Research

GIS is involved in research projects focusing on the development of remote sensing and GIS techniques for environmental assessment and monitoring. The projects are done in collaboration with other research institutions such as the CSIR, universities and international organizations.

 

  A key component of forest productivity is forest health and the effect it has on producing a sustainable yield. It has been reported that forests affected by pests and pathogens account for losses of millions of Rands due to timber damage or tree mortality, excluding the impacts due to loss of growth. Management practices are now increasingly designed to maintain and enhance the long term health of forest ecosystems while providing, socio-cultural and environmental opportunities for present and future generations. Technologies such as Remote Sensing and Geographic Information Science provide a scientific framework for the detection and monitoring of forest health condition rather than relying on broad scale visual assessments.

Sirex noctilio caused considerable mortality in commercial pine forests in KwaZulu-Natal, South Africa. In this project, the ability to remotely detect S. noctilio infestations was crucial for monitoring the spread of the wasp and for the effective deployment of suppression activities. This project focused on the development of techniques based on remote sensing technology to accurately detect and map S. noctilio infestations. The project evaluated the robustness and accuracy of various machine learning algorithms in identifying spectral, spatial and environmental parameters that allowed for the successful detection of S. noctilio infestations. The project was jointly supported by the NRF and SAPPI forests. To date, the project was very successful and is implemented on an industry wide basis by the Institute of Commercial Forestry as a national strategy to reduce the spread of the pest. 



   
 Adult female Gonipterus scutellatus and b. defoliated leaf (Carbone et al., 2006)

Other health projects focus on detecting and mapping physical damage of vegetation by pests such as the Gonipterus scutellatus weevil and mopane worms using remotely sensed data.

Vegetation Mapping

In Southern Africa, natural forest and rangeland resources form an important resource base for food and medicinal products that form part of people’s subsistence as well as their economic base and well-being. Recent studies has however showed that the long term sustainability of the natural forest in Southern Africa and its resources is under threat because of lack of information on the distribution of the species. Mapping tree species in the natural forest of Southern Africa is therefore extremely important for forest management purposes. Earth observation and GIS provides a cost-effective tool to help forest managers better understand forest characteristics, such as forest area, locations, species, even down to the level of characterizing individual trees and health status.

Recent advancements in remotely sensed technology, such as hyper spectral and new multispectral sensors such as Worldview with strategically positioned bands have encouraged the development of more robust approaches towards forest species mapping including the discrimination of alien invasive plants within commercial forest ecosystems. This research group has conducted a number of vegetation mapping projects using such tools, ranging from mopane discrimination, brackern fern mapping, rangeland grass species mapping to commercial pine varieties mapping.

 
 Assessing Distribution and Quality of Vegetation Species


 
Global climate change is expected to cause shifts in the percentage cover and abundance of grass species following C3 and C4 photosynthetic pathways. These two groups differ in a number of physiological, structural and biochemical properties. Remote sensing high spectral and spatial resolution systems have been explored to detect, classify and map C3 and C4 grasses. This study also investigated an innovative method to address the challenges associated with spectral dimensionality and the related multicollinearity in the spectral response of vegetation. It involves developing a user-defined inter-band correlation filter function to resample hyperspectral data. The utility of the new resampling technique was assessed for discriminating C3 and C4 grasses and for predicting their nutrient content (nitrogen, protein, moisture, and fibre) using field spectroscopy and Worldview-2 images in the Cathedral Peak region of the Drakensberg Mountains 



Contact Webmaster | View the Promotion of Access to Information Act | View our Privacy Policy
© University of KwaZulu-Natal: All Rights Reserved