Orientador: Virginia Sanches Uieda / Resumo: Os ecossistemas aquáticos servem de habitat e local de alimentação para uma fauna bastante diversificada. Apesar da grande importância desses ambientes, o crescente aumento das populações humanas e o uso da terra para diferentes fins têm levado a uma intensa alteração desses locais e trazido diversos impactos para a fauna de invertebrados bentônicos. O objetivo deste trabalho foi analisar a possível influência de diferentes estados de conservação da mata ripária sobre a estrutura do habitat e da fauna de macroinvertebrados bentônicos. O trabalho foi realizado em 36 riachos localizados em quatro microbacias, na área suburbana e rural do município de Avaré, SP, sendo em cada microbacia estudados 3 riachos apresentando mata ripária nas duas margens (denominado perfil Mata), 3 com a mata totalmente excluída (perfil Pasto) e 3 com mata somente em uma das margens (perfil Intermediário). Entre outubro e dezembro de 2015 foi realizada a caracterização dos riachos utilizando um protocolo de análise física do habitat, que possibilita uma análise quantitativa ampla e, assim, uma caracterização detalhada do riacho, e a coleta dos invertebrados com amostrador surber (malha 250 µm). A análise dos dados ambientais mostrou uma separação nítida entre os riachos de mata e pasto, em função do substrato e mesohabitat, com os riachos de mata sendo caracterizados pelo substrato grosseiro e abrigo de folhas e os riachos de pasto pelo substrato fino e abrigo de macrófitas. Para a análise da composiç... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: Aquatic ecosystems serve as habitat and feeding place for a diversified fauna. Despite the great importance of these environments, the increase in human population and the use of land for different purposes had led to an intense alteration of these places and brought diverse impacts to the fauna of benthic invertebrates. This study aims to analyze the possible influence of different states of riparian vegetation conservation on the habitat and the structure of benthic invertebrates. The work was carried out in 36 streams located in four microbasins, in the suburban and rural area of the municipality of Avaré, SP. In each microbasin was studied 3 streams presenting riparian vegetation on both banks (Forest profile), 3 with forest completely excluded (Pasture profile) and 3 with this vegetation only in one margin (Intermediate profile). Between October and December 2015, the characterization of the streams was carried out using a protocol of physical analysis of the habitat, which allows a comprehensive quantitative analysis and, thus, a detailed stream characterization, and the invertebrates were sampled with a surber sampler (250 µm net). The analysis of the environmental data showed a clear separation between forest and pasture streams, as a function of the substrate and mesohabitat, with forest streams represented by coarse substrate and leaves shelter and pasture streams represented by fine substrate and macrophyte shelter. In the analysis of the fauna composition (presenc... (Complete abstract click electronic access below) / Mestre

Many aspects concerning the association of riverine fish with in-channel habitat remain poorly understood, greatly hindering the ability of researchers and managers to address declines in fish assemblages. Recent insights gained from landscape ecology suggest that small, uni-scalar approaches are unlikely to effectively determine those factors that influence riverine structure and function and mediate fish-habitat associations. There appears to be merit in using multiple-scale designs built upon a geomorphologically-derived hierarchy to bridge small, intermediate and large spatial scales in large rivers. This thesis employs a hierarchical design encompassing functional process zones (referred to hereafter as zones), reaches and mesohabitats to investigate fish-habitat associations as well as explore patterns of in-channel habitat structure in one of Australia's largest dryland river systems; the Barwon-Darling River. In this thesis, empirical evidence is presented showing that large dryland rivers are inherently complex in structure and different facets of existing conceptual models of landscape ecology must be refined when applied to these systems. In-channel habitat and fish exist within a hierarchical arrangement of spatial scales in the riverscape, displaying properties of discontinuities, longitudinal patterns and patch mosaics. During low flows that predominate for the majority of time in the Barwon-Darling River there is a significant difference in fish assemblage composition among mesohabitats. There is a strong association between large wood and golden perch, Murray cod and carp, but only a weak association with bony herring. Golden perch and Murray cod are large wood specialists, whereas carp are more general in there use of mesohabitats. Bony herring are strongly associated with smooth and irregular banks but are ubiquitous in most mesohabitats. Open water (mid-channel and deep pool) mesohabitats are characterised by relatively low abundances of all species and a particularly weak association with golden perch, Murray cod and carp. Murray cod are weakly associated with matted bank, whereas carp and bony herring associate with this mesohabitat patch in low abundance. Nocturnal sampling provided useful information on size-related use of habitat that was not evident from day sampling. Both bony herring and carp exhibited a variety of habitat use patterns throughout the die1 period and throughout their lifetime, with temporal partitioning of habitat use by juvenile bony herring and carp evident. Much of the strong association between bony herring and smooth and irregular banks was due to the abundance of juveniles (&lt100mm in length) in these mesohabitats. Adult bony herring (&gt100mm length) occupied large wood more than smooth and irregular banks. At night, juvenile bony herring were not captured, suggesting the use of deeper water habitats. Adult bony herring were captured at night and occupjed large wood, smooth bank and irregular bank. Juvenile carp (&lt200mm length) were more abundant at night and aggregated in smooth and irregular banks more than any other mesohabitat patch. Adult carp (&gt200mm length) occupied large wood during both day and night. There is a downstream pattern of change in the fish assemblage among river zones, with reaches in Zone 2 containing a larger proportion of introduced species (carp and goldfish) because of a significantly lower abundance of native species (bony herring, golden perch and Murray cod) than all other zones. In comparison, the fish assemblage of Zone 3 was characterised by a comparatively higher abundance of the native species bony herring, golden perch and Murray cod. A significant proportion of the amongreach variability in fish assemblage composition was explained at the zone scale, suggesting that geomorphological influences may impose some degree of top-down constraint over fish assemblage distribution. Although mesohabitat composition among reaches in the Barwon-Darling River also changed throughout the study area, this pattern explained very little of the large-scale distribution of the fish assemblage, with most of the variability in assemblage distribution remaining unexplained. Therefore, although mesohabitat patches strongly influence the distribution of species within reaches, they explain very little of assemblage composition at intermediate zone and larger river scales. These findings suggest that small scale mesohabitat rehabilitation projects within reaches are unlikely to produce measurable benefits for the fish assemblage over intermediate and large spatial scales in the Barwon-Darling River. This indicates the importance taking a holistic approach to river rehabilitation that correctly identifies and targets limiting processes at the correct scales. The variable nature of flow-pulse dynamics in the Barwon-Darling River creates a shifting habitat mosaic that serves to maintain an ever-changing arrangement of habitat patches. The inundation dynamics of large wood habitat described in this thesis highlights the fragmented nature of mesohabitat patches, with the largest proportion of total in-channel large wood remaining unavailable to fish for the majority of the time. At low flows there is a mosaic of large wood habitat and with increasing discharge more potential large wood habitat becomes available and does so in a complex spatial manner. What results in this dryland river is a dynamic pattern of spatio-temporal patchiness in large wood habitat availability that is seen both longitudinally among different river zones and vertically among different heights in the river channel. Water resource development impacts on this shifting habitat mosaic. Projects undertaking both fish habitat assessment and rehabilitation need to carefully consider spatial scale since the drivers of fish assemblage structure can occur at scales well beyond that of the reach. Fish-habitat associations occurring at small spatial scales can become decoupled by process occurring across large spatial scales, making responses in the fish assemblage hard to predict. As rivers become increasingly channelised, there is an urgent need to apply research such as that conducted in this thesis to better understand the role that in-channel habitats play in supporting fish and other ecosystem processes. Habitat rehabilitation projects need to be refined to consider the appropriate scales at which fish assemblages associate with habitat. Failure to do so risks wasting resources and forgoes valuable opportunities for addressing declines in native fish populations. Adopting multi-scalar approaches to understanding ecological processes in aquatic ecosystems, as developed in this thesis, should be a priority of research and management. To do so will enable more effective determination of those factors that influence riverine structure and function at the approariate scale.

Most northeastern river basins are stressed by the effects of development but the complexity of evaluating rivers often hinders the establishment of effective management regulations. Many methodologies have been proposed for assessing instream habitat, determining critical flow levels, and evaluating biological communities, but no one approach is universal. The overall objective of this thesis is to move towards standardizing components of river modeling. Rather than examine a full model, I investigated individual steps of MesoHABSIM, an instream habitat modeling approach. The two components studied involved applying the Reference Fish Community (RFC) method to identify a biological reference; and using depth and velocity data to standardize the description of hydraulic types. The RFC approach identifies the fish species and the expected proportions that should be present in a less impacted version of a river system. The Eightmile River watershed, was the focus of the study in which the RFC approach was employed to determine whether the fish community of this rural watershed, meets or exceeds a community developed using reference rivers. Similarity indices were used to identify differences between the existing (field-sampled) and expected communities. While the analysis of the Eightmile community indicates that it is in a better condition than the majority of rivers studied, it also shows some deviation from the reference, most likely due to elevated water temperatures and regional declines. The hydraulic type characterization study was developed to reduce the effort needed in depth and velocity measurement after this was identified as the most time consuming portion of the MesoHABSIM methodology. I used a series of pair-wise, independent-sample, Kolmogorov-Smirnov tests on a large bank of depth and velocity data to determine if patterns could be confirmed for each type of hydromorphological unit (HMU) across various streamflows. Few of the data sets were statistically similar enough to be combined and the mapping effort could not be simplified based on this investigation. Neither investigation provided the intended reference for the particular component of river modeling, further emphasizing the complexity in this area of study. However, the information gathered can be used as pioneering steps in future investigations.

With a new focus on flow regulation by the Tennessee Valley Authority (TVA) in reservoir tailwaters, it is now possible to recover many mussel species that once occurred in these reaches. Before flows can be modified to create habitat for freshwater mussels, suitable microhabitat conditions must be defined. In this study, I used multiple approaches to define suitable microhabitats for species in the free-flowing upper Clinch River, Virginia and Tennessee, where reproducing mussel populations persist. During summer low flows in 2003 and 2004, I measured flow and substrate conditions in over 1000 microhabitat patches (0.25 m² quadrat samples) across five river reaches. Flow characteristics and embeddedness were significantly different between microhabitats occupied and unoccupied by the most abundant species (MRPP, p < 0.05). Comparison of simple and multiple logistic regression models with Akaike's Information Criteria (AIC) demonstrated that increasing Fleisswasserstammtisch (FST) hemisphere number (a measure of shear stress), decreasing degree of embeddedness, and increasing mean column velocity best explained species occurrences in a microhabitat patch. Subtle differentiation in habitat use among species was observed; however, most species appeared to be microhabitat generalists. Species were grouped into three habitat guilds using corresponding canonical analysis and cluster analysis: fast-flow specialists (FFS), fast-flow generalists (FFG), and slow-flow tolerant (SFT). I used the same data set to develop and test transferability of Habitat Suitability Criteria (HSC) for three habitat guilds and seven species of adult freshwater mussels. Nonparametric tolerance limits were used to define the range of suitable and optimal habitat during summer low flows. Optimal habitat was defined as those ranges of FST hemisphere number, mean column velocity, and embeddedness occupied by the central 50% of independent observations for a species or guild, whereas suitable habitat was defined by those ranges occupied by the central 90% of observations. The transferability of criteria to other reaches of the Clinch River was assessed using one-sided Chi-square tests. Criteria developed for the fast-flow specialist (FFS) and fast-flow generalist (FFG) guilds, as well as most criteria for species in those guilds, transferred to destination reaches. In contrast, criteria developed for the slow-flow tolerant (SFT) guild and individual constituent species consistently failed to transfer. Criteria for FFS and FFG guilds and their constituent species should be incorporated into flow simulation models such as PHABSIM to gauge the effect of minimum flows on mussel habitat quality and quantity. These criteria could also be used to determine suitable sites for mussel translocations. However, my criteria require further testing in other rivers before they can be transferred beyond the Clinch River. Behavior and physiological responses to laboratory manipulations of flow velocity and substrate particle size were used to elucidate microhabitat preferences of Actinonaias pectorosa, Potamilus alatus, and Ptychobranchus subtentum. These species appeared less stressed in the fastest flow treatment, demonstrating significantly higher oxygen consumption and oxygen-to-nitrogen (O:N) ratios than in slower flow treatments. Only P. alatus demonstrated a preference for substrate particle size, and consistently selected finer particle sizes. Actinonaias pectorosa and P. subtentum demonstrated preference for fast-flow microhabitats by readily burrowing in those conditions, while abandoning slow-flow conditions. The lack of preference for substrate particle size demonstrated by A. pectorosa and P. subtentum supports conclusions of previous studies that substrate particle size is of little or secondary importance for explaining mussel microhabitat use. These results, along with previous studies in the Clinch River, demonstrate that the stable habitats of riffles and runs; characterized by fast flows during summer low flows, low percent bedrock, and low embeddedness, are the most suitable habitats for mussel assemblages. To create and maintain suitable habitat conditions in tailwaters, releases should maintain flow over riffles at a minimum depth of no less than 30 cm in riffles that provide higher shear stress conditions (FST number > 7) and velocities (> 0.70 m/s). Periodic releases that are sufficient to transport silt and sand, but not high enough to transport larger substrate should be adequate to maintain substrates with a low degree of embeddedness. Doing so would create suitable habitat for all mussels, from the most to least specialized. Additionally, HSC developed for FFS and FFG guilds can be used to determine suitable destination sites for translocations of species belonging to these guilds. / Master of Science

This dissertation focused in the comprehensive analysis of the capabilities of some non-tested types of Artificial Neural Networks, specifically: the Probabilistic Neural Networks (PNN) and the Multi-Layer Perceptron (MLP) Ensembles. The analysis of the capabilities of these techniques was performed using the native brown trout (Salmo trutta; Linnaeus, 1758), the bermejuela (Achondrostoma arcasii; Robalo, Almada, Levy & Doadrio, 2006) and the redfin barbel (Barbus haasi; Mertens, 1925) as target species. The analyses focused in the predictive capabilities, the interpretability of the models and the effect of the excess of zeros in the training datasets, which for presence-absence models is directly related to the concept of data prevalence (i.e. proportion of presence instances in the training dataset). Finally, the effect of the spatial scale (i.e. micro-scale or microhabitat scale and meso-scale) in the habitat suitability models and consequently in the e-flow assessment was studied in the last chapter. / Esta tesis se centra en el análisis comprensivo de las capacidades de algunos tipos de Red Neuronal Artificial aún no testados: las Redes Neuronales Probabilísticas (PNN) y los Conjuntos de Perceptrones Multicapa (MLP Ensembles). Los análisis sobre las capacidades de estas técnicas se desarrollaron utilizando la trucha común (Salmo trutta; Linnaeus, 1758), la bermejuela (Achondrostoma arcasii; Robalo, Almada, Levy & Doadrio, 2006) y el barbo colirrojo (Barbus haasi; Mertens, 1925) como especies nativas objetivo. Los análisis se centraron en la capacidad de predicción, la interpretabilidad de los modelos y el efecto del exceso de ceros en las bases de datos de entrenamiento, la así llamada prevalencia de los datos (i.e. la proporción de casos de presencia sobre el conjunto total). Finalmente, el efecto de la escala (micro-escala o escala de microhábitat y meso-escala) en los modelos de idoneidad del hábitat y consecuentemente en la evaluación de caudales ambientales se estudió en el último capítulo. / Aquesta tesis se centra en l'anàlisi comprensiu de les capacitats d'alguns tipus de Xarxa Neuronal Artificial que encara no han estat testats: les Xarxes Neuronal Probabilístiques (PNN) i els Conjunts de Perceptrons Multicapa (MLP Ensembles). Les anàlisis sobre les capacitats d'aquestes tècniques es varen desenvolupar emprant la truita comuna (Salmo trutta; Linnaeus, 1758), la madrilla roja (Achondrostoma arcasii; Robalo, Almada, Levy & Doadrio, 2006) i el barb cua-roig (Barbus haasi; Mertens, 1925) com a especies objecte d'estudi. Les anàlisi se centraren en la capacitat predictiva, interpretabilitat dels models i en l'efecte de l'excés de zeros a la base de dades d'entrenament, l'anomenada prevalença de les dades (i.e. la proporció de casos de presència sobre el conjunt total). Finalment, l'efecte de la escala (micro-escala o microhàbitat i meso-escala) en els models d'idoneïtat de l'hàbitat i conseqüentment en l'avaluació de cabals ambientals es va estudiar a l'últim capítol. / Muñoz Mas, R. (2016). Multivariate approaches in species distribution modelling: Application to native fish species in Mediterranean Rivers [Tesis doctoral no publicada]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/76168 / TESIS