The use of on-site wastewater treatment systems for the treatment and dispersal of domestic effluent is common in urban fringe areas which are not serviced by centralised wastewater collection systems. However, due to inappropriate siting and inadequate evaluation of soil characteristics, the failure of these systems has become a common scenario. The current standards and guidelines adopted by many local authorities for assessing suitable site and soil conditions for on-site dispersal areas are coming under increasing scrutiny due to the public health and environmental impacts caused by poorly performing systems, in particular septic tank-soil adsorption systems. In order to achieve sustainable on-site wastewater treatment with minimal impacts on the environment and public health, more appropriate means of assessment of long term performance of on-site dispersal areas are required. The research described in the thesis details the investigations undertaken for the development of robust assessment criteria for on-site dispersal area siting and design and assessment of the long term performance of soil dispersal areas. The research undertaken focused on three key research areas; (i) assessment of site and soil suitability for providing adequate treatment and dispersal of domestic wastewater; (ii) understanding sorption, purification and transport processes influencing retention and release of pollutants and the natural controls governing these processes and (iii) the development of assessment criteria for long term behaviour of soils under effluent dispersal. The research conducted was multidisciplinary in nature, with detailed investigations of the physical and chemical processes involved in on-site wastewater treatment and dispersal. This involved extensive field investigations, sampling and monitoring, laboratory and soil column testing and detailed data analysis across the fields of soil science, groundwater quality, subsurface hydrology, chemical contamination, and contaminant fate and transport processes. The interactions between these different disciplines can be complex which resulted in substantial amounts of data being generated from the numerous field and laboratory investigations and sampling undertaken. In order to understand the complex relationships that can occur, multivariate statistical techniques were utilised. The use of these techniques was extremely beneficial. These techniques not only allowed not only the respective relationships between investigated parameters to be identified, but also adequate decisions based on the correlations were able to be formulated. This allowed a more appropriate assessment of the influential factors, and the prediction of ongoing changes to soil properties due to effluent disposal. The primary outcomes for this research were disseminated through a series of peer reviewed scientific papers centred on these key disciplines. The assessment of site and soil suitability was achieved through extensive soil sampling throughout the study areas and detailed laboratory testing and data analysis. The study identified and investigated the role of influential site and soil characteristics in the treatment performance of subsurface effluent dispersal areas. The extent of effluent travel and the ability of the soil to remove pollutants contained in the effluent by adsorption and/or nutrient uptake were investigated. A framework for assessing the renovation ability of the major soil groups located throughout Southeast Queensland was also developed. The outcomes provide a more rigorous scientific basis for assessing the ability of soil and evaluating site factors to develop more reliable methods for siting effluent dispersal areas. The resulting assessment criteria developed was compared with soil column studies to determine the robustness and validity of the outcomes. This allowed refinement of the assessment criteria in developing a more reliable approach to predicting long term behaviour of soils under sewage effluent dispersal. Multivariate techniques assisted in characterising appropriate soils and to determine their long-term suitability for effluent treatment and dispersal. The assessment criteria developed included physical, chemical and sub-surface hydrological properties of a site and soil which can be used to predict suitability for long term effluent treatment and dispersal. These include:  Moderate to slow drainage (permeability) to assist the movement of effluent (percolation) through the soil profile and allow adequate time for treatment and dispersal to occur. With longer percolation times, the opportunity for exchange and transport processes increase.  Significant soil cation exchange capacity and dominance of exchangeable Ca2+ or exchangeable Mg2+ over exchangeable Na+. Although a soil dominated by Mg2+ is found to promote dispersion of soil particles to some extent, its impact is far less than that of Na+. A stable soil would have a Ca: Mg ratio > 0.5.  Low exchangeable Na+ content to maintain soil stability.  Minimum depth of 400mm of potentially unsaturated soil before encountering a restrictive horizon, to permit adequate purification to take place.  Clay type with Illite and mixed mineralogy soils being the most sensitive to Na+. In general, significant increases in ESP occur in soils with 30 to 40% clay and in the presence of illite clay. Small amounts of smectite clays enhance treatment potential of a soil. The research outcomes have significantly contributed to the knowledge base on best practice in on-site dispersal area siting and design. The developed predictive site and soil suitability assessment criteria allows more appropriate evaluation of site and soil characteristics for providing long term effluent renovation. This is generally not done in the current assessment techniques for on-site dispersal areas. The processes and techniques used in the site and soil suitability assessment, although based on the common soil types typical of South East Queensland, can be implemented in other regions, provided appropriate soil information is collected or available. The predictive assessment criteria have been developed at a generic level, allowing easy implementation into most assessment processes. This gives the framework the flexibility to be developed for other areas specifically targeting the most influential on-site dispersal area siting and design factors, and assessment of long term performance under wastewater application.

Thesis (MScAgric (Viticulture and Oenology))--University of Stellenbosch, 2010. / ENGLISH ABSTRACT: Because of long-term climate changes, apparently associated with higher temperatures and fewer rainfall events, factors such as water-use efficiency and site selection for new cultivars are a matter of increasing importance for viticulture. Within this context, the root system is expected to play a key role. Its relevance to grapevine functioning is due to the numerous functions in which it is involved. In the light of this, the development of the root system is highly relevant to the viticulturist because of the fact that grapevine growth and functioning are dependent on the development of the root system. Differences can, therefore, be expected in terms of berry ripening on single grapevines of the same scion for situations with differing development of root systems, despite being grafted on the same rootstock. Root growth is influenced by several factors, among the ecological aspects. Soil parameters have a predominant influence on root growth and distribution but also annual root production can be altered by canopy manipulation. Due to the importance of root growth to the aboveground development of the vine, it is critical to gain understanding of the relationship between soil factors and root growth and distribution, and the central role that the subterranean environment plays in the concept of terroir. This study aimed to investigate the effect of selected soil physical and chemical parameters on root growth and distribution and to investigate whether having very different canopies influences root growth. In order to achieve these goals, two experiments were conducted; the first was performed in two commercial Sauvignon blanc vineyards each grafted onto Richter 110, non-irrigated, with two treatments: undisturbed lateral growth and complete lateral removal. The second study included the analysis of eight commercial Sauvignon blanc vineyards grafted onto Richter 99 and Richter 110 located in the Stellenbosch Wine of Origin District. Measurements of physical and chemical soil parameters, root growth and distribution, canopy growth and functioning, vine water status and berry composition were performed. The edaphic factors appeared to be one of the most important parameters that affected root development by changing soil water availability and possibly causing physical or chemical limitations on root growth. From the results of this study, it is clear that severe water stress and a pH (KCl) lower than 4.5 play a key role in the limitation of root growth. Due to the fact that most of the soils from the Stellenbosch Wine of Origin District, especially the subsoils, are acidic, this is a factor to consider before planting. On the other hand, the combination of favourable edaphic conditions, such as a subsoil pH of higher than 5.0, light- to mediumtextured subsoil and moderate water stress, allow increased growth of thin roots. However, the effect of canopy management on root growth cannot be discounted due to its importance in the variation of carbohydrate demand by competing sinks. This study showed that lateral removal done from when the berries are at pea size results in an increase in the number of thin roots (0.5-2.0 mm). The secondary leaf area represents at least the same leaf area as the primary leaf area in all the vineyards evaluated, which reveals the relative importance of the laterals in the total leaf area of the vine and the potential importance in terms of microclimate and leaf area available for photosynthesis. Studies of root growth should take the vineyard canopy architecture into account. / AFRIKAANSE OPSOMMING: As gevolg van langtermyn klimaatsveranderinge wat toegeskryf kan word aan die voorkoms van hoër temperature en laer reënval, is faktore soos effektiwiteit van waterverbruik en liggingseleksie vir nuwe kultivars van kardinale belang vir wingerdkunde. Binne hierdie konteks, speel die wortelsisteem ‘n belangrike rol. Die belangrikheid hiervan vir wingerdfunksionering kan toegeskryf word aan die talle funksies waarby dit betrokke is. Die ontwikkeling van die wortelsisteem is dus hoogs relevant vir die wingerdkundige, omdat wingerdgroei en funksionering afhanklik is van die ontwikkeling van die wortelsisteem. Verskille kan daarom dus verwag word in terme van korrelrypwording op ‘n enkele wingerdstok van dieselfde onderstok vir gevalle met verskillende ontwikkeling van die wortelsisteem, ten spyte daarvan dat dit op dieselfde onderstok geënt is. Wortelgroei word, onder ekologiese aspekte, deur verskillende faktore beïnvloed. Grondfaktore het meerendeels ‘n predominante invloed op wortelgroei en -verspreiding, terwyl jaarlikse wortelproduksie deur lowermanipulasie beïnvloed kan word. Weens die belangrikheid van wortelgroei vir die bogrondse ontwikkeling van die wingerd, is dit krities om kennis op te doen oor die verhouding tussen grondfaktore en wortelgroei en –verspreiding, asook die sentrale rol wat die subterreinomgewing op die terroir-konsep speel. Die studie was daarop gemik om die invloed van geselekteerde fisiese en chemiese parameters van grond op wortelgroei en -verspreiding vas te stel, en ook te ondersoek of verskillende lowers wortelgroei sal beïnvloed. Om laasgenoemde doelwitte te bereik, is twee eksperimente uitgevoer. Die eerste is uitgevoer in ‘n kommersïele Sauvignon blanc-wingerd wat geënt is op Richter 110, sonder besproeïng en met twee behandelings, naamlik onversteurde sêkondere lootgroei en volledige sêkondere lootverwydering. Die tweede studie het die analise van agt kommersïele Sauvignon blancwingerde geënt op Richter 99 en Richter 110 in die Stellenbosch Wyn van Oorsprong Distrik. Metings van fisiese en chemiese grondfaktore, wortelgroei en -verspreiding, lowergroei en -funksionering, plantwaterstatus en korrelsamestelling is uitgevoer. Dit blyk dat edafiese faktore een van die belangrikste parameters is wat wortelontwikkeling beïnvloed deur beskikbaarheid van grondwater te verander, en wat moontlik fisiese en chemiese beperkings op wortelgroei kan veroorsaak. Uit die resultate van die studie is dit duidelik dat intense waterspanning en ‘n pH (KCl) laer as 4.5 ‘n belangrike rol in die beperking van wortelgroei speel. Aangesien die meeste van die grondsoorte in die Stellenbosch Wyn van Oorsprong Distrik, veral al die subgronde, suur is, is dit ‘n faktor wat in oorweging geneem moet word voor aanplantings. Die kombinasie van gunstige edafiese toestande, soos ‘n subgrond met ‘n pH hoër as 5.0, ‘n lig tot medium tekstuur en matige waterspanning, sal dus aanleiding gee tot ‘n toename in die groei van dun wortels. Die effek van lowerbestuur op wortelgroei kan egter nie buite rekening gelaat word nie weens die belangrikheid daarvan in die variasie van koolhidraataanvraag deur kompeterende vraagpunte. Hierdie studie toon dat, indien sêkondere lootverwydering tydens ertjiekorrelgrootte toegepas is, dit aanleiding gee tot ‘n toename in die dun wortels (0.5 tot 2.0 mm). Die sêkondere blaaroppervlakte verteenwoordig minstens dieselfde blaaroppervlakte as die primêre blaaroppervlakte in al die wingerde wat ondersoek is, wat dui op die belangrikheid van sêkondere lote in die totale blaaroppervlakte van die wingerd en die potensiële belangrikheid daarvan in terme van mikroklimaat en blaaroppervlakte wat vir fotosintese beskikbaar is. Studies van wortelgroei moet lowerargitektuur in ag neem.

O teor de água no solo é um dos fatores determinantes nos processos de transferência entre o solo e a atmosfera, contribuindo nos balanços de água e de energia. Esse teor é influenciado pelas entradas de água na bacia hidrográfica, por características climáticas, topográficas, de cobertura vegetal, práticas de manejo agrícola e propriedades do solo. A grande heterogeneidade desses fatores faz com que a caracterização desse teor seja ainda um grande desafio. Essa pesquisa objetivou desenvolver abordagens baseadas em Redes Neurais Artificiais (RNAs) para determinação da variação espacial e temporal do teor de água no solo, utilizando informações climáticas, propriedade físicas do solo e variáveis topográficas de uma bacia hidrográfica, com área aproximada de 78 km², localizada na Região Sul do Brasil (bacia do Taboão). A RNA adotada é uma rede de duas camadas, com 25 neurônios na camada intermediária, sendo o treinamento realizado por meio do algoritmo retropropagativo, considerando16 iterações iniciais dos pesos sinápticos, e número máximo de ciclos igual a 30.000. No total foram testadas 40 variáveis de entrada, sendo quatro referentes à topografia (altitude, declividade, distância do ponto ao trecho do rio mais próximo e desnível do ponto ao trecho mais próximo do rio); oito relacionadas ao solo (tipo de solo, densidade do solo, resistência à penetração no solo para as camadas de 0 a 20 cm e 20 a 40 cm, tensão da água no solo em apenas um ponto na bacia e percentual de argila, silte e areia), 10 relativas ao clima (clima, evapotranspiração de referência, temperatura do ar máxima e temperatura do ar, umidade relativa do ar máxima e umidade relativa do ar mínima, pressão atmosférica, radiação solar global, velocidade do vento e temperatura na relva), e 18 variáveis de chuva (chuva de 1, 2, 3, 4, 5, 6 e 12h; chuva de 1, 2, 3, 5, 10, 15, 20, 25 e 30 dias; chuva média ponderada horária; chuva média ponderada diária). A saída dos modelos foi comparada com valores de umidade gravimétrica determinados por amostras coletadas em 26 pontos da bacia, distribuídos espacialmente na bacia, no período compreendido entre 15/01 e 10/08/2013. Neste período o teor de água no solo (umidade gravimétrica) variou entre 13,73 e 33,75%. Os resultados demonstram que é possível estimar o teor de água no solo, com distribuição espacial e temporal, com boa eficiência (NSverificação = 0,77), empregando dados topográficos da bacia, propriedades físicas do solo e dados de chuva. As informações climáticas, por outro lado, não afetam significativamente essa estimativa (NSv=0,28), podendo até diminuir a eficiência do modelo (NSv=0,77 para NSv=0,68). O emprego de muitas variáveis não gera necessariamente o melhor desempenho do modelo, pois uma variável pode mascarar a outra e, até mesmo, interferir a eficiência do modelo (NSv=0,70 e NS=0,61 para os modelos onde foram utilizadas 38 variáveis de entrada), além de aumentar o custo e o tempo para aquisição dessas variáveis, e a dificuldade de interpretação dos resultados em relação às várias entradas. Alternativamente, pode-se estimar o teor de água no solo utilizando modelos mais simplificados que empregam dados de chuva monitorados e informações extraídas de mapas (topografia e tipo de solo), mas o desempenho desses modelos é menor (NSv 0,66). A análise de importância das variáveis de entrada delimitou a tensão da água no solo e a chuva como as variáveis mais influentes nos modelos de melhor desempenho, e a densidade do solo como a menos importante. Nos modelos mais simples, a variável menos relevante é a declividade e a mais importante é a chuva. A análise de sensibilidade demonstrou que nem sempre os modelos conseguem reproduzir o que deveria ocorrer no ambiente natural. / The water content in the soil is one of the determining factors in the transfer processes between the soil and the atmosphere, contributing to the balances of water and energy. This content is influenced by inputs to the basin, climate characteristics, topography, land cover characteristics, agricultural practices, and soil properties. These wide heterogeneity factors make the soil water content characterization still a challenge. This research aimed to develop an Artificial Neural Network (ANN) model to determine the spatial and temporal variation of the water content in the soil, using climate data, physical properties of soil, and topographic variables, of a basin with an area of approximately 78 km2, located in Brazil`s southern region (Taboão basin). The model adopted is a double layer feedforward neural network with 25 neurons in the hidden layer. The learning method is the back propagation algorithm, with 16 interactions to avoid local minima, and the maximum number of cycles chosen was 30,000. A total of 40 input variables were tested, including four of topography (altitude, slope, distance from the point to the nearest stretch of river and unevenness of the point closest to the stretch of the river), eight of soil related variables (soil type, soil density, soil penetration resistance for layers from 0 to 20 cm and from 20 to 40 cm, soil water tension at a single point in the basin and percentage of clay, silt and sand), 10 climate-related variables (climate, evapotranspiration reference, maximum and minimum air temperature, maximum and minimum air relative humidity, atmospheric pressure, global solar radiation, wind speed and temperature on grass) and 18 variables related to rain (accumulated precipitation in 1, 2, 3, 4, 5, 6 e 12h; accumulated precipitation in 2, 3, 5, 10, 15, 20, 25 and 30 days; weighted hourly accumulated precipitation; weighted daily accumulated precipitation). The outputs of the models were compared with values determined by gravimetric moisture samples collected from 26 points spatially distributed in the basin, in the period between 15/01 and 10/08/2013. During this period the soil water content (gravimetric water content) ranged from 13.73 to 33.75%. The results show that it is possible to estimate the water content of the soil, temporal and spatial distribution, with good efficiency (NSverication = 0.77), using topographic data from the basin, soil physical properties and precipitation data. The weather information, on the other hand, did not significantly affect the estimate (NSv = 0.28) and may even decrease the efficiency (NSv) of the model (from 0.77 to 0.68). The use of many variables not necessarily generates the best performance of the model as a variable may mask another and even disrupt the efficiency of the model (NSv = 0.70 and NSv = 0.61, where 38 input variables were used), besides increasing the cost and the time to acquire these variables, and the difficulty of interpreting the results in relation to the various inputs. Alternatively, one can estimate the water content in soil using more simplified models, employing monitored rainfall data and information extracted from maps (topography and soil type), but the performance of these models is smaller (NSv 0.66). The analysis of the importance of input variables delimited the soil water tension and the rain as the most influential variables in the best models, and the density of the soil as the least important. In the simplest models, the less relevant variable is the slope and the most important is the rain. The sensitivity analysis showed that the models cannot always play what should occur in the natural environment.

O teor de água no solo é um dos fatores determinantes nos processos de transferência entre o solo e a atmosfera, contribuindo nos balanços de água e de energia. Esse teor é influenciado pelas entradas de água na bacia hidrográfica, por características climáticas, topográficas, de cobertura vegetal, práticas de manejo agrícola e propriedades do solo. A grande heterogeneidade desses fatores faz com que a caracterização desse teor seja ainda um grande desafio. Essa pesquisa objetivou desenvolver abordagens baseadas em Redes Neurais Artificiais (RNAs) para determinação da variação espacial e temporal do teor de água no solo, utilizando informações climáticas, propriedade físicas do solo e variáveis topográficas de uma bacia hidrográfica, com área aproximada de 78 km², localizada na Região Sul do Brasil (bacia do Taboão). A RNA adotada é uma rede de duas camadas, com 25 neurônios na camada intermediária, sendo o treinamento realizado por meio do algoritmo retropropagativo, considerando16 iterações iniciais dos pesos sinápticos, e número máximo de ciclos igual a 30.000. No total foram testadas 40 variáveis de entrada, sendo quatro referentes à topografia (altitude, declividade, distância do ponto ao trecho do rio mais próximo e desnível do ponto ao trecho mais próximo do rio); oito relacionadas ao solo (tipo de solo, densidade do solo, resistência à penetração no solo para as camadas de 0 a 20 cm e 20 a 40 cm, tensão da água no solo em apenas um ponto na bacia e percentual de argila, silte e areia), 10 relativas ao clima (clima, evapotranspiração de referência, temperatura do ar máxima e temperatura do ar, umidade relativa do ar máxima e umidade relativa do ar mínima, pressão atmosférica, radiação solar global, velocidade do vento e temperatura na relva), e 18 variáveis de chuva (chuva de 1, 2, 3, 4, 5, 6 e 12h; chuva de 1, 2, 3, 5, 10, 15, 20, 25 e 30 dias; chuva média ponderada horária; chuva média ponderada diária). A saída dos modelos foi comparada com valores de umidade gravimétrica determinados por amostras coletadas em 26 pontos da bacia, distribuídos espacialmente na bacia, no período compreendido entre 15/01 e 10/08/2013. Neste período o teor de água no solo (umidade gravimétrica) variou entre 13,73 e 33,75%. Os resultados demonstram que é possível estimar o teor de água no solo, com distribuição espacial e temporal, com boa eficiência (NSverificação = 0,77), empregando dados topográficos da bacia, propriedades físicas do solo e dados de chuva. As informações climáticas, por outro lado, não afetam significativamente essa estimativa (NSv=0,28), podendo até diminuir a eficiência do modelo (NSv=0,77 para NSv=0,68). O emprego de muitas variáveis não gera necessariamente o melhor desempenho do modelo, pois uma variável pode mascarar a outra e, até mesmo, interferir a eficiência do modelo (NSv=0,70 e NS=0,61 para os modelos onde foram utilizadas 38 variáveis de entrada), além de aumentar o custo e o tempo para aquisição dessas variáveis, e a dificuldade de interpretação dos resultados em relação às várias entradas. Alternativamente, pode-se estimar o teor de água no solo utilizando modelos mais simplificados que empregam dados de chuva monitorados e informações extraídas de mapas (topografia e tipo de solo), mas o desempenho desses modelos é menor (NSv 0,66). A análise de importância das variáveis de entrada delimitou a tensão da água no solo e a chuva como as variáveis mais influentes nos modelos de melhor desempenho, e a densidade do solo como a menos importante. Nos modelos mais simples, a variável menos relevante é a declividade e a mais importante é a chuva. A análise de sensibilidade demonstrou que nem sempre os modelos conseguem reproduzir o que deveria ocorrer no ambiente natural. / The water content in the soil is one of the determining factors in the transfer processes between the soil and the atmosphere, contributing to the balances of water and energy. This content is influenced by inputs to the basin, climate characteristics, topography, land cover characteristics, agricultural practices, and soil properties. These wide heterogeneity factors make the soil water content characterization still a challenge. This research aimed to develop an Artificial Neural Network (ANN) model to determine the spatial and temporal variation of the water content in the soil, using climate data, physical properties of soil, and topographic variables, of a basin with an area of approximately 78 km2, located in Brazil`s southern region (Taboão basin). The model adopted is a double layer feedforward neural network with 25 neurons in the hidden layer. The learning method is the back propagation algorithm, with 16 interactions to avoid local minima, and the maximum number of cycles chosen was 30,000. A total of 40 input variables were tested, including four of topography (altitude, slope, distance from the point to the nearest stretch of river and unevenness of the point closest to the stretch of the river), eight of soil related variables (soil type, soil density, soil penetration resistance for layers from 0 to 20 cm and from 20 to 40 cm, soil water tension at a single point in the basin and percentage of clay, silt and sand), 10 climate-related variables (climate, evapotranspiration reference, maximum and minimum air temperature, maximum and minimum air relative humidity, atmospheric pressure, global solar radiation, wind speed and temperature on grass) and 18 variables related to rain (accumulated precipitation in 1, 2, 3, 4, 5, 6 e 12h; accumulated precipitation in 2, 3, 5, 10, 15, 20, 25 and 30 days; weighted hourly accumulated precipitation; weighted daily accumulated precipitation). The outputs of the models were compared with values determined by gravimetric moisture samples collected from 26 points spatially distributed in the basin, in the period between 15/01 and 10/08/2013. During this period the soil water content (gravimetric water content) ranged from 13.73 to 33.75%. The results show that it is possible to estimate the water content of the soil, temporal and spatial distribution, with good efficiency (NSverication = 0.77), using topographic data from the basin, soil physical properties and precipitation data. The weather information, on the other hand, did not significantly affect the estimate (NSv = 0.28) and may even decrease the efficiency (NSv) of the model (from 0.77 to 0.68). The use of many variables not necessarily generates the best performance of the model as a variable may mask another and even disrupt the efficiency of the model (NSv = 0.70 and NSv = 0.61, where 38 input variables were used), besides increasing the cost and the time to acquire these variables, and the difficulty of interpreting the results in relation to the various inputs. Alternatively, one can estimate the water content in soil using more simplified models, employing monitored rainfall data and information extracted from maps (topography and soil type), but the performance of these models is smaller (NSv 0.66). The analysis of the importance of input variables delimited the soil water tension and the rain as the most influential variables in the best models, and the density of the soil as the least important. In the simplest models, the less relevant variable is the slope and the most important is the rain. The sensitivity analysis showed that the models cannot always play what should occur in the natural environment.

O teor de água no solo é um dos fatores determinantes nos processos de transferência entre o solo e a atmosfera, contribuindo nos balanços de água e de energia. Esse teor é influenciado pelas entradas de água na bacia hidrográfica, por características climáticas, topográficas, de cobertura vegetal, práticas de manejo agrícola e propriedades do solo. A grande heterogeneidade desses fatores faz com que a caracterização desse teor seja ainda um grande desafio. Essa pesquisa objetivou desenvolver abordagens baseadas em Redes Neurais Artificiais (RNAs) para determinação da variação espacial e temporal do teor de água no solo, utilizando informações climáticas, propriedade físicas do solo e variáveis topográficas de uma bacia hidrográfica, com área aproximada de 78 km², localizada na Região Sul do Brasil (bacia do Taboão). A RNA adotada é uma rede de duas camadas, com 25 neurônios na camada intermediária, sendo o treinamento realizado por meio do algoritmo retropropagativo, considerando16 iterações iniciais dos pesos sinápticos, e número máximo de ciclos igual a 30.000. No total foram testadas 40 variáveis de entrada, sendo quatro referentes à topografia (altitude, declividade, distância do ponto ao trecho do rio mais próximo e desnível do ponto ao trecho mais próximo do rio); oito relacionadas ao solo (tipo de solo, densidade do solo, resistência à penetração no solo para as camadas de 0 a 20 cm e 20 a 40 cm, tensão da água no solo em apenas um ponto na bacia e percentual de argila, silte e areia), 10 relativas ao clima (clima, evapotranspiração de referência, temperatura do ar máxima e temperatura do ar, umidade relativa do ar máxima e umidade relativa do ar mínima, pressão atmosférica, radiação solar global, velocidade do vento e temperatura na relva), e 18 variáveis de chuva (chuva de 1, 2, 3, 4, 5, 6 e 12h; chuva de 1, 2, 3, 5, 10, 15, 20, 25 e 30 dias; chuva média ponderada horária; chuva média ponderada diária). A saída dos modelos foi comparada com valores de umidade gravimétrica determinados por amostras coletadas em 26 pontos da bacia, distribuídos espacialmente na bacia, no período compreendido entre 15/01 e 10/08/2013. Neste período o teor de água no solo (umidade gravimétrica) variou entre 13,73 e 33,75%. Os resultados demonstram que é possível estimar o teor de água no solo, com distribuição espacial e temporal, com boa eficiência (NSverificação = 0,77), empregando dados topográficos da bacia, propriedades físicas do solo e dados de chuva. As informações climáticas, por outro lado, não afetam significativamente essa estimativa (NSv=0,28), podendo até diminuir a eficiência do modelo (NSv=0,77 para NSv=0,68). O emprego de muitas variáveis não gera necessariamente o melhor desempenho do modelo, pois uma variável pode mascarar a outra e, até mesmo, interferir a eficiência do modelo (NSv=0,70 e NS=0,61 para os modelos onde foram utilizadas 38 variáveis de entrada), além de aumentar o custo e o tempo para aquisição dessas variáveis, e a dificuldade de interpretação dos resultados em relação às várias entradas. Alternativamente, pode-se estimar o teor de água no solo utilizando modelos mais simplificados que empregam dados de chuva monitorados e informações extraídas de mapas (topografia e tipo de solo), mas o desempenho desses modelos é menor (NSv 0,66). A análise de importância das variáveis de entrada delimitou a tensão da água no solo e a chuva como as variáveis mais influentes nos modelos de melhor desempenho, e a densidade do solo como a menos importante. Nos modelos mais simples, a variável menos relevante é a declividade e a mais importante é a chuva. A análise de sensibilidade demonstrou que nem sempre os modelos conseguem reproduzir o que deveria ocorrer no ambiente natural. / The water content in the soil is one of the determining factors in the transfer processes between the soil and the atmosphere, contributing to the balances of water and energy. This content is influenced by inputs to the basin, climate characteristics, topography, land cover characteristics, agricultural practices, and soil properties. These wide heterogeneity factors make the soil water content characterization still a challenge. This research aimed to develop an Artificial Neural Network (ANN) model to determine the spatial and temporal variation of the water content in the soil, using climate data, physical properties of soil, and topographic variables, of a basin with an area of approximately 78 km2, located in Brazil`s southern region (Taboão basin). The model adopted is a double layer feedforward neural network with 25 neurons in the hidden layer. The learning method is the back propagation algorithm, with 16 interactions to avoid local minima, and the maximum number of cycles chosen was 30,000. A total of 40 input variables were tested, including four of topography (altitude, slope, distance from the point to the nearest stretch of river and unevenness of the point closest to the stretch of the river), eight of soil related variables (soil type, soil density, soil penetration resistance for layers from 0 to 20 cm and from 20 to 40 cm, soil water tension at a single point in the basin and percentage of clay, silt and sand), 10 climate-related variables (climate, evapotranspiration reference, maximum and minimum air temperature, maximum and minimum air relative humidity, atmospheric pressure, global solar radiation, wind speed and temperature on grass) and 18 variables related to rain (accumulated precipitation in 1, 2, 3, 4, 5, 6 e 12h; accumulated precipitation in 2, 3, 5, 10, 15, 20, 25 and 30 days; weighted hourly accumulated precipitation; weighted daily accumulated precipitation). The outputs of the models were compared with values determined by gravimetric moisture samples collected from 26 points spatially distributed in the basin, in the period between 15/01 and 10/08/2013. During this period the soil water content (gravimetric water content) ranged from 13.73 to 33.75%. The results show that it is possible to estimate the water content of the soil, temporal and spatial distribution, with good efficiency (NSverication = 0.77), using topographic data from the basin, soil physical properties and precipitation data. The weather information, on the other hand, did not significantly affect the estimate (NSv = 0.28) and may even decrease the efficiency (NSv) of the model (from 0.77 to 0.68). The use of many variables not necessarily generates the best performance of the model as a variable may mask another and even disrupt the efficiency of the model (NSv = 0.70 and NSv = 0.61, where 38 input variables were used), besides increasing the cost and the time to acquire these variables, and the difficulty of interpreting the results in relation to the various inputs. Alternatively, one can estimate the water content in soil using more simplified models, employing monitored rainfall data and information extracted from maps (topography and soil type), but the performance of these models is smaller (NSv 0.66). The analysis of the importance of input variables delimited the soil water tension and the rain as the most influential variables in the best models, and the density of the soil as the least important. In the simplest models, the less relevant variable is the slope and the most important is the rain. The sensitivity analysis showed that the models cannot always play what should occur in the natural environment.

The formation of plastic strains in non-cohesive soils due to large number of loading cycles is a phenomenon of great importance in geotechnical and civil engineering. It constitutes a considerable cause for failures and deformations in various types of engineering applications including pavements. Strain accumulation due to cyclic loading has been studied for years through different models. This thesis reviews various models and focuses on the Bochum model through which, the most contributing soil and traffic parameters on permanent strains formation in pavement subgrades can be figured out. This represents the base for studying the serviceability of increasing the gross weights of vehicles that affect the behavior and size of cyclic loading. This was discussed through investigating the efficacy of increasing the number of vehicle axles and through increasing the vehicle gross weight while keeping the number of axles to check their impacts at the levels of strain formation in soil and consequently on its deformation. The results showed a considerable difference in settlements after changing the axle configurations of vehicles through increasing its number of axles. The work is expected to open a new area of scientific research in pavement designs seeking for ideal configurations of vehicle axles and to provide an advanced approach for studying soil deformations due to higher cyclic loadings.

[...] L’objectif decette thèse est l’étude des mécanismes modulant la biodisponibilité des métaux pour l’escargotCantareus aspersus (syn. Helix aspersa), invertébré vivant à l’interface sol‐plante‐air (maillonintermédiaire, saprophage, phytophage, de chaînes alimentaires). La biodisponibilité est principalementévaluée ici en mesurant l’accumulation (concentrations internes en métaux après 28 jours d’exposition) etles flux d’assimilation. L’influence de paramètres édaphiques sur la disponibilité et la biodisponibilité desmétaux des sols et la contribution des sources de contamination (sols/plantes) des escargots constituentles variables étudiées dans deux conditions d'exposition :[...] L’ensemble des résultats souligne la nécessité de prendre en compte les facteurs et lesmécanismes qui modulent la biodisponibilité des métaux pour modéliser au mieux leur accumulation etleur assimilation par les escargots. Aucune méthode chimique unique d’estimation de la biodisponibilitédes métaux n’ayant pu être déterminée, nous préconisons l’utilisation de mesures biologiques quireflètent mieux la réelle biodisponibilité. / [...] The aim of this thesis is to study the mechanisms that modulate metal bioavailabilityfor the garden snail Cantareus aspersus (= Helix aspersa) a soil invertebrate living at the interfacesoil‐plant‐air (saprophagous and phytophagous intermediate link in the food chain).[...] These data underline the necessity to take into account the factors and mechanisms that modulate themetal bioavailability for snails to better model accumulation and assimilation of metal by snails. As nounique chemical method to assess metal bioavailability was determined, we recommend the use ofbiological measures that identify the real metal bioavailability rather than the use of chemical measures.

No description available.

Les modèles de culture constituent des outils indispensables pour comprendre l’influence des conditions agropédoclimatiques sur le système sol-plante à différentes échelles spatiales et temporelles. A l’échelle locale de la parcelle agricole, le modèle peut être utilisé dans le cadre de l’agriculture de précision pour optimiser les pratiques de fertilisation azotée de façon à maximiser le rendement ou le revenu tout en minimisant le lessivage des nitrates vers la nappe. Cependant, la pertinence de l’utilisation du modèle repose sur la qualité des prédictions réalisées, basée entre autres sur une bonne détermination des paramètres d’entrée du modèle. Dans le cadre de l’agriculture de précision, les paramètres concernant les propriétés des sols sont les plus délicates à connaître en tout point de la parcelle et il existe très peu de cartes de sols permettant de les déterminer de manière précise. Néanmoins, dans ce contexte, on peut disposer d’observations acquises automatiquement sur l’état du système sol-plante, telles que des images de télédétection, les cartes de rendement ou les mesures de résistivité électrique du sol. Il existe alors une alternative intéressante pour estimer les propriétés des sols à l’échelle de la parcelle qui consiste à inverser le modèle de culture à partir de ces observations pour retrouver les valeurs des propriétés des sols. L’objectif de cette thèse consiste (i) dans un premier temps à analyser les performances d’estimation des propriétés des sols par inversion du modèle STICS à partir de différents jeux d’observations sur des cultures de blé et de betterave sucrière, en mettant en oeuvre une méthode bayésienne de type Importance Sampling, (ii) dans un second temps à mesurer l’amélioration des prédictions de variables agro-environnementales réalisées par le modèle à partir des valeurs estimées des paramètres. Nous montrons que l’analyse de sensibilité globale permet de quantifier la quantité d’information contenue dans les jeux d’observations et les performances réalisées en matière d’estimation des paramètres. Ce sont les propriétés liées au fonctionnement hydrique du sol (humidité à la capacité au champ, profondeur de sol, conditions initiales) qui bénéficient globalement de la meilleure performance d’estimation par inversion. La performance d’estimation, évaluée par comparaison avec l’estimation fournie par l’information a priori, dépend fortement du jeu d’observation et est significativement améliorée lorsque les observations sont faites sur une culture de betterave, les conditions climatiques sont sèches ou la profondeur de sol est faible. Les prédictions agro-environnementales, notamment la quantité et la qualité du rendement, peuvent être grandement améliorées lorsque les propriétés du sol sont estimées par inversion, car les variables prédites par le modèle sont également sensibles aux propriétés liées à l’état hydrique du sol. Pour finir, nous montrons dans un travail exploratoire que la prise en compte d’une information sur la structure spatiale des propriétés du sol fournie par les mesures de résistivité électrique, peut permettre d’améliorer l’estimation spatialisée des propriétés du sol. Les observations acquises automatiquement sur le couvert végétal et la résistivité électrique du sol se révèlent être pertinentes pour estimer les propriétés du sol par inversion du modèle et améliorer les prédictions des variables agro-environnementales sur lesquelles reposent les règles de choix des pratiques agricoles / Dynamic crop models are very useful to predict the behavior of crops in their environment and are widely used in a lot of agro-environmental work. These models have many parameters and their spatial application require a good knowledge of these parameters,especially of the soil parameters. These parameters can be estimated from soil analysis at different points but this is very costly and requires a lot of experimental work. Nevertheless,observations on crops provided by new techniques like remote sensing or yield monitoring, is a possibility for estimating soil parameters through the inversion of crop models. In my work, the STICS crop model is studied for the wheat and the sugar beet and it includes more than 200 parameters. After a previous work based on a large experimental database for calibrate parameters related to the characteristics of the crop, I started my study with a global sensitivity analysis of the observed variables (leaf area index LAI and absorbed nitrogen QN provided by remote sensing data, and yield at harvest provided by yield monitoring) to the soil parameters, in order to determine which of them have to be estimated. This study was made in different climatic and agronomic conditions and it reveals that 7 soil parameters (4 related to the water and 3 related to the nitrogen) have a clearly influence on the variance of the observed variables and have to be therefore estimated. For estimating these 7 soil parameters, I chose a Bayesian data assimilation method (because I have prior information on these parameters) named Importance Sampling by using observations, on wheat and sugar beet crop, of LAI and QN at various dates and yield at harvest acquired on different climatic and agronomic conditions. The quality of parameter estimation is then determined by comparing the result of parameter estimation with only prio rinformation and the result with the posterior information provided by the Bayesian data assimilation method. The result of the parameter estimation show that the whole set of parameter has a better quality of estimation when observations on sugar beet are assimilated. At the same time, global sensitivity analysis of the observed variables to the 7 soil parameters have been performed, allowing me to build a criterion based on sensitivity indices (provided by the global sensitivity analysis) able to rank the parameters with respect to their quality of estimate. This criterion constitutes an interesting tool for determining which parameters it is possible to estimate to reduce probably the uncertainties on the predictions. The prediction of the crop behaviour when estimating the soil parameters is then studied. Indeed, the quality of prediction of agro-environmental variables of the STICS crop model (yield, protein of the grain and nitrogen balance at harvest) is determined by comparing the result of the prediction using the prior information on the parameters and the result using the posterior information. As for the estimation of soil parameters, the prediction of the variable is made on different climatic and agronomic conditions. According to the result of parameter estimation, assimilating observations on sugar beet lead to a better quality ofprediction of the variables than observations on wheat. It was also shown that the number ofcrop seasons observed and the number of observations improve the quality of the prediction