Traditionally, hypoxia has been defined as the situation where DO levels have fallen below 2.0 mg O2 L-1, but increasing evidence suggests that this low level of DO is inadequate to describe the onset of hypoxia impacts for many organisms. Consequently, there is a need for a greater understanding of how ‘moderate’ alterations in DO levels will affect ecosystem processes and functionality, specifically through behavioural and physiological alterations at the organism and community level. This thesis reports on mesocosm experiments which were conducted to examine the effects of moderate ( > 3.0 mg O2 L-1) hypoxia on firstly, a key ecosystem engineer, the brittlestar Amphiura filiformis, and secondly, on the Station L4 infaunal macrobenthic community. Station L4 is a longstanding marine biodiversity and MSFD reference site and forms part of the Western Channel Observatory. At the organism level, short-term (14 d) exposure to moderate hypoxia significantly reduced oxygen uptake rates, oocyte diameter and oocyte development in A. filiformis. However, these physiological affects occurred irrespective of brittlestar population density. Additionally, moderate hypoxia reduced brittlestar activity, in terms of bioturbation behaviour, consequentially having an effect on ammonium and silicate fluxes. These observations were only detected when brittlestar population density was high. It was concluded that denser populations of A. filiformis may therefore exhibit the greatest changes in behaviour and shifts in ecosystem function as competition for resources and oxygen heightens. The benthic community at Station L4, displayed considerable tolerance to medium-term (6 wk.) exposure to moderate hypoxia, in terms of structure, diversity and bioturbatory behaviour, but these results may be different if exposure was longer or more severe. Alterations in nutrient fluxes were detected, but there was little evidence to suggest these changes were due to macrofaunal behavioural alterations. Additionally, results from this study revealed that bringing complex natural communities into the mesocosm caused a substantial loss of individuals and species, mainly due to translocation and disturbance effects. This important insight into the effects of bringing community assemblages into the mesocosm confirms that even with a loss of diversity, the L4 community maintained functionality and was resilient to alterations in DO. This suggests that the L4 benthic community does not depend on any one specific species for the provision of important ecosystem processes, resulting in considerable functional resilience within the L4 system. However, vulnerability to benthic systems may increase if functionality is dominated by species such as A. filiformis. Consequently, moderate hypoxia may not immediately affect benthic communities in terms of structure and diversity, but the physiological effects on individuals, especially to reproductive development, may cause alterations in the quality and quantity of planktonic propagules supplied by benthic species to the pelagic environment. This could affect benthic community diversity and functionality in the long term if repeated hypoxic events occur.

Ce travail de thèse s'inscrit dans l'effort actuel de construction d'une écologie intégrative. J'y étudie les mécanismes d'interaction entre ressources abiotiques, réseaux trophiques et propriétés des écosystèmes, au moyen d'une expérience d'évolution, d'un modèle de méta-écosystème et d'un modèle bioénergétique d'assemblage d'écosystèmes. Les organismes modifient la disponibilité des ressources en les prélevant pour leur croissance. Inversement, la disponibilité des ressources influence la diversité et la composition en espèces du réseau trophique, en agissant comme force de sélection sur les traits d'acquisition des ressources (chap. 1, 5). Les propriétés de l'écosystème, telles que stabilité et productivité, dérivent des interactions entre la dynamique des ressources et celle du réseau trophique (chap. 2). Enfin, le fonctionnement de l'écosystème rétroagit sur les ressources abiotiques via le recyclage de la biomasse (chap. 2, 5). Ces processus interviennent lors de l'assemblage des réseaux trophiques et structurent le développement des écosystèmes (chap. 3-5). Dans cette thèse j'analyse ces mécanismes de rétroaction biotique-abiotique sur plusieurs échelles d'organisation, d'espace et de temps. Notamment, les modèles développés ici fournissent des outils novateurs pour étudier les mécanismes de construction des écosystèmes, en mettant en évidence les liens entre métabolisme des espèces, structure du réseau trophique et fonctionnement de l'écosystème, et leur variation au cours du temps. Ce travail ouvre de vastes perspectives de recherche en combinant les derniers progrès d'une écologie intégrative dans une conception mécaniste du développement des écosystèmes. / This thesis participates to the current effort towards the construction of an integrative ecology. I study the feedback mechanisms between abiotic resources, food webs and ecosystem properties, through an evolution experiment, a model of metaecosystem, and a bioenergetic ecosystem assembly model. Organisms modify resource availability by consuming them for their growth. Conversely, resource availability influences the species diversity and composition of the food web, by acting as a selection pressure on traits for resource acquisition (chap. 1, 5). Ecosystem properties, such as stability and productivity, derive from the interactions between resource and food web dynamics (chap. 2). Finally, ecosystem functioning feeds back on abiotic resources through the recycling of biomass (chap. 2 and 5). These processes occur during the food web assembly and drive the development of ecosystems (chap. 3-5). In this thesis I analyze these biotic-abiotic feedback mechanisms on several scales of organization, space and time. The models developed here provide innovative tools to study the mechanisms of ecosystem construction by pointing out the links between species metabolism, food web structure and ecosystem functioning, and their variation through time. This work opens wide research perspectives, as it combines the most recent progress of an integrative ecology into a mechanistic framework of ecosystem development.

Le drainage minier acide de Carnoulès est caractérisé par un pH très acide et une forte concentration en métaux et en arsenic. Par différentes approches moléculaires, des études précédentes ont montré une faible biodiversité et ont permis d'établir un modèle de fonctionnement de la communauté bactérienne. Le but de ce travail de thèse a été de préciser la composition et le fonctionnement de cette communauté bactérienne, en utilisant pour cela des approches culturales et fonctionnelles, en se focalisant particulièrement sur le recyclage de la matière organique. L'élaboration de différents milieux a permis l'isolement de 49 souches bactériennes appartenant à 19 genres, augmentant ainsi de 10 % la diversité bactérienne détectée à Carnoulès par rapport aux approches métagénomiques précédentes. Parmi les 19 genres, 3 sont nouveaux dont un, inféodé aux écosystèmes acides, a été caractérisé taxonomiquement et dénommé Acidiminas carnoulesii. La capacité de l'isolat Q8 appartenant au genre Paenibacillus à dégrader I'amidon et la xylane, dans de larges gammes de pH et de concentrations en arsenic, a permis d'attribuer à Paenibacillus un rôle dans la résilience de la communauté pour ces fonctions. Un criblage fonctionnel de I'ADN de Q8 dans Escherichia coli apermis d'isoler les gènes codant les protéines de dégradation de ces polymères. Par ailleurs, un criblage de 80000 clones de la banque d'ADN métagénomique de Carnoulès a permis la détection de 28 clones positifs pour l'activité amylolytique. Deux protéines ne présentant aucune similarité de séquence avec des amylases connues ont été caractérisées in vitro, confirmant leurs activités amylolyiques et démontrant que la bioprospection dans des sites a priori incongrus, autorise des découvertes insoupçonnées. Ces travaux ont ainsi permis de montrer que les approches culturales et fonctionnelles apportent des informations nouvelles par rapport à celles obtenues par les approches moléculaires. La complémentarité de ces approches est vérifiée, et elle apparaît indispensable dans l'analyse de la complexité des écosystèmes. Cependant, la compréhension de leur fonctionnement exigera des efforts redoublés. / The Carnoulès acid mine drainage is characterized by an acidic pH and high metal and arsenic concentrations. Based on the low bacterial biodiversity, molecular approaches allowed the determination of a bacterial community functioning model. The aim of the PhD work was to clariff both the composition andthe functioning of this community, using cultural and functional approaches, focusing on the organic matter recycling. Different media were designed that allowed the isolation of 49 bacterial strains belonging to 19 genera,leading to a 70 o  increase in the bacterial diversity compared with previous metagenomic approaches. Among the 19 genera,3 are new, one of which, a previously uncultured genus frequently detected in acidic environments, has been taxonomically characterized and named Acidiminos cornoulesii.The ability of strain Q8, belonging to the genus Paenibacillus, to degrade starch and xylan over a wide rangeof pH values as well as 4rsenic concentrations allows to assign to Paenibacillus a role in the resilience of the community for these functions. A function-based screening of the Q8-DNA in Escherichia coli led to isolatethe genes encoding the polymer-degrading proteins.Moreover, a function-based screening of 80,000 clones from a Carnoulès metagenomic DNA library led to the detection of 28 positive clones for the amylolyic activity. Two proteins sharing no sequence similarity with known amylases were characterized in vitro, demonstrating that bioprospecting in a priori incongruous sites can lead to unsuspected discoveries.These works show that new informations using cultural and functional approaches can be obtained compared to those gained with molecular approaches. Results confirmed the complementarity between both approachesis crucial to analyze complex ecosystems. Nevertheless, more research efforts still have to be undertaken tounderstand their functioning.

The Kruger National Park (KNP) management has recently made considerable progress in enacting new policies to address biodiversity conservation and management challenges. However, the interlinkages among the new policies to support biodiversity conservation and management have not been evaluated, particularly the principles and approaches, and how these interact in terms of policy decision-making at the economic and political levels. This study seeks to evaluate the interlinkages among the new policies and how they affect one another, as an avenue to promote a more integrated and comprehensive policy implementation. These policies include the tourism policy, the elephant management policy, and the water provision policy. At policy-making level, the integration of these three policies enhances the possibilities for balancing and controlling the pressures exerted on environmental resources. At the implementation level, it offers a framework for the coordination of the interventions of the policies in space so as to recognise and capitalize on their synergies. Given the inevitably multifarious and departmentalized nature of policy formulation in the KNP, there is a need for the analysis of interlinkages amongst policies. The conceptual framework underpinning this study derives much from Briassoulis’ (2004) policy integration. The research utilizes both qualitative and quantitative research methods, and focuses on selected camp sites within the Park. The findings indicate that current procedures for tourism policy-making and implementation are weak, and little is done by way of impact assessment. This has been attributed to the lack of capacity at the KNP. The findings also reconfirm that policy-making is too fragmented. As a result, policy coordination and cooperation among park managers is weak. This study suggests that an environmental policy integration approach can lead to improved policy-making and implementation. Informed by the data collected from interviews, questionnaires and document analyses, a management framework has been developed to demonstrate how an integrated approach to Environmental Policy Integration (EPI) or management can help sustain the practice of wildlife tourism and support biodiversity conservation. It is concluded that greater realisation of integrated policy-making and implementation in the KNP can be achieved by establishing a formal coordinating office. However, a special feature of KNP policy-making is its widespread consultation system that can provide a fertile ground for enhancing EPI.

My thesis is focused on valuation methods intensively used agricultural landscapes according to the importance of near-natural habitats in terms of their natural values and ecosystem services. This is done at four sites using pitfall traps, where there was a catch beetles (Carabidae). To evaluate the data I used indicators such as the Shannon index serenity. Procedures for calculating the ecological value of the area and the ecosystem services provided I processed according to the proposed method Seják et al., (2010). Rating I performed by BVM (,, habitat valuation methods "). Replacement cost method I used for the calculation of ecosystem services.

This thesis deals with the phenomenon of ecological recovery of ecotone backyard, a specific phenomenon in rural areas of the Czech Republic, focusing on the one hand synchronously, partly diachronic. Theoretically, based on the scientific literature discusses the topic first as a phenomenon in itself, then focuses on the description of the possibilities that may be related to the maintenance or revitalization of the ecotone recover or regain, both in the dimension of eco-environmentalist, and in the dimension of socio-ethnographic.

Signal crayfish (Pacifastacus leniusculus) is important invasive species in European freshwaters. Its influence on other freshwater organisms is well known from the literature, as well as direct and indirect impact on fish assemblages. This work was focused on the experimental evaluation of non-indigenous signal crayfish as a predator of salmonids compared to indigenous noble crayfish (Astacus astacus). Moreover, the possible importance of young-of-the-year signal crayfish as a prey for salmonids was assessed. There were carried out experiments using eggs and hatchings of grayling (Thymallus thymallus) as a prey for adult and subadult specimens of both, signal and noble crayfish. Next experiment used young-of-the-year signal crayfish as a prey for young-of-the-year of brown trout (Salmo trutta). Results showed that the danger of signal crayfish for grayling eggs is slightly higher. However, detected differences were surprisingly lower and in the majority of parameters even insignificant. In the case of grayling hatchings, was not detected any significant difference at all. The abilities of both tested species to prey on eggs and hatchings are therefore very similar, comparable. Even so, the effect of signal crayfish can be importantly higher in natural conditions because of its more dense populations, higher growth rate and fast maturation. In accordance to our findings, it is evident that brown trout has no so high effect of crayfish juveniles compared with crayfish effect on salmonids early developmental stages. Crayfish are therefore more likely able to limit salmonids than conversely.

During this PhD. thesis, the importance of methane-oxidizing bacteria (methanotrophs) and their ecological demands were studied on the longitudinal transect along an important European river - the River Elbe. However, it was necessary to adjust methodologies for precise measurements of methane oxidation in such a variable aquatic environment. Based on laboratory experiments and field measurements, several key methodological recommendations for future planning of methane oxidation rate estimations in an unknown environment have been identified or specified. In line with the variability of the river habitats, considerable heterogeneity was also found in the obtained data on methane concentration and methanotrophical activity. Probably, some of the most important information gathered during many field sampling campaigns is that sites with the highest methane concentration usually showed a very low activity of methanotrophic bacteria (resulting in higher methane emissions). These sites are predominantly human modified sections of the river, such as locks, weirs, harbors and canals. On the contrary, the free-flowing parts of the river, modified only by groynes, showed low level of methane concentration. And so groynes could represent a more effective solution and "natural-close" habitats of navigability of rivers.

O conceito de ‘ecossistema’ emergiu da necessidade de compreender o caráter extremamente dinâmico da vegetação, interpretado a partir daí como o resultado da interação recíproca entre um dado complexo de organismos e seu conjunto amplo de fatores do ambiente físico. Um ramo das ciências ecológicas desenvolveu-se desse conceito, visando examinar o resultado de tais interações em termos de fluxos de energia, matéria e informação. Desenvolvimentos conceituais recentes apontam para uma concepção do ecossistema sob a ótica de um novo paradigma, para o qual aninhamento, hierarquia, decomposabilidade relativa, probabilidade e dependência de escala são critérios chave. Outro desenvolvimento importante, a análise de trajetórias, abriu a possibilidade de tratar a dinâmica e o funcionamento do ecossistema como fenômenos em múltiplas escalas. Incertezas metodológicas e ecológicas decorrem numa visão pouco nítida de como o funcionamento e a estrutura do ecossistema interagem sob a influência de um determinado conjunto de fatores de uso e do ambiente físico. A situação demanda uma abordagem analítica na qual classificações funcionais e estruturais sejam implementadas independentemente, com o fim de estabelecer ‘a posteriori’ quanto e como as classificações estão interconectadas. A tarefa é ainda mais desafiante, em termos de método e interpretação, quando consideramos o contexto hierárquico e complexo em que a análise deve ser feita e a dependência de definição dos resultados. Esta tese refere-se ao desenvolvimento de ferramentas conceituais e metodológicas para analisar a heterogeneidade funcional dos ecossistemas no espaço, em relação a fatores significativos de uso e do ambiente, e aos diferentes tipos de vegetação presentes numa determinada região. Com esse objetivo, adotamos o conceito de ‘Tipos Funcionais de Ecossistemas’ (TFEs), os quais reúnem unidades espaciais com padrão de funcionamento similar, sem considerar seus atributos estruturais, e avançamos num esquema classificatório de TFEs que permite capturar as respostas funcionais de curto prazo dos ecossistemas em cenários de mudanças ambientais e de uso altamente dinâmicas. Também examinamos a sensibilidade dos tipos funcionais de ecossistemas a diferentes definições de funcionamento e parâmetros de escala espacial. Os TFEs provaram ser sensíveis a estas variáveis analíticas, oferecendo assim a possibilidade de indagar a natureza multidimensional e multi-escala dos fenômenos do ecossistema. Os TFEs capturam eficientemente os aspectos mais relevantes da resposta sazonal da vegetação aos fatores do ambiente biofísico, provendo assim uma ferramenta útil para descrever a heterogeneidade espacial do funcionamento dos ecossistemas em domínios temporais e geográficos específicos. Nesta tese avançamos no reconhecimento e descrição dos principais tipos de paisagem no planalto basáltico do Rio Grande do Sul, e propomos mecanismos e controles responsáveis desses padrões característicos. Da associação espacial entre feições do terreno, solos, tipos de uso e vegetação, identificamos três tipos básicos de paisagens e definimos preliminarmente seu domínio espacial. Os resultados descrevem um forte relacionamento entre a distribuição dos grandes tipos fisionômicos de vegetação, os solos e os processos formadores de relevo. Assim sendo, os campos dominam onde relevo e solos indicam a ocorrência de remanescentes de uma antiga superfície de pediplanação, em quanto as florestas prevalecem onde os agentes geomorfológicos têm rejuvenescido a paisagem. Porém, com o objetivo de compreender os processos responsáveis destes padrões, é essencial fazer ‘downscaling’ desde a escala regional na qual os processos formadores de relevo e de solos dominam a diferenciação espacial de variáveis ecológicas, até a escala local na qual fatores biológicos e relacionados com o regime de distúrbio adquirem maior importância na produção de padrões de heterogeneidade espacial. Identificamos que a abordagem ecossistêmica funcional é a maneira mais promissora de relacionar processos de natureza tão divergente. / The ‘ecosystem’ concept emerged from the need for understanding the highly dynamic nature of the vegetation, interpreted from thereon as the reciprocal interaction among the organism-complex and a wide array of factors of the physical environment. A full branch of the ecological sciences developed from this concept, aimed to assessing the outcome of such interactions as flows of energy, matter and information. Recent conceptual developments points to a conception of ecosystem as an entity evolving under the influence of a novel paradigm, for which nestedness, hierarchy, relative decomposability, probability and scale-dependency are central. Another important development, trajectory analysis, opens the possibility to treat ecosystem dynamics and ecosystem functioning as multi-scale phenomena. Methodological and ecological uncertainties determine a rather fuzzy picture of how ecosystem function and structure interplay under the influence of some set of drivers of the physical environment and land use. The whole situation waits for an analytical path to be designed in which functional and structural classifications are carried out independently, in order to establish a posteriori whether they are connected and how they are connected. The task is even more defiant, both in terms of methods and interpretation, if we consider the already complex hierarchical context in which the analysis should be set and the definition-dependency of the outcome. This thesis is about the development of conceptual and analytical tools for analyzing the functional heterogeneity of the ecosystems in the space, in relation to meaningful environmental and land-use factors and to the different types of vegetation present over a given region. To that aim, we adopt the concept of Ecosystem Functional Types (EFTs), which enclose spatial units with similar functional patterns, no attention paid to their structure, and advance on an EFT classificatory scheme that allows capturing the short-term functional response of the ecosystems to environmental and land-use changes. Furthermore, we examine the effect of using different surrogates of ecosystem functioning on the resulting picture of functional patchiness. The effect of changing parameters of spatial scale is also tested. The Ecosystem Functional Types proved to be heavily definition-dependent and sensitive to spatial scale, which allows exploring the multi-dimensional and multi-scale nature of ecosystem phenomena. The EFTs efficiently capture the most relevant features of the seasonal response of the vegetation to the drivers of the biophysical environment, providing so a useful tool for depicting the spatial heterogeneity of ecosystem functioning in a given geographic and temporal domain. In this report we also accomplished the recognition and description of main landscape types in the basaltic tablelands of Rio Grande do Sul, and proposed mechanisms and controls responsible for their characteristic patterns. From the spatial association of terrain features, soils, land-use and vegetation, we identified three basic landscape types and broadly defined their spatial domain. The picture described tells of a rather close relationship among the distribution of the major physiognomic types of the vegetation, soils, land-use and land-forming processes. In this picture, the grasslands prevail where terrain and soil features suggest there are the remnants of an old pediplanation surface, while forests seems to dominate wherever geomorphic agents have rejuvenated the landscape. However, in order to understand the processes responsible of these patterns it is then essential to downscale from the regional realm where terrain and soil-forming phenomena dominate spatial differentiation, to the fine-scale processes at which biological and disturbance-related factors are most influential in the production of patterns of spatial heterogeneity. We identify the functional approach to the ecosystems as the most promising way to correlate processes of such a different nature.

Made available in DSpace on 2015-03-26T13:50:12Z (GMT). No. of bitstreams: 1 texto completo.pdf: 2283755 bytes, checksum: 7b17f9a16dc883593adc1fe443b6a654 (MD5) Previous issue date: 2013-03-20 / Conselho Nacional de Desenvolvimento Científico e Tecnológico / The performance of LSEMs (Land surface and ecosystem models) depends on the parameters of the equations representing the simulated process. However, the measurement of some parameters can be impractical or even impossible; therefore, they need to be estimated, or preferably optimized specifically for each ecosystem. When the parameters are calibrated to a single variable (mono-objective problem) they may not represent the reality, because the complexity of the model and its dependence on several variables (multi-objective problem). Thus, simultaneous multi-objective optimizations are indispensable. However, the optimization performance decreases as the number of variables to be optimized simultaneously increases. Furthermore, the study of simultaneous optimization using more than three objectives is a new area and not yet sufficiently studied. For simultaneous optimization of a large number of variables, there is a method that uses concepts of hierarchical systems theory in which the optimization occurs from the fastest (radiative fluxes) to the slowest process (carbon allocation). This study evaluates the performance of the hierarchical optimization using the index D (the average of the ratios between the individual outputs of multi-objective optimization and monoobjective). Understanding how the performance index D varies with respect to the number of objective functions optimized and to the number of hierarchical levels is important for the development of this research area. Two steps are necessary to achieve the study goals. First, a sensitivity analysis was performed to determine the output variables sensitivity to the model parameters. After, simulations were made using all possible combinations among the seven micrometeorological variables available (PARo, fAPAR, Rn, u *, H, LE, NEE) taking into account the hierarchy of processes. The results indicate that for up to three objective functions, hierarchical multi-objective optimization generates better results than the simultaneous multiobjective optimization (one hierarchical level), provided that the parameters distribution among hierarchical levels is consistent with the sensitivity analysis. Another important result shows that for the same number of outputs optimized, the greater the number of hierarchical levels the better the performance of the optimized model. However, the model performance falls quickly as the number of objective functions increases, evidencing that the power of hierarchy calibration that use a high number of objective functions is highly dependent on the removal of some constraints for model s performance. / O desempenho de um LSEM (Modelo de superfície terrestre e ecossistema) depende dos parâmetros das equações que representam os processos simulados. Contudo, a mensuração de alguns destes parâmetros pode ser impraticável ou até mesmo impossível; por isso, necessitam ser estimados ou, preferencialmente, otimizados para cada ecossistema. Quando os parâmetros são calibrados para uma única variável (problema mono-objetivo) eles podem não representar bem a realidade, dado a complexidade do modelo e sua dependência de diversas variáveis (problema multiobjetivo). Por isso, há a necessidade de uma otimização simultânea multiobjetiva. Porém, o desempenho da otimização diminui com o aumento do número de variáveis otimizadas simultaneamente e, além disso, o estudo da otimização simultânea de mais de três objetivos é uma área relativamente nova e não suficientemente estudada. Para a otimização simultânea de um grande número de variáveis, existe uma metodologia na qual se utiliza conceitos de teoria hierárquica de sistemas em que a otimização ocorre dos processos mais rápidos (fluxos radiativos) para os mais lentos (alocação de carbono). Este trabalho avalia o desempenho da otimização hierárquica do modelo, utilizando o índice D (a média das razões individuais entre as saídas das otimizações multiobjetiva e monoobjetiva). Entender como o índice de desempenho D do algoritmo de otimização hierárquico varia em relação ao número de funções objetivo otimizadas é de extrema importância para o desenvolvimento desta área de pesquisa. Para fazer atingir os objetivos, foram necessárias duas etapas. Primeiramente, foi feita uma análise de sensibilidade, a fim de conhecer a sensibilidade das variáveis de saída aos parâmetros do modelo. Depois, foram feitas simulações com todas as combinações possíveis entre as sete variáveis micrometeorológicas disponíveis (PARo, fAPAR, Rn, u*, H, LE, NEE) levando em consideração a hierarquia dos processos. Os resultados encontrados indicam que, para até três funções objetivo, a otimização multiobjetiva hierárquica pode gerar resultados melhores do que a otimização multiobjetiva tradicional (um único nível hierárquico), desde que a distribuição dos parâmetros entre as variáveis seja feita de forma coerente com a análise de sensibilidade. Outro resultado importante revela que para um mesmo número de saídas otimizadas, quanto maior o número de níveis hierárquicos melhor o desempenho do modelo otimizado. Porém, o desempenho do modelo diminui rapidamente quando o número de funções objetivo aumenta, evidenciando que o poder da calibração hierárquica para o uso de um grande número de funções objetivo é altamente dependente de algumas restrições que o modelo possui e um alto desempenho do modelo para muitas funções objetivo será possível somente após a remoção delas.