Increased functional diversity has been linked to an increase in ecosystem level processes (ELPs), such as productivity, ecosystem reliability and invasion resistance. However, there has been no exact definition of functional diversity and it is not known which indices are appropriate for its measurement. Consequently, continuous indices have rarely been applied in examination of relationships between functional diversity and ELPs and little is known of the mechanisms linking functional diversity to ELPs. This thesis begins by providing an exact definition of functional diversity, identifying its primary components and devising appropriate continuous indices for the measurement of these components. These indices are used to examine relationships between functional diversity and three ELPs - biomass production, the reliability of biomass production and invasion resistance. Initially these examinations are conducted using a mechanistic community assembly and dynamic model. This model is based on physiological and morphological character data for species occurring in the short-tussock grassland communities of the Luggate field experiment, in the southern South Island, New Zealand. Finally, relationships between functional diversity, mean annual community cover and the reliability of cover are examined in the Luggate field experiment. Functional composition (i.e. the actual functional characters of the species within a community) appeared to exercise the greatest influence on ecosystem reliability in the community assembly and dynamic model. However, there was evidence that functional diversity increased the reliability of productivity via the co-variance effect. Functional composition also exercised the greatest influence over mean annual productivity in the model, almost completely accounting for the negative relationship between functional diversity and mean annual productivity. These results are respectively analogous to the positive and negative selection effects seen in species richness / ELP relationships. Both functional diversity and functional composition influenced community invasion resistance in the model. Here, increased functional diversity acted to increase community resistance to invasion. In the Luggate field experiment, neither functional diversity nor functional composition was related to mean annual cover, though species richness was positively related to it. Increased functional diversity acted to increase the reliability of cover. However, this did not appear to occur via the co-variance effect, but as a result of increased functional diversity increasing consistency in the species composition of experimental plots through time. These results demonstrate that studies examining functional diversity must account for the effects of species identity. A framework is proposed to accommodate selection effects associated with functional diversity / ELP relationships. The usefulness of the co-variance effect as a testable mechanism linking functional diversity to ecosystem reliability in the field is questionable, as there is no objective way of measuring it.

The vegetation communities that existed in the semi-arid intermontane basins and gorges of Central Otago prior to human settlement ~750 years B.P. are poorly understood. This is because of a lack of fossil evidence and complex restructuring by anthropogenic factors, especially increased fire frequency, and more recently mammalian grazing. There is also little information regarding the effect of the lost fauna on maintaining and structuring presettlement communities, both in Central Otago and throughout the eastern South Island dryland zone. This study aims to provide a clearer understanding of the functioning of pre-settlement ecosystems in dryland Central Otago, particularly the role of the largest vertebrate herbivores, the moa (Aves: Dinornithiformes), and to explore the implications of the extinct fauna for land conservation management across New Zealand. Late Pleistocene and Holocene vegetation communities of the Central Otago lowlands were reconstructed from plant macrofossils, including seeds, leaves, and wood, excavated from rockshelter, cave, and swamp deposits throughout the region. The macrofossils represent three main vegetation types: late Pleistocene to mid (late?) Holocene basin floor wetland herb associations, Olearia-shrublands surrounding these wetlands, and mid to late Holocene open scrubland and woodland in gorges and on low altitude slopes, dominated by filiramulate Olearia, Coprosma, and Corokia, with abundant lianes (Muehlenbeckia spp. and Rubus spp.) and understorey herbs. Many native tree and shrub species that are presently widespread in the Central Otago lowlands were rare or absent prior to anthropogenic fires (e.g. Discaria toumatou, Kunzea ericoides, Leptospermum scoparium). Other tree and shrub species once present are now extinct in the region (e.g. Coprosma obconica, Plagianthus regius, Pseudopanax ferox). The loss of these indigenous woody vegetation communities was a major factor contributing to the extirpation of many small bird species, and undoubtedly also reptile and invertebrate species, from the region. Plant macrofossils from rockshelters included remains of bird nests, identifiable by desiccated feathers and eggshell amongst them. These macrofossils include the first described plant remains from the nests of moa, which were constructed from a shallow bed of twigs of locally available shrubs and lianes. Many of the twigs are 25-30 mm in length and show evidence of having been clipped by moa bills. Desiccated coprolites, mostly of moa, but also specimens attributed to Finsch�s duck (Chenonetta finschi) and red-crowned parakeet (Cyanoramphus novaezelandiae), were recovered from rockshelter excavations. Moa species associated with a sample of coprolites were identified using ancient DNA analysis, and plant macrofossils from these were examined, together with previously unexamined moa gizzard content samples excavated from mires in the eastern South Island dryland zone. The results indicate that, in addition to previously reported browsing, upland moa (Megalapteryx didinus) and heavy-footed moa (Pachyornis elephantopus) also functioned as grazers, and seeds in their coprolites are dominantly of low shrubs and ground-cover herbs. Of particular interest was the higher than expected frequency of seeds from the currently rare and threatened 'spring annual' herbs; Ceratocephala pungens and Myosurus minimus subsp. novae-zelandiae (Ranunculaceae), suggesting further research on potential ecological relationships between moa and these plants would be worthwhile. The results of this study have provided a baseline for future conservation and restoration projects in the Central Otago lowlands.

The vegetation communities that existed in the semi-arid intermontane basins and gorges of Central Otago prior to human settlement ~750 years B.P. are poorly understood. This is because of a lack of fossil evidence and complex restructuring by anthropogenic factors, especially increased fire frequency, and more recently mammalian grazing. There is also little information regarding the effect of the lost fauna on maintaining and structuring presettlement communities, both in Central Otago and throughout the eastern South Island dryland zone. This study aims to provide a clearer understanding of the functioning of pre-settlement ecosystems in dryland Central Otago, particularly the role of the largest vertebrate herbivores, the moa (Aves: Dinornithiformes), and to explore the implications of the extinct fauna for land conservation management across New Zealand. Late Pleistocene and Holocene vegetation communities of the Central Otago lowlands were reconstructed from plant macrofossils, including seeds, leaves, and wood, excavated from rockshelter, cave, and swamp deposits throughout the region. The macrofossils represent three main vegetation types: late Pleistocene to mid (late?) Holocene basin floor wetland herb associations, Olearia-shrublands surrounding these wetlands, and mid to late Holocene open scrubland and woodland in gorges and on low altitude slopes, dominated by filiramulate Olearia, Coprosma, and Corokia, with abundant lianes (Muehlenbeckia spp. and Rubus spp.) and understorey herbs. Many native tree and shrub species that are presently widespread in the Central Otago lowlands were rare or absent prior to anthropogenic fires (e.g. Discaria toumatou, Kunzea ericoides, Leptospermum scoparium). Other tree and shrub species once present are now extinct in the region (e.g. Coprosma obconica, Plagianthus regius, Pseudopanax ferox). The loss of these indigenous woody vegetation communities was a major factor contributing to the extirpation of many small bird species, and undoubtedly also reptile and invertebrate species, from the region. Plant macrofossils from rockshelters included remains of bird nests, identifiable by desiccated feathers and eggshell amongst them. These macrofossils include the first described plant remains from the nests of moa, which were constructed from a shallow bed of twigs of locally available shrubs and lianes. Many of the twigs are 25-30 mm in length and show evidence of having been clipped by moa bills. Desiccated coprolites, mostly of moa, but also specimens attributed to Finsch�s duck (Chenonetta finschi) and red-crowned parakeet (Cyanoramphus novaezelandiae), were recovered from rockshelter excavations. Moa species associated with a sample of coprolites were identified using ancient DNA analysis, and plant macrofossils from these were examined, together with previously unexamined moa gizzard content samples excavated from mires in the eastern South Island dryland zone. The results indicate that, in addition to previously reported browsing, upland moa (Megalapteryx didinus) and heavy-footed moa (Pachyornis elephantopus) also functioned as grazers, and seeds in their coprolites are dominantly of low shrubs and ground-cover herbs. Of particular interest was the higher than expected frequency of seeds from the currently rare and threatened 'spring annual' herbs; Ceratocephala pungens and Myosurus minimus subsp. novae-zelandiae (Ranunculaceae), suggesting further research on potential ecological relationships between moa and these plants would be worthwhile. The results of this study have provided a baseline for future conservation and restoration projects in the Central Otago lowlands.

This thesis set out to examine the soils, plants, lichens, invertebrates and common skinks along a gradient of habitat modification. This gradient was represented by ten study sites, of varying management histories, in a mid altitude tall tussock (Chionochloa spp.) grassland in eastern Central Otago. One extreme of the gradient was represented by intact tall tussock plants with a high density of inter tussock vegetation. The other extreme was represented by intensively managed exotic pasture. Native plants, invertebrates and reptiles were present along the entire gradient of modification and there was a clear threshold in the diversity of native plants along the gradient. This threshold was reached after the management practice of ploughing was applied to a site. Sites that had not been ploughed were found to be similar in their native plant diversity, and all unploughed sites had significantly higher native plant diversity than unploughed sites. There was considerable variation in the abundance of individual plant and Coleoptera species along the modification gradient. The most intensively modified sites had the lowest plant diversity, the lowest abundance and diversity of Orthoptera, as well as the lowest abundance of common skinks. Sites with low levels of modification contained a higher abundance of common skinks, and, native forbs, lichens and bryophytes than the more intensively modified sites. With regard to individual plant and Coleoptera species recorded, there were generally not clear thresholds for their presence and absence along the modification gradient. Rather, there was a gradual turnover for most species along the gradient such that the plant and Coleoptera community at opposing ends of the gradient had less species in common sites of similar modification levels. A high proportion of the Coleoptera species observed were present in all sites, suggesting that disturbance has selected for an adaptive generalist life trait, and that the species remaining comprise the resilient portion of the Coleoptera fauna. The relationship between Coleoptera communities and the gradient of modification was scale dependent, with the strongest relationship being observed at the largest scale measured. The hypothesis that faunal diversity and abundance would be best predicted by the structural diversity of the vegetation rather than by plant species diversity was tested. Structural diversity was strongly correlated (p < 0.01) with Coleoptera diversity, and Coleoptera abundance whereas plant species diversity was not. However, plant species diversity was a stronger correlate of common skink abundance than plant structural diversity. The structural diversity of the vegetation has the potential to be measured remotely and could be a correlate for faunal diversity and abundance when undertaking landscape scale studies. This thesis demonstrates that native biota exists within the agricultural environment of the mid-altitude tall tussock (Chionochloa rigida) grasslands of eastern Central Otago despite intensive modification in some areas. The findings suggest that it is possible to integrate the objectives of agriculture and the conservation of some faunal groups especially for the more resilient species. However, this study also indicates that relatively unmodified sites are valuable reservoirs of biodiversity in the mid altitude zone and it is recommended that the limited number of lightly modified sites that remain in the mid altitude zone be conserved.

This study is an interdisciplinary ecological study addressing the fine-scale relationships between plants, invertebrates and the environment in an alpine ecosystem. Alpine environments are marked by steep environmental gradients and complex habitat mosaics at various spatial scales. Regular forming periglacial patterned ground landforms on the Old Man Range, Central Otago, South Island, New Zealand present an ideal medium for studying plant-invertebrate-environment relationships due to their partitioning of the landscape into discrete units of contrasting environmental conditions, and the existence of some baseline knowledge of the soil, microclimate, vegetation and flora. The study was conducted in three types of patterned ground (hummocks, stripes and solifluction terraces) on the Old Man Range. Each component of the study was sampled at the same spatial scale for comparison. Temperature was recorded in the soil and ground surface from April 2001 to March 2004 in microtopographic subunits (microsites) of each patterned ground landform. Plant species cover was sampled within each microsite; invertebrates were sampled from soil cores taken from the same locations as plant samples in April 2001 and September 2001. The two sampling occasions coincided with autumn before the soil freezes, and winter when maximum freezing was expected. Fine-scale changes in the topographic relief of the patterned ground led to notable differences in the timing and duration of snow. The steepest environmental gradients existed during periods of uneven snow distribution. The soil in exposed or south-facing microsites froze first, beginning in May, and typically froze to more than 40cm depth. Least exposed microsites rarely froze. Within the microtopography, patterns of freezing at specific locations were consistent between years with only minor differences in the timing or depths of freezing; however, notable variation in freezing existed between similar microsites. Within the microtopography, different assemblages of organisms were associated with different microsites. In total, 84 plant and lichen species were recorded, grouping into six community types. Species composition was best explained by growing degree-days, freeze-thaw cycles, time frozen and snow-free days; species diversity and richness increased with increasing environmental stress as indicated by freeze-thaw cycles, time frozen and exposure. In total 20,494 invertebrates, representing four Phyla, 12 Classes, 23 Orders and 295 morpho-taxa were collected from 0.17m� of soil. Acari, Collembola and Pseudococcidae were the most abundant invertebrates. Over 95% of the invertebrates were found in the plant material and first 10cm depth of soil. Few significant relationships were found between diversity, richness or abundance of invertebrate taxa and the microsites; however, multivariate analyses identified distinct invertebrate assemblages based on abundance. Invertebrate composition was best explained by recent low temperature and moisture, particularly in winter; however, plant composition also explained invertebrate composition, but more so in autumn. This research has shown that organisms in the alpine environment of the Old Man Range are sensitive to fine-scale changes in their environment. These results have implications as to how historical changes to the ecosystem may have had long-lasting influences on the biota, as well as how a currently changing climate may have further impacts on the composition and distribution of organisms.

This study is an interdisciplinary ecological study addressing the fine-scale relationships between plants, invertebrates and the environment in an alpine ecosystem. Alpine environments are marked by steep environmental gradients and complex habitat mosaics at various spatial scales. Regular forming periglacial patterned ground landforms on the Old Man Range, Central Otago, South Island, New Zealand present an ideal medium for studying plant-invertebrate-environment relationships due to their partitioning of the landscape into discrete units of contrasting environmental conditions, and the existence of some baseline knowledge of the soil, microclimate, vegetation and flora. The study was conducted in three types of patterned ground (hummocks, stripes and solifluction terraces) on the Old Man Range. Each component of the study was sampled at the same spatial scale for comparison. Temperature was recorded in the soil and ground surface from April 2001 to March 2004 in microtopographic subunits (microsites) of each patterned ground landform. Plant species cover was sampled within each microsite; invertebrates were sampled from soil cores taken from the same locations as plant samples in April 2001 and September 2001. The two sampling occasions coincided with autumn before the soil freezes, and winter when maximum freezing was expected. Fine-scale changes in the topographic relief of the patterned ground led to notable differences in the timing and duration of snow. The steepest environmental gradients existed during periods of uneven snow distribution. The soil in exposed or south-facing microsites froze first, beginning in May, and typically froze to more than 40cm depth. Least exposed microsites rarely froze. Within the microtopography, patterns of freezing at specific locations were consistent between years with only minor differences in the timing or depths of freezing; however, notable variation in freezing existed between similar microsites. Within the microtopography, different assemblages of organisms were associated with different microsites. In total, 84 plant and lichen species were recorded, grouping into six community types. Species composition was best explained by growing degree-days, freeze-thaw cycles, time frozen and snow-free days; species diversity and richness increased with increasing environmental stress as indicated by freeze-thaw cycles, time frozen and exposure. In total 20,494 invertebrates, representing four Phyla, 12 Classes, 23 Orders and 295 morpho-taxa were collected from 0.17m� of soil. Acari, Collembola and Pseudococcidae were the most abundant invertebrates. Over 95% of the invertebrates were found in the plant material and first 10cm depth of soil. Few significant relationships were found between diversity, richness or abundance of invertebrate taxa and the microsites; however, multivariate analyses identified distinct invertebrate assemblages based on abundance. Invertebrate composition was best explained by recent low temperature and moisture, particularly in winter; however, plant composition also explained invertebrate composition, but more so in autumn. This research has shown that organisms in the alpine environment of the Old Man Range are sensitive to fine-scale changes in their environment. These results have implications as to how historical changes to the ecosystem may have had long-lasting influences on the biota, as well as how a currently changing climate may have further impacts on the composition and distribution of organisms.