different biodiversity profile from that of the mainland and its oceans. The islands are volcanic in origin and experience a cold temperate or polar climate with a strong oceanic influ¬ence; with five terrestrial ecosystem types described. There are 29 marine ecosystem types covering the shore, the territorial waters and Exclusive Economic Zone, and these range from temperate ecoregions in the north to polar ecoregions in the south. As part of the Southern Ocean, our sub-Antarctic marine ecosystems contribute to a globally important carbon sink and play an integral role in climate regulation.

PLACEHOLDER SECTION IF NEEDED PART D: RECOMMENDED PROCESSES, STRUCTURES AND PROTOCOLS The following key recommendations of processes, structures and protocols have emerged out of several decades of development. They are A few recommendations are described in more detail below. Other processes such as a) an integrated approach has been discussed previously in section 2.1 . a) An integrated approach b)a) A governance structure c)b) Transparent mapping and classification protocols including protocols for proposing changes d)c) A managed network of contributors from diverse sectors

Governance structures National-scale ecosystem classification initiatives in South Africa have progressed as collaborative, iterative bodies of work since the early 2000’s. The national vegetation map committee was set up in 2007 following the publication of The Vegetation of South Africa, Lesotho and Swaziland (Mucina and Rutherford 2006); after a period of inactivity, it continues to meet annually to guide national vegetation map updates and revisions. In the marine, inland aquatic and estuarine realms the National (Spatial) Biodiversity Assessment processes in 2004 and 2011 provided impetus and resources for the improvement of national scale ecosystem classification systems and maps. During this initial “NBA era”, conceptual consistency across the realms (terrestrial, inland aquatic, marine, and estuarine) became important, as has the integration and “edge-matching” of the different maps.
A major recommendation of the NBA 2011 was to formalise the national ecosystem classification and mapping committees for each realm, and for SANBI to facilitate this process. In 2015 four new Ecosystem Classification Committees (ECC) were set up: Marine and Coastal ECC, Estuarine ECC, Inland Wetland ECC, and the River ECC. The aim of the ECCs is to oversee and guide revisions and updates of national ecosystem classification systems and maps. The ECCs are a powerful way of building collaborative relationships in the sector, and a way to ensure wide support and buy in for the maps and classification systems that result. In addition to the realm specific ECCs, an overarching national ecosystem classification committee (NECC) was initiated in 2016 to facilitate inter realm collaboration and ensure that the ECCs operate in unison (Table 6). Table 6. National ecosystem classification (and mapping) committees Committee Secretariat Funding National Ecosystem Classification Committee SANBI SANBI National Vegetation Map Committee SANBI SANBI Marine & Coastal Ecosystem Classification Committee SANBI SANBI Estuarine Ecosystem Classification Committee CSIRSANBI (seconded to the CSIR for the NBA 2018) SANBI River Ecosystem Classification Committee CSIRSANBI (seconded to the CSIR for the NBA 2018) SANBI Wetland Ecosystem Classification Committee SANBI SANBI

Documented mapping and classification approaches including protocols for proposing changes Documented protocols ensure that rules for defining and delineating classes within the hierarchy are available for validation by reviewers. Documentation also allows transparency in the classification and mapping process, which can be viewed by potential contributors and is an important step in creating and enabling environment for feedback between users of the system and system curators. Protocols need to be explicit and provide a framework for the mapping and classification to be repeatable and for changes to be implemented using corresponding approaches to avoid mismatches between the original and updated versions.

A managed network of contributors from diverse sectors Creating a network of willing contributors between national-, provincial- and local-scale ecosystem classifiers and mappers ensures that data from across different authoritative scales are cross-walked into a single national system. This allows for assessment, conservation planning and prioritisation that harnesses knowledge from all sectors in the community of practice while minimising conflict between spatial scales.

Current research and future research needs Research needs to improve the classification systems and maps of each realm, and the integration of the realms - to provide strategic guidance for the range of organisations and institutions whose work contributes to or could contribute to building this science foundation. Field validation of ecosystem maps is required. In particular, the marine maps require in field validation using trawl surveys, ROVs, grab samples and other methods. Research priorities are reflected in the NBA 2018 and there is a need to advance to a more data driven classification. Terrestrial systems benefit for have good high-resolution imagery coverage that allows for “desk top” validation. Analysis of species data from field surveys to validate the ecosystem type delineations made using remote sensing and modelling approaches. For the vegetation map, a large database of underutilised plot data is currently being prepared (National Vegetation Database) for this purposes; for other ecosystem classifications field surveys are required to test many of the types for the first time. For estuaries, in addition to key datasets on key abiotic parameters (seasonal salinity regimes, mouth state/connectivity, water clarity and sediment structure) and biotic assemblages (plants, invertebrate, fish, and birds) critical spatial information is needed on estuary topography / bathymetry. Coastal LiDAR data that is accurately corrected to mean seal level (by means of ground truthing) with the vegetation signal removed would go a long way to addressing mapping uncertainty. Remote sensing should also be investigated as a means of generating a long term open water area record (hydro period). The use of drone technology should be investigated as a means of improving mapping accuracies in the upper reaches of estuaries where current topological maps under estimate the estuary topography and extent of tidal influence.
REFERENCES

1. Acocks, J.P.H. (1988) Veld types of South Africa, 3rd edn. Memoirs of the Botanical Survey of South Africa, No. 57. Botanical Research Institute, Pretoria
2. Berliner, D. (2005) Systematic conservation planning for the forest biome of South Africa –Approach, methods and results of the selection of priority forests for conservation action. Department of Water Affairs and Forestry, Pretoria. 108p.
3. Berliner, D. (2008) Revision of forest conservation targets and data preparation for the refinement of SANBI criteria for the listing of threatened forest ecosystems under the Biodiversity Act: Phase 1 and 2. Report WP976. Department of Water Affairs and Forestry, Pretoria. 28p.
4. Bogaart, P., Chan, J.Y., Horlings, H., Keith, D., Larson, T., Sayre, R., Schenau, S., Soulard, F. (2019). Discussion paper 1.1: An ecosystem type classification for the SEEA EEA. Paper submitted to the SEEA EEA Technical Committee as input to the revision of the technical recommendations in support of the System on Environmental-Economic Accounting. Version of 29 April 2019.
5. Botts, E.A., Skowno, A., Driver, A., Holness, S., Maze, K., Smith, T., Daniels, F., Desmet, P., Sink, K., Botha, M., Nel, J. & Manuel, J. (under review) Integration of South Africa’s threatened ecosystems into conservation planning and environmental policy.
6. Dayaram, A., Powrie, L., Rebelo, T. & Skowno, A. (2017) Vegetation Map of South Africa, Lesotho and Swaziland 2009 and 2012: A description of changes from 2006. Bothalia-African Biodiversity & Conservation, 47(1):1-10.
7. Dayaram, A., Harris, L.R., Grobler, B.A., Van der Merwe, S., Powrie, L.W., Rebelo, A.G., Vlok, J.H.J., Desmet, P.G., Qabaqaba, M., Hlahane, K.M., & Skowno, A. (2019) Vegetation Map of South Africa, Lesotho and Swaziland 2018: A description of changes from 2006. Bothalia, 49(1), a2452. https://doi.org/10.4102/abc.v49i1.2452.
8. Driver, A., Maze, K., Rouget, M., Lombard, A.T., Nel, J., Turpie, J.K., Cowling, R.M., Desmet, P., Goodman, P., Harris, J., Jonas, Z., Reyers, B., Sink, K. & Strauss, T. (2005) National Spatial Biodiversity Assessment 2004: priorities for biodiversity conservation in South Africa. Strelitzia 17. South African National Biodiversity Institute, Pretoria, South Africa.
9. Driver, A., Sink, K.J., Nel, J.N., Holness, S., Van Niekerk, L., Daniels, F., Jonas, Z., Majiedt, P.A., Harris, L. & Maze, K. (2012) National Biodiversity Assessment 2011: An assessment of South Africa’s biodiversity and ecosystems. Synthesis Report. South African National Biodiversity Institute and Department of Environmental Affairs, Pretoria.
10. Guo, D., Desmet, P.G. and Powrie, L.W. (2017) Impact of the Future Changing Climate on the Southern Africa Biomes, and the Importance of Geology. Journal of Geoscience and Environment Protection, 5, 1-9.
11. Gunn, M.D. 1971. Illtyd Buller Pole Evans (1879- 1968). Bothalia, African Biodiversity and Conservation, 10, 2, 131-135. DOI:10.4102/abc.v10i2.1525
12. Harris, L.R., Bessinger, M., Dayaram, A., Holness, S., Kirkman, S., Livingstone, T., Lombard, A.T., Lück-Vogel, M., Pfaff, M., Sink, K.J., Skowno, A. L., Van Niekerk, L., 2019. Advancing land-sea integration for ecologically meaningful coastal conservation and management. Biological Conservation, 237, 81-89.
13. Harris, L.R., Nel, R., Schoeman, D.S., 2011. Mapping beach morphodynamics remotely: A novel application tested on South African sandy shores. Estuarine, Coastal and Shelf Science 92, 78-89.
14. IUCN (2016). An Introduction to the IUCN Red List of Ecosystems: The Categories and Criteria for Assessing Risks to Ecosystems. Gland, Switzerland: IUCN. vi + 14pp.
15. Keith, D. A., J. P. Rodríguez, K. M. Rodríguez-Clark, E. Nicholson, K. Aapala, A. Alonso, M. Asmüssen, S. Bachman, A. Basset, E. G. Barrow, J. S. Benson, M. J. Bishop, R. Bonifacio, T. M. Brooks, M. A. Burgman, P. Comer, F. A. Comín, F. Essl, D. Faber-Langendoen, P. G. Fairweather, R. J. Holdaway, M. Jennings, R. T. Kingsford, R. E. Lester, R. M. Nally, M. A. McCarthy, J. Moat, M. A. Oliveira-Miranda, P. Pisanu, B. Poulin, T. J. Regan, U. Riecken, M. D. Spalding and S. Zambrano-Martínez (2013) Scientific Foundations for an IUCN Red List of Ecosystems. PLoS ONE 8(5)
16. Sink KJ, Harris LR, Skowno AL, Livingstone T, Franken M, Porter S, Atkinson LJ, Bernard A, Cawthra H, Currie J, Dayaram A, de Wet W, Dunga LV, Filander Z, Green A, Herbert D, Karenyi N, Palmer R, Pfaff M, Makwela M, Mackay F, van Niekerk L, van Zyl W, Bessinger M, Holness S, Kirkman SP, Lamberth S, Lück-Vogel M. 2019. Chapter 3: Marine Ecosystem Classification and Mapping. In: Sink KJ, van der Bank MG, Majiedt PA, Harris LR, Atkinson LJ, Kirkman SP, Karenyi N (eds). 2019. South African National Biodiversity Assessment 2018 Technical Report Volume 4: Marine Realm. South African National Biodiversity Institute, Pretoria. South Africa. http://hdl.handle.net/20.500.12143/6372
17. Lagabrielle, E. 2009. Preliminary report: National Pelagic Bioregionalisation of South Africa. Cape Town: South African National Biodiversity Institute.
18. Lombard, A.T., Strauss, T., Harris, J., Sink, K., Attwood, C. & Hutchings, L. 2004: South African National Spatial Biodiversity Assessment 2004: Technical Report. Volume 4: Marine Component. Pretoria: South African National Biodiversity Institute.
19. Low, A.B. and Rebelo, A.T.G (eds) (1996) Vegetation of South Africa, Lesotho and Swaziland. Department of Environmental Affairs and Tourism, Pretoria.
20. Mucina, L., 2019. Biome; evolution of a crucial ecological and biogeographical concept. New Phytologist.222, 97-114. doi: 10.1111/nph.15609
21. Mucina, L. & Rutherford, M.C., 2006. The vegetation of South Africa, Lesotho and Swaziland. Strelitzia 19. South African National Biodiversity Institute, Pretoria.
22. Nel, J.L., Maree, G., Roux, D., Moolman, J., Kleynhans, N., Silberbauer, M. & Driver, A., (2004). South African National Spatial Biodiversity Assessment 2004: Technical Report. Volume 2: River Component. CSIR Report Number ENV-S-I-2004-063. Stellenbosch: Council for Scientific and Industrial Research.
23. Nel, J.L. & Driver, A., 2012. National Biodiversity Assessment 2011: Technical Report. Volume 2: Freshwater Component. CSIR Report Number CSIR/NRE/ECO/IR/2012/0022/A. Council for Scientific and Industrial Research, Stellenbosch.
24. Nel, J.L., Driver, A., Strydom, W.F., Maherry, A., Petersen, C., Hill, L., Roux, D.J., Nienaber, S., Van Deventer, H., Swartz, S. & Smith‐Adao, L.B. 2011. Atlas of Freshwater Ecosystem Priority Areas in South Africa. WRC Report No. TT 500/11. Water Research Commission, Pretoria.
25. Ollis, D.J., Ewart-Smith, J.L., Day, J.A., Job, N.M., Macfarlane, D.M., Snaddon, C.D., Sieben, J.J., Dini, J.A., & Mbona, N. (2015) The development of a classification system for inland aquatic ecosystems in South Africa. Water SA 41 (5) 727–745.
26. Ollis, D.J., Snaddon, C.D., Job, N.M. & Mbona, N. 2013. Classification System for Wetlands and other Aquatic Ecosystems in South Africa. User Manual: Inland Systems. SANBI Biodiversity Series 22. South African National Biodiversity Institute, Pretoria.
27. SANBI & UNEP-WCMC. 2016. Mapping biodiversity priorities: A practical, science-based approach to national biodiversity assessment and prioritisation to inform strategy and action planning. UNEP-WCMC, Cambridge, UK.
28. Sink, K.J., Holness, S., Harris, L., Majiedt, P.A., Atkinson, L., Robinson, T., Kirkman, S., Hutchings, L., Leslie, R., Lamberth, S., Kerwath, S., Von der Heyden, S., Lombard, A.T., Attwood, C., Branch, G., Fairweather, T., Taljaard, S., Weerts, S., Cowley, P., Awad, A., Halpern, B., Grantham, H. & Wolf, T. 2012. National Biodiversity Assessment 2011: Technical Report. Volume 4: Marine and Coastal Component. South African National Biodiversity Institute, Pretoria
29. Skowno, A.L., Poole, C.J., Raimondo, D.C., Sink, K.J., Van Deventer, H., Van Niekerk, L., Harris, L.R., Smith-Adao, L.B., Tolley, K.A., Zengeya, T.A., Foden, W.B., Midgley, G.F. & Driver, A. 2019. National Biodiversity Assessment 2018: The status of South Africa’s ecosystems and biodiversity. Synthesis Report. South African National Biodiversity Institute, an entity of the Department of Environment, Forestry and Fisheries, Pretoria. pp. xx–xx. http://hdl.handle.net/20.500.12143/6362
30. Turpie, J.K. 2004. South African National Spatial Biodiversity Assessment 2004: Technical Report. Volume 3: Estuary Component. Pretoria: South African National Biodiversity Institute. Van Niekerk and Turpie 2011 NBA Estuaries
31. Van Deventer, H., Nel, J.L., Mbona, N., Job, N. Ewart-Smith, J. and Maherry, A. (2016) Desktop classification of inland wetlands for systematic conservation planning in data-scarce countries: mapping wetland ecosystem types, disturbance indices and threatened species associations at country-wide scale. Aquatic Conservation: Marine and Freshwater Ecosystems 26 57-75.
32. Van Deventer, H, Smith-Adao, L., Collins, N.B., Grenfell, M., Grundling, A., Grundling, P-L., Impson, D., Job, N., Lotterr, M., Ollis, D., Peterson, C., Scherman, P., Sieben, E.J.J., Tererai, F. and Van der Colff, D. (2019) South African National Biodiversity Assessment 2018: Technical Report. 3. Volume 2: Inland Aquatic (Freshwater) Component. Council for Scientific and Industrial Research (CSIR) and South African National Biodiversity Institute (SANBI), Pretoria, South Africa. CSIR report number CSIR/NRE/ECOS/IR/2019/0004/A. SANBI handle: http://hdl.handle.net/20.500.12143/6230.
33. Van Deventer, H.; Smith-Adao, L.; Mbona, N.; Petersen, C.; Skowno, A.; Collins, N.B.; Grenfell, M.; Job, N.; Lötter, M.; Ollis, D.; Scherman, P.; Sieben, E.; Snaddon, K. 2018a. South African Inventory of Inland Aquatic Ecosystems (SAIIAE). Version 2, released on 2018/11/06. South African National Biodiversity Institute, Pretoria. Report Number: CSIR report number CSIR/NRE/ECOS/IR/2018/0001/A; SANBI report number http://hdl.handle.net/20.500.12143/5847.
34. Van Deventer, H, Smith-Adao, L., Peterson, C., Mbona, N., Skowno. & Nel, J.L. (2018b) Review of available data for a South African Inventory of Inland Aquatic Ecosystems (SAIIAE). Water SA 44 (2) 184–199. DOI: http://dx.doi.org/10.4314/wsa.v44i2.05.
35. Van Deventer, H., Van Niekerk, L., Adams, J., Dinala, M.K., Ramjukadh, C-L., Skowno, A. and Weerts, S.P. (In review) National Wetland Map 5 – An improved spatial extent and representation of inland aquatic and estuarine ecosystems in South Africa. BioRxiv DOI: http://biorxiv.org/cgi/content/short/640441v1.
36. Vlok, J.H.J. and Euston-Brown, D.I.W. 2002. The patterns within, and the ecological processes that sustain, the subtropical thicket vegetation in the planning domain for the Subtropical Thicket Ecosystem Planning (STEP) project. TERU Report 40, University of Port Elizabeth.
37. Von Maltitz, G., Mucina, L., Geldenhuys, C., Lawes, M., Eeley, H., Adie, H., Vink, D., Flemming, G. and Bailey, C. 2003. Classification system for South African Indigenous Forests. An objective classification for the Department of Water Affairs and Forestry. Environmentek Report ENV-P-C 2003-017, CSIR, Pretoria.
38. Van Niekerk, L. & Turpie, J.K. (eds). 2012. National Biodiversity Assessment 2011: Technical Report. Volume 3: Estuary Component. CSIR Report Number CSIR/NRE/ECOS/ER/2011/0045/B. Council for Scientific and Industrial Research, Stellenbosch.
39. Van Niekerk, L., Adams, J.B., James, N.C., Lamberth, S.J., MacKay, C.F., Turpie, J.K., Rajkaran, A., Weerts, S.P. and Whitfield, A.K., 2019, An Estuary Ecosystem Classification that encompasses biogeography and a high diversity of types in support of protection and management, African Journal of Aquatic Science, x,
40. Van Niekerk L., Taljaard S., Adams J.B., Fundisi D., Huizinga P., Lamberth S.J., Mallory S., Snow G.C., Turpie J.K., Whitfield A.K., Wooldridge T.H. 2015. Desktop provisional eco-classification of the temperate estuaries of South Africa. Water Research Commission Report No. 2187/1/15. Pretoria, South Africa.

APPENDIX 1: SUMMARY OF KEY NATIONAL ECOSYSTEM CLASSIFICATION SYSTEMS AND MAPS DEVELOPED PER REALM. TEXT HIGHLIGHTED IN BOLD REPRESENT DATA THAT WAS USED IN THE 2018 NATIONAL BIODIVERSITY ASSESSMENT. Realm Year Reference Description Terrestrial 1929 Pole-Evans (1929) Map, peer reviewed publication and published volume based on one of the first coordinated country-wide botanical surveys, entitled ‘The vegetation of South Africa’ with a focus on agricultural potential but triggered by the needs to describe degradation in the landscape. Number of units: 19; mapping scale: 1:3 000 000. 1988 Acocks (1988) Report and map: Veld types of South Africa. Compiled from extensive field work starting in the 1950’s and focused on agriculture potential and biome shifts in arid parts of country. Number of units:70 ; mapping scale: 1:250 000 1996 Low and Rebelo (1996) Report and map: Vegetation Map of South Africa, Lesotho and Swaziland, based on Acocks (1988) with a revised classification system applied. 2006 Mucina and Rutherford (2006) Book, maps and GIS: Vegetation of South Africa, Lesotho and Swaziland. Detailed descriptions and maps compiled from existing work of numerous authors, including Acocks (1988). A classification and mapping of the vegetation types on Prince Edward Islands was also included. Each biome-focused chapter has the full list of contributors. Number of units: 435; mapping scale: 1:250 000. This version was used in the first ecosystem assessment (Driver et al., 2005) 2011 Update of Mucina and Rutherford (2006). Referred to as the 2009 version. Minor update of Mucina and Rutherford (2006) including refinements of vegetation unit boundaries and descriptions for fine-scale planning work between 2006 and 2009 (all changes described in detail by Dayaram et al., 2017). Number of units: 438; mapping scale: 1:250 000 2011 National terrestrial ecosystem map for the Schedule of Threatened Ecosystems (Government Gazette no. 34809, 2011) This map was a combination of Mucina and Rutherford (2006), the national forest inventory (Von Maltitz et al. 2003, Berliner 2005), selected priority areas from the systematic biodiversity plans, and high irreplaceability forests patches or clusters systematically identified by in Berliner (2008). This version was used in the National Biodiversity Assessment 2011 (Driver et al., 2012). 2016 Update of Mucina and Rutherford (2006). Referred to as the 2012 version. Minor update of Mucina and Rutherford (2006) including refinements of vegetation unit boundaries and descriptions for KwaZulu Natal Province between 2009 and 2012 (all changes described in detail by Dayaram et al., 2017). Number of units: 442; mapping scale: 1:3000 to 250 000 2018 Update of Mucina and Rutherford (2006). Referred to as the 2018 version Major update of Mucina and Rutherford (2006) including refinements of vegetation unit boundaries in the Albany thicket biome based on Volk et al. 2003, revised seaward boundaries of coastal units (Harris et al. 2019), revised forests biome polygon boundaries (Dayaram et al. 2019), revised estuary boundaries (Van Niekerk et al., in press) and removal of wetland polygons (Dayaram et al. in press). Number of units: 465; mapping scale: 1:3000 to 250 000 Realm Year Reference Description Marine & Coastal 2004 Lombard et al. (2004) New marine biozones (broad, sub-biome level units) mapped for NSBA 2004 2011 Sink et al. (2011), Harris (2011) Marine benthic and coastal map developed for the NBA 2011, combined with pelagic biozone map from Lagabrielle (2009)
2018 Sink et al. (in prep) Revised marine classification and map (with benthic and pelagic combined) edge matched with a revised coastal classification and map. 2018 Harris et al. (2019) Revised coastal classification and map edge matched with the 2018 version of the national vegetation map, estuarine ecosystem map and marine ecosystem map. Realm Year Reference Description River 2004 Nel et al. (2004) River heterogeneity signatures (geomorphic province combined with hydrology, based on national scale 1:500 000 scale rivers lines with stream order and names added. 2011 Nel et al. (2011) River ecosystem types mapped for the NFEPA project and the NBA 2011, based on same 1:500 000 scale rivers database as for 2004 with additional river reaches added to align with large estuaries 2018 Nel et al. (2011) River ecosystem types mapped for the NFEPA project and the NBA 2011, based on modified 1:500 000 scale rivers layer with new names added, additional river reaches added to align with small estuaries Wetland 2006 Mbona et al. (unpublished) based on van den Berg et al. (2008) National Wetland Map versions 1-3: spatial data on the location and shape of wetlands based on remote sensing at a national scale, no classification system applied. 2011 Nel et al. (2012) National Wetland Map version 4: wetland classification system and map developed for the NBA 2011 and NFEPA 2010 based on a combination of biogeography and hydrogeomorphology. 2013 Ollis et al. (2011) National wetland classification system: wetland classification system 2018 Van Deventer et al. (2018) National Wetland Map version 5: wetland classification system and map developed for the NBA 2018 and based on a combination of biogeography and hydrogeomorphology with spatial data from a wide range sources Realm Year Reference Description Estuarine 2004 Turpie (2004) National map large estuaries. 225 individual estuaries were mapped (as points) with no classification system applied (Turpie 2004) 2011 van Niekerk and Turpie (2012) National map and classification system of large estuaries. 300 individual estuaries were mapped (as polygons) and classified for the NBA 2011. 2018 van Niekerk et al. (in press) Revised map and classification system of large estuaries (as polygons) with new (point-based) map of micro estuaries developed for NBA 2018 based on various existing data sources.

APPENDIX 2: REALM ECOSYSTEM MAPS AND LISTS OF ECOSYSTEM TYPES. Terrestrial Realm

Marine Realm

Freshwater Realm River

Freshwater Realm Wetlands*

*Classification as it appears in the NBA 2018 i.e. adaptation of Ollis et al. (2013)

Estuarine Realm

Coastal Zone

Estuarine Coastal Cool Temperate Arid Predominantly Cool Temperate Estuarine Lagoon Cool Temperate Estuarine Lake Cool Temperate Large Fluvially Dominated Cool Temperate Large Temporarily Closed Cool Temperate Micro-estuary Cool Temperate Predominantly Open Cool Temperate Small Fluvially Dominated Cool Temperate Small Temporarily Closed Subtropical Estuarine Ba Subtropical Estuarine Lake Subtropical Large Fluvially Dominated Subtropical Large Temporarily Closed Subtropical Micro-estuary Subtropical Predominantly Open Subtropical Small Temporarily Closed Tropical Estuarine Lake Warm Temperate Estuarine Bay Warm Temperate Estuarine Lake Warm Temperate Large Fluvially Dominated Warm Temperate Large Temporarily Closed Warm Temperate Micro-estuary Warm Temperate Predominantly Open Warm Temperate Small Fluvially Dominated Warm Temperate Small Temporarily Closed

Marine Coastal

Agulhas Boulder Shore Agulhas Dissipative Intermediate Sandy Shore Agulhas Dissipative Sandy Shore Agulhas Exposed Rocky Shore Agulhas Exposed Stromatolite Rocky Shore Agulhas Inner Shelf Mosaic Agulhas Inner Shelf Reef Agulhas Intermediate Sandy Shore Agulhas Island Agulhas Kelp Forest Agulhas Mixed Shore Agulhas Muddy Mid Shelf Agulhas Reflective Sandy Shore Agulhas Sandy Inner Shelf Agulhas Sheltered Rocky Shore Agulhas Stromatolite Mixed Shore Agulhas Very Exposed Rocky Shore Agulhas Very Exposed Stromatolite Rocky Shore Aliwal Shoal Reef Complex Cape Bay Cape Boulder Shore Cape Exposed Rocky Shore Cape Island Cape Kelp Forest Cape Mixed Shore Cape Rocky Inner Shelf Cape Sandy Inner Shelf Cape Sheltered Rocky Shore Cape Very Exposed Rocky Shore Delagoa Mixed Shore Delagoa Sandy Inner Shelf Delagoa Very Exposed Rocky Shore Durnford Inner Shelf Reef Complex Durnford Mid Shelf Reef Complex Eastern Agulhas Bay False and Walker Bay Kei Fluvial Fan Kei Reef Mosaic Kosi Coral Community KZN Bight Deep Shelf Edge KZN Bight Mid Shelf Mosaic KZN Bight Mid Shelf Reef Complex KZN Bight Muddy Inner Shelf KZN Bight Muddy Shelf Edge KZN Bight Outer Shelf Mosaic KZN Bight Sandy Inner Shelf Leadsman Coral Community Namaqua Exposed Rocky Shore Namaqua Kelp Forest Namaqua Mid Shelf Fossils Namaqua Mixed Shore Namaqua Muddy Mid Shelf Mosaic Namaqua Muddy Sands Namaqua Sandy Inner Shelf Namaqua Sandy Mid Shelf Namaqua Sheltered Rocky Shore Namaqua Very Exposed Rocky Shore Natal Boulder Shore Natal Delagoa Dissipative Intermediate Sandy Shore Natal Delagoa Dissipative Sandy Shore Natal Delagoa Intermediate Sandy Shore Natal Delagoa Reflective Sandy Shore Natal Exposed Rocky Shore Natal Mixed Shore Natal Very Exposed Rocky Shore Orange Cone Inner Shelf Mud Reef Mosaic Orange Cone Muddy Mid Shelf Port St Johns Inner Shelf Mosaic Port St Johns Muddy Mid Shelf Port St Johns Muddy Shelf Edge Sodwana Coral Community Southern Benguela Dissipative Intermediate Sandy Shore Southern Benguela Dissipative Sandy Shore Southern Benguela Intermediate Sandy Shore Southern Benguela Reflective Sandy Shore Southern KZN Inner Shelf Mosaic St Helena Bay St Lucia Sandy Inner Shelf St Lucia Sandy Mid Shelf Trafalgar Reef Complex uThukela Mid Shelf Mosaic uThukela Mid Shelf Mud Coarse Sediment Mosaic uThukela Outer Shelf Muddy Reef Mosaic Western Agulhas Bay Wild Coast Inner Shelf Mosaic

Terrestrial Coastal Agulhas Limestone Fynbos Agulhas Sand Fynbos Albany Mesic Thicket Albertinia Sand Fynbos Alexander Bay Coastal Duneveld Algoa Sandstone Fynbos Atlantis Sand Fynbos Bethelsdorp Bontveld Blombos Strandveld Canca Limestone Fynbos Cape Flats Dune Strandveld Cape Flats Sand Fynbos Cape Seashore Vegetation Cape Winelands Shale Fynbos De Hoop Limestone Fynbos Eastern Coastal Shale Band Vegetation Elands Forest Thicket Elim Ferricrete Fynbos Garden Route Granite Fynbos Garden Route Shale Fynbos Goukamma Dune Thicket Gouritz Valley Thicket Grassridge Bontveld Groot Brak Dune Strandveld Hamburg Dune Thicket Hangklip Sand Fynbos Hartenbos Dune Thicket Hopefield Sand Fynbos Humansdorp Shale Renosterveld Kasouga Dune Thicket Knysna Sand Fynbos Kogelberg Sandstone Fynbos KwaZulu-Natal Coastal Belt Grassland Lambert's Bay Strandveld Langebaan Dune Strandveld Lourensford Alluvium Fynbos Mangrove Forest Maputaland Coastal Belt Maputaland Wooded Grassland Mossel Bay Shale Renosterveld Motherwell Karroid Thicket Namaqualand Coastal Duneveld Namaqualand Heuweltjie Strandveld Namaqualand Inland Duneveld Namaqualand Riviere Namaqualand Sand Fynbos Namaqualand Seashore Vegetation Namaqualand Strandveld Namib Lichen Fields Namib Seashore Vegetation Northern Coastal Forest Northern Richtersveld Yellow Duneveld Overberg Dune Strandveld Overberg Sandstone Fynbos Peninsula Granite Fynbos Peninsula Sandstone Fynbos Peninsula Shale Fynbos Peninsula Shale Renosterveld Pondoland-Ugu Sandstone Coastal Sourveld Potberg Ferricrete Fynbos Potberg Sandstone Fynbos Richtersveld Coastal Duneveld Richtersveld Sandy Coastal Scorpionstailveld Saldanha Flats Strandveld Saldanha Granite Strandveld Saldanha Limestone Strandveld Sardinia Forest Thicket South Eastern Coastal Thornveld Southern Afrotemperate Forest Southern Cape Dune Fynbos Southern Coastal Forest Southern Richtersveld Yellow Duneveld St Francis Dune Thicket Subtropical Dune Thicket Subtropical Seashore Vegetation Swamp Forest Transkei Coastal Belt Tsitsikamma Sandstone Fynbos Umtiza Forest Thicket Western Gariep Plains Desert