In southern Scotland, human-mediated deforestation for pasture since the Neolithic has dramatically altered the landscape. We intensively surveyed a single catchment (Moffat Dale) for severely fragmented populations of <i>Sorbus aucuparia </i>L., an insect pollinated bird dispersed species, and <i>Fraxinus excelsior </i>L., which is wind pollinated and wind dispersed. These remnants are being considered for seed collection in a native woodland restoration programme currently being implemented. Quantifying genetic variation at isozyme and chloroplast DNA markers in <i>S. aucuparia </i>remnants revealed that high levels of genetic diversity are maintained. However, genetic differentiation among remnants was detected for both types of marker and the estimated ratio of pollen flow to seed flow between fragments is close to one (<i>r=</i>1.36) suggesting reduced historical pollen-mediated gene flow but efficient seed dispersal. Similarly, <i>F. excelsior </i>remnants maintain high levels of genetic diversity at nuclear microsatellite markers and low interpopulation differentiation (q=0.080). Using the neighbourhood model, it was estimated from open-pollinated progeny arrays that contemporary pollen flow is extensive and that effective pollen dispersal distance with in the catchment averages 328 m. A detailed paternity analysis conducted on progeny arrays confirmed these results. Although pollen flow is an important component of realised gene flow, a parentage analysis showed that it is not predominant as 56.6% of the seedlings that recently established in Moffat Dale immigrated into the catchment. <i>S. aucuparia </i>and <i>F. excelsior </i>remnants in a severely deforested landscape are part of the wide reproductive network. Genetic diversity within remnants and gene exchange among them have been maintained by efficient long distance seed and pollen-mediated dispersal, making them an appropriate seed source for planting stock.

A framework for modelling the dynamics of tropical forests is described. The framework makes use of simulation models to predict the long term growth and yield of forests under different management regimes. It is designed to have practical application for the sustainable management of forest resources in tropical countries. The framework comprises a suite of simulation models, each of which is appropriate in particular circumstances. Each model uses a disaggregated representation of a forest stand. An individual-based representation is used for trees in the forest stand above a threshold size, while a more aggregate representation (such as a cohort representation) is used for seedlings and saplings. Processes of stand disturbance and recovery from disturbance are captured. Disturbance results from tree falls and from stand harvesting. Recovery from disturbance may involve seed production, seedling establishment, and competition between trees. Local interactions within the stand are captured, as is species-specific behaviour. The content of each model in the framework is represented in a text-based model design. Details of the content are specified using a formal representation language. The language has semantics and syntax for specifying how a set of generic modelling concepts is employed in an individual model. Details specified in this way include the names and kinds of model variables, and the algorithms used in model calculations. A formal-representation (an 'ontology') of the set of generic modelling concepts developed for use in the framework was created using the Ontolingua server. This provides an unambiguous specification of the modelling concepts used. This assists communication and may also make it easier to design rules and procedures for translating the model content into forms compatible with different modelling systems.

No description available.

No description available.

No description available.

This study utilises breeding bird data collected by Patterson et al. 1995 (1991 and 1992) and C.A.McSorley (1998 and 1999), in Kielder Forest, collected using the point count survey methodology. Survey sites in 1998 and 1999 were selected utilising a stratified random technique. Repeated counts of birds generate seasonally and annually correlated bird densities, however there are some significant differences in density from early to late spring. Thus, it is justifiable to use the maximum density over early and late spring for further analyses. Annual correlations are discussed in terms of resource availability. Small-scale density patterns are affected by the interactions between patch tree age, edge comparison and the distance from the patch boundary. Patch boundary avoidance or preference is observed for many species. Low densities are generally found in hard edges. These results are discussed in the context of the trade-off between higher food availability and higher predation risk at patch edges. The occurrence and density of passerines in 1999 are modelled using patch variables generated using a geographical information system (GIS) and a spatial quantification program, FRAGSTATS. The power of associations are mediocre to good (r(^2)evalues are generally approx. 0.2). The performances of the models using 1998 data for validation are mixed. However, some of these models could be used as tools for predicting the effects of forest restructuring on the passerines of Kielder. The factors determining why some species are widespread and locally common, and vice versa are investigated. The results show that resource availability affects the interspecific abundance-distribution relationship. Willow warblers aggregate more readily than residents do, perhaps as a result of utilisation of heterospecific and conspecific cues to quantify habitat 'quality’. The results from all chapters are put into a national ornithological context and discussed in terms of forest management. Further work is also proposed in the final chapter.

No description available.

No description available.

No description available.

No description available.