An analysis of landscape connectivity of the Grassland Biome in Mpumalanga using graph theory. MSc Project Louise Fourie Supervisor: Prof. M.

Transcription

1 An analysis of landscape connectivity of the Grassland Biome in Mpumalanga using graph theory MSc Project Louise Fourie Supervisor: Prof. M. Rouget

2 Introduction South African Grassland Biome Second largest biome after savanna Most threatened and least protected biome in South Africa Only 1.6% formally protected

3 Introduction Habitat loss and fragmentation: Two primary threats to biodiversity Connectivity helps maintain viable populations in fragmented landscapes Important in biodiversity conservation

4 Aims To analyse the connectivity of grassland habitat patches in Mpumalanga using graph theory To investigate the importance of abandoned croplands for maintaining overall connectivity

5 Connectivity The degree to which the landscape facilitates or prevents movement of organisms among available habitat patches

6 Connectivity Can be either structural or functional Structural connectivity: Arrangement of habitat and landscape without reference to specific species Functional connectivity: The behavioural responses of an organism to the arrangement of habitat in the landscape

7 Structural and functional connectivity Structural connectivity between patch 1 and 2 relatively good Patch 1 Functional connectivity Movement of a specific species between patch 1 and 2 can be either easy or difficult depending on properties of the species Patch 2

8 Measuring Connectivity Different Measures available Nearest Neighbour distance (use patch occupancy data and inter patch distance) Spatial pattern indices (number, size, extent, shape or aspect of habitat patches) Observed immigration, emigration or dispersal rates (actual observed movements of species)

9 Measuring Connectivity Calabrese & Fagan (2004)

10 Graph theory Represent landscape as a set of nodes and edges Edge (connection between nodes) Node (habitat patch)

11 Advantages : Graph theory Provide a detailed picture of connectivity Modest data requirements Very suitable for large scale landscape analysis needed for conservation planning

12 Graph theory Can use both structural and dispersal data Unify multiple aspects of habitat connectivity Can be applied at patch or landscape levels Many graph theory based indices available to evaluate structural or functional connectivity

13 Data Data used Recent land cover for Mpumalanga (2008) Old fields identified from old 1:50,000 topographic maps Major roads

14 Methods Natural grassland habitat patches in Mpumalanga Patches smaller than 5 ha removed grassland habitat patches Total area of habitat patches km 2 Of which km 2 is old fields 40% of the Grassland Biome in Mpumalanga transformed

15 Methods Natural grassland habitat patches in Mpumalanga 500 m Natural F4 Old fields

16 Methods Input Output Size of habitat patches Importance of patches for overall connectivity Distances between patches Conefor Sensinode programme Integral Index of Connectivity Number of components

17 Methods Number of Components (Clusters): Component = set of nodes with a path between every pair of nodes No connection between nodes of different components Component 1 Component 2

18 Results Number of Components Number of Components Without old fields With old fields Above a distance threshold of 1000m there are less than 50 components left Distance threshold (m)

19 Method Integral Index of Connectivity (IIC) Seen as the best binary Index to measure connectivity Include habitat area in measurement Ranges from 0 to 1 Increase with improved connectivity

20 Results Integral index of connectivity (IIC) 25% improvement when old fields are added Without old fields With old fields Threshold distance (km)

21 Importance of habitat patches for overall connectivity Distance threshold: 500 m

22 Weighted average of importance of patches in different vegetation types KwaZulu Natal Highland Thornveld** Northern Escarpment Quartzite Sourveld** ( (diic x Area))/Total area of veg type Northern Escarpment Dolomite Grassland** Barberton Montane Grassland** Frankfort Highveld Grassland Andesite Mountain Bushveld Low Escarpment Moist Grassland Soweto Highveld Grassland Tsakane Clay Grassland Rand Highveld Grassland Eastern Highveld Grassland KaNgwane Montane Grassland** Northern Free State Shrubland Long Tom Pass Montane Grassland** Paulpietersburg Moist Grassland Sekhukhune Montane Grassland** Amersfoort Highveld Clay Grassland Lydenburg Thornveld** Steenkampsberg Montane Grassland** Wakkerstroom Montane Grassland **

23 Average connectivity per vegetation type

24 Conservation implications The grassland habitat patches of Mpumalanga are well connected at a distance threshold of 500m The most connected vegetation types are: Wakkerstroom montane grassland Eastern Temparate freshwater wetlands Steenkampsberg montane grassland The abandoned croplands present in this landscape increase the connectivity by 25% Landscape connectivity influence conservation value

25 Thank You