La mobilisation et la gestion de l’eau de surface constitue la clé de voûte du bien-être social et économique des populations des zones arides d’Afrique de l’Ouest. Cependant l’utilisation des outils de gestion et de prédétermination hydrologique est entravée par les changements climatiques et environnementaux que connait l’Afrique de l’Ouest depuis les années 1970. Les impacts de ces changements sont assez bien documentés sur les caractéristiques moyennes des régimes hydrologiques mais le sont moins lorsqu’on s’intéresse aux caractéristiques des extrêmes. Cette thèse s’inscrit dans le cadre d’une contribution à un projet d’actualisation des normes hydrologiques en Afrique de l’Ouest, et constitue le premier pas essentiel pour ce projet. Son objectif général est de caractériser les régimes de crues dans la région en répondant aux questions suivantes : Quelles sont les évolutions observées sur les régimes des crues en Afrique de l’Ouest ? Ces évolutions sont-elles en rapport avec les changements climatiques et/ou les modifications environnementales ? Quels sont les scénarii possibles de l’évolution des crues au regard des simulations climatiques futures ? La difficulté majeure de ce travail est liée à la disponibilité des données hydro climatiques et environnementales dans la région, aussi, l’analyse s’est basée sur des outils reconnus robustes dans la littérature. Dans un premier temps, l’utilisation des données de 14 stations hydrométriques de la région... / Water resources plays a key role in the social progress and economic development of west african countries. But the mobilization of water is hampered by climate and environmental changes that undergoes the region since 1970. Extremes parts of hydrological regimes are also impacted, but less studies have focus on their evolution, in relation with climate and environmental changes. The main objective of this thesis is to characterize hydrological extreme events in West Africa, we tried to answer the following questions:- What are the trends of maximum discharge in west Africa?- Are these trends due to climate changes or enviromental changes?- Which speculations can be made from these evolutions, according to future climate simulations?In the first part of the work, we analyzed the trends of floods over 14 watersheds of the region. This analysis allowed us to highlight a clustering behavior of flood according to the climatic region the catchments belong to. Increasing trends have been found on flood magnitude and flood frequency of the 3 sahelian catchments used, and decreasing trends of flood magnitude were found on three sudanian catchments. Finally, the remaining catchments did not showed significant trend in their flood regime...

É fato que a crescente urbanização e a conseqüente impermeabilização do solo são alguns dos principais fatores que interferem no escoamento das águas de chuva. Isso se retrata na aceleração do escoamento, com a diminuição dos tempos de trânsito e aumento dos picos de vazão. É cada vez mais freqüente a ocorrência de inundações em áreas que anteriormente não eram atingidas. A impermeabilização de grandes áreas que possam vir a interferir nas condições hidrológicas de determinada região deve ser analisada pelos órgãos fiscalizadores. O meio técnico e acadêmico se vê frente a frente com o desafio de pesquisar, estudar, analisar e propor técnicas que possam atenuar os efeitos causados pelas chuvas em áreas densamente urbanizadas. Este trabalho intenciona dar uma contribuição à drenagem urbana e vem propor a adoção de uma técnica compensatória que se mostrou muito eficiente. Trata-se do pavimento permeável do tipo sem infiltração no solo. Este dispositivo foi implantado de maneira convencional em um estacionamento na Universidade de São Paulo e contou com o suporte financeiro da Prefeitura do Município de São Paulo. Foram monitorados dois tipos de estrutura com revestimentos de BCP e CPA, durante o período de 6 meses. A instrumentação constou de uma estação pluviométrica e 4 sensores de nível instalados em caixas coletoras dotadas de vertedouros. Durante a pesquisa foi desenvolvido um modelo matemático para validação do modelo físico. Ao final dos estudos concluiu-se que o desempenho do pavimento permeável para os dois tipos de estruturas monitoradas resultou em amortecimento da vazão afluente entre 28% e 87% no BCP e 56% e 85% no CPA. Portanto, esse tipo de dispositivo se mostrou muito eficiente. / The increasing urbanization and the consequent imperviousness of the soil are the main factors that may interfere with the flow of stormwater. It causes the accelerating of the flow with the decrease in transit times and the increase of peak flows. The occurrence of flooding in areas not previously affected is getting more frequently. The imperviousness of large areas that may interfere with the hydrological conditions of a region should be analyzed by the technicians. The technicians and academics find themselves face to face with the challenge of researching, studying, analyzing and proposing techniques that can mitigate the effects caused by rain in densely urbanized areas. This study intends to give a contribution to the urban drainage and is proposing the adoption of a compensatory technique that has proved very efficient. This is the type of permeable pavement without infiltration into the soil. This device was implanted in a conventional way in a parking lot at the University of Sao Paulo with the financial support of the Sao Paulo City Hall. During 6 months two types of permeable pavements covered with permeable interlocking concrete pavement and porous pavement were monitored. The instrumentation consisted of a rain collector and four level sensors installed in collection boxes provided with spillways. During the research it was developed a mathematical model to validate the physical model. At the end of the studies it was concluded that the performance of permeable pavement for the two types of structures monitored resulted in weakening of flow between 28% and 87% in the permeable interlocking concrete pavement and 56% and 85% in the porous pavement. Therefore, this type of device was very effective.

A capacidade adaptativa às enchentes diz respeito à capacidade inerente de indivíduos ou de um sistema de se ajustar aos efeitos desse evento e lidar com ele, de modo a moderar seus danos potenciais. A cidade de São Paulo é particularmente vulnerável às enchentes devido ao seu histórico de uso e ocupação do solo. O objetivo deste trabalho é analisar a capacidade adaptativa a partir da realidade local de moradores do Jardim Pantanal, localizado na zona leste do município de São Paulo às várzeas do rio Tietê, a fim de propor ações que possam contribuir na construção dessa capacidade. A pesquisa foi desenvolvida por meio de levantamento documental e bibliográfico, entrevistas semiestruturadas, análise das transcrições, codificação, e categorização dos dados. As capacidades adaptativas genérica e específica nos níveis organizacionais individual e de sistema são baixas, e entre os determinantes da capacidade adaptativa às enchentes os recursos financeiros, a vulnerabilidade urbana e as estratégias de enfrentamento foram considerados os mais importantes, em nível individual. A falta de recursos, a irregularidade de rendimentos e a ausência de diversificação na fonte de renda limitam as opções disponíveis de moradia em áreas regulares e dificultam a mobilização de recursos para a adoção de medidas preventivas e de recuperação pós-evento. A vulnerabilidade urbana expressa-se pela ocupação em área irregular, onde não são realizados investimentos em medidas de infraestrutura por parte dos moradores, que poderiam reduzir a exposição aos impactos das enchentes, pois não se sabe até quando poderão permanecer na área. As estratégias de enfrentamento demonstram ter caráter apenas reativo sem qualquer planejamento, sendo decididas e tomadas reativamente quando a água sobe. Tendo em vista os aspectos observados, a construção da capacidade adaptativa às enchentes no Jardim Pantanal requer: a) entrosamento entre as medidas de adaptação autônomas (do indivíduo) e as planejadas (do sistema); b) ações de adaptação antecipatórias, mais do que responsivas; e c) medidas de adaptação de curto e longo prazos que considerem as vulnerabilidades que surgiram durante o período de adaptação. / Adaptive capacity to floods means the inherent capacity of individuals or systems to adjust and cope with its effects, in order to moderate potential damages. The city of São Paulo is particularly vulnerable to floods due to its history of occupation and land use. The goal of this research was to analyze the adaptive capacity in the local reality of the Jardim Pantanal residents, located in the east zone of the municipality of São Paulo, at the bed of Tietê River, in order to propose actions that can contribute in building this capacity. The research was developed through bibliographical and documental research, semi-structured interviews, transcript analysis, coding and categorization of data. Generic and specific adaptative capacities in the individual and system organizational levels are low, and among the determinants of adaptive capacity to floods, financial resources, urban vulnerability and coping strategies are considered the most important at the individual level. The lack of resources and the instability and lack of diversification of income sources limit the options available in regular housing areas and make it difficult to mobilize resources for the adoption of preventive and post-event recovery measures. Urban vulnerability is due to illegal land occupation, where investments in infrastructure, which could reduce the exposure to the impacts of floods, are not made by residents, because they dont know for how long they will be able to stay in the area. Coping strategies were found to have only a reactive character, without any planning, being decided and taken reactively when the water rises. Given the observed aspects, building adaptive capacity to floods in Jardim Pantanal requires a) understanding between autonomous (individual) and planned (system) adaptation measures; b) anticipatory rather than responsive adaptation actions; and c) short and long-term adaptation measures to consider the vulnerabilities arising during the adaptation period.

As inundações podem ter um grande impacto na saúde das populações. Doenças produzidas pela exposição direta ou indireta às águas, podem resultar em surtos de doenças gastrointestinais, infecções respiratórias, leptospirose entre outras. Já que os cenários de mudança climática predizem o aumento na frequência e severidade das inundações devido ao aquecimento global, é necessário ter claridade nos impactos gerados sobre a saúde humana como reflexo dos efeitos combinados das mudanças climáticas sobre os ecossistemas, o sistema econômico e o sistema social. A pesquisa visa analisar a vulnerabilidade da população do município de São Paulo às doenças infecciosas relacionadas com inundações. O processo tem como fio condutor o modelo de avaliação da vulnerabilidade conhecido como modelo de vulnerabilidade expandida. Dez variáveis consideradas como representativas para o município de São Paulo, foram escolhidas para a análise da vulnerabilidade social, correlacionadas e analisadas dentro da unidade espacial bacia hidrográfica para o ano 2010, mediante a análise de componentes principais. Bases de dados de precipitação da Rede Telemétrica de Pluviômetros de Superfície do Departamento de Águas e Energia Elétrica (DAAE), foram os dados usados para a avaliação dos níveis de precipitação. Adicionalmente, as estatísticas das bases de dados de Autorização de Internação Hospitalar do Sistema Universal de Saúde (AIH/SUS) foram usadas para a avaliação das variáveis epidemiológicas. Os resultados mostram que a vulnerabilidade social aumenta na direção centro-periferia e que as aglomerações de valores altos, evidentemente mais abundantes do que as aglomerações de valores baixos, aparecem nas bacias de vulnerabilidades sociais significativamente altas. / Floods can have an important impact on people\'s health. Diseases caused by direct or indirect exposure to flood water can result in outbreaks of gastrointestinal diseases, respiratory infections and leptospirosis among others. Since the climate change scenarios predict an increase in the frequency and severity of floods due to global warming, it is necessary to have clarity of the impacts on human health resulting by the combined effects of climate change on ecosystems and the economic and social systems. The study aims to analyze the vulnerability of the city of São Paulos population to infectious diseases related to floods. The process has the vulnerability assessment model known as expanded vulnerability model as its conducting wire. Ten variables considered as representative for the city of São Paulo, are the material for the analysis of social vulnerability. Their correlation and variance analysis is measure within the watershed spatial unit for the year 2010, by the principal component analysis methodology. The databases of precipitation made by the Telemetric Rain Gauges network of the Department of Water and Power (DAAE) are the values used to assess precipitation levels. In addition, the databases of Authorization for Hospitalization of the Brazilian Universal Health System (AIH/SUS) are use too for the evaluation of epidemiological variables. The results show that social vulnerability increases in downtown-periphery direction and that the clusters of high values, clearly more abundant than those of low levels, appear in the watersheds classified with significantly high social vulnerability.

The current standard of simulating flood flow in natural river reaches is based on solving the 1-D or 2-D St. Venant equations or using hybrid 1-D/2-D models based on the same equations. These models are not always able to accurately predict floodwave propagation, especially around and downstream of regions where 3-D effects become important, or at times when the main assumptions associated with these models are violated (e.g. flow becomes pressurized due to presence of a hydraulic structure like a bridge or a culvert). A 3-D modeling approach, though computationally much more expensive, is not subject to such limitations and should be able to predict accurately predict floodwave propagation even in regions where 3-D effects are expected to be significant. This dissertation describes the development and validation of a 3-D time-accurate RANS-based model to study flood-related problems in natural environments. It also discusses how results from these 3-D simulations can be used to better calibrate lower order models. Applications are included where the flow becomes pressurized during high flow conditions and the sediment entrainment potential of the flow during the flooding event is estimated. Another important category of applications discussed in the present study are floodwave propagation induced by a sudden dam break failure. Results show that 2-D models show fairly large differences with 3-D model predictions especially in regions where 3-D effects are expected to be significant (e.g. near channel-floodplain transition, in highly curved channels, near hydraulic structures). The study also discusses the use of the validated 3-D model as an engineering design tool to identify the optimum solution for flood protection measures intended to reduce flooding in the Iowa River near Iowa City. 3-D simulation results are also used to discuss hysteresis effects in the relationship between bed shear stress and the stage/discharge. Such effects need to be taken into consideration to accurately estimate erosion associated with the passage of a floodwave.

This thesis examines the role of rainfall variability and uncertainties on the spatial scaling structure of peak flows using the Whitewater River basin in Kansas, and Iowa River basin in Iowa as illustrations. We illustrate why considering individual hydrographs at the outlet of a basin can lead to misleading interpretations of the effects of rainfall variability. The variability of rainfall is characterized in terms of storm intensity, duration, advection velocity, zero-rain intermittency, variance and spatial correlation structure. We begin with the simple scenario of a basin receiving spatially uniform rainfall of varying intensities and durations, and advection velocities. We then use a realistic space-time rainfall field obtained from a popular rainfall model that can reproduce desired storm variability and spatial structure. We employ a recent formulation of flow velocity for a network of channels and calculate peak flow scaling exponents, which are then compared to the scaling exponent of the channel network width function maxima. The study then investigates the role of hillslope characteristics on the peak flow scaling structure. The basin response at the smaller scales is driven by the rainfall intensities (and spatial variability), while the larger scale response is dominated by the rainfall volume as the river network aggregates the variability at the smaller scales. The results obtained from simulation scenarios can be used to make rigorous interpretations of the peak flow scaling structure obtained from actual space-time model, and actual radar-rainfall events measured by the NEXRAD weather radar network. An ensemble of probable rainfall fields conditioned on the given radar-rainfall field is then generated using a radar-rainfall error model and probable rainfall generator. The statistical structure of ensemble fields is then compared with that of given radar-rainfall field to quantify the impact of radar-rainfall errors on 1) spatial characterization of the rainfall events and 2) scaling structure of the peak flows. The effect of radar-rainfall errors is to introduce spurious correlations in the radar-rainfall fields, particularly at the smaller scales. However, preliminary results indicated that the radar-rainfall errors do not significantly affect the peak flow scaling exponents.

East Iowa experienced large flooding during June of 2008. This study used Cuencas hydrological model to simulate the discharges of June 2008 at Eldorado and Elkader, in the Turkey River Watershed, in North East Iowa. The results of this study were used to test the performance of Cuencas modeling this flood event and to explore the role of land cover change in the floods of 2008 at Elkader, Iowa. Cuencas was found to be a suitable tool to predict this event, that requires relatively low resources. The total time to run each simulation was around two hours which is reasonable for such large watershed (900 mi2), but a computer cluster was needed to run these simulations. The results from this study suggest that the role of land cover change from pre-settlement to current conditions was significant when using the rainfall conditions of 2008. The discharges simulated at Elkader, Iowa were almost twice as large when using the 2001 land cover, than when using the land cover found during 1832-1859, recorded during the General Land Office (GLO) survey. These results need to be taken only as preliminary results, since there is no data to validate the model at the time of the GLO survey, and since it is the first time that Cuencas is used to model the effects of land cover in Iowa's hydrology. However, the potential large reduction on discharge of the pre-settlement land cover is an incentive to investigate this issue further and continue developing Cuencas to capture the effects of less drastic land cover changes.

O objetivo desta pesquisa foi estudar a ocorrência histórica de desastres associados a enchentes/inundações/alagamentos que afetaram o município de Conde, no Litoral Norte da Bahia, no período entre 1966 e 2016, a partir do levantamento dos eventos extremos, os principais fatores deflagradores (condições de contorno) e danos e prejuízos associados, bem como da análise da percepção de risco de desastres por parte dos diferentes atores envolvidos e as estratégias de gestão do risco. A abordagem metodológica adotada foi quali-quantitativa, e contou com as seguintes etapas: (a) cadastro histórico eventos elaborado principalmente pelo método hemerográfico, que permitiu também a identificação das áreas atingidas e dos danos e prejuízos sofridos; (b) identificação das condições de contorno de cada evento - com base em variáveis hidrológicas, climatológicas e oceanográficas coletadas nos bancos de dados públicos disponíveis para a região; (c) mapa de uso e ocupação da terra com a espacialização dos setores de risco alto e muito alto previamente mapeados, em ambiente SIG, para identificação das possíveis causas antrópicas em escala local, da ocorrência dos desastres; (d) análise de percepção de risco de desastres - aplicação de questionários orientados por meio da realização de entrevistas em moradores das áreas de risco muito alto/alto; (e) avaliação da gestão de risco por parte de agentes públicos municipais e estadual - por meio de entrevistas orientadas para essa finalidade. Os resultados indicam que as causas dos desastres em Conde são naturais, de origem local e regional, mas fortemente influenciadas por ocupações inadequadas em áreas de risco, que amplificam a magnitude dos eventos e os danos e prejuízos causados por eles. Estes desastres não parecem seguir um padrão exatamente rítmico, variando sua ocorrência em intervalos entre 2 e 7 anos, porém predominado (9 vezes) os intervalos entre 3 e 4 anos, que deslocam a média geral para ciclos de 4,2 anos. Pela análise da percepção de risco de desastres ficou evidenciado que a memória seletiva ou fraca em relação à baixa frequência e à irregularidade de ocorrência dos eventos, interfere no nível de preocupação e de preparação da população e dos tomadores de decisão quanto ao enfrentamento de futuros desastres. Deste modo, recomenda-se que os gestores municipais melhorem a gestão de risco de desastres, adotando uma postura mais proativa, estabelecendo parcerias técnico-científicas com universidades e centros de pesquisa em desastres, com o objetivo de construir um aparato de monitoramentos ambientais, promover iniciativas educativas e dar suporte à implementação de medidas e ações de adaptação para o enfrentamento de futuros eventos. / The objective of this thesis is to understand the behavior of flood-related flood / flooding / flooding factors that affected the municipality of Conde (Litoral Norte da Bahia) between 1966 and 2016. This study also sought to know and analyze the perception of risks of disasters by the actors involved. Elaborating a register of extreme events, raising their boundary conditions and analyzing how subjects experience and decision makers perceive these events are steps that help the implementation of disaster risk management in municipalities that present hazardous areas, that is, with risk of disaster. In this sense, to reach the objective, it was important to follow the steps, for that the methodological strategy was the qualitativequantitative, based on a case study, based on an integrated approach of physical-natural and social variables. The use of the hemerographic method to survey the history, the analysis of hydrological, climatological and oceanographic variables, the use of GIS for spatialisation of disaster risks, questionnaire application and interviewing, and field empirical research were satisfactory and successful . The results indicate that the Conde events have local and regional boundary conditions and that occupations in high and very high risk areas can increase the magnitude of future events with greater damages and losses. These disasters do not seem to follow an exactly rhythmic pattern, varying their occurrence in intervals between 2 and 7 years, but predominate (9 times) intervals between 3 and 4 years, which shift the general average to cycles of 4.2 years. It was evidenced through the analysis of the perception that, the selective or weak memory in relation to low frequency and irregular occurrence of the events, interferes in the level of concern and preparation of the population and of the decision makers regarding the confrontation of future disasters. Thus, it is recommended that municipal managers adopt a proactive stance, and seek to establish technical-scientific partnerships with Universities and Disaster Monitoring and Prevention Research Centers and Centers (CEMADEN, Geological Survey of Brazil - CPRM, SUDEC / Bahia etc. .), with the objective of building a geotechnical and legal instrument apparatus, promoting educational actions and support for the implementation of measures and actions in the area of civil protection and defense and disaster risk management.

Design flood estimation in small to medium sized ungauged catchments is frequently required in hydrologic analysis and design and is of notable economic significance. For this task Australian Rainfall and Runoff (ARR) 1987, the National Guideline for Design Flow Estimation, recommends the Probabilistic Rational Method (PRM) for general use in South- East Australia. However, there have been recent developments that indicated significant potential to provide more meaningful and accurate design flood estimation in small to medium sized ungauged catchments. These include the L moments based index flood method and a range of quantile regression techniques. This thesis focuses on the quantile regression techniques and compares two methods: ordinary least squares (OLS) and generalised least squares (GLS) based regression techniques. It also makes comparison with the currently recommended Probabilistic Rational Method. The OLS model is used by hydrologists to estimate the parameters of regional hydrological models. However, more recent studies have indicated that the parameter estimates are usually unstable and that the OLS procedure often violates the assumption of homoskedasticity. The GLS based regression procedure accounts for the varying sampling error, correlation between concurrent flows, correlations between the residuals and the fitted quantiles and model error in the regional model, thus one would expect more accurate flood quantile estimation by this method. This thesis uses data from 133 catchments in the state of Victoria to develop prediction equations involving readily obtainable catchment characteristics data. The GLS regression procedure is explored further by carrying out a 4-stage generalised least squares analysis where the development of the prediction equations is based on relating hydrological statistics such as mean flows, standard deviations, skewness and flow quantiles to catchment characteristics. This study also presents the validation of the two techniques by carrying out a split-sample validation on a set of independent test catchments. The PRM is also tested by deriving an updated PRM technique with the new data set and carrying out a split sample validation on the test catchments. The results show that GLS based regression provides more accurate design flood estimates than the OLS regression procedure and the PRM. Based on the average variance of prediction, standard error of estimate, traditional statistics and new statistics, rankings and the median relative error values, the GLS method provided more accurate flood frequency estimates especially for the smaller catchments in the range of 1-300 km2. The predictive ability of the GLS model is also evident in the regression coefficient values when comparing with the OLS method. However, the performance of the PRM method, particularly for the larger catchments appears to be satisfactory as well. / Master of Engineering (Honours)

Using a multi-scale experimental approach, the research undertaken in this thesis investigated the role of the flood pulse in ephemeral floodplain wetlands of the River Murray, in order to better understand the impact of river regulation (and permanent inundation) on these wetlands. An ecosystem-based experiment was conducted on the River Murray floodplain, to compare changes in nutrient availability and phytoplankton productivity in three ephemeral wetlands (over a drying/reflooding cycle) with three permanently inundated wetlands. In the ephemeral wetlands, both drying and re-flooding phases were associated with significant increases in nutrient availability and, in some cases, phytoplankton productivity. It was demonstrated that the ?flood pulse?, as described by the Flood Pulse Concept (FPC), can occur in ephemeral wetlands in dryland river-floodplain systems, although considerable variation in the nature of the pulse existed amongst these wetlands. Results of this experiment suggest that factors such as the degree of drying and length of isolation during the dry phase, the rate of re-filling, timing of re-flooding and the number of drying/re-flooding cycles may be potentially important in producing the variation observed. Permanent inundation of ephemeral wetlands effectively removed these periods of peak nutrient availability and phytoplankton productivity, resulting in continuously low levels (of nutrient availability and phytoplankton productivity). It was concluded that alteration of the natural hydrological cycle in this way can significantly reduce nutrient availability, primary production and secondary production, essentially changing the structure and function, the ecology, of these wetlands. Equally, the results of this experiment indicate that some of the changes resulting from river regulation and permanent inundation can be somewhat reversed, within a relatively short period of time, given re-instatement of a more natural hydrological regime. A mesocosm experiment was used to examine the influence of the dry phase, specifically the effect of the degree of wetland drying, on patterns of nutrient availability and primary productivity comprising the flood pulse. Compared to permanent inundation, re-flooding of completely desiccated sediments increased carbon (C) and nitrogen (N) availability while partial drying generally decreased, or had little effect on, C and N availability after re-flooding. However, degree of drying had little effect on phosphorus availability or rates of primary production measured after re-flooding, and it is possible that these two factors are related. Partial drying reduced rates of community respiration after reflooding, possibly a reflection of the reduced carbon concentrations measured in these mesocosms in this phase of the experiment. Degree of drying also influenced the macrophyte community (measured after three months of flooding), with plant biomass generally decreasing and species diversity increasing as the degree of drying increased (with the exception of complete sediment desiccation which had lasting negative effects on both macrophyte biomass and species diversity). The results of the ecosystem and mesocosm experiments were utilised, in addition to results collected from the same experiment conducted at two smaller scales (minicosms and microcosms), to assess whether the effects of hydrological regime on nutrient availability at the ?wetland? scale could be replicated in smaller-scale experiments. None of the smaller-scaled experiments included in this investigation were able to replicate the specific response to hydrological regime recorded at the ecosystem scale, however the mesocosm experiment did produce results that were more similar to those at the ecosystem scale than those produced by the mini and microcosm experiments. The results of this study indicated that extrapolation of results from small-scale experiments should be undertaken with caution, and confirmed that a multi-scale approach to ecological research is wise, where large-scale field experimentation and/or monitoring provides a check on the accuracy, and hence relevance, of conclusions reached via mesocosm experiments.