ResearchSpace

Physio-climatic classification of South Africa's woodland biome

Show simple item record

dc.contributor.author Fairbanks, DHK en_US
dc.date.accessioned 2007-01-22T06:25:56Z en_US
dc.date.accessioned 2007-06-07T10:05:04Z
dc.date.available 2007-01-22T06:25:56Z en_US
dc.date.available 2007-06-07T10:05:04Z
dc.date.issued 2000-07 en_US
dc.identifier.citation Fairbanks, DHK. 2000. Physio-climatic classification of South Africa's woodland biome. Plant Ecology, vol. 149(1), pp 71-89 en_US
dc.identifier.issn 1385-0237 en_US
dc.identifier.uri http://hdl.handle.net/10204/1436 en_US
dc.identifier.uri http://hdl.handle.net/10204/1436
dc.description.abstract In an effort to develop more holistic ecosystem approaches to resource assessment and management, landscapes need to be stratified into homogeneous geographic regions. These regions can then be used in a monitoring framework to develop reliable estimates of ecosystem productivity. A regional characterization of the woodland biome has been developed for South Africa, delineated by satellite imagery and using environmental data and a rigorous statistical methodology. Distribution maps of key environmental variables are analyzed by factor analysis, an iterative clustering technique and maximum likelihood classification to quantify and identify homogeneous physio-climatic units. A spatial clustering technique was used to identify regions, which are statistically different with regard to five physiographic, climatic and edaphic variables deemed important within southern African savanna woodlands. The woodland biome of South Africa at one kilometre resolution was successively divided. Thirty year mean monthly temperature, total plant-available water balance of soil, elevation, landscape topographic position, and landscape soil fertility were used as input classification variables. The map data were submitted to a factor analysis and varimax axis rotation. The factor analysis removes correlations from the input variables, reduces the dimensionality, and normalizes the axis measurements. A cluster analysis was performed on the three principal factor scores using a modified iterative optimization clustering procedure to determine the finest level of classes statistically permitable. Twenty-seven identified unimodal cluster signatures were then submitted to a maximum likelihood classification where the statistical probability of the GIS cell assignment is carried out to determine class membership. The final map of custom physio-climatic regions is described, and these custom regions are compared with a vegetation potential map of the woodland types identified in the South African summer rainfall zone. en_US
dc.format.extent 224606 bytes en_US
dc.format.mimetype application/pdf en_US
dc.language.iso en en_US
dc.publisher Kluwer Academic Publishers en_US
dc.rights Copyright: 2000 Kluwer Academic Publishers en_US
dc.subject Abiotic variables en_US
dc.subject Ecoregions en_US
dc.subject GIS en_US
dc.subject Geographic information systems en_US
dc.subject Ecology en_US
dc.subject Forestry en_US
dc.subject Climatic variables en_US
dc.subject Savannas en_US
dc.title Physio-climatic classification of South Africa's woodland biome en_US
dc.type Article en_US
dc.identifier.apacitation Fairbanks, D. (2000). Physio-climatic classification of South Africa's woodland biome. http://hdl.handle.net/10204/1436 en_ZA
dc.identifier.chicagocitation Fairbanks, DHK "Physio-climatic classification of South Africa's woodland biome." (2000) http://hdl.handle.net/10204/1436 en_ZA
dc.identifier.vancouvercitation Fairbanks D. Physio-climatic classification of South Africa's woodland biome. 2000; http://hdl.handle.net/10204/1436. en_ZA
dc.identifier.ris TY - Article AU - Fairbanks, DHK AB - In an effort to develop more holistic ecosystem approaches to resource assessment and management, landscapes need to be stratified into homogeneous geographic regions. These regions can then be used in a monitoring framework to develop reliable estimates of ecosystem productivity. A regional characterization of the woodland biome has been developed for South Africa, delineated by satellite imagery and using environmental data and a rigorous statistical methodology. Distribution maps of key environmental variables are analyzed by factor analysis, an iterative clustering technique and maximum likelihood classification to quantify and identify homogeneous physio-climatic units. A spatial clustering technique was used to identify regions, which are statistically different with regard to five physiographic, climatic and edaphic variables deemed important within southern African savanna woodlands. The woodland biome of South Africa at one kilometre resolution was successively divided. Thirty year mean monthly temperature, total plant-available water balance of soil, elevation, landscape topographic position, and landscape soil fertility were used as input classification variables. The map data were submitted to a factor analysis and varimax axis rotation. The factor analysis removes correlations from the input variables, reduces the dimensionality, and normalizes the axis measurements. A cluster analysis was performed on the three principal factor scores using a modified iterative optimization clustering procedure to determine the finest level of classes statistically permitable. Twenty-seven identified unimodal cluster signatures were then submitted to a maximum likelihood classification where the statistical probability of the GIS cell assignment is carried out to determine class membership. The final map of custom physio-climatic regions is described, and these custom regions are compared with a vegetation potential map of the woodland types identified in the South African summer rainfall zone. DA - 2000-07 DB - ResearchSpace DP - CSIR KW - Abiotic variables KW - Ecoregions KW - GIS KW - Geographic information systems KW - Ecology KW - Forestry KW - Climatic variables KW - Savannas LK - https://researchspace.csir.co.za PY - 2000 SM - 1385-0237 T1 - Physio-climatic classification of South Africa's woodland biome TI - Physio-climatic classification of South Africa's woodland biome UR - http://hdl.handle.net/10204/1436 ER - en_ZA


Files in this item

This item appears in the following Collection(s)

Show simple item record