about us | careers | terms & conditions | intranet | extranet | sitemap | contact us
   
Skip Navigation Links
Home
Skip Navigation Links
Knowledge Hub
Skip Navigation Links
Research
Skip Navigation Links
Resources & Tools
Skip Navigation Links
Learning
Skip Navigation Links
Events
Skip Navigation Links
News & Media
Skip Navigation Links
FET Water
Skip Navigation Links
SCM
Login | Register
Go Search
     
Probabilistic modelling of wetland condition
Expanded Title:This study intended to contribute a tool to identify wetland condition with the view to safeguard them. The growing levels of catchment degradation occurring nationally necessitate a higher-level approach be taken to assess wetland conditions per wetland type, in order to begin prioritizing conservation actions at a regional and national level. There are usable hierarchical wetland classifications and tools to estimate individual wetland condition and level of function. The aims of this study were to prepare a model to predict wetland condition statistically (probability of degradation per wetland type), based on the hydro-geomorphic type and position of the wetland within landscape, and then to validate the prediction model developed, by reference to the known wetland ecosystem integrity status. Field and existing data for a total of 463 seep, floodplain and valley bottom (channelled and unchannelled) hydrogeomorphic (HGM) units (focused only in KZN) were used to develop predictive models of HGM unit condition. Statistical models were successfully developed to estimate both condition and probability of degradation for all four HGM types. Data on land cover, infrastructure and catchment physiography covering twenty one potential explanatory variables, were linked to individual HGM units. Multiple linear regression models were developed to estimate A-F Present Ecological Status (PES) scores, while multiple logistic regression models were developed to estimate the probability of degradation per HGM type. Models were developed at tertiary and quaternary catchment scales, and fine scale using a 1000m buffer radius around each HGM unit. Average wetland conditions per HGM type per quaternary catchment were successfully mapped to provide images on spatial patterns of degradation within KwaZulu-Natal. The models show promise, and would be useful at a regional planning level to identify priority catchments where field assessments of wetland condition would be required. Models can be further developed and refined using more data, and covering a wider area. These are intended to complement existing initiatives such as NFEPA, and NWCS.
Date Published:01/08/2012
Document Type:Research Report
Document Subjects:Ecosystem - Biodiversity protection
Document Format:Report
Document File Type:pdf
Research Report Type:Consultant
WRC Report No:KV 298/12
ISBN No:978-1-4312-0307-9
Authors:Rivers-moore N; Cowden C
Project No:K8/928
Organizations:Freshwater Research Centre; CapeNature
Document Size:806 KB
Copyright 2016 - Water Research Commission Designed By: Ceenex