Human impact on land in terms of transformations in land cover and land use have increased rapidly over the years. The most dominant and visible change was found in the greater extent of the Transport and Urban land cover classes by 2012.
Background
According to several authors, land cover can be described as the types of features found on the Earth's surface, such as water bodies, vegetation and man-made structures (Ellis, 2013; Lillesand et al., 2008). As mentioned earlier, the impact of human activities on the Earth's land cover and land use has increased rapidly over the years and as a result of these ongoing land transformations, planning sustainable urban development becomes a challenge.
Use of Remote Sensing and Change Detection
Motivation for this study
Umhlanga Ridge
However, one of the most widely used land cover classification systems was designed by the United States Geological Survey (USGS) in the mid-1970s (Anderson et al., 1976; Lillesand et al., 2014). Furthermore, it was later seen that the areas of the grassland land cover class in 2006 (Figure 6D) were used in the expansion of urban landscapes by 2012 (Figure 6E).
Aim
Research Objectives
To identify the dominant land use and land cover changes that occurred between 2006 and 2012 in the Umhlanga Ridge.
Methodology
Thesis Outline
Summary
Introduction
Sustainable Development
To gain a complete understanding of SD and all that it entails, models of SD must be analyzed. With the process of SD requiring effective, efficient and transparent communication and production between all relevant government or public entities; the two most popular and well-known models for the SD discourse will be examined in more detail (Blewitt, 2010; Jabareen, 2006; Jain et al., 2007; Pezzey, 1992).
Models of Sustainable Development
The Three Ring Model
- Challenges of the Three Ring Model
As depicted in Figure 2B, this new dimension creates the 'Four Pillar Sustainable Development' model and highlights the direct link between the ecological and economic sectors through the decision-making processes at regional, local and global levels (Ndah, 2015 ; Waas et al., 2011). One of the most recurring and often mentioned flaws is that this model compartmentalizes the three different entities of SD and therefore underestimates the importance of the interconnected relationship between these entities (Giddings et al., 2002; Waas et al., 2011).
The Nest Sustainable Development Model
- Challenges of the Nest Sustainable Development Model . 14
The first of the five principles we will discuss is the principle of 'intergenerational equality'. This principle recognizes the important role that biodiversity and the overall integrity of the environment play in the functioning of society and all of humanity (Haughton, 1999).
Sustainable Development and Urban Areas
The first concept in SUFM that will be looked at is 'compactness'. The concept of greening aims to incorporate the natural element into the city or urban area by adding to the city's aesthetic value and overall attractiveness (Dumreicher et al., 2000; Nassauer and Faust, 2013).
Sustainable Development in Umhlanga Ridge
The document produced by the company ASM Consortium explains how all Umhlanga Ridge developments were declared to be observed from a sustainable approach, thus incorporating the natural environment into the urban area (ASM Consortium, 2008). This has resulted in alienation between those who own private vehicles and those who do not,” (Beires Similarly the Umhlanga Ridge development adopted SUFM's principles to displace the 'car culture'.
Summary
Finally, one of the key aspects included in the Umhlanga Ridge development that will ensure the sustainability of the area is that of increased public participation and also the recognition of the need for more public-private partnerships (Bredell, 2012; ASM Consortium, 2008; Yli-Pelkonen and Kohl, 2005). The acceptance of this mutually beneficial relationship will ensure that there is total transparency in the process and implementation of developments, that all concerns - from all stakeholders - will be heard and that all committees - at local, regional and national levels - will are heard. confident and proud of the step taken towards a sustainable society (Beires, 2010; Nomico and Sanders, 2003).
Introduction
Land Use and Land Cover (LULC) Changes
The study identified that the negative effects far outweigh the benefits of urban expansion (Schirmer et al., 2008). Furthermore, it was stated that TIR generated better results in terms of improved detail and exceptional accuracy of the LULC changes (Chi et al., 2009).
Umhlanga Ridge
The Gateway Shopping Centre
As the Tongaat Hulett Company and its Moreland Developments sector focused on the conversion and development of former agricultural land to urban land use, the emergence of the Gateway Shopping Center was seen as one of the prominent developments in Umhlanga (Kitchin, 2003; Michel and Scott , 2005). Being the first major development of modern times in the Umhlanga area, the Gateway Shopping Center was located in the ideal area to allow a constant flow of energy to the surrounding developments (Bredell, 2012; Michel and Scott , 2005).
Climatic Conditions of Umhlanga
As highlighted in the previous sections, Umhlanga was once an area dominated by agriculture, sugarcane cultivation and minimal development. As evident from the above discussions, the development of the Umhlanga suburb has made Umhlanga something of a commercial, residential and tourist destination. destination (ASM Consortium, 2008; Michel and Scott, 2005; Tongaat Hulett Development, 2011; Wood, 2008). Many environmentalists had protested the development of the Gateway shopping center due to its size; the fact that the presence of the mall has led to a serious loss of agricultural land in the Umhlanga areas and even more land has been used for the construction of roads and other commercial and residential sites surrounding the mall (Yusuf and Allopi, 2004) .
Location of the Study Area
Summary
Introduction
LULC Classification Schemes
Identification of each LULC Class
This land cover is composed of all non-vegetated areas or areas consisting of minimal vegetation. All areas composed of buildings and urban development; this includes residential, commercial, industrial areas as well as communication systems.
![Table 5A: The Land Cover Classification Scheme utilised in this study [Adapted from (Thompson, 1996)]](https://thumb-ap.123doks.com/thumbv2/pubpdfnet/10639030.0/78.918.257.620.114.408/table-land-cover-classification-scheme-utilised-adapted-thompson.webp)
Field Survey
This area consists of land that has been plowed and/or prepared for the cultivation of various crops, but timber production is excluded. It is composed of the dominant grassy, non-woody and rooted herbaceous plant species.
Data Collection
Acquisition of the Satellite Imagery and related information
Overall, 60 GCPs were collected for the purpose of this research from different locations throughout the study area in order to document each different land use and class. For the purposes of this research, high-resolution SPOT 5 images were obtained from the SANSA archives.
Acquisition of Additional Data
Then Mr. Mnguni proposed by Mr. Rory Wilkinson (Director of Urban Planning for the Tongaat Hulett Development Sector) as an ideal candidate to participate in this study, which in turn led to the selection of the participant, Ms. Karen Petersen (Director of Development for the Tongaat Huletts Development Sector) to be included in the study as a key informant. Furthermore, this remained in line with the snowball sampling method, which stated that the process of collecting referrals to the sample group should continue "until the required number or saturation point in terms of the information sought is reached," (Kumar.
Types of Satellite Imagery Pre-processing Methods
However, there have been numerous arguments surrounding the use of both or either of the radiometric and geometric correction procedures. Agapiou et al. 2011) believes that radiometric correction can be extremely tedious, time-consuming and a challenging process compared to geometric correction. 2010) state that radiometric corrections and associated atmospheric corrections performed on remotely sensed images are one of the most profound and important preprocessing procedures.
Classification Comparison
Parallelepiped Classifier
Based on the simple Boolean 'and/or' function, the threshold of each class's signature is used to determine whether the pixel in question belongs to a specific class (Jensen, 2005; Teodoro et al., 2009). Therefore, pixels that are able to fall above the low threshold and below the high threshold of a particular class of parallelepiped are assigned to that class (Lillesand et al., 2014; . Schowengerdt, 2006; Teodoro et al., 2009) .
Minimum Distance to Mean (MDM) Classifier
Perumal and Bhaskaran (2010) also found that while using two bands, the parallelepiped classifier determines the training area of pixels in each successive band based on the minimum and maximum pixel values in the image. While one disadvantage of this method is that a single pixel may fall in the overlapping region between two or more parallelepipeds of a class, another disadvantage is that the pixels also cannot fall within any of the parallelepipeds of that class, leaving that pixel unclassified (Lillesand et al., 2014; Teodoro et al., 2009).
Maximum Likelihood (ML) Classifier
The algorithm states that all training data collected for each class within each band tends to follow a normal (Gaussian) distance (Jensen, 2005; Keuchel et al., 2003; Pal and Mather, 2003). Due to the computationally intensive nature of the ML classifier, the algorithm has a significantly slower processing time compared to other classifiers, especially when processing larger images (Aronoff, 2005; Patil et al., 2012).
Post Classification Procedure
Accuracy Assessment
Thus, the error matrix can report the relationship between the true land cover classes and the classes that were mapped (Congalton and Green, 2008; Patil et al., 2012). Using the error matrix report, the classification that produced the highest K and OA coefficient was used to classify the images in question for this study.
Change Detection
Summary
Introduction
Finding a Suitable Classifier
Maps and Accuracy Assessment Results generated
Results attained from each Classification
Image Classification
Considered a 'hard' classifier, the ML algorithm assigns digital numbers or pixels according to the land cover class to which they are mutually exclusive (Weng, 2011). This limitation – although small – has the potential to affect the accuracy and authenticity of the urban land cover class represented in the generated land cover maps (Figure 6D and Figure 6E).
Accuracy Assessment
Furthermore, Foody (2002; Foody, 2010) emphasizes that the universal minimum standard required for thematic land cover maps is not universally acceptable, when practical actions are taken into account or considered. However, it has been widely recognized that to improve the OA of a classification, the use of more ground truth data or reference data for the accuracy assessment, as the probability of the pixels belonging to land cover classes is increased.
Changes in the Land Cover Classes
- Urban/Built Up Lands
- Transport Networks
- Grasslands
- Forest and Woodlands
- Cultivated Lands
- Barren
Umhlanga's transformation began in the early 1860s with one of the key participants in this transformation being that of the Tongaat Hulett Company. Referring to the modified land cover classification scheme (Thompson, 1996) illustrated in Chapter Five of this study (Table 5A), transportation networks are referred to as a subclass of the urban land cover class.
Change Detection Patterns
Socio-economic Implications
It has also been added that notices will be placed at the location, in newspapers and that letters will be sent to residents of the proposed location. The public will then have thirty days to comment on the upcoming development and those comments will be passed on to the project team assigned to the development to best mitigate the issue.
Future Changes for Umhlanga and Umhlanga Ridge
The first phase of the construction of the Izinga area has already been completed, and the second and final phase of its development will provide a significant expansion of the currently implemented mixed land use design. According to the details of the Ridgeside development published on the Tongaat Hulett Development Sectors website, this specific area contains four development phases that provide a link between Umhlanga Ridge and Umhlanga Rocks Village (Tongaat Hulett Development, 2014).
Summary
Similarly, while the development of Cornubia Park has just begun, the next phases of development have been designed, approved and will begin construction soon after the first phase is complete. Repeated by Ms. Petersen, was told that other sub-zones will emerge within Ridgeside Park in order to effectively implement its mixed land use criteria and reduce the use of private vehicles.
Introduction
General Conclusions
However, due to the expansion of the urban and grassland land cover classes, the once prominent land cover class of cultivated land suffered a significant decline of about 11.05% in its spatial extent. In the year 2012, it was clear that the intervention in the urban landscape on both of these land cover classes has led to a highly stressed biodiversity and overall ecological organisms.
Limitations
- Classifier algorithm limitations
- Lack of historical remotely sensed data
- Limitations of the interview process
- Change detection limitations
In order to complement this study, interviews were conducted with members of the Tongaat Hulett Development Sector using a Snowball sampling method. One of the deepest limitations experienced during this process was the fact that although more than ten members were approached to conduct interviews, only four participants were interested in participating in the study.
Recommendations
Improve the spatial resolution
Models to determine/predict future LULC transformations
Longer time periods
Mustard JF, Defries RS, et al. 2004) Land use and land cover change pathways and impacts. From pattern to process: landscape fragmentation and the analysis of land use/land cover change.
Access to more CD techniques
