Habilitationsschrift--Universität Münster with title: Die Wanderviehwirtschaft in den Hochgebirgs- und Beckenlandschaften der westlichen USA und ihre Auswirkungen im Naturraum. / Includes English summary. Includes bibliographical references (p. 419-469).

Thesis (M. Sc. Agric(Pasture science))--University of Pretoria, 2005. / Includes bibliographical references. Available on the Internet via the World Wide Web.

To proffer a sustainable solution to ecological degradation in rangeland ecosystems as a consequence of fire, grazing and browsing, an understanding of rangeland ecological processes is vital. Due to the complexity of ecological processes and their interrelationships, it is usually difficult or expensive to directly measure status of ecological processes. Therefore, biological and physical characteristics are often used to indicate the functionality of ecological processes and site integrity. Long-term effects of fire, grazing and browsing on characteristics of the vegetation and organic matter and their subsequent effects on selected rangelands ecosystem ecological processes was conducted at Honeydale section of the University of Fort Hare farm in the Eastern Cape Province of South Africa and Matopos Research Station in Zimbabwe. In this study, attributes of biotic community integrity (species richness, composition and diversity), soil stability (basal cover, standing dead grass biomass, tuft to tuft distance, tufts diameter, canopy distance and stem to stem distance), productivity and plant vigour (grass yield, total canopy volume, plant height, canopy height, canopy diameter, main stem diameter, sprouts diameter and number of sprouts) and hydrologic function and nutrient cycling (grass litter biomass, soil organic carbon and microbial biomass carbon) were used to estimate long-term effects of burning, grazing and browsing by goats on the functionality of ecological processes in the rangeland ecosystem. Burning did not have differential effect on grass species richness (P>0.05), woody species diversity as well as compositional percentage for D.eriatha, C.plurinodis, S.fimbriatus, A.karro and E.rigida. Burning increased decreasers and increaser II species proportions and reduced (P ≤ 0.05) grass yield, total canopy volume, tree height, canopy height main stem diameter and sprouts diameter. Long-term burning, grazing, and goats browsing had differential effects on site stability. The effects on basal cover, tuft to tuft distance, tufts diameter, canopy distance and basal distance as a consequence of long-term burning, grazing, and goats browsing were not significantly different, whereas the effects on standing dead grass biomass as a result of long-term burning frequencies were significantly different. Long-term effects of burning followed by ten-year period of fire exclusion had significantly different effects on tuft-tuft distance but did not have statistically different effects on tufts diameter, canopy distance and basal distance. Long-term burning grazing and browsing had significantly different effects on attributes of hydrologic functions and nutrient cycling in the rangeland ecosystem (grass litter biomass, SOC and BMC). Long-term effects of burning followed by ten-year period of fire exclusion had significantly different effects on grass litter biomass, and SOC. Through their effect on vegetation and organic matter characteristics, burning, grazing and browsing could influence functionality of selected rangeland ecological processes such as biological community integrity, productivity and plant vigour, site stability, hydrologic function and nutrient cycling.

Vegetational responses to changes in exposures within a constant slope range were studied on the Shallow Upland Range Sites and Granitic Hills Range Site in the Chihuahuan semidesert grasslands in Central Arizona. Sixteen exposures with slopes between 11 and 17 degrees were chosen for subsample sites. Environmental, complete soil descriptions, and vegetational composition data were taken. All data were analyzed using analyses of variance, ordination programs, and regression analyses to determine climate, soils, and vegetational relationships among exposures. The resulting data indicated that the geological lithologic unit on which the soils formed was the most important factor affecting apparent vegetational type. In this study, the data obtained from the complete soil profile descriptions contributed little information to the understanding of vegetational responses. Soil surface characteristics and surface soil horizon properties influenced soil moisture relationships. The conservation of soil moisture appeared to be more important to plant communities than did the total moisture holding capacity of the soil continuum. Monthly precipitation reliability and soil surface reflectances were environmental factors affecting plant communities occurring on different exposures. Fall/spring, winter/spring, and spring soil temperature interactions were the most important environmental factors affecting vegetation on different sloping exposures. All exposures within each of the four sample locations had vegetational components that were similar to the vegetational components of other exposures but all exposures were found to have different plant communities. Each exposure within a given slope range is a phase and/or subphase of currently used range site descriptions. A range site that is based on a potential natural community at one type location cannot be extrapolated across broad geographical expanses to define vegetative potentials for other areas having similar vegetative aspects. Range site descriptions must be site specific for one geographical rangeland that has had the same historical uses.

Thesis (M.S.)--University of Wyoming, 2008. / Title from PDF title page (viewed on June 24, 2009). Includes bibliographical references (p. 93-104).

The world we live in is changing rapidly. Ecological, economic and social aspects and understandings are all undergoing paradigm shifts. Communities, farmers and individuals in arid zones are experiencing climate changes, more so than city dwellers. A better understanding of the current thinking in range ecology and management, especially of arid environments, is critical to the management of these delicate, complex systems. Ecosystems in equilibrium or in disequilibrium react differently to management and reclamation efforts. An understanding of the basic principles and how they evolved is important in order to apply these principles correctly in the management of arid zones. The use of keystone species and simple technologies, such as water harvesting and mulching, can all be used to reclaim and manage the arid zones. Understanding the differences between systems in equilibrium and disequilibrium can be used as a guide for planning appropriate future research in the arid zones. Searching for new indigenous species to help in the reclamation of arid zones is of the utmost importance. An ecological criterion was used to identify potential plant species for reclamation of degraded arid rangelands of southern Africa. Tripteris sinuatum and Sutherlandia microphylla were identified as possible candidates. Germination studies, with seeds harvested from naturally occurring plants, were conducted for both species. Treatments were based on the natural seed dispersal mechanisms for both species. S. microphylla has the potential to become an important plant species for reclamation purposes in arid zones of southern Africa. Not only potential new species should be sought, but also the management of species, already in use, are of critical importance. The relative palatability and survival of 16 different Atriplex species and accessions were determined at two different localities in the arid Northern Cape Province of South Africa. Significant differences were found between species at both localities. It is believed that because of the variety of species in the Atriplex genus, relative palatability and survival should be used to determine which of the different species could be useful under specific climatic and soil conditions. The establishment and reaction of Atriplex nummularia and Cassia sturti were tested for season of planting and the use of a stone mulch at two locations in the Northern Cape Province of South Africa. A. nummularia reacted the best to a moderate pruning treatments, while Cassia sturtii reacted best to severe pruning. These results should aid in the management of planted fodder plantations. Numerous attempts at improving natural veld have failed in the past. An examination of landscape function and the potential to harvest water in localized areas should drive veld improvement in arid zones. Seeds from two different plants species (Tetragonia calycina and Tripteris sinuatum) were used to inter-seed a bare patch in the Northern Cape Province of South Africa. The two species, with two treatments, (brush packing or not) in two different locally occurring eco-topes, were used to determine plant establishment. The water run-on eco-tope showed a significantly higher plant establishment percentage than the water run-off eco-tope. The establishment of perennial grass species was also found on the water run on eco-tope, three years after establishing the reclamation site. Such sites could form an important link in biodiversity conservation. / Dissertation (MSc Agric (Pasture Science))--University of Pretoria, 2007. / Plant Production and Soil Science / unrestricted

Thesis (MSc)--University of Stellenbosch, 2003. / ENGLISH ABSTRACT: Arid and semi-arid rangelands of the Nama-Karoo Biome are believed to have changed considerably since the arrival of domestic livestock in the veld. Severe grazing pressure is considered to be one of the prime factors responsible for the perceived degradation of vegetation and concurrent soil degradation. To understand the process of degradation and to make further recommendations for future veld restoration, a large-scale project was undertaken in the Eastern Cape. This project focused on the role that landscape heterogeneity plays in providing refuges for plant species. The key question asked in this project was: What role do these hypothetically less grazed mesas play in the conservation of rangelands in the Nama-Karoo of the Eastern Cape? This study, a component of the larger project, investigated grazing and soil landscape gradients on- and off- the three mesas (Tafelberg, Folminkskop and Buffelskop) in the Middelburg District of the Eastern Cape, South Africa and the possible influences that they might have on the veld. The flats surrounding the mesas were mostly used as grazing camps in contrast to the plateaux of the mesas, which had varied levels of accessibility. In the absence of direct observations, the primary objective of the study was to test the differential dung pellet abundance and impact of animals on different parts of the landscape. As such, dung pellet counts in this study were correlated with surrogates of soil physical properties including bare ground, trampling and litter cover. Variation in dung pellet density was found at the different habitats (flats, slopes and plateaux) of all mesas. The flats to the north-eastern and south-western of Tafelberg mesa were found to be more heavily utilized by livestock and herbivores, while the plateaux and southeastern slopes of Folminkskop and Buffelskop were also utilized by grazers. The Tafelberg mesa was the only study site that was consistent with the hypothesis which stated that grazers would be less concentrated on the plateau compared to the surrounding flats due to its inaccessibility, whilst the high mean dung pellet density on the plateaux of the smaller Folminkskop and Buffelskop mesas due to easier access contradicted the original hypothesis. It appeared that dung pellet density did not clearly turn out to be an indicator of habitat use in this study, but showed where slopes and plateaux were accessible to herbivores, as in the case for the Buffelskop mesa, a higher abundance of dung pellets were found suggesting that higher intensities of habitat use took place. During this study a strong pattern was observed of bare soil patches on the flats surrounding the mesas. There was a decrease in percentage of bare soil along the gradient of the three mesas with a high percentage of bare soil on the plateaux of the mesas. A significant correlation was found between bare soil and dung pellet density. However, the plateaux of Folminkskop and Buffelskop had a high percentage of bare soil compared to the plateau of Tafelberg mesa. Farmers mainly used these smaller mesas as grazing camps for their livestock and herbivores. A positive correlation between bare soil and litter cover of the different habitats was evident in this study. A lower percentage of litter cover at these sites was associated with a high percentage of bare soil. Litter is very important in a healthy vegetation community in terms of nutrient cycling and fertile patches. A detailed assessment of soil chemical and physical properties would reveal, firstly, if vegetation change is better explained by soil or grazing effects and, secondly, if changes in soil have resulted from land use. Differences in macro- and micro-site variations between open-canopy (between shrubs) and closed-canopy (under shrubs) sites for each habitat were determined to differentiate between local scales due to land use and landscape scales due to geomorphology. The results suggested that carbon, phosphorus, calcium, magnesium, copper and manganese levels in soils at landscape scale better explain vegetation changes between habitats. At a local scale (open- and closed-canopy sites) land use was responsible for little changes in soils. Changes in only soil potassium, zinc and boron elements were actually a consequence of local scales due to land use. The soil nutrient content on the slopes appeared to be intermediate between the flats and plateaux of all three mesas. It appeared that dolerite capped Tafelberg and Folminkskop mesas had high silt and clay content, while Buffelskop (sandstone) mesa had a lower silt and clay content. Consequently, the texture and parent material of the soils contributed to the variations in soil nutrient composition between these mesas. High infiltration rate together with low nutrient content on the flats clearly showed that these flats, surrounding the mesas were degraded. The high infiltration rates were caused by high activity by livestock and other indigenous animals on the flats which breaks the surface crusting of bare soil and improve infiltration. It was concluded that high levels of grazing at these sites have also altered the textural and soil properties. Endozoochory dispersal and the deposition of dung pellets in areas of small patch disturbances play an important role in veld regeneration in degraded areas. Dung pellets collected from permanent study sites on the southeastern and northwestern flats and slopes, and all the study sites on the plateau of Tafelberg mesa, was sown in seedling trays, watered and monitored for seedling germination. Species list were then compared to below-ground soil seed bank data and above-ground vegetation data collected by other researchers at the same permanent study sites. Higher seedling percentages were recorded from dung pellets collected on the flats than on the plateau. A total of sixteen species were found to germinate in dung pellets collected on the flats compared to ten species germinating in dung pellets collected on the slopes and two species on the plateau of Tafelberg mesa. The seeds that germinated represent a variety of palatable grasses and shrub species. Aristida sp., Eragrostis bicolor, Eragrostis ch/orome/as and Eragrostis obtusa were palatable grasses recorded for dung pellets collected on the flats. Of the species recorded, Aristida sp., Chenopodium sp. and Pentzia sp. were found in dung pellets but were not recorded in parallel soil seedbank and vegetation studies. Successful restoration of veld conditions requires strict grazing management practices. Germination of seed in dung pellets might be considered to be a valuable means of indicating restoration potential and rangeland conditions for the identification of both degraded and conservation worthy areas. With appropriate land management skills and restoration measures, these challenges can be constructively and creatively faced. / AFRIKAANSE OPSOMMING: Daar word geglo dat die ariede en sernl-arlede weivelde van die Nama-Karoo bioom aansienlik verander het sedert die aankoms van mak lewende hawe in die veld. Swaar weidingsdruk word beskou as een van die vernaamste faktore verantwoordelik vir die waarneembare oorbeweiding van plantegroei en die gelyktydige grond agteruitgang (degradasie). Ten einde die proses van degradasie beter te verstaan, en om verdere aanbevelings te kan maak vir toekomstige veld herstel, is huidiglik 'n grootskaalse projek in die Oos-Kaap aan die gang oor die rol wat landskap ongelyksoortigheid speel in die voorsiening van skuiling aan plant spesies. Hierdie studie ondersoek weiding en grondlandskap gradiënte (op en af) van die drie mesas (Tafelberg, Folminkskop en Buffelskop) in die Middelburg streek van die Oos- Kaap, Suid-Afrika en die moontlike invloed wat hierdie gradiënte op die veld het. Die vlaktes was meestal gebruik vir weidingskampe in teenstelling met die kruin van die mesas. Hierdie studie het duidelike patrone vir habitat-gebruik op die vlaktes, hange en kruine van al die mesas aangetoon. Die vraag wat gestel was, was watter rol speel hierdie hipoteties minder beweide mesas in die bewaring van weivelde in die Nama- Karoo van die Oos-Kaap? Hierdie studie toets nie direk vir mis verspreiding as 'n plaasvervanger vir dier verspreiding in die landskap sedert werklike dier getalle nie oorweeg was nie. In die awesigheid van direkte waarneming, die primêre doelwit was om te toets die differensiaal misdigtheid en die impak van diere op verskillende dele van die landskap. Mis getalle in hierdie studie was gekorreleer met plaasvervangers van fisiese eienskappe insluitent onbedekte grond, vertrapping en droë plant material decking. Variasie in misdigtheid was gevind by die verskillende habitate (vlaktes, hange en kruine) van all die mesas. Die vlaktes aan die noord-oostelike en suid-westelike kante van die Tafelberg mesa was meer hewig benut deur lewende hawe en hêrbivore, terwyl die kruine en suid-oostelike hange van Folminkskop en Buffelskop ook gebruik was deur weidende diere. Die Tafelberg mesa was die enigste studie area wat konsekwent was met die hipotese, terwyl die hoë gemiddeld misdigtheid op die kruine van die kleiner Folminkskop en Buffelskop mesas, as gevolg van makliker toegang weerspreek die oorspronklike hipotese. Dit blyk dat misdigtheid duidelik opkom as 'n nie aanwyser habitat-gebruik in hierdie studie, maar wys well waar hange en kruine toeganklik was vir herbivore, soos in die geval van Buffelskop mesa, oorvloed van mis was gevind wat voorstel dat hoer intensiteit van habitat-gebruik voorgekom het. 'n Duidelike patroon is waargeneem tydens hierdie studie van onbedekte grond areas op die vlaktes rondom mesas. Daar was 'n afname in persentasie onbedekte grond langs die gradiënt van die drie mesas. Die kruine van Folminkskop en Buffelskop toon hoër persentasies onbedekte grond in vergelyking met die kruin van die Tafelberg mesa. Daar is 'n betekenisvolle korrelasie tussen onbedekte grond en misdigtheid gevind. Boere gebruik hierdie kleiner mesas hoofsaaklik as weidingskampe vir hulle lewende hawe en hêrbivore. In hierdie studie is daar 'n positiewe korrelasie gevind tussen onbedekte grond en droë plant material decking van die verskillende habitatte. Droë plantmateriaal bedekking speel 'n baie belangrik rol in 'n gesonde plant gemeenskap in terme van voedingsirkulering en vrugbare kolle. 'n Breedvoerige skatting van grond chemise en fisiese eienskappe wou openbaar eerstens of plantegroei verandering beter verklaar word deur grond of weiding effekte en tweedens, of verandering in grond veroorsaak was deur land-gebruik. Verskille in makro- en mikroterrein-wisseling tussen oop-blaredak (tussenplantruimtes) en toeblaredak (onderplantdekking) terreine vir elke habitat was bepaal om te onderskei tussen lokaal effek as gevolg van land-gebruik en landskap effek as gevolg van geomorfologie. Die resultate stel voor dat koolstof, fosfor, kalsium, magnesium, koper en mangaan vlakke in die grond by landskap effek beter plantegroei verandering tussen habitate verduidelik. By 'n lokaal effek (oop-en toe-blaredak terreine) land-gebruik was verantwoordelik vir min veranderinge in grond. Verandering in grond kalium, sink en baron elemente was werklik 'n gevolg van lokaal effek as gevolg van land-gebruik. Die nutrient inhoud van grond op die hange blyk intermediêr te wees tussen die van die vlaktes en kruine van al drie mesas. Hierdie studie toon ook dat die geologie en topografie van hierdie mesas die grond tekstuur en nutrient samestelling op en af van die mesas beinvloed. Dit wil voorkom asof die dolerite bedekte Tafelberg en Folminkskop mesas, hoë slik en klei materiaal bevat, terwyl Buffelskop (sandsteen) mesa 'n laer slik en klei inhoud het. Gevolglik, tekstueel en ouer materiaal grond dra by tot die variasie in grond nutrient samestelling tussen hierdie mesas. Hoë infiltrasie tempo tesame met lae voedingstof van die vlaktes dui daarop dat die vlaktes rondom die mesas oorbewei was. Hierdie hoë infiltrasie tempo word veroorsaak deur hoë dierlike aktiwieteite op die vlaktes deur lewende hawe en ander inheemse diere wat die oppervlakke van onbedekte grond versteur en infiltrasie verbeter. Die gevolgtrekking word gemaak dat die hoë weidingsvlakke van hierdie terreine ook die tekstuur en ander grond eienskappe verander het. Endozoochoreeuse verspreiding en die mis-afsetting in gebiede van klein areaversteurings, speel 'n belangrike rol in veld herlewing in oorbeweide gebiede. Mis versamel by permanente studie terreine op die SO en NW vlaktes en hange, as ook al die studie terreine op die kruin van die Tafelberg mesa, is gesaai in saailing bakke, besproei en gekontroleer vir saailing ontkieming. 'n Lys van spesies is vergelyk met onder-grondse saadbank gegewens en bo-grondse plantegroei gegewens wat deur ander navorsers by dieselfde studie terreine versamel was. Hoër saailing persentasies is gevind vir mis wat op die vlaktes versamel is as op die kruine. 'n Totaal van sestien geidentifiseerde spesies het uit die mis op die vlaktes ontkiem in vergelyking met die tien spesies wat ontkiem het uit die mis van die hange en twee spesies iut die mis van die kruin van Tafelberg mesa. Die ontkiemde saad verteenwoordig 'n verskeidenheid van vreetbare gras en struik spesies. Aristide sp., Eragrostis bicolor, Eragrostis chloromelas en Eragrostis obtusa is vreetbare grasse wat gevind was in die mis versamelop die vlaktes. Die spesies opname wys dat Aristida sp., Chenopodium sp. en Pentzia sp. Gevind was in die mis, maar nie in die parallelle grond saadbank en plantegroei studies nie. Die suksesvolle herstel van veldtoestande vereis "n streng weidingsbestuur toepassing. Die ontkieming van saad in mis kan beskou word as "n bekostigbare manier om veld herstel potentiaal te bepaal, asook weiveld toestande vir die indentifiseering van beide degradasie en bewaring van waardevolle areas. Met gepaste veldbestuur vaardighede en herstel maatreels, kan hierdie uitdagings konstruktief en kreatief aangespreek word.

This study investigated the vegetation and soil characteristics in relation to distance from water points under different land management systems in semi-arid rangelands of South Africa. Six study sites, two each from communal grazing, commercial farming and game reserves were selected. Two watering points were selected in each study site. Two 500 m transects were laid from the selected watering points. Each transect was divided into sub-transects at 25 m, 50m, 100m, 200m, 300m and 500m from water points. A 100 m2 plot was marked in each sub-transect to record grass biomas, species composition, structure and distribution of woody vegetation and physical and chemical soil properties. Data were collected for two seasons 2012/13 (winter and summer). Thirty and 41 grass and woody species respectively were identified in all study areas. The most common and dominating grass species include Cynodon dactylon, Digitaria eriantha, Eragrostis obtusa, Setaria sphacelata and Sporobolus fimbriatus. Cynodon dactylon and S. sphacelata occurred more abundantly (p < 0.05) in the game reserves than in the other land management catergories. All the grass species had similar (p > 0.05) abundance along distance gradient from water points. Grass dry matter (GDM) showed significant differences (p < 0.05) between and within land management systems. However, GDM was not significantly affected by season, location of water point within each farm or reserve and distance along water points. Acacia karoo, Coddia rudis and Ehretia rigida were the most dominant woody species. Tree equivalent (TE) density of all encroaching woody plants combined was significantly (p < 0.05) higher on the communal area (1732 TE ha-1) than the commercial ranches (1136 TE ha-1) and game reserves (857 TE ha-1), but with no marked variations along distance from water points under all the land management systems. The electric conductivity (EC) was significantly (P < 0.01) higher in game reserves than in communal grazing areas and ranches. Soil organic matter percentage showed greatest and lowest values in the game reserves and commercial respectively. Soil pH and bulk density did not vary but soil organic matter (SOM), EC, bulk density and soil compaction were significant different with no increasing or decreasing trends. Soil properties were affected by herbivore pressure and trampling around water points with inconsistence magnitude and direction. In conclusion, grass species composition and GDM did not respond to distance from water points because either grazing gradient was absent or the length of transects was not enough to explain the absence or presence of gradients. High proportion of A. karoo and high densities of seedlings and saplings would seem as very good indicators of the woody vegetation changes in the different land management systems and distance from water points. The soil quality indicators around the water-points showed that livestock and game affected soil parameters.

No description available.

Past research suggests that social organisation in Australian rodent species is determined by the predictability of resources in the environment (Happold 1976a). Notomys alexis (the spinifex hopping mouse or tarrkawarra) is widely distributed throughout the Australian arid-zone (Breed 1998a; Watts & Aslin 1981). Large groups of animals have been found sharing burrows in the wild and laboratory observations suggest that the species is highly social (Happold 1976a; Stanley 1971). A closely related species, Notomys mitchelli (Mitchell’s hopping mouse or pankot) is relatively common throughout the southern semi-arid zone (Watts 1998a; Watts & Aslin 1981). Much less is known about N. mitchelli; field studies have been subject to low recapture rates and few laboratory studies have involved this species (Baverstock 1979; Cockburn 1981a; Crichton 1974). Following Happold (1976a), it was hypothesised that the socio-ecology of N. mitchelli will be qualitatively different to that of N. alexis. Studying wild populations of Notomys mitchelli in the Middleback Ranges, South Australia and N. alexis outside the desert township of Roxby Downs, S.A., this research aimed to describe the socio-ecology of these species, with reference to the predictability of their environments. Uniquely, bioclimatic modelling of the species’ known distributions was also conducted to confirm that the study’s underlying assumption that the two rodents inhabit essentially different environments was correct. These studies were complemented by the observation of captive groups of N. alexis. This work confirmed that the habitats of Notomys mitchelli and N. alexis are distinct; the more arid habitat of N. alexis is subject to greater environmental fluctuations than that of the semi-arid dwelling N. mitchelli. Contrary to expectations, however, observation of free-living animals characterised N. mitchelli social groups as highly unstable; while particular individuals remained in the population for long periods, many animals appeared to be transients. Further, burrow groups appeared to be much smaller than predicted by Happold (1976a), and based on loose aggregations of male animals rather than small groups of females. While decidedly social in the laboratory, free-living N. alexis lived in groups no bigger than N. mitchelli and these groups were equally ephemeral in constitution. Moreover, free-living N. alexis appeared to utilise activity areas that were no larger than those used by N. mitchelli, despite occupying a more unpredictable and apparently less well-resourced habitat. As a whole, this research represents a comprehensive examination of the principal behavioural theory commonly applied to Australian native rodent species, untested since its publication three decades ago. From the data collected during this work, it cannot be said that the environmental predictability hypothesis for native rodent social organisations as proposed by Happold (1976a) is adequate to differentiate the social behaviour of these semi-arid and arid-dwelling species. Instead, it is suggested that, while both N. mitchelli and N. alexis are undoubtedly socially tolerant, predation and/or parasite load, driving burrowing behaviour, has a greater influence on the social behaviour of free-living rodent populations than habitat predictability. / http://proxy.library.adelaide.edu.au/login?url= http://library.adelaide.edu.au/cgi-bin/Pwebrecon.cgi?BBID=1373743 / Thesis (Ph.D.) - University of Adelaide, School of Earth and Environmental Sciences, 2009