Thesis (MSc)--University of Stellenbosch, 2005. / ENGLISH ABSTRACT: This study investigated the composition of Themeda triandra Renosterveld in part of the Grootvadersbosch Conservancy and the effects of selected environmental and management variables to provide guidelines for promoting the presence of Themeda triandra in the veld. The Zürich-Montpellier phytosociological method was used to determine the composition of the Renosterveld communities. The point quadrat method was used to determine the cover of Themeda triandra at three grass dominated sites and compare cover from one site with past cover measurements at the specific site. Ordination was used to examine the effects of the environmental and management variables on the plant communities. Two community groups, five communities and five subcommunities were identified and described. The Themeda triandra – Stoebe phyllostachys Grassland Community Group consists of two communities of which one has two subcommunities. The Themeda triandra – Elytropappus rhinocerotis Shrubland Community Group consists of three communities of which one has three subcommunities. The vegetation units described in this study have not been described previously in the literature. One community belongs to Silcrete Fynbos while another subcommunity is transitional between Renosterveld and thicket and gallery forest. The rest of the plant communities fit the definition for Renosterveld (the local Renosterveld type is Eastern Rûens Shale Renosterveld). Cover of Themeda triandra did not differ significantly either between sites or between years. The comparison of Themeda triandra cover between years was done at a site that had been burnt between the last two sampling times yet the cover was not significantly different. This indicates that fire and other management practices did not have a negative impact on Themeda triandra at the site. Unconstrained ordination of the dataset in which cover/abundance values were included, grouped the relevés by moisture and disturbance but the presence/absence dataset indicates that the two community groups have a slight transitional overlap. Constrained ordination of both datasets with a) soil variables and b) management variables, both showed a tendency to separate the relevés into community groups, that did not happen with topographic and vegetation variables. Ordination did not separate the community groups into their subdivisions. The soil variables (both nutrients and texture) influence the vegetation structure and the community distribution. Under the existing grazing management regime, fire at three to five year intervals promoted the dominance of Themeda triandra by affecting the structure of the plant communities, the abundance of species other than Themeda triandra, and influencing which plant community is present. The use of fire as a management tool was regulated by the importance of the natural veld pastures to the farmers. The natural veld pastures are not suitable for dairy cattle in milk. Thus dairy farmers are less likely to burn the natural veld (no planned burns only chance fires) than those who farm with mutton or beef (planned burns on a three to five year interval). / AFRIKAANSE OPSOMMING: Die studie het die samestelling van Themeda triandra Renosterveld in ’n deel van die Grootvadersbosch-Bewaria en die uitwerking van geselekteerde omgewings- en bestuursveranderlikes ondersoek om riglyne vir die bevordering van Themeda triandra in die veld daar te stel. Die Zürich-Montpellier fitososiologiese metode is gebruik om die samestelling van die Renosterveldgemeenskappe te bepaal. Die puntkwadraat-metode is gebruik om ’n skatting van die dekking van Themeda triandra by drie gras-gedomineerde persele te bepaal en om ’n vergelyking te maak tussen die huidige en vorige dekking van ’n enkele perseel. Ordinasie is gebruik om die invloed van omgewings- en bestuursveranderlikes op die plantgemeenskappe te bepaal. Twee gemeenskapsgroepe, vyf gemeenskappe en vyf subgemeenskappe is geïdentifiseer en gedefinieer. Die Themeda triandra – Stoebe phyllostachys Grasland-gemeenskapsgroep bestaan uit twee gemeenskappe waarvan een in twee subgemeenskappe onderverdeel is. Die Themeda triandra – Elytropappus rhinocerotis Struik-gemeenskapsgroep bestaan uit drie gemeenskappe waarvan een in drie subgemeenskappe onderverdeel is. Die plantegroei-eenhede wat in die studie beskryf is, is nie voorheen in die literatuur beskryf nie. Een gemeenskap behoort aan Silkreet-fynbos en ’n ander subgemeenskap is ’n oorgangsfase tussen Renosterveld en struikbosveld of woud, terwyl die res van die plantgemeenskappe binne die definisie van Renosterveld val (die plaaslike Renosterveld tipe staan bekend as Oostelike Rûens Skalierenosterveld). Die bedekking van Themeda triandra het nie betekenisvol gevarieer tussen óf die verskillende lokaliteite óf die verskillende jare nie. Die vergelyking van Themeda triandra-bedekking oor tyd is onderneem in ’n gebied wat tussen opnames gebrand is. Die bedekking het nie betekenisvol verskil nie. Dit dui aan dat vuur en ander bestuurspraktyke nie ’n negatiewe invloed op Themeda triandra in hierdie gebied het nie. Onbeperkte ordinasie van die datastel met die vergelyking van bedekking/volopheidwaardes, groepeer die relevés volgens vogtigheid en versteuring, terwyl die datastel ten opsigte van teenwoordigheid/ afwesigheid aandui dat die twee gemeenskapsgroepe ’n effense oorgangs-oorvleueling het. Beperkte ordinasie van beide datastelle met a) grondveranderlikes en b) bestuursveranderlikes, toon albei ’n neiging om die gemeenskapsgroepe te skei, wat nie gebeur het met die topografiese- en plantegroeiveranderlikes nie. Ordinasie het nie die gemeenskapsgroepe onderverdeel in gemeenskappe of subgemeenskappe nie. Die grondveranderlikes (beide voedingstowwe en tekstuur) beïnvloed die struktuur en die verspreiding van die plantegroei. Met die bestaande weidingsbestuur bevorder brande met ’n interval van tussen drie tot vyf jaar die oorheersing van Themeda triandra deur die struktuur van die teenwoordige plantgemeenskap te beïnvloed, deur die getal van die verskillende plantsoorte te beïnvloed en selfs deur die plantgemeenskap se voorkoms te beïnvloed. Die waarde wat die boer aan die natuurlike veld as weiveld heg bepaal die mate waartoe hulle veldbrand gebruik as ’n deel van bestuur. Die natuurlike veld is nie geskik vir melkkoeie wat in die melkproduksiestadium is nie. Die melkboere is dus minder geneig om die natuurlike veld te brand (geen beplande brande nie net kans brande) as die boere wat met vleisbeeste of skape boer (beplande brande elke drie tot vyf jaar).

Maize (Zea mays L.) is the most important grain crop in the country. Approximately 12 million tons of maize grain is produced annually on approximately 2.5 million ha of land. However, increased farming intensity can lead to fragmentation of habitat and has a tendency to decrease the biodiversity of an area. Therefore, to ensure the continued functionality of agro-ecosystems, methods in agriculture must be assessed and adapted when necessary to ensure the persistence of biological diversity. Unfortunately, the effect of crop production on species diversity and composition in South Africa is still relatively unknown, and no baseline data exists with which to gauge the possibility of unknown extinction risks of important biological elements. The objectives of this study were to compare plant and arthropod diversity patterns and species turnover of maize agro-ecosystems between biomes (grassland and savanna) and along a maize field-field margin gradient (MAFFMAG). Surveys of maize agro-ecosystems were conducted in six provinces of South Africa, namely North-West, Mpumalanga, KwaZulu-Natal, Limpopo, Free State and the Eastern Cape. Repeated measures ANOVA revealed a significantly lower plant and arthropod species diversity and richness in maize fields compared to field margins. Non-metric multidimensional scaling revealed that arthropod species composition differed between biomes although not along MAFFMAGs, indicating that arthropod species composition is dependent on biome rather than distance from maize field. Floristic data revealed unique species compositions for maize fields and field margins and also for biomes. Furthermore, maize fields and field margins of grassland sites were more similar in plant species composition than the savanna localities, suggesting higher regional beta diversity for savanna regions. Spearman‘s rank order correlations revealed generally positive but weak or no relationships between plant and arthropod diversity. This study provides baseline data for identification, monitoring and conservation of priority species and will allow the future evaluation of ecosystem services provided by plants and associated arthropods, especially natural enemies of pests, in maize agro-ecosystems. / MSc (Environmental Sciences), North-West University, Potchefstroom Campus, 2014

The Kafue Flats floodplain wetland system in southern Zambia is under increasing climate and human pressures. Firstly, drought episodes appear more prevalent in recent years in the region and secondly, two dams were built on the lower and upper ends of the wetland in 1972 and 1978, respectively, across the Kafue River which flows through the wetland. The study uses multi-temporal remote sensing to assess change in extent and vigour of green vegetation, and extent of water bodies and dry land cover on the Kafue Flats. The change detection's management value is assessed. Four normalised, co-registered digital Landsat images from 24 September 1984, 3 September 1988, 12 September 1991 and 20 September 1994 were used. The main change detection method used was comparison of classifications, supplemented by Normalised Difference Vegetation Index (NDVI) and Principal Component Analysis (PCA) change detection. Ancillary land use and environmental data were used in interpreting the change in the context of cause and effect. The results indicate inconsistent trends in the changes of most land cover classes, as a result of manipulation of the wetland by man through annual variations in the timing and magnitude of regulated flows into the wetland, as well as burning. However, the results also show spatial reduction in the wetland's dry season dense green reed-grass vegetation in upstream sections which are not affected by the water backing-up above of the lower dam. Sparse green vegetation is replacing the dense green vegetation in these upstream areas. It is inferred that this dry season degradation of the wetland threatens bird species which may use the reeds for dry season nesting. It is proposed that ground surveying and monitoring work at the micro-habitat level is necessary to ascertain the implications of the losses. It is concluded that, in spite of difficulties, multi-temporal remote sensing has a potential role in wetland change assessment on the Kafue Flats at the community level, but that it needs to be supplemented by targeted, micro-habitat level ground surveys.

En Europe, la prairie occupe près de 40% de la surface agricole utile et fournit un ensemble de services environnementaux et agricoles, tout en constituant un réservoir de diversité végétale et animale. Cet écosystème herbacé, plurispécifique et multifonctionnel est un système biologique complexe qui fait interagir l’atmosphère, la végétation et le sol, via les cycles biogéochimiques, notamment ceux du carbone et de l’azote. Motivées par le maintien des biens et services des prairies face aux changements climatiques et atmosphériques, les recherches actuelles sur l’écosystème prairial s’attachent à étudier l’évolution des processus clés du système prairial (i .e. production, échanges gazeux, changements d’espèce) sous changement climatique complexe. Ce projet de thèse a pour objectif d’étudier in situ les impacts des principales composantes du changement climatique (température de l’air, précipitations, concentration atmosphérique en gaz carbonique) sur des prairies extensives de moyenne montagne. Nous cherchons à mettre en évidence les changements de structure et de fonctionnement de l’écosystème prairial sous l’influence d’un scénario de changement climatique prévu à l’horizon 2080 pour le centre de la France. Ce scénario (ACCACIA A2) prévoit une augmentation de 3.5°C des températures de l’air, une augmentation des concentrations atmosphériques en CO2 de 200 ppm et une réduction des précipitations estivales de 20 %. Nos résultats indiquent qu’à moyen terme (trois ans de traitements expérimentaux) le réchauffement a des effets néfastes sur la production annuelle du couvert végétal. L’effet bénéfique d’une élévation des teneurs en CO2 sur la production aérienne n’apparaît qu’à partir de la troisième année. La richesse spécifique (nombre d’espèces) et les indices de diversité taxonomique n’ont pas montré de variations significatives sous changement climatique. Cependant après trois années de réchauffement, l’abondance des graminées semble être altérée. Contrairement à la production, les traits sont plus affectés par la concentration en CO2 élevée que par le réchauffement. Après trois ans de traitements, des mesures d’échanges gazeux (CO2) à l’échelle du couvert végétal pendant la saison de croissance ont montré un effet négatif du réchauffement sur l’activité photosynthétique du couvert et une acclimatation de la photosynthèse au cours de la saison de croissance sous CO2 élevé. Ces tendances ont aussi été trouvées sur la photosynthèse foliaire d’une des espèces dominantes du couvert (Festuca arundinacea). L’effet négatif direct du réchauffement à l’échelle foliaire semble être associé à une diminution des sucres dans les limbes. L’acclimatation à l’enrichissement enCO2 à l’échelle foliaire, quant à elle, semble être indirectement dépendante du statu hydrique du sol. Notre étude a aussi porté sur l’analyse des échanges gazeux sol-atmosphère d’un des principaux gaz à effet de serre trace des prairies, l’oxyde nitreux (N2O). Malgré une forte variabilité inter- et intra- annuelle, les flux de N2O semblent être favorisés sous réchauffement. L’augmentation de la température affecte aussi positivement les taux de nitrification et leur pool microbien associé (AOB), et les rejets de N2O via dénitrification. De plus, les flux de N2O mesurés aux champs ont montré une corrélation plus forte à la taille des populations microbiennes (nitrifiantes et dénitrifiantes) en traitement réchauffé qu’en traitement témoin. En conclusion, la température semble être le facteur principal dans les réponses de cette prairie aux changements climatiques futurs. De plus, nos résultats suggèrent que le fonctionnement (production, émissions de N2O) des prairies extensives de moyenne montagne est plus vulnérable aux changements climatiques que la structure de la communauté végétale. / In France, the grassland ecosystem represents an important part of the total of agricultural landscape and provides important economic and ecological services. This multifunctional ecosystem is a complex biological system where atmosphere, plants and soil interact together,via the biogeochemical cycles (particularly carbon and nitrogen cycles). In order to maintain goods and services from grasslands in changing environmental conditions, current research on the grassland ecosystem focus on the evolution of key grassland processes (i.e. production,gaseous exchanges, biodiversity) under multiple and simultaneous climate change.This thesis addresses the impacts of the three main climate change drivers (air temperature, precipitation and atmospheric carbon dioxide concentrations) on an extensively-managed upland grassland in situ. We investigated changes in ecosystem function and structure under the influence of a projected climate scenario for 2080 for central France. This scenario (ACCACIA A2) comprises : air warming of 3.5°C, 20 % reduction of the summer precipitation and an increase of 200 ppm in atmospheric carbon dioxide (CO2).Our results indicate that in the medium term (after three years of experimental treatments), warming had negative effects on the annual aboveground production. Elevated CO2 had no significant effects on aboveground production initially, but positive effects on biomass from the third year onwards. Species richness and the indices of species diversity did not show significant differences in response to climate change, but warming was associated with a decline in grass abundance after three years. Contrary to biomass production, plant traits showed a stronger response to elevated CO2 than to warming. After three years of study, canopy-level photosynthesis showed a negative effect of warming but an acclimation to elevated CO2 during the growing season. This pattern was also found for leaf-level photosynthetic rates measured on a dominant grass species (Festuca arundinacea). For Festuca, the direct negative effect of warming was associated with a decrease in leaf fructan metabolism. In contrast, the photosynthetic acclimation under elevated CO2 observed in Festuca seemed closely linked to the indirect effect of soil water content. Our study also examined effects of climate change on one of the main trace greenhouse gases in grasslands, nitrous oxide (N2O). During our study, N2O fluxes showed significant inter-and intra-annual variability. Nevertheless, mean annual N2O fluxes increased in response to warming. Warming had a positive effect on nitrification rates, denitrification rates and the population size of nitrifying bacteria (AOB). Furthermore, field N2O fluxes showed a stronger correlation with the microbial population size in the warmed compared with the control treatment. Overall, warming seems to be the main factor driving ecosystem responses to projected climate change conditions for this cool, upland grassland. In addition, our results suggest that grassland function (aboveground production, N2O emissions) are more vulnerable to complex climate change than grassland community structure for our study system.

An analysis utilizing both ArcGIS and ethnographic interviews from private land owners and environmental professionals examined how man-made landscape changes affected plant and tree species in the LBJ National Grasslands in Wise County, Texas north of Decatur. From the late 1800s to the Dust Bowl Era the land was used for crop production and cattle grazing resulting in erosion and loss of soil nutrients. The research indicated by 2001 that cattle grazing and population increase resulted in land disturbance within the administrative boundary of the national grasslands. Participants expressed concern over the population increase and expansion of 5 to 10 acre ranchettes for cattle grazing common in modern times. Recommendations for the future included utilizing and expanding the resources already existing with environmental professionals to continue controlling erosion.

En Afrique Sub-Saharienne, les zones humides forment d'importantes zones de pâturages pour bon nombre de pastoralistes. Le régime d'inondation de ces plaines influence leur fonc tionnement et leur productivité. Pour comprendre et maitriser ces régimes, il est donc essentiel d'évaluer l'impact de l'infrastructure hydro-électrique sur les dynamiques d'inondations en aval des barrages. Pourtant, le manque de données disponibles, l'absence de modèles validés et la forte variabilité environnementale rendent cet exercice difficile. Cette thèse contribue à quantifier l'impact du changement des ressources hydriques sur la productivité fourragère des prairies inondables dans le Delta du fleuve Tana, au Kenya. 1/ Les caractéristiques de croissance et la productivité d'une prairie inondable à Echinochloa stagnina (Retz) P. Beauv. ont été déterminées pour différents régimes de fauche, d'irrigation et conditions d'inondations. Ensuite, un modèle de croissance adapté à des Graminées en C4, tropicales et pérennes, de prairies inondables a été développé. Il constitue, à notre connaissance, le premier modèle éco-physiologique adapté à ce type de prairies. 2/ Des processus hydrologiques importants pour le fonctionnement des écosystèmes (étendue, période, durée et fréquence d'inondation) ont été caractérisés grâce à un modèle de bilan hydrologique et à l'utilisation de techniques de télédétection, et cela en dépit de la faible instrumentalisation du bassin, du peu de données topographiques et d'un fort couvert nuageux. 3/ Une analyse préliminaire de l'impact de différents scénarios d'inondations sur la production fourragère a été effectuée en utilisant des indicateurs de services écosystémiques. Cette thèse participe à l'amélioration de nos connaissances des services écosystémiques des zones humides par la construction de modèles et par l'évaluation de scénarios dans une région du monde où ce type de données est rare. / Wetlands are a vital resource for many pastoralists in Sub-Saharan Africa as they provide dry-season grazing zones. As floods are essential for wetland ecosystems, the assessment of water abstraction and hydroelectric infrastructure on downstream flooding dynamics is crucial. Yet, scarce data, environmental variability and the lack of models make this challenging. This research contributes to quantifying the influence of changing water resources on fodder pro- duction of floodplain grasslands in the Tana River Delta, Kenya. 1/ Growth characteristics of floodplain grasslands of Echinochloa stagnina (Retz) P. Beauv. for different flood and manage- ment options were determined and a quantification of their productivity achieved. This kind of data is scarce for floodplain grasslands. A plant growth model adapted to tropical floodplain conditions and perennial C4 grasses was developed, and is the first known physiologically based model for floodplain grasslands. 2/ Hydrological processes of ecological importance (flood extent, timing, duration, frequency) were characterized in a poorly gauged basin using a water-balance model combined with remote-sensing techniques, despite precise knowledge of discharge rates, topography and a high cloud cover. 3/ A preliminary analysis explored different flooding scenarios and their impact on fodder production through the use of simple ecosystem service indicators. This PhD contributes to the repertoire of wetland ecosystem ser- vices by building biophysically based simulation models and exploring possible scenarios in a region of the world and an ecosystem where these type of evaluations are rare.

Doctor of Philosophy / Department of Agronomy / Ganga M. Hettiarachchi / Although phosphorus (P) is an essential nutrient for the growth of plants, it is one of the most limiting nutrients in terms of availability as a high proportion of applied P rapidly transforms into insoluble forms with low solubility in soils. To further understand the fate of P applied to soils, two separate but related studies using three high P-fixing soil types each were used for which the objectives were to investigate the mobility, availability, and reaction products from two granular and one liquid P fertilizer alone or plus a fertilizer enhancement product. Energy dispersive spectroscopy showed a substantial amount of P remained in the granule following a 5-week incubation. At the end of the 35-day incubation period there was evidence that the fluid fertilizer was superior over the granular sources in terms of enhanced diffusion and extractability of P for three calcareous soils with varying levels of CaCO₃. Phosphorus x-ray absorption near-edge structure (XANES) spectroscopy results in conjunction with resin-extractable P indicated a strong negative correlation between Ca-P solids formed and P extractability, suggesting that degree of Ca-P formation limits P solubility. For the three acidic P-fixing soils the results were complex. In two out of three acid soils, liquid P treatments diffused farther from the application point than the granular treatments. Phosphorus XANES results suggested that Fe-P or Al-P interactions control the overall P solubility. Integration of pH, resin extractable-P and XANES results suggested the P retention mechanism was either dominated by adsorption or precipitation depending on soil pH. More acidic soil conditions favored precipitation. The objectives of the third study were to observe how long-term (14 years) addition of P with or without N influences the inorganic and organic P pools in a native grassland soil using sequential fractionation, XANES, and ³¹P-nuclear magnetic resonance (NMR) spectroscopy. The overall results suggested that P and N fertilization and associated changes in plant productivity induced significant changes in soil P pools such as Ca-P, phytic acid, monoesters, and residual forms of P. The addition of P alone induced formation of inorganic P forms while the addition of P and N induced transformation of residual P forms into more labile and/or organic P forms.

Le carbone organique des sols (COS) joue un rôle majeur dans le maintien des propriétés des sols, et constitue un important réservoir de carbone sensible aux perturbations anthropiques dont les changements d'usage ou de gestion des terres. En Amazonie, la déforestation libère des gaz à effet de serre (GES) par le brûlis de la végétation mais les stocks de COS sont également susceptibles d'évoluer et de contribuer aux émissions de GES. Nous montrons que ces variations de stocks sont cependant mal comprises à l'échelle du biome, en raison de l'approche par chrono-séquence qui induit de nombreuses incertitudes et du manque de données sur la gestion des agrosystèmes implantés après déforestation. L'étude a été conduite sur un site agronomique diachronique en Guyane, déboisé sans brûlis et avec des restitutions de biomasse forestière au sol, sur lequel ont été implantés 3 systèmes de culture : une prairie et deux systèmes maïs/soja avec ou sans travail du sol. L'objectif a été de déterminer le devenir du carbone forestier et du COS des agrosystèmes. Les fluctuations des stocks ont été mesurées jusqu'à 5 ans après déforestation dans la couche 0-30 cm, un bilan est proposé pour la couche 0-100 cm à 5 ans. La décomposition des débris de bois, apportés au sol suite à la déforestation, a été étudiée via leur perte de masse et leur caractérisation par pyrolyse Rock-Eval. La répartition granulométrique du COS a été mesurée 4 ans après déforestation. L'isotopie δ13C a été utilisée dans le sol sous prairie pour distinguer le carbone d'origine prairiale. L'apport de carbone issu de la déforestation a entrainé une augmentation des stocks de COS, mais de courte durée car les débris de bois se sont rapidement décomposés et n'ont pas induit de stockage durable de COS. Cinq ans après déforestation les stocks de COS sous prairie sont similaires à ceux observés sous forêt, grâce à des apports de carbone importants par les racines, alors que sous cultures les stocks diminuent d'environ 18 %, sans que l'on ait distingué un effet du travail du sol. La décroissance du carbone forestier, qui concerne l'ensemble des fractions granulométriques du sol, a donc été compensée par les apports de carbone sous prairie, ce qui n'est pas le cas sous cultures annuelles. Le modèle RothC a pu être validé dans notre situation même s'il a surestimé légèrement les stocks sous cultures. Nos résultats, replacés dans le contexte amazonien montrent que les diminutions de COS observées ici sont moins importantes que pour l'ensemble des tropiques humides, probablement en raison de la gestion optimale du site et de la courte durée du temps d'observation. / Soil organic carbon is a key component of soil quality, and represents a large part of the terrestrial carbon stock, sensitive to human perturbations including land-use change. In Amazonia, deforestation induces greenhouse gases (GHG) emissions due to vegetation burning, but SOC stocks also change, which can induce GHG emissions. We show that these changes are misunderstood at the biome scale, because of the chronosequence approach that induces uncertainties, and because of the lack of management data of the agrosystems established after deforestation. We studied here an agronomic trial with a diachronic approach in French Guiana, deforested with a fire-free method that returned large amount of forest organic matter. Three agrosystems were set up: a grassland and two annual crop systems (maize/soybean) with and without soil tillage. We aimed to measure the fate of forest carbon and of SOC in the agrosystems. SOC stocks fluctuations were assessed up to 5 years after deforestation in the layer 0-30 cm, and a comparison forest-agroecosystems in the layer 0-100 cm was done at 5 years. Decomposition of woody debris buried in the soil after deforestation was assessed by mass loss approach and Rock-Eval pyrolysis. SOC distribution in granulometric fractions was measured 4 years after deforestation. δ13C methods were used in the grassland soil to distinguish the carbon derived from forest or grassland. We found that carbon inputs from deforestation increased SOC stocks, but only at short-term because woody debris decomposition was fast and did not induce a mid-term SOC storage. Five years after deforestation SOC stocks in grassland are similar to the forest, thanks to carbon inputs from root activity. In the annual crops SOC stocks decrease of about 18 %, and no difference is found according to the soil tillage. The decay of forest soil carbon, which affected the whole granulometric fractions of SOC, is thus offset in grassland but not in annual crops. RothC model could be validated in our study, but slightly overestimated SOC stocks in annual crops. Replaced in the Amazonian context, our results showed that the SOC decrease here was lower than other studies across humid tropics. This can probably be explained by the optimal management of the agrosystems, and the short time lapse studied.

A research report submitted to the Faculty of Science, University of the Witwatersrand, and Johannesburg in partial fulfillment of the requirements for the degree Master o f Science (Geographical Information System and Remote Sensing) Johannesburg, 2017. / The purpose of this analysis was to map the frequency of veld fires using remote sensing data from 2012 to 2016. The analysis successfully answered three objectives which are mapping the spatio-temporal pattern of veld fires in Mazowe district from 2012 to 2016, assessing the strength of association between burnt area and land cover classes and to determine the degree of veld fires in Mazowe district. The analysis of the veld fire frequency was undertaken in ENVI 5.3 and Arc Map software. Landsat imagery and MODIS fire products were collected and processed. For each year two Landsat images were used, one image before the season of fire starts which was used to classify the land cover classes and one image after fire season which was used to classify the burnt and unburnt classes. The MODIS fire products data were used to validate the burnt and the unburnt classification. The evaluations of the classifiers were done through accuracy assessment using confusion matrix and the results ranged from 85 to 95%.The study quantified the relationship between burnt areas land cover classes. The study also calculated the fire frequency. The results revealed that the veld fire frequency was high in A1 farms which measures 5 hectares, A2 farms which measures average of 318 hectares and grasslands compared to other land tenure and land cover classes. Areas with high frequency were observed in south, south west and some central parts of Mazowe district. There was high fire occurrence in 2012 and 2014. The study also noted that the fire occurrence was gradually decreasing, however the levels of fire occurrence remains high. The study concluded that A1, A2 farms and grasslands are prone to veld fires. The study recommends adaption of remote sensing techniques in eradicating veld fires. / MT 2017

Les prairies représentent 45% de la surface agricole en France et 40% en Europe, ce qui montre qu’il s’agit d’un secteur important particulièrement dans un contexte de changement climatique où les prairies contribuent d’un côté aux émissions de gaz à effet de serre et en sont impactées de l’autre côté. L’enjeu de cette thèse a été de contribuer à l’évaluation des incertitudes dans les sorties de modèles de simulation de prairies (et utilisés dans les études d’impact aux changements climatiques) dépendant du paramétrage du modèle. Nous avons fait appel aux méthodes de la statistique Bayésienne, basées sur le théorème de Bayes, afin de calibrer les paramètres d’un modèle référent et améliorer ainsi ses résultats en réduisant l’incertitude liée à ses paramètres et, par conséquent, à ses sorties. Notre démarche s’est basée essentiellement sur l’utilisation du modèle d’écosystème prairial PaSim, déjà utilisé dans plusieurs projets européens pour simuler l’impact des changements climatiques sur les prairies. L’originalité de notre travail de thèse a été d’adapter la méthode Bayésienne à un modèle d’écosystème complexe comme PaSim (appliqué dans un contexte de climat altéré et à l’échelle du territoire européen) et de montrer ses avantages potentiels dans la réduction d’incertitudes et l’amélioration des résultats, en combinant notamment méthodes statistiques (technique Bayésienne et analyse de sensibilité avec la méthode de Morris) et outils informatiques (couplage code R-PaSim et utilisation d’un cluster de calcul). Cela nous a conduit à produire d’abord un nouveau paramétrage pour des sites prairiaux soumis à des conditions de sécheresse, et ensuite à un paramétrage commun pour les prairies européennes. Nous avons également fourni un outil informatique de calibration générique pouvant être réutilisé avec d’autres modèles et sur d’autres sites. Enfin, nous avons évalué la performance du modèle calibré par le biais de la technique Bayésienne sur des sites de validation, et dont les résultats ont confirmé l’efficacité de cette technique pour la réduction d’incertitude et l’amélioration de la fiabilité des sorties. / Grasslands cover 45% of the agricultural area in France and 40% in Europe. Grassland ecosystems have a central role in the climate change context, not only because they are impacted by climate changes but also because grasslands contribute to greenhouse gas emissions. The aim of this thesis was to contribute to the assessment of uncertainties in the outputs of grassland simulation models, which are used in impact studies, with focus on model parameterization. In particular, we used the Bayesian statistical method, based on Bayes’ theorem, to calibrate the parameters of a reference model, and thus improve performance by reducing the uncertainty in the parameters and, consequently, in the outputs provided by models. Our approach is essentially based on the use of the grassland ecosystem model PaSim (Pasture Simulation model) already applied in a variety of international projects to simulate the impact of climate changes on grassland systems. The originality of this thesis was to adapt the Bayesian method to a complex ecosystem model such as PaSim (applied in the context of altered climate and across the European territory) and show its potential benefits in reducing uncertainty and improving the quality of model outputs. This was obtained by combining statistical methods (Bayesian techniques and sensitivity analysis with the method of Morris) and computing tools (R code -PaSim coupling and use of cluster computing resources). We have first produced a new parameterization for grassland sites under drought conditions, and then a common parameterization for European grasslands. We have also provided a generic software tool for calibration for reuse with other models and sites. Finally, we have evaluated the performance of the calibrated model through the Bayesian technique against data from validation sites. The results have confirmed the efficiency of this technique for reducing uncertainty and improving the reliability of simulation outputs.