Thesis (PhD)--Stellenbosch University, 1975. / Some digitised pages may appear illegible due to the condition of the original hard copy. / ENGLISH ABSTRACT: A study was made of certain relations operating within a South West African woodland savanna ecosystem which was characterised by simplification within most of its basic components. . Main attention was given to the relatively unexplored field of spatial and structural organisation of the primary producer component in relation to its quantified intra-seasonal functional changes including its phenology. Simple mathematical model descriptions of seasonal mass change of the herbaceous parts of the primary producer component were derived. Various parts of the primary producer component were also evaluated in terms of some of the main climatic driving forces of the ecosystem, decomposition relations in the form of disappearance rates, soil-vegetation mineral relations and consumer (mainly insect) relations. Gross effects of manipulation of the ecosystem type by fire and management is included.

Thesis (Ph. D.)--University of Wisconsin--Madison, 1955. / Typescript. Includes abstract and vita. Title from title screen (viewed May 9, 2006). Includes bibliographical references (leaves 167-174). Online version of the print original.

The detection of hierarchically nested structure in a semi-arid savanna as predicted by complex systems theory requires a method that detects context specific multi-scaled pattern in a proxy that represents the net effect of system processes. Statistical assumptions preclude the use of many traditional methods in the detection of hierarchical structure in heterogeneous landscapes so to circumvent statistical barriers to inference I developed a linear scale-space based application to represent multi-scaled woody vegetation structure in a spatially explicit manner. Analysis of a scale-space representation of woody cover across multiple scales explicitly recognizes landscape context and emergent pattern due to the causality principle inherent linear scale-space generation. As a proxy for process in scale and space I utilize the merge events of woody canopy cover, which should theoretically be considered the point at which processes shift domain. Scale-space representations were analyzed using a spatially explicit discontinuity analysis that compares the distribution of structure across the dimension of scale to that of a neutral model specific to the landscape in question. The application was tested for rigor and ability to detect multi-scaled, context dependent pattern in test datasets. The effects of fire and herbivory on the multi-scaled structure of a semi-arid savanna landscape were compared using the merge events from scale-spaces generated from a 33 year herbivore browser exclosure. No more hierarchical structure is present in real world savannas than can be expected from random. Hypotheses put forward to explain the results include: procedural and philosophical bias, errors in the application, or that the landscapes are not hierarchically nested. Each hypothesis is discussed in the light of the evidence and after synthesis I discuss that savanna landscapes may have more randomness within the pattern and process than previously acknowledged

A thesis submitted to the Faculty of Science, University of the Witwatersrand, Johannesburg in fulfilment of the requirements for the degree of Doctor of Philosophy, 2017 / Plant phenology, the study of the timing of biological events such as leaf flush, has been identified as a key tool in monitoring and understanding the impacts that changing climates may have on the world. Unfortunately there is a noticeable lack of phenological research in Africa’s savannas in comparison to other global ecosystems, such as temperate forests. Savannas are known for their complexity in structure and composition and the phenology of their flora is no exception to this. Owing to the highly seasonal climates in this system, plants have had to evolve a range of phenological strategies to cope with the limited window of suitable growth conditions during the periodic wet season in the austral summer. One of these phenological strategies is that of early-greening; where deciduous trees will produce new flush prior to the onset of seasonal rainfall when the environmental conditions are at their driest. There is likely no other ecosystem in the world where the early-greening phenomenon occurs, however, it has been recorded in savannas from Africa, Australia, Asia and South America. The global prevalence of this strategy suggests that early-greening trees must gain some benefit from utilising this seemingly risky leaf flushing strategy. The main aim of this study was to investigate two of the hypothesised benefits of an earlygreening strategy for semi-arid savanna trees. The first hypothesis suggested that earlygreeners are extending their growing season beyond that of grasses and late-greeners – whose leaf flush is limited by the availability of water through the onset of seasonal rainfall. The second hypothesis is that early-greeners produce new leaves ahead of the emergence of invertebrate herbivores with the onset of the first rains, therefore decreasing the risk of damage to the vulnerable tissues and allowing the defence mechanisms within/upon the leaf to develop. This study took place in the Nylsvley Nature Reserve (NNR) over three consecutive austral growing seasons (August 2012 to April 2015) and was conducted on the two dominant deciduous broad-leaved trees: Burkea africana and Terminalia sericea. Burkea africana is known to flush prior to the onset of seasonal rainfall and was considered the earlygreening species in this study. Terminalia sericea has been recorded as flushing leaves prior to the onset of rainfall, but is more often a facultative-greener – rapidly producing new leaves with the onset of seasonal rainfall. This study experienced high seasonal variability between years with early-greening only occurring during the final (2014-2015) season. Nevertheless, I was able to assess the weekly green-up phenology of both species during the first three months of green-up and found that the green-up of the trees was largely disconnected from the main system driver – water – while grass green-up was closely linked to the onset of rainfall. This complements the theory of temporal niche separation; however, when assessing the functionality of the new leaves produced, carbon gain only occurred after the first two weeks post bud-break. In a system such as the NNR where rainfall onset variability is high, trees will only gain the advantage of an extended growing season if the onset of rainfall is more than two weeks after the start of flush. Using historical rainfall records (1980-2014), I estimated that 46% of years could have potentially experienced early-greening with rainfall commencing after the 15th October – the earliest date of green-up prior to rainfall onset in the NNR during this study. One of the benefits tested relating to early-greening in this study was that early-greeners avoid invertebrate herbivore damage on vulnerable new leaves. This study provided evidence for the use of a phenological defence strategy to cope with invertebrate herbivory pressure. Leaves which emerged before the rains had lower rates of herbivore damage than those which emerged after. Moreover, it was demonstrated that the constant turnover of leaves with high photosynthetic rates (T. sericea) is a reasonable mechanism for dealing with high leaf herbivory, and can result in equivalent end of season leaf area (and carbon gain) to species which invest in defence and have slower turnover rates (B. africana). Using the ground-based phenological measures in conjunction with remotely sensed NDVI imagery, the frequency of early-greening across seven comparable broad-leaved woodland sites in southern Africa from 2002 to 2014 was quantified. Of the environmental variables considered, the predictability of early-season rain (rather than total rainfall amount) was best correlated with early-greening. In savannas where rainfall onset and annual amounts were highly variable (such as the NNR), early-greening was less frequent (20% of the years) while in savannas closer to the equator where rainfall amounts were consistently >900 mm per annum and the onset dates began within a two week window each season, early-greening occurred in 80-90% of all years. The decrease in the proportion of early-greening events in the NNR from the 1980s to the past decade could be driven by the changing rainfall regimes over South Africa – with a predicted decrease in the number of precipitation events, but an increase in the storm intensity and rainfall amounts in each of these events. Fewer precipitation events may increase the risks associated with the early-greening strategy and this may be driving the NNR trees to use this strategy less frequently. This study has highlighted the need for a long-term phenological monitoring network within southern Africa’s savannas and has illustrated how early-greening species can benefit over other flora when environmental conditions are suitable for them to commence early leaf flush. This thesis has shown that early-greening broad-leaved savannas trees in South African savannas are more likely to avoid invertebrate herbivory than extend their growing seasons. / XL2017

A dissertation presented to the Faculty of Science in fulfilment of the requirements for the Master of Environmental Sciences degree University of the Witwatersrand, School of Animal, Plant and Environmental Sciences. South Africa, Johannesburg. March, 2016 / Understanding how trees limit light will advance theories of bush encroachment and tree-grass coexistence in savannas and it will enable us to explore how we can best conserve fire maintained, fire-dependent vegetation systems in South Africa and how our biome boundaries may shift. In order to understand how fire maintains the boundary between closed canopies and open grassy environments, the aim of this research was to test the hypothesis that; tree traits affect light that passes through the canopies of trees to the ground, by in uencing the threshold tree density at which grass stops growing under trees. Using the Point-Centered-Quarter method I estimated tree density using a proxy, mean PCQ distance, to refl ect changes in inter-tree distance across the gradient of tree density. At the same plots I used dry-weight ranking to record changes in grass cover, species composition and abundance, a balance scale to weigh wet and dry grass biomass, in tandem with densiometer measurements and hemispherical images to obtain canopy cover and the amount of sunlight received by the ground, in addition to the tree density data. To elucidate the interaction between tree traits and sunlight (correlation, boxplots, ANOVA, post-hoc test, MRD, quartile-regression analyses and linear modelling) as well as response of grass species to light (regression analysis, boxplots, ANOVA and GAMs), respective passable analyses were employed using R. The results show that biomass declines with changes in grass species composition and abundance, that are in turn driven by decline in light transmittance; when horizontal and compound leafed trees, especially on short, umbrella shaped trees with horizontal spheroids prevail. Consequently reducing light loving grass species with characteristics that lead to high fire frequency and intensity; high cover, height and biomass, that prevail when tall trees with vertically angled leaves and spheroids dominate, leading to higher light transmittance. This succinct understanding highlights variation in the tree-grass relationship across savanna ecosystems, suggesting site-specific interventions and recommendations to bush encroachment. Findings reveal value in monitoring tree canopy cover across savanna ecosystems; using it as an early warning proxy for changes in primary production and fire regimes. Conclusions challenge assumptions about the minimum light tolerance of grass, and provide important clues to help disantangle mechanisms by which grasses may persist at low light levels. Key words: Savanna, South Africa, light transmittance, trees, grass, bush encroachment / LG2017

The goals of this research are to identify the major constraints to plant succession in degraded cerrado vegetation. The area selected for this research, is located on the right bank of the Araguaia River in Goias State (Brazil). Floristic composition survey, seed rain and soil seed bank assessments were conducted in four forest fragments and in a degraded area. Planted seedlings, Soil translocation, Seed translocation and Artificial perches were also investigated. Floristic composition survey showed a richness of 145 species for the whole area and absolute density of 623.89 individuals.ha-1 for the degraded area and a range from 1333.72 to 2247.27 individuals.ha-1 for forest fragments. Seed rain survey resulted in 87 species identified, and annual seed arrival was 2.36 seeds.m-2 for the degraded area and 103.75 seeds.m-2 to 236.25 seed.m-2 to forest fragments. Soil seedbank analysis found a richness of 33 species and 87% of germinated seeds were from non-woody plants. The mean seed density of woody plants was 0 for the degraded area and ranged from 25 seeds.m-2 to 257.50 seed.m2 for forest fragments. Introducing seedlings showed a seedling survival rate of 38% after 24 months observation. Enterolobium sp and Inga spp showed higher rates of growth and positive interaction with fertilizer. The application of Tanglefoot, showed no significant results in any treatment. Treatments with fertilizer had less herbivory registered. Luehea candicans showed higher rate of herbivory. Seedling mortality was lower in fenced blocks and indicated Inga spp with lowest rate. Amongst nucleation strategies Seed translocation recruited more individuals after two years, followed by perches and soil seed bank translocation, but perches recruited more different species and had the same rate for seedling survival when compared to seed rain translocation. The results from this study show that succession in degraded Cerrado may be constrained by biotic and abiotic barriers.

Faculty of Science School of Animal,Plants and Enviromental Science 0204279a craigm.kznwildlife.com / Fire is a key determinant of savanna dynamics, and would thus have a major influence on the vegetation dynamics of Mkuzi Game Reserve. Given this logic, it is an important and commonly used management tool in this reserve. Its main uses in the reserve are for either removing moribund material or for reducing woody plant encroachment, both of which normally entail dry season burns. As a consequence, fire often results in a green flush of vegetation that is highly favoured by grazing herbivores. A further management goal is maintaining or improving biological diversity by promoting vegetation heterogeneity. Current policy prescribes this should be achieved through point-source ignitions rather than by block-burning, which was the earlier practice. This study explores spatial and temporal fire patterns at a landscape scale in Mkuzi Game Reserve using Geographic Information Systems (GIS). Much of our understanding of the dynamics of fire has previously been determined at a plot scale and scaling up of these insights to a landscape scale is problematic, hence this project aimed to contribute to our understanding of the dynamics of fire at a landscape scale. The study also specifically examined how the fire regime in the reserve has changed with a change in the burning philosophy and strategy, namely from block burning to the point source ignition (PSI) strategy, which began to be implemented in the mid-1980's. Fire frequency was related to both geological type and vegetation type. The fact that geology was related to fire frequency was not surprising because the relationship between geology and vegetation in the reserve has previously been established. The varying amount of herbaceous material per vegetation type apparently influenced fire frequency. Spatial variation in fire frequency was also positively related to rainfall variation over the reserve, while the total area burnt per annum was positively related to the preceding wet season rainfall, but not for years with a high dry season rainfall. The influence of rainfall on grass production and thus fuel load explained these relationships. In addition, there was some evidence of a carry over effect of rainfall where the previous wet season rainfall together with the preceding wet season rainfall influenced total annual area burnt, but this was only significant for years when dry season rainfall was low. Contrary to an expected negative influence, dry season rainfall had no effect on the total annual area burnt. Grazer biomass had a significant limiting effect on fire frequency over the reserve (spatially), most likely due to consumption of herbaceous ii material, but there was no relationship between grazer biomass and total annual area burnt (temporally). Dry season burns were significantly larger than wet season burns and can be attributed to the more favourable fuel condition during the dry season. Intense burns were also generally larger than the cooler burns, namely those rated as patchy/very patchy and clean. This was mainly attributed to a high fuel load which is critical for intense fires but also positively influences the spread of fire. The comparison of the block burning strategy and the point source ignition (PSI) strategy showed that fire frequency was greater during the PSI burning period than during the block burning period. The total area burnt per annum was greater during the PSI burning period than during the block burning period, but individual burn sizes were not significantly different between the two strategies. Evidence showed that individual burns that occurred during the PSI period had boundaries that were more irregular than those of block burns. Fires were most common during the dry season for both burning strategies, but the proportion of the burns that occurred during the dry season was greater for the PSI burning period than for the block burning period. Evidence also showed that a much greater emphasis was put on applying dry season prescribed burns during the PSI period than during the block burning period. A greater effort was also made during the PSI period to burn firebreaks, which were only implemented during the dry season. Arson fires (started deliberately or accidentally by neighbours) were more common during the block burning period than during the PSI period, while under both burning strategies, they were more common during the dry season than the wet season. There was no distinguishable difference in the burn intensity patterns between block and PSI burning, that is, the proportions of burns in the different burn intensity classes were not significantly different between the two burning strategies. Although the contribution of the individual fire barrier types showed some change with a change from block burning to a PSI strategy, the combined contribution of natural barriers did not increase, and that of management barriers did not decrease, as would have been expected. In addition, natural and management barriers were apparently of equivalent importance during both burning strategies.

Thesis (Ph.D.)--University of Wyoming, 2008. / Title from PDF title page (viewed on Mar. 9, 2010). Includes bibliographical references.

Thesis (M.S. - Renewable Natural Resources)--University of Arizona. / Includes bibliographical references (leaves 96-107).

Thesis (Ph.D.) - James Cook University, 2002. / Typescript (photocopy) Bibliography: leaves 271-298.