To obtain a better understanding of the functioning of a globally important biome, the BOREAS project set out to make measurements of the interaction between a number of boreal ecosystems and the atmosphere. This study is concerned with measurements made in one such ecosystem, the old-growth black spruce (OBS) site of the BOREAS Southern study area (SSA), located in Saskatchewan, Canada, during the period 1993-1996. This thesis focuses on efforts to understand the net ecosystem exchange of CO<SUB>2</SUB> and water vapour in terms of the sub-component fluxes operating within the ecosystem. The component fluxes of primary interest are photosynthesis, respiration and evapotranspiration of foliage, and the efflux of CO<SUB>2</SUB> from the forest floor. Woody biomass respiration and changes in the storage of CO<SUB>2</SUB> in the air mass within the ecosystem are also considered. Methodologies were developed to study these fluxes on a continuous basis. An "open" system gas exchange chamber for measuring soil CO<SUB>2</SUB> efflux was designed that eliminated the major problems that have been associated with this methodology in the past. The system was used to investigate temporal and spatial variation in soil CO<SUB>2</SUB> efflux at the field site. A scheme that integrated this temporal and spatial variation was used to estimate the CO<SUB>2</SUB> efflux from the forest floor for an entire year. Spatial variability in soil CO<SUB>2</SUB> efflux was high, and was related empirically to the thickness of the dead moss layer. Hour to hour variation was well described as an exponential function of soil temperature, and was significantly related to atmospheric turbulence.

A “real-time” eddy covariance (EC) system (<i>EdiSol</i>) was developed to measure fluxes of CO₂, momentum, sensible heat and latent heat of a boreal black spruce stand as part of the BOReal Ecosystem-Atmosphere Study (BOREAS). BOREAS is a multi-scale study using satellites, aircraft, tower and leaf-scale measurements over the major vegetation types of the Canadian boreal forest. The measurements reported in this thesis are an essential component of BOREAS being the stand scale (tower) measurements over one of the major vegetation types. <i>EdiSol </i>uses a commercially available three-axis ultrasonic anemometer and a closed path infrared gas analyser (air is ducted from close to the anemometer to the optical bench). Software was written to acquire the output data from the anemometer and gas analyser and concurrently calculate CO₂, water vapour, momentum, sensible heat and latent heat fluxes. The high frequency, primary, data is stored for further analysis. The calculations made by the software and the instruments and their connection are described. The corrections for frequency response of the system and tests of the software output are presented. The observations were made from 25 March to 28 November 1996 over the black spruce stand. There was no CO₂ assimilation when soil temperature at 5 cm and 10 cm depth was less than ca 0°C (corresponding to an air temperature of -3°C on average). Under these conditions, carbon exchange averaged 0.8 g C m^-2day^-1 (0.77 mmol CO₂ m^-2s^-1). Mean half-hourly net ecosystem flux reached -14 mmol m^-2 s^-1 in June during the day and 10 mmol m^-2s^-1 at night when the atmosphere was well mixed (u_* > 0.4 m s^-1). The optimum temperature for CO₂ assimilation was 30°C, close to the maximum air temperature observed at 26 m. When the soil temperature above 10 cm depth was less than 0°C assimilation was zero but quantum efficiency of assimilation (a) reached 0.66 in July.

No description available.

<I>Zea mays</I> L. is an important food crop and common intercrop species grown in semi-arid agroforestry systems in Kenya, East Africa. A preliminary field study showed that AMF spores occurred in particularly high number in this system compared with that reported for other less disturbed ecosystems. The same spore types occurred in tree and alley soil, however spore number and composition were variable, between host species, in fallow and cropped areas of the system and between wet and dry seasons. At the end of the dry season (February, 1991), fallow areas had significantly higher spore numbers than cropped areas of the system. Lower numbers of spores were found in soil in the dry season than in the wet season, when spore numbers were significantly higher in tree soil in the cropped area than in soil from other areas. The first experiment was established to determine the influence of host and fallowing on the effectiveness of AMF populations occurring in this system on mycorrhizal formation, growth and nutrition of plants. These results showed that plants inoculated with tree soil from the cropped area of the system had significantly higher mycorrhizal infection formation in roots, however this was not reflected in the growth and nutrient uptake of plants which were unaffected by AMF inoculation. The effects of soil type, soil volume and soil phosphorus were thought to have influenced the response of plants to inoculation in the previous experiment and so the influence of these factors on growth and nutrition of inoculated and uninoculated plants were investigated. Results from this experiment showed that plants had significantly different mycorrhizal infection in roots when grown in different soils. Mycorrhizal infection was higher in plants grown in smaller pots, decreasing significantly with incremental increases (1, 4 and 10 litre pots) in soil volume.

Co-flowering plant species (species that flower together in space and seasonal time) that share pollinators could compete for pollination. This can result in a negative effect on reproductive success for one or more species. Such plant species could compete for one or both of pollinator visits or pollen load quality. Plant species could minimise competition for pollinators (i) by using different pollinators, (ii) through spatial separation, (iii) by flowering at different seasonal times, (iv) by placing pollen on different parts of the same pollinator and (v) by segregating shared pollinators in daily time. In this thesis I investigate pollination interactions between plant species in (i) a guild of ten acacia species and (ii) an entire flowering plant community in Laikipia, Kenya. I examined the acacia species’ flowering phonologies to identify species that regularly co-flower. I compared daily patterns of dehiscence between co-flowering acacia species for evidence of character displacement on a daily timescale and assessed the extent to which (i) acacia species shared pollinators and (ii) shared pollinators are partitioned in daily time in response to patterns of pollen availability. A number of pollinator species were shared among acacia species making competition for pollination possible. Although up to five species of acacia regularly co-flowered, there was no consistent evidence to suggest that character displacement of dehiscence time had occurred among these species. Previous analyses have only considered the dynamics of co-flowering within guilds of related species. In this thesis I examined evidence for the daily partitioning of shared pollinators in a savannah plant community using a null model approach to analyse plant-pollinator interaction networks.

This study set up the hypothesis that branches are shaped so as to maximise total photosynthetically active radiation interception per unit of assimilate expended in branch structural material. Measurements of branch geometry, combined with computer modelling were used to develop a deeper understanding of form-function relations. The thesis seeks to discover whether light interception (as a measure of benefit) per unit invested material (as a cost) is maximised in natural branches by the optimisation of allocation of structural material among the shoots of branches. Thus a cost-benefit analysis paradigm is applied to the problem of structural carbon allocation among shoots. Topology of branches has been regarded as an invariant property of branch age, so providing an appropriate constraint. A topological classification scheme was devised for branch elements (links and shoots) so that form-function relations could be resolved. Geometric attributes of 125 individual branches from trees of <I>Picea sitchensis</I> Bong. (Carr.) were analysed to find evidence for form-function relationships. Multivariate differences in geometric design among genotypes were not considered sufficient to motivate form optimisation through a breeding program. Within branch shoot length distribution was related to position in the topology which may be a reflection of light interception potential. Diameters of branch elements were proportional to lengths, but proportionality constants varied with topological position. An allocation model based on mechanical design of shoots was more successful in explaining shoot allometry than one based on the pipe model. The mechanical load safety factor varied among branch elements according to their expected potential for light interception. Hydraulic (pine model) design of branches reinforced the pattern of distribution in mechanical safety factors. This provides possible support for the theory of shoot autonomy. It is concluded that tree branches are likely a result of competing optimisation goals acting within genetic and ontogenic constraints and that shoot size may be related to the potential for competition for light interception, with shoots that show greater potential taking a disproportionate share of resources within the branch.

We tested several hypotheses which all link the carbon cycling in European forests using the Soil-Plant-Atmosphere model (SPA) and eddy covariance measurements from the CARBOEUROPE project. Firstly we run the model against flux data and obtained the optimised set of parameters for ten flux sites at a variety of climatic conditions. The inter-site model parameters are investigated to test whether the variations in parameters display a pattern which can be used in up-scaling of model application. It was found that maximum carbonylation rate (V_cmax) displays a simple linear relationship with latitude (R² = 0.72) and the ratio of autotrophic respiration to GPP remains at a constant (0.47). Furthermore we applied a Bayesian calibration method, which quantifies both flux measurement error and parameter priori distribution, to the same ten flux sites to tackle the uncertainty of model outputs. It was found that given amount of data and associated errors, model parameters and the confidence intervals that are related to plant functional type (PFT) can be identified. The third hypothesis we tested in the thesis is that carbon uptake by plant is enhanced under diffuse radiation. We compared the light use efficiency (LUE) under both direct and diffuse regime for eight contrasting ecosystems using both flux data and modelled outputs. Our results show that diffuse radiation results in a higher LUE for both forests and grasslands and response of forests to diffuse radiation is even higher. The model produced strikingly similar results. Finally we applied the model to a specific flux site experiencing drought to test the hypothesis that soil moisture and precipitation may influence the carbon uptake of forests and further have an impact on the carbon cycling.

The pollination biology and reproduction of a woody plant community of cerrado, the neotropical savanna vegetation in Brazil, was studied. A delimited area of relatively dense cerrado which included 59 woody species was studied from July 1988 to September 1989 in the Ecological Reserve area of Brasilia Botanic Garden. Aspects of the phenology, floral biology, breeding systems and fruit-set patterns were studied for the whole community or for a subsample of the species in the area. Some species, group of species or aspects of the reproductive biology of these plants were further analyzed in separate case history studies. The species in the community showed varied phenological behaviour with vegetative growth and reproduction distributed throughout the year. It seems that the phenology of adult individuals is, to a certain extent, independent of the markedly seasonal wet and dry climate of the area. Only the seedling establishment seems to be confined to a specific period at the beginning of the rains. Wind and animal dispersal mechanisms are partially constrained showing two different peaks but flowering phenology seems to be much more flexible and even for congeneric species flowering periods may differ markedly. The phenological strategy of each species seems to be a more or less independent assemblage of the possible options in each phenophase. The plants could be grouped in seven pollination guilds based on their main pollinators. These guilds were basically similar to the ones described for other tropical communities. Unspecialized flowers pollinated by small insects, including flies, wasps and small bees formed the most Common group. The more conspicuous and specialized flowers pollinated by medium to large, long-tongued bees formed the second most important group. Other characteristically tropical pollination systems such as bat and beetle pollination were also present. Hummingbird pollination was rare and wind and butterfly pollinated flowers were absent. The fauna of visitors, their foraging habits and possibly their seasonal segregation is similar to that described for Costa Rican seasonal forests, particularly in the case of the large bee fauna. No pollination system seems to be seasonally limited although a peak of large bee pollinated flowers, for example, could be identified at the onset of the rains. Breeding systems could be established for a sample of 22 species with different pollination systems, the majority (86%) of which presented self-incompatibility mechanisms. This sample and the frequency of dioecious species (15% of the total) permitted an estimate of 84% as the frequency of obligatory outcrossing species in the area. Such frequency is similar to those obtained for other lowland tropical forest communities and much higher than the estimates for altitudinal cloud forests in the tropics. Apomixis was present in two species in the community and is possibly present in another two, which indicate that, although rare, this may be an important phenomenon amongst the cerrado woody species. The site of the incompatibility reaction was studied for most of the self-incompatible species and a majority presented &quot;lateacting self-incompatibility&quot;, with self-pollen tubes reaching the ovary or even penetrating the ovule. More &quot;classical&quot; self-pollen tube arrest in the style was also observed for some groups. Fruit-set was usually low but the data should be regarded with caution since they varied both spatially and temporally. Nevertheless, some differences between pollination system groups and relationships with breeding system are suggested by the results. The case history studies involved a self-incompatible shrub species, Yellozia squamata (Velloziaceae), which presented a distinct flowering phenology; a bat-pollinated tree, Hymenaea stigonocarpa (Leg. Caesalpinioideae), which presented 'late acting self-incompatibility'; another legume tree, Sclerolobium paniculatum, which present two varieties segregated ecologically but which were not clearly isolated reproductively; contrasting breeding systems, self-incompatibility and apomixis, in two cerrado Eriotheca spp. (Bombacaceae); six species of Vochysia (Vochysiaceae) which occur in different vegetation physiognomies in the Ecological Reserve of the Brasilia Botanic Garden but have very similar floral biology and outbreeding system; a study on some small &quot;settling&quot; moth pollinated species in the study area which have similar outbreeding systems as their large hawkmoth pollinated counterpart; and a general study on the occurrence of dioecy which is less important in cerrado areas than in the contiguous forest habitats. The breeding features emerging from this study support the idea of the cerrado vegetation as stable communities where biomass output is possibly limited by availability of nutrients and restrict establishment conditions, but not restricted by seasonality or disturbance in terms of opportunity and predictability for the reproductive process of the woody plants. If outbreeding systems are indicative of environmental stability, then cerrado conditions seems to be comparable to those in the lowland tropical forest.

Recent developments in NextGen sequencing and DNA barcoding have been exploited to enhance capacity to study pollinator service in a focussed and holistic manner. Initial efforts focussed on developing robust molecular systems for quantification of two- and multi-species pollen mixtures recovered from pollinators. A simple twospecies system used pyrosequencing to provide quantitative information on the relative abundance of Brassica napus and B. rapa pollen. The more demanding multi-species system used rbcL barcode amplicons sequenced on the 454 XLR platform. Sequence tags allowed for all resultant sequences to be assigned to the insect from which they derived and so provide quantitative pollen carriage data. The pyrosequencing system was deployed on pollen DNA recovered from a pollinator guild collected from either the crop (B. napus) or from nearby populations of its close relative, B. rapa. The aim was to compare the propensity of each pollinator to carry crop pollen to the wild recipient over a linear transect. Surprisingly, the most studied pollinator with regards to gene flow, honey bees (Apis mellifera), was least prone to carry crop pollen to wild B. rapa. The understudied hoverflies (Syrpidae), however, often carried significant proportions of crop pollen >200 m. The 454 analysis system characterised changes in pollen carriage with phenological progression and according to species, site and gender. All factors were found to be significant in shaping pollen carriage but to differing degrees. The most striking was the extent to which migrations to other communities varied between pollinators and was most evident during a flowering transition period when resources appeared limiting. This study provides a platform for more detailed works to track pollen carriage within and between plant communities in a specific landscape as the flowering season progresses. They also provide methodological tools to test theoretical models constructed to predict pollinator behaviour on a landscape scale.

No description available.