Thesis (MSc)--Stellenbosch University, 2012. / ENGLISH ABSTRACT: Supervised classifiers are commonly used in remote sensing to extract land cover information. They are, however, limited in their ability to cost-effectively produce sufficiently accurate land cover maps. Various factors affect the accuracy of supervised classifiers. Notably, the number of available training samples is known to significantly influence classifier performance and to obtain a sufficient number of samples is not always practical. The support vector machine (SVM) does perform well with a limited number of training samples. But little research has been done to evaluate SVM’s performance for geographical object-based image analysis (GEOBIA). GEOBIA also allows the easy integration of additional features into the classification process, a factor which may significantly influence classification accuracies. As such, two experiments were developed and implemented in this research. The first compared the performances of object-based SVM, maximum likelihood (ML) and nearest neighbour (NN) classifiers using varying training set sizes. The effect of feature dimensionality on classifier accuracy was investigated in the second experiment. A SPOT 5 subscene and a four-class classification scheme were used. For the first experiment, training set sizes ranging from 4-20 per land cover class were tested. The performance of all the classifiers improved significantly as the training set size was increased. The ML classifier performed poorly when few (<10 per class) training samples were used and the NN classifier performed poorly compared to SVM throughout the experiment. SVM was the superior classifier for all training set sizes although ML achieved competitive results for sets of 12 or more training samples per class. Training sets were kept constant (20 and 10 samples per class) for the second experiment while an increasing number of features (1 to 22) were included. SVM consistently produced superior classification results. SVM and NN were not significantly (negatively) affected by an increase in feature dimensionality, but ML’s ability to perform under conditions of large feature dimensionalities and few training areas was limited. Further investigations using a variety of imagery types, classification schemes and additional features; finding optimal combinations of training set size and number of features; and determining the effect of specific features should prove valuable in developing more costeffective ways to process large volumes of satellite imagery. KEYWORDS Supervised classification, land cover, support vector machine, nearest neighbour classification maximum likelihood classification, geographic object-based image analysis / AFRIKAANSE OPSOMMING: Gerigte klassifiseerders word gereeld aangewend in afstandswaarneming om inligting oor landdekking te onttrek. Sulke klassifiseerders het egter beperkte vermoëns om akkurate landdekkingskaarte koste-effektief te produseer. Verskeie faktore het ʼn uitwerking op die akkuraatheid van gerigte klassifiseerders. Dit is veral bekend dat die getal beskikbare opleidingseenhede ʼn beduidende invloed op klassifiseerderakkuraatheid het en dit is nie altyd prakties om voldoende getalle te bekom nie. Die steunvektormasjien (SVM) werk goed met beperkte getalle opleidingseenhede. Min navorsing is egter gedoen om SVM se verrigting vir geografiese objek-gebaseerde beeldanalise (GEOBIA) te evalueer. GEOBIA vergemaklik die integrasie van addisionele kenmerke in die klassifikasie proses, ʼn faktor wat klassifikasie akkuraathede aansienlik kan beïnvloed. Twee eksperimente is gevolglik ontwikkel en geïmplementeer in hierdie navorsing. Die eerste eksperiment het objekgebaseerde SVM, maksimum waarskynlikheids- (ML) en naaste naburige (NN) klassifiseerders se verrigtings met verskillende groottes van opleidingstelle vergelyk. Die effek van kenmerkdimensionaliteit is in die tweede eksperiment ondersoek. ʼn SPOT 5 subbeeld en ʼn vier-klas klassifikasieskema is aangewend. Opleidingstelgroottes van 4-20 per landdekkingsklas is in die eerste eksperiment getoets. Die verrigting van die klassifiseerders het beduidend met ʼn toename in die grootte van die opleidingstelle verbeter. ML het swak presteer wanneer min (<10 per klas) opleidingseenhede gebruik is en NN het, in vergelyking met SVM, deurgaans swak presteer. SVM het die beste presteer vir alle groottes van opleidingstelle alhoewel ML kompeterend was vir stelle van 12 of meer opleidingseenhede per klas. Die grootte van die opleidingstelle is konstant gehou (20 en 10 eenhede per klas) in die tweede eksperiment waarin ʼn toenemende getal kenmerke (1 tot 22) toegevoeg is. SVM het deurgaans beter klassifikasieresultate gelewer. SVM en NN was nie beduidend (negatief) beïnvloed deur ʼn toename in kenmerkdimensionaliteit nie, maar ML se vermoë om te presteer onder toestande van groot kenmerkdimensionaliteite en min opleidingsareas was beperk. Verdere ondersoeke met ʼn verskeidenheid beelde, klassifikasie skemas en addisionele kenmerke; die vind van optimale kombinasies van opleidingstelgrootte en getal kenmerke; en die bepaling van die effek van spesifieke kenmerke sal waardevol wees in die ontwikkelling van meer koste effektiewe metodes om groot volumes satellietbeelde te prosesseer. TREFWOORDE Gerigte klassifikasie, landdekking, steunvektormasjien, naaste naburige klassifikasie, maksimum waarskynlikheidsklassifikasie, geografiese objekgebaseerde beeldanalise

The topic of this study is to evaluate data sources from the perspective of the information about landscape they provide to their users. The aim is comparison of thematic content of Urban Atlas database and data from Czech cadastre of real estate in Prague metropolitan region between years 2006 and 2012 with focus at meaning of classification systems used by both datasets. The data are processed by evaluation of thematic similarity and statistical tools which quantify similarity between researched data. Results are further verified by using validation data. Important results are visualized by charts, tables and maps. The areas with high degree of dissimilarity were found using chosen methods and their thematic characteristics were further examined as well as their major causes. It was proved that differences between both datasets are significant and they share certain characteristics. It was also proved that cadastral data are to high extent out-of-date. Keywords Urban Atlas, cadastre of real estate, land use, land cover, Prague

Master of Arts / Geography / Marcellus M. Caldas / As concerns increase over climate change, energy independence, and higher fuel prices, Brazilian sugarcane ethanol is seen as a part of a clean energy future. Brazilian sugarcane ethanol was developed with a long history of government support, and with the introduction of the flex fueled car in 2003, demand rose dramatically. These factors have helped sugarcane expand beyond its traditional regions of Brazil into the Cerrado. More recently however, private capital from both domestic and foreign companies have started investing in Brazilian agriculture and these investments have helped fuel the sugarcane expansion into the Cerrado in the last 15 years. Over 22 sugarcane mills have been constructed in the Brazilian state of Goiás, located in the heart of the Cerrado. The increased investments driving the expansion of sugarcane into the Cerrado brings numerous questions regarding its environmental and social impacts. Thus, the goal of this thesis is to understand how the structural organization of the sugarcane ethanol mills’ affects development at a municipality level in the state of Goiás, Brazil. More specifically, this thesis has two objectives; to evaluate the effects of the sugarcane mills’ influence on land use and land cover change in these municipalities; and to compare how domestic owned mills, foreign owned mills, and jointly owned mills affect socioeconomic development on the municipalities. Three municipalities were analyzed, Edéia, Caçu, and Quirinópolis. Results showed that land use and land cover change varied by municipality. The majority of Edéia’s sugarcane expansion came from lands already in agricultural use. On the other hand, Caçu’s and Quirinópolis’s sugarcane expansion came from pasture lands. However, throughout all the municipalities, sugarcane expansion over native vegetation was small. All three municipalities increased their socioeconomic development levels over the past 20 years as reported on the Human Development Index. In addition, urban survey responses revealed that the residents of Edéia perceived the sugarcane mill had made their lives better than respondents in either Quirinópolis or Caçu. However, this analysis covers only a brief period of time, and future analysis of these, and other municipalities that host sugarcane mills throughout the Cerrado will be needed.

The human-induced changes of the Earth's land surfaces have been unprecedented, with outcomes often indicating degradation and loss of environmental quality. Mato Grosso State in Brazil, location of the study area, underwent extensive land-use and land-cover changes in recent decades with the rates, patterns and consequences poorly documented until now. In this context, the aim of the present research is to propose a multidisciplinary approach for quantifying historical land-use and environmental changes in the southeast part of this State, where the Cerrado biome (Brazilian savannas) has been intensively converted into agricultural lands. The methodology includes three parts: remote sensing change detection, land vulnerability mapping, and identification of key environmental indicators. Land-use/cover information was extracted from a temporal remote sensing dataset using an object-oriented classification approach, and the changes quantified employing a post-classification method. In addition, the study area was assessed for its vulnerabilities, focusing mainly on erosion risks, wetlands, and areas with limited or no suitability for crops. Finally, key environmental indicators were identified from the preceding steps and analyzed within the Organisation for Economic Co-operation and Development (OECD) Pressure-State-Response (PSR) framework. The results provided an improved mapping of the Cerrados natural vegetation conversion into crops and pastures, and indicate that the Cerrado vegetation was intensively converted and also became more fragmented in the time frame studied. Between 1985 and 2005 the area lost approximately 6491 km 2 of Cerrados (42 %). Modeling based on the Universal Soil Loss Equation indicated significant increase in erosion risk from 1985 to 2005 mainly related to the increase in crop areas and the crops' encroachment into more fragile lands.The identification of environmental indicators rendered complex environmental information more generally accessible by structuring it within the PSR framework.The indicators captured key information about land-use and environmental changes in the area, showing that agricultural expansion is the major human activity exerting pressure on natural resources at a landscape scale, and that the pattern of change included high rates of crop expansion and the use of fragile environments such as wetlands and sandy erodable soils.

This thesis aims to establish analytical frameworks to map urban growth patterns with spaceborne remote sensing data and to evaluate environmental impacts through Landscape Metrics and Ecosystem Services. Urbanization patterns at regional scale were evaluated in China's largest urban agglomerations and at metropolitan scale in Shanghai, Stockholm and Beijing using medium resolution optical satellite data. High-resolution data was used to investigate changes in Shanghai’s urban core. The images were co-registered and mosaicked. Tasseled Cap transformations and texture features were used to increase class separabilities prior to pixel-based Random Forest and SVM classifications. Urban land cover in Shanghai and Beijing were derived through object-based SVM classification in KTH-SEG. After post-classification refinements, urbanization indices, Ecosystem Services and Landscape Metrics were used to quantify and characterize environmental impact. Urban growth was observed in all studies. China's urban agglomerations showed most prominent urbanization trends. Stockholm’s urban extent increased only little with minor environmental implications. On a regional/metropolitan scale, urban expansion progressed predominately at the expense of agriculture. Investigating urbanization patterns at higher detail revealed trends that counteracted negative urbanization effects in Shanghai's core and Beijing's urban-rural fringe. Beijing's growth resulted in Ecosystem Services losses through landscape structural changes, i.e. service area decreases, edge contamination or fragmentation. Methodological frameworks to characterize urbanization trends at different scales based on remotely sensed data were developed. For detailed urban analyses high-resolution data are recommended whereas medium-resolution data at metropolitan/regional scales is suggested. The Ecosystem Service concept was extended with Landscape Metrics to create a more differentiated picture of urbanization effects.​ / <p>QC 20160205</p>

Sustainable Built Environments Senior Capstone Project / This study analyzes 15 years of wildlife tracking data across more than 40 transects in the Sky Islands surrounding landscape to investigate how human-wildlife encounters may respond to a decade of land development. The average detection of species per visit (ADPV), the quantification for human-wildlife encounters and indicator of species richness, was calculated for each transect across two sample periods 2001-2011 and 2011-2015. ArcMap was used to visualize the ADPV across sampling sites in the Sky Islands region. The p-value was then calculated to determine whether there was a significant difference between the ADPV of all species and of focal species before and after 2011. The results concluded there was no significant difference and the null hypothesis was accepted.

Thesis (MScConsEcol)--Stellenbosch University, 2012. / ENGLISH ABSTRACT: Wetland systems provide vital hydrological ecosystem goods and services to mankind. When wetlands are transformed, through invasion by alien plants or replaced with agriculture, natural capital is lost, and the system is no longer able to provide the same quality of hydrological ecosystem services. Natural capital can be restored, but it involves substantial financial investment, and there is no guarantee that these hydrological ecosystem services will be fully recovered. This thesis aimed to investigate the hydrological impact of the land-cover changes in the Kromme River Catchment over the last 50 years, by using a combination of mapping and hydrological modelling techniques. We hypothesized that wetland loss in the Kromme has resulted in a shift in the flow regime, greater responsiveness to floods as a result of less storage, lower baseflow, and reduced water quality. We also hypothesised that the riparian invasion by Acacia mearnsii has caused flow reductions as a result of increased evaporation relative to the wetlands. Modelling results predict that over the past 50 years, the transformation of the floodplain wetlands in the Kromme River has shifted the flow regime, reducing baseflows and increasing the responsiveness of the catchment to extreme rainfall events. The invasion of A. mearnsii over time has also been predicted to have caused a reduction in river flow. Various restoration scenarios were considered, however if the Kromme were to be restored back to a land-cover state comparable to the 1950’s, 26.9 km2 (65.1%) of A. mearnsii would have to be cleared, and 5.2 km2 (34.2%) of the wetlands would have to be restored. The hydrological benefits would include a predicted increase in riverflow (42 mm/a), baseflow (2.9 mm/a), an increase in flood protection and improved water quality. This restoration strategy could be regarded as a type of insurance plan, and the benefits gained in terms of increased ecosystem service delivery would be the insurance premium. In conclusion it appears that restoration, insuring natural capital in the Kromme River, would provide significant economic returns on investment. / AFRIKAANSE OPSOMMING: Moeraslandstelsels voorsien die mens van noodsaaklike hidrologiese ekosisteemgoedere en -dienste. Wanneer moeraslande verander word, hetsy deur die indringing van uitheemse plante of vervanging met landboubedrywighede, gaan natuurlike kapitaal verlore en kan die stelsel nie meer dieselfde gehalte hidrologiese ekosisteemdienste lewer nie. Hoewel natuurlike kapitaal herwin kan word, behels dit beduidende finansiële belegging, en is daar boonop geen waarborg dat die hidrologiese ekosisteemdienste ten volle sal herstel nie. Hierdie tesis het ten doel gehad om die hidrologiese impak van die grondbedekkingsveranderinge in die Krommerivier-toeloopgebied oor die afgelope 50 jaar met behulp van ’n kombinasie van karterings- en hidrologiese modelleringstegnieke te ondersoek. Die hipotese was dat moeraslandverlies in die Kromme tot ’n verandering in die vloei-regime, hoër responsiwiteit op erge reënval as gevolg van minder bergingsruimte, ’n laer basisvloei en swakker watergehalte gelei het. Daar is voorts gehipoteseer dat die oewerindringing deur Acacia mearnsii ’n verlaging in vloei veroorsaak het weens ’n toename in verdamping uit die moeraslande. Modelleringsresultate dui daarop dat die transformasie van die vloedvlakte-moeraslande in die Krommerivier oor die afgelope 50 jaar die vloei-regime verander het, basisvloei verminder het en die toeloopgebied se responsiwiteit op erge reënval verhoog het. Die indringing van A. mearnsii het ook volgens aanduidings mettertyd ’n vermindering in riviervloei tot gevolg gehad. Verskeie herstelscenario’s is oorweeg. Om die grondbedekking in die Kromme te herstel tot wat dit in die 1950’s was, moet 26,9 km2 (65,1%) van die A. mearnsii verwyder en 5,2 km2 (34,2%) van die moerasland herwin word. Die hidrologiese voordele kan ’n verwagte toename in riviervloei (42 mm/a) en basisvloei (2,9 mm/a), ’n toename in vloedbeskerming sowel as beter watergehalte insluit. Hierdie herstelstrategie kan as ’n soort versekeringspolis beskou word, en die voordele verbonde aan beter ekosisteemdienslewering as die versekeringsuitbetalung. Ten slotte blyk dit dat die herstel van die Kromme, en die gepaardgaande versekering van natuurlike kapitaal, beduidende ekonomiese opbrengste op belegging sal meebring.

This dissertation focuses on the lived, material realities of rural women, men, girls, and boys struggling to make a living in the context of changing national development priorities and changing environmental conditions in Southern Province, Zambia. Over the last 20 years, Gwembe Tonga migrants living in the frontier farming area of Kulaale have witnessed significant declines in non-cultivated “bush” resources due to the conversion of forest and grassland to agricultural uses. This dissertation seeks to understand how women, men, boys, and girls differently experience these declines according to local gender- and age-based divisions of subsistence labor. Drawing on a variety of theoretical lenses—including Feminist Geography, Feminist Political Ecology, African Feminisms, the Anthropology of Childhood, and the Anthropology of the State—and utilizing a unique blend of qualitative/ethnographic and quantitative/geospatial research methods, this study finds that the “extractive workloads” (the average annual distance traveled for the collection of key bush resources) associated with women, men, girls, and boys are both unequal and contrary to recent speculations about the distinctive vulnerability of adult women to environmental change. The unequal labor burdens associated with the extraction of bush resources in this changing frontier landscape are but one of several missing “links” that this dissertation identifies within current theorizing about the gendered dimensions of environmental change. Other “links” include the social organization and religious life of Gwembe Tonga migrants, the demographic structure of Kulaale homesteads (their organization on the landscape and their demographic composition), the interplay between agency and vulnerability in children’s daily lives, and the role of the state in shaping Kulaale residents’ perceptions of and interactions with the surrounding environment. This story of Gwembe Tonga migrants’ gendered and aged experiences of environmental change unfolds in the context of competing national economic strategies—frontier development wildlife conservation. This dissertation concludes that women, men, girls, and boys are all physically and economically vulnerable to the changes associated with frontier development, conservation policy, and environmental change, with social, political, and economic factors prompting them to experience vulnerability in aged and gendered ways.

This dissertation addresses gender dimensions of Land Use and Land Cover Change (GLULCC) in the last few decades in a collective land titled to Afro-descendant communities in the Pacific region of Colombia, South America, and examines socio-economic and political signifiers affecting land use decisions, rights, and responsibilities. It shows how contrasting but complementary subfields of investigation, Political Ecology and Land Use Science, have contributed ontological, epistemological and practical scholarly works to help better understand the Gender Dimensions of Land Use and Land Cover Change (GLULCC). Historical and current information on environmental, socioeconomic and settlement processes provided a comprehensive portrait of the study area. The remote sensing process (a mainstream method for identifying land use and land cover change) helped exploring the spatial setting of land cover/use, and to reflect on the opportunities and constrains of the steps undertaken during this procedure under the lenses of researching their gendered dimensions. Statistical analyses on both census data (secondary data) and survey sample data (fieldwork data) allowed to establish a set of three groups of gendered land uses, namely, women-akin, men-akin, and gender-blind uses. Exploratory statistics, pairwise correlations, and binary and multinomial logit regression models helped to reassert the latter gendered categories’ assertions. A concluding narrative perspective of GLULCC seeks to further contribute to work streaming/ mainstreaming what I consider may be a scholarly-fertile research line. It hopes to bond, with another perspective, previous theoretical, spatial and quantitative outcomes, under the lenses of the practical experience of fieldwork, which also by way of participatory observation and semi-unstructured interviews brought to the researcher (me) valuable insights and information besides the previous outcomes. Empirical evidence allowed identifying gender-based time allocation, resource-use power relations, and reproductive strategies. Finally, the found rearrangement of settlement spaces and production systems provides practical indications that women´s role on LULCC is well beyond the establishment of small gardens and orchards, or the collection of fuel wood to provide for their families. In contrast, inside this collective title, women’s decisions/strategies have also restructured settlement patterns, and thus, land use dynamics of larger areas at heterogeneous spatial and temporal scales. / text

New forms of disturbance, and alteration of current disturbance regimes in arid and semiarid ecosystems, have resulted in the modification and degradation of large regions. This research explores vegetation response as a consequence of two different disturbance events in the southwestern US and northwestern Mexico. This topic was explored in this dissertation utilizing remotely sensed geospatial information in three separate studies.The first study explores the development of methods to assess the effectiveness of pre-fire restoration efforts, by evaluating vegetation response as a function of local environmental variables. Here I evaluated three fire locations at Bandelier National Monument (New Mexico). My models explain post-fire vegetation response as a function of environmental inputs and pre-fire site conditions (restored, unrestored and control areas). However, further analysis will be needed to better understand the effect of pre-fire restoration techniques on post-fire vegetation response.My second study explores the development of monitoring practices using remotely sensed data to assess land cover dynamics through time. The study area was the arid agro-ecosystem of La Costa de Hermosillo (LCH) in northwestern Mexico. My results show a continuous tendency towards a decrease in agriculture from 1988 until 2009. Detailed change detection demonstrates high rates of change from agriculture to other land cover classes in areas with dense agricultural developments. Implementation of these monitoring protocols would help with the application of restoration practices.The third study we used remote sensing time series data to assess phenological trends and variability among land cover types in relation to climatic variability within communities present in a heavily impacted agro-ecosystem (LCH). My analysis comprised three different agricultural land cover types including abandoned agricultural fields, and four additional natural land cover classes. I found that productivity has not increased in abandoned fields (since abandonment). Furthermore, I found that the models developed in this study significantly explain phenological variability as a function of climatic variability.These studies suggest that the use of remote sensing tools could effectively contribute to our ability to monitor vegetation dynamics in arid ecosystems. The implementation of methodologies generated in this work would significantly inform managers in decision making processes.