No description available.

No description available.

The overall aim of this thesis is to investigate long-term (multi-century) vegetation development in a high-latitude setting, with a particular emphasis on the emergence and persistence of spatial structure in plant communities. The spatial distribution of plants within a community influences vegetation dynamics and the functioning of terrestrial ecosystems. Knowledge of the spatiotemporal dynamics of vegetation is therefore crucial to understanding ecosystem response to disturbance, and to successful ecosystem management. Studies of spatiotemporal dynamics from high-latitude settings are rare, despite these regions being among the most sensitive to warming and subject to ongoing environmental change. The study was based on a primary succession on Mt Hekla in south-central Iceland. The chronosequence approach was used to infer 850 years of vegetation development from a suite of 14 lava flows (five of which had been disturbed by the deposition of volcanic tephra). The thesis is organised around four main research themes: 1) Trajectories of development- How do the global (i.e. non-spatial) properties of a community (e.g. species diversity) change with terrain age in a primary succession? Plant species frequency data from 12 transect surveys (each comprising 400 contiguous 10 cm x 10 cm quadrats) were used to test the predictions of classical models of primary succession; 2) Initial colonisation- How do plants first establish on newly-created terrain? Photographic surveys and point-pattern analysis were used to assess the processes by which pioneer species colonised ‘safe sites’ on lava surfaces six and fifteen years old; 3) Spatial scale and structure- How and why does the spatial structure of vegetation vary over long timescales? Vegetation and soil data from transect surveys (item 1, above) were analysed using a variety of spatial statistics, in order to test three models of spatiotemporal dynamics; 4) Temporal changes in environmental gradients- To what extent does biotic reaction feed-back into the small- (metre-) scale distribution of plants? Randomised sampling of vegetation on 42 sites was combined with continuous microclimatic monitoring to assess the changing steepness of environmental gradients, and the impact this process had on vegetation development. The analyses revealed a robust, multi-century trajectory of vegetation development on undisturbed sites. The development of vegetation on newly-emplaced flows was rapid, as colonists randomly exploited small- (millimetre) scale surface irregularities in the lava. A thick (up to 20 cm), spatially homogeneous ‘carpet’ of moss formed within 50 years of flow emplacement. Thereafter, the vegetation became progressively more structured with increasing terrain age, as patches of vegetation formed and expanded. On the oldest sites, differentiation of the vegetation according to metre-scale variations in elevation was apparent. Overall, the results emphasise the limitations of classical models of succession in high-latitude habitats and suggest widespread applicability for the nucleation model of primary succession. The study also suggests that the spatiotemporal response of vegetation to environmental gradients is mediated by short-range positive feedback. These findings have implications for modelling vegetation development on other primary substrates (e.g. glacial forelands) and the response of spatially patchy, high-latitude vegetation to future climate change.

This study was undertaken to reconstruct the regional vegetational history of Holderness, in south-east Yorkshire. It is an area with great potential for palaeo-ecological investigations and yet it has been largely ignored during the great increase of research in this field in recent years. Being on the east coast, Holderness is well placed to provide information on the migration of plants into this country, particularly in Late-glacial and early Flandrian times, when the region was joined to the continent by dry land. As a distinct and somewhat isolated region of low-lying land which must have been rather waterlogged for much of Post-glacial time it would have provided a considerable challenge to prehistoric man. Therefore an examination of the extent to which early man was able to emploit this difficult environment seems to be worthwhile. Among the most interesting advances in the study of vegetational history have been the introduction of the evaluation of pollen concentration and absolute pollen frequencies in an attempt to give a more quantitative picture of post vegetational communities, and also the considerable reassessment of climatic conditions during the Late glacial period. Consequently, in the elucidation of the vegetational history of Holderness, the intention has also been to throw some light on these current problems. The Late-glacial of Holderness has therefore been studied in terms of pollen concentrations and absolute pollen frequencies, and the results obtained compared with other recent evidence, much of which has come from north-western England, to try to establish to what extent conditions varied within northern England during this period. Somewhat more approximate estimates of absolute pollen frequencies in Post-glacial times have been made, and the value that this information adds to the understanding of vegetational history of the area, has been considered.

Pollen diagrams prepared from the sediments of Loch Lomond and the Dubh Lochan have provided a general overview of the vegetational changes which have occurred in the Loch Lomond area during the Flandrian period. Progressive deforestation since 5000 B.P. and particularly during the last few centuries has been high-lighted with the period of maximum extent of mixed Quercus woodland in the Loch Lomond area having occurred between c.6000 B.P. to c.5000 B.P. The Alnus rise horizon has been demonstrated to have occurred later than expected in the Loch Lomond area. Absolute pollen analysis has enhanced the interpretation of the relative percentage data and has shown that the pollen concentration has varied widely in the Loch Lomond sediment during the Flandrian. Pollen deposition data has been calculated from the Dubh Lochan sediments allowing a comparison to be made with deposition data from other lakes in north west England and Scotland, It has been concluded that concentration of pollen within the Dubh Lochan has occurred. Dinoflagellate analysis has provided the first positive evidence of a marine phase in Loch Lomond's mid Flandrian history. A period of meromixis of ectogenic origin has been hypothesised to account for a laminated band following the marine sediment within which pollen and macrofossil preservation was found to be uniquely good, with pollen counting of the Loch Lomond sediment having been generally taxing. Derived Carboniferous spores were noted throughout the length of the Loch Lomond profile. Analysis of selected terrestrial sites to compliment the picture of the status of Pinus sylvestris gained from the Loch Lomond sediments has been undertaken. The locality of the post Ulmus decline Pinus curve increase demonstrated by the Loch Lomond pollen data has been traced to Glen Falloch. The Glen Falloch peat has recorded a change from Betula woodland to Pinus-Betula woodland at a time generally recognised to be one during which Pinus was in decline in Scotland and Northern Ireland. Fire attributable to man has been considered to account for the initiation of growth of Pinus on the Glen Falloch peat surface. Radiocarbon dating of Pinus wood has demonstrated the continuous presence of Pinus in Glen Falloch until C.1600 B.P., with the present day Pinus trees having been considered to be of natural origin. Data from the Ptarmigan has suggested that in at least one part of the southern Loch Lomond area more Pinus exist today than at any time during the last six thousand years.

This thesis considers some of the main issues surrounding the quantitative models that have been developed to reconstruct vegetation from pollen assemblages. Conducted within a pine dominated woodland, a palynologically difficult landscape, to determine vegetation changes over the late Holocene the results highlight the complexities of undertaking such studies in these contexts. Pollen productivity estimates were calculated from moss samples over the woodland using complete sets and derived subsets to detect influences of sampling design on resultant model output. Differences in the PPE sets were compared using reconstructions from simulation models in comparison to observed vegetation patterns. The results indicate that both parameter calculation and model reconstructions were influenced by the landscape form and composition. Sensitivity of the models to such small variations in parameter values heightens the need for robust data generation and increased investigation to controlling factors on pollen productivity. Performance of the reconstruction models experienced variation with respect to deposition basin size and site specific characteristics. Overall, the regional reconstructions proved to generate more confident estimates of vegetation cover whilst local scale reconstructions were subject to greater variability. Comparison of the quantitative modelling to standard interpretation and the modern analogue approach shows contrasts between the results obtained with respect to limitations associated with each method and the time frames, recent (ca. 100 years) and longer (ca. 3000 years), over which they were applied. Consequently, no one quantitative approach could be identified as being superior as site specific variations were recognised in relation to the most suitable approach. In response, a hierarchical technique is proposed to utilise the benefits of each technique and to obtain detailed information to strengthen interpretations. However, it is stressed study specific constrains that determine the available resources will influence the ability to fully apply this composite approach.

For the first time in northern Scotland, Modern tree-ring chronologies for Scots pine growing on peat are compared against those growing on mineral substrates. Mean tree-ring growth of pine on active bogs/mires is found to be limited to 0.5 to 1 mm yr-1, compared to ≥1.5 mm yr-1 on adjacent mineral sites. Almost instant change of radial growth rates in response to changes in water levels highlights the potential use of pine in reconstructions of lake levels and water tables in bog and mire. Dendroclimatological analysis identifies January and February temperatures to often be more important than summer temperature. Positive correlation of ring-width and North Atlantic Oscillation (NAO) indices also occur in January and February. Lower winter temperatures, due to increased altitude and distance from the moderating effects of the ocean, may be important in limiting tree growth near its northern margin. Moving correlation functions identify a widespread reduction in the response of pine growing on both substrates from the 1920s. Nine subfossil pine site chronologies located beyond the species current northern limit are cross-matched to form a chronology called WRATH-9. This chronology is tentatively crossdated against Irish pine chronologies to provide the first picture of Neolithic Scots pines 200 year expansion from c. 3200 BC and subsequent 250 year retreat across northern Scotland at annual resolution. The mean orientation of maximum radial growth at eleven modern pine sites is found to coincide well with the W/SW prevailing wind, suggesting Scots pine may provide a good proxy indicator of wind. Six coeval Neolithic sites indicate a broadly consistent northerly prevailing wind. This provides tentative evidence for a change of prevailing wind that may be related to a southward incursion of the polar front in the eastern N. Atlantic. The potential of this exciting subfield of dendroclimatological analysis is called dendroaeology and is highlighted for further research.

This thesis looks into the use of historical macrophyte records to assess long term changes in macrophyte communities in lakes and potential reasons for these changes. In particular it uses historical records to assess changes in macrophyte communities in the Norfolk Broads and West Midland Meres, two sets of lowland, eutrophic lakes in England. It provides a critical examination of the use of historical records, highlighting some of the constraints common to such data such as variations in recording effort, and bias in species recording and site selection. Having acknowledged these issues we then go on to develop a robust way to interpret such data, using a “change index” based on species persistence over the last 200 years within individual lakes. Species with high change index values, which represented species which had persisted or increased within the lake districts, were those known to be characteristic of eutrophic lakes. Conversely species with low index scores, which had declined in both the broads and meres over the last 200 years, included species associated with less fertile conditions but also a selection of typically eutrophic species. Averaging of change index scores in present day survey data served to identify the historically least changed lakes and to rank lakes in order of degree of botanical change over the last century. We then analysed the ecological basis of the change index in order to better understand the processes behind the decline of some species and survival of others in the Norfolk Broads and West Midland Meres. Functional groups determined from morphological and regenerative traits displayed significant differences in change index values in both groups of lakes, but declining taxa occurred across a wide range of plant growth forms. Non-hierarchical clustering of species based on their ecological preferences, obtained from published literature, resulted in groups with distinct change index values, indicating that changes in the status of species could be partly explained by these preferences. Of these, trophic preference was consistently the most important, with species of less fertile habitats consistently experiencing the greatest declines. However, some characteristically eutrophic species have also declined significantly, particularly in the broads. In these cases increasing loss of shallow water, low energy habitats in the broads, or loss of fluctuating water levels and less alkaline backwaters in the meres, appear to have been contributory factors. In addition to the change index approach, we also used historical records at a site level to complement palaeolimnological analysis and investigate the change in macrophyte community composition and structure at Barton Broad, Norfolk. Sediment samples were extracted from the bottom of the broad and analysed for sub-fossil remains and pollen of macrophytes. The historical records and palaeolimnological analysis combined showed that early communities did not consist entirely of low growing, oligotrophic and mesotrophic species as previously thought, but in fact comprised a mixture of these and other more characteristically high nutrient species associated with taller, or free-floating growth habit. As eutrophication progressed throughout the last century, the community was increasingly dominated by these latter growth forms. Diversity was maintained, however, since encroaching reedswamp generated a mosaic of low energy habitats which supported a range of species unable to withstand the hydraulic forces associated with more open water habitat. When the reedswamp disappeared in the 1950s, many of the dependent aquatic macrophytes also declined resulting in widespread macrophyte loss. The thesis demonstrates not just the complexities of using historical records, but also ways in which these can be overcome to make useful observations about macrophyte community change and lake ecological integrity to inform conservation and lake management, both on a site and lake district level.