Many stream restoration projects, as part of returning a degraded ecosystem to a healthier state, aim to restore aquatic invertebrate populations. Unfortunately, many attempts only „beautify‟ streams without achieving improvements in biodiversity. Lack of connectivity of a restoration site to a regional species pool may explain some failures. I tested this by collecting larval and adult aquatic insects from an agriculturally impacted Canterbury high country stream to evaluate connectivity of the regional species pool. The stream was surrounded by high-quality habitat in an adjacent National Park. Surrounding streams contained diverse assemblages of aquatic insects, but processes in the environment and limitations of in-stream habitat meant their adults did not always arrive at the target. In addition, oviposition habitat for hydrobiosid caddisflies was added to sections of stream and compared to un-manipulated control sections to test oviposition site limitation. The addition of oviposition habitat led to more hydrobiosid egg masses in comparison to control reaches. However, oviposition was also limited by in-stream habitat conditions, particularly the abundance of fine sediments. Sedimentation is a common pollutant in streams and is linked to decreases in habitat, food resources, and invertebrate populations. Moreover, common restoration methods, such as riparian management, have little success at reversing already high sediment levels, and are therefore insufficient to bring improvements to in-stream communities or sought-after habitat conditions. Therefore, after determining sediment was restricting sensitive invertebrate recovery at Riversdale Stream, by adding patches of high quality habitat I experimentally compared the factorial effects of sediment flushing and channel narrowing on sediment removal. Treatments improved habitat and prompted recovery of sensitive invertebrates, but an interactive effect where both flushing and channel narrowing combined created the most improved habitat conditions and the greatest improvements of invertebrate communities. Thus, while habitat improvements are an important part of restoration, features limiting species recovery such as connectivity and sedimentation, are particularly important.

Dendritic stream networks are inherently spatially and hierarchically structured, but the effects of this structure on stream communities are largely unknown. My aim was to investigate spatial patterns in stream networks using extensive spatial sampling of both adult and benthic macroinvertebrates in four stream networks on the West Coast of the South Island, New Zealand. Using spatial modelling and analyses, I answered questions about appropriate spatial measurements to capture ecological processes in stream networks, metacommunity processes at different scales in space and time, and how local and regional processes interact to structure metacommunities in stream networks. Spatial eigenfunction analyses showed that distance measures that explained most variance in stream macroinvertebrate communities were stream distance and weighted stream distance measures. They performed better than Euclidean distance to measure spatial structure that is ecologically relevant to stream network communities. The spatial pattern of benthic stream macroinvertebrates was stable over time, whereas community composition changed significantly, as shown by space-time interactions modelled by MANOVA-like redundancy analysis. Thus, spatial processes structuring stream metacommunities remained constant, in agreement with neutral model predictions. Network-scale properties, particularly flood disturbances, influenced the relative importance of spatial and environmental variation in stream network metacommunities. Additionally, quantile regression indicated that three key variables, habitat size, isolation and local habitat conditions, jointly limited community structure in stream networks, providing empirical support for both island biogeography and metacommunity theories. My study indicated that spatial structuring has an important influence on stream macroinvertebrate communities. The results contribute to broader ecological theory and understanding of community assembly by relating empirical results to theoretical predictions. In particular, they advance understanding of spatial processes in stream networks. The research also highlights a number of new methods, which were successfully applied to stream systems to elucidate complex spatial patterns.

Examensarbete är utfört på GKN Aerospace Applied Composites (ACAB) där produktionsflödet för fyra produkter har kartlagts. Idag upplever företaget svårigheter med att hålla leveranstiden för dessa produkter, dessutom upplevs verkstäderna som ostrukturerade. GKN Aerospace Applied Composites tillverkar idag flertalet produkter i komposit till både den civila och militära marknaden. För detta examensarbete har flödet av tre stycken radomer och en storlek av tryckkärl kartlagts. Tryckkärl är en produkt som tillverkas i större volymer och har ett flödesorganiserat produktionsupplägg. Radomerna är kundorderspecifika produkter som endast tillverkas i enstyck eller mindre volymer. Detta innebär att radomverkstaden är funktionellt organiserad. Med ett varierande och växande produktsortiment ökar kraven på produktionsplaneringen och hantera kombinationen av de två olika produktionsuppläggen. För att kunna ta fram förbättringsförslag för framtiden har en analys genomförts av produktionen. Förbättringsförslagen resulterar i studiens rekommendationer och förbättringsplan till ACAB. Den metod som använts är Value Stream Mapping (VSM), som är en metod i Lean filosofin. Denna metod används för att eliminera alla former av slöseri genom att identifiera värdeskapande och icke värdeskapande tid. För att sammanställa en VSM för respektive produkt genomfördes en datainsamling som bestod av intervjuer, observationer och uppmätningar av flödenas avstånd. För varje produkt identifierades problemområden med tillhörande förbättringsförslag, dessa förslag på åtgärder och rekommendationer resulterade i ett framtida alternativt flöde för produktionen. Historisk data har analyserats av genomloppstider, en medeltid för varje process har beräknats ur äldre order för respektive produkt. Som en riktlinje för dagens produktion har ACAB idag standardtider som de planerar leveransdatum efter. Beräknade medeltider har satts i förhållande till standardtiderna för att kunna urskilja vilka processtider som överstiger eller understiger dagens riktlinjer. I de alternativa flödena presenteras åtgärdsförslag för att kunna förbättra verkstädernas layout genom att förkorta de uppmätta avstånden. Utöver jämförelse av tider och avstånd presenteras övergripande åtgärder som 5S och kontinuerlig utbildning av personal. Generellt för båda produktionerna har jämförelsen mellan medeltider och standardtider visat att standardtiderna behöver korrigeras. Korrigeringen av tider medför inte kortare genomloppstider utan bidrar till en bättre planering av leveransdatum. Däremot förväntas de förändringar av layout och organisering som föreslagits för respektive verkstad reducera genomloppstiderna.

Our objective was to quantify if macroinvertebrate assemblages in Indiana streams were better predicted from co-occurring fish assemblages or environmental variables. We used Canonical Correspondence Analysis (CCA) with forward-selection of variables to identify significant environmental predictor variables for macroinvertebrate and fish assemblages. A partial Mantel test was used to determine if fish assemblage composition and macroinvertebrate assemblage composition significantly co-vary while controlling for environmental effects. The CCAs resulted in two significant predictors of macroinvertebrate distribution and relative abundance, and four significant predictors of fish distribution and relative abundance. Similarity matrices of fish and macroinvertebrates were significantly correlated in the Mantel (r = 0.22, p = 0.019) and partial Mantel tests (r = 0.23, p = 0.013). Our results suggest that macroinvertebrates respond to local and regional environmental variation, and less to local presence of fishes. / Department of Biology

Access to abstract restricted until 05/2015. / Access to thesis restricted until 05/2015. / Department of Biology

This thesis presents an investigation into the phenomena of scale effects on cavitation inception and noise of marine propellers. The overall aim is to extend the understanding of these phenomena and improve predicting methods. The investigations, which are largely experimental in nature, are restricted to the tip vortex and sheet types of cavitation. Chapter 1 includes a state-of-the-art review of the scale effect studies based on published papers to form the basis for the main objectives and structure of thesis. The objectives require systematic tests in a cavitation tunnel to explore the viscous scale effects contributing to the phenomena, particularly for the effect of the free-stream turbulence, and to include this effect in extrapolation procedures. Chapter 2 is concernedw ith the background flow measurementsin the cavitation tunnel under the effect of systematically varying levels of the free-stream turbulence generated by using wire meshes. This background information is obtained using a Laser Doppler Anemometry; measurements made with the latter provide a systematic basis on which the analyses of the cavitation inception and noise experiments can be performed. In Chapter 3, a set of cavitation inception tests is described with a NACA66 rectangular foil whose cross-section represents a typical blade section of a marine propeller. The inception measurements for systematically varying levels of the free-stream turbulence and that of the leading edge roughness are presented for different angles of attack and the results are discussed. Chapter 4 includes another set of cavitation inception experiments with a 5-bladed of model propeller of the Meridian Series. The measurements are taken for varying levels of the free-stream turbulence, blade roughness and dissolved gas contents. The results are analysed and discussed with a specific emphasis on the similarities between the effects of the free-stream turbulence and blade roughness. Chapter 5 presents a set of systematic noise measurements, with the same test propeller under the similar effects of the free-stream turbulence, blade roughness and dissolved gas content, using a single external hydrophone. The analyseso f these measurements,in terms of the tunnel background noise and net propeller noise, are presented and discussed for two operating conditions representing a typical non-cavitating and cavitating noise spectrum. In Chapter 6, a semi-empirical tool is developed to predict the inception of cavitation including the effect of the free-stream turbulence based on Lighthill's Leading Edge Correction factor (Lighthill, 1951). This tool is correlated with the inception tests results of the model propeller and its potential to be used as an extrapolator for the full-scale prediction is discussed. An attempt is made to establish a correspondence between the level of the free-stream turbulence and that of the blade roughness and its impact on the current test procedures is discussed. This chapter also includes an analysis of the similarity criteria to incorporate the effect of the free-stream turbulence in the inception of cavitation using the Dimensional Analysis procedure. In Chapter 7, a general review of the study together with the main conclusions from the thesis are presented and some recommendations for future work are made.

Streaming data have gained considerable attention in database and data mining communities because of the emergence of a class of applications, such as financial marketing, sensor networks, internet IP monitoring, and telecommunications that produce these data. Data streams have some unique characteristics that are not exhibited by traditional data: unbounded, fast-arriving, and time-changing. Traditional data mining techniques that make multiple passes over data or that ignore distribution changes are not applicable to dynamic data streams. Mining data streams has been an active research area to address requirements of the streaming applications. This thesis focuses on developing techniques for distribution change detection and mining time-changing data streams. Two techniques are proposed that can detect distribution changes in generic data streams. One approach for tackling one of the most popular stream mining tasks, frequent itemsets mining, is also presented in this thesis. All the proposed techniques are implemented and empirically studied. Experimental results show that the proposed techniques can achieve promising performance for detecting changes and mining dynamic data streams.

From the advent of association rule mining, it has become one of the most researched areas of data exploration schemes. In recent years, implementing association rule mining methods in extracting rules from a continuous flow of voluminous data, known as Data Stream has generated immense interest due to its emerging applications such as network-traffic analysis, sensor-network data analysis. For such typical kinds of application domains, the facility to process such enormous amount of stream data in a single pass is critical.

Stream ecology is dominated by studies that describe the distribution of invertebrate populations and communities. Many of these studies make untested assumptions about the various ecological processes that influence these distributions. The most common assumption, which is often unstated, is that distribution patterns reflect habitat selection by individual animals. However, very few studies have directly tested the relative importance of this process or considered other processes that may also affect the distribution of stream invertebrates. In this thesis I use a correlative survey in combination with manipulative field experiments and behavioural observations to investigate the factors and ecological processes that are associated with the distribution of hydropsychid caddisfly larvae in rocky upland streams in southeastern Australia.

[Abstract]: Outlier detection is an important research problem in data mining that aims to discover useful abnormal and irregular patterns hidden in large data sets. Most existing outlier detection methods only deal with static data with relatively low dimensionality.Recently, outlier detection for high-dimensional stream data became a new emerging research problem. A key observation that motivates this research is that outliersin high-dimensional data are projected outliers, i.e., they are embedded in lower-dimensional subspaces. Detecting projected outliers from high-dimensional streamdata is a very challenging task for several reasons. First, detecting projected outliers is difficult even for high-dimensional static data. The exhaustive search for the out-lying subspaces where projected outliers are embedded is a NP problem. Second, the algorithms for handling data streams are constrained to take only one pass to process the streaming data with the conditions of space limitation and time criticality. The currently existing methods for outlier detection are found to be ineffective for detecting projected outliers in high-dimensional data streams.In this thesis, we present a new technique, called the Stream Project Outlier deTector (SPOT), which attempts to detect projected outliers in high-dimensionaldata streams. SPOT employs an innovative window-based time model in capturing dynamic statistics from stream data, and a novel data structure containing a set oftop sparse subspaces to detect projected outliers effectively. SPOT also employs a multi-objective genetic algorithm as an effective search method for finding theoutlying subspaces where most projected outliers are embedded. The experimental results demonstrate that SPOT is efficient and effective in detecting projected outliersfor high-dimensional data streams. The main contribution of this thesis is that it provides a backbone in tackling the challenging problem of outlier detection for high-dimensional data streams. SPOT can facilitate the discovery of useful abnormal patterns and can be potentially applied to a variety of high demand applications, such as for sensor network data monitoring, online transaction protection, etc.