Análise de sensibilidade e propagação de incerteza em modelos hidrossedimentológicos: contribuição à modelagem de bacias hidrográficas
Carregando...
Arquivos
Data
2016-10-28
Autores
Pereira, Luiz Henrique [UNESP]
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
Atenuar incertezas nos resultados de modelos dinâmicos que estimam a produção de sedimento na vertente e o transporte do material no canal fluvial torna-se fundamental quando se considera a premência de políticas territoriais em minimizar o risco de sub ou super exploração dos recursos naturais, bem como indicar a disponibilidade de água em bacias hidrográficas. A aplicação de modelos de simulação de processos ambientais tem sido amplamente favorecida pelo avanço das geotecnologias, em especial dos Sistemas de Informações Geográficas, que viabilizam a extração, tratamento, análise e integração de dados geoespaciais. No entanto, observa-se que pouca atenção tem sido dada à análise e avaliação dos fatores responsáveis pela discrepância entre estimativas e observações. Diante o exposto, este trabalho apresentou como principal objetivo caracterizar a variabilidade espacial da incerteza propagada pela aplicação dos modelos hidrossedimentológicos EUPS, MEUPS e REUPS, e indicar sua correlação espacial com características geomorfométricas da área em análise. As atividades foram desenvolvidas com o escopo teórico da modelagem de sistemas ambientais, e baseadas em técnicas de geoprocessamento e sensoriamento remoto. Os resultados obtidos demonstram que a sensibilidade dos parâmetros é específica para cada tipo de bacia modelada, sendo os Fatores C, P os mais sensíveis para bacia do Monjolo Grande, e os Fatores C e LS os mais sensíveis para a bacia do ribeirão Jacutinga. As incertezas possuem maior expressividade nas áreas predominantemente de solos arenosos, e há correlação significativa do grau de incerteza dos resultados dos modelos com as características geomorfológicas, sobretudo em áreas de vertentes côncavas.
Lessening the uncertainties in the results from geospatial dynamic models, considering those that estimate the sediment production in the hillslope and the transportation of sediments over the river-channel, becomes essential when considering the current need to gather trustworthy quantitative information. In this sense, the contribution of hydro-sedimentological modeling is a significant part on the landscape planning phase, effectively composing the process of agricultural land management. The application of geospatial modeling has been widely benefited by the improvement on geotechnologies. However, it´s application has been applied indiscriminately from the procedures and methods for gathering entrance data. Differences in spatial scale of analysis, the characteristics of the geographical area of interest and the evaluation of the trustworthiness of the results are not taken into account. With these issues exposed this paper aimed to characterize the spatial variables and the uncertainties programed by the applications of the hydro-sedimentological models USLE, MUSLE e RUSLE. Indicating it´s spatial correlation with the geomorphometry characteristics of the analyzed areas it was possible to propose an objective criteria for the selection of the models based on the area´s geomorphological characteristics searching to minimize the statistical uncertainties thus offering measurements of trustworthiness in the final results. The activities were developed with the theoretical scope of environmental systems modeling and based on geoprocessing and remote sensing techniques. Results gathered show that the sensitivity of the parameters is specific to each type of watershed that was modeled, C and P factors being the most sensitives for Monjolo Grande river basin (sandy soil), and C and LS factors were the most sensitives for Jacutinga river basin (clay soil). The uncertainties are more prominent in the areas where the soil is predominantly sandy. There was a significant correlation between the level of uncertainty and the results from the models with geomorphological characteristic, especially in concave hillslope areas.
Lessening the uncertainties in the results from geospatial dynamic models, considering those that estimate the sediment production in the hillslope and the transportation of sediments over the river-channel, becomes essential when considering the current need to gather trustworthy quantitative information. In this sense, the contribution of hydro-sedimentological modeling is a significant part on the landscape planning phase, effectively composing the process of agricultural land management. The application of geospatial modeling has been widely benefited by the improvement on geotechnologies. However, it´s application has been applied indiscriminately from the procedures and methods for gathering entrance data. Differences in spatial scale of analysis, the characteristics of the geographical area of interest and the evaluation of the trustworthiness of the results are not taken into account. With these issues exposed this paper aimed to characterize the spatial variables and the uncertainties programed by the applications of the hydro-sedimentological models USLE, MUSLE e RUSLE. Indicating it´s spatial correlation with the geomorphometry characteristics of the analyzed areas it was possible to propose an objective criteria for the selection of the models based on the area´s geomorphological characteristics searching to minimize the statistical uncertainties thus offering measurements of trustworthiness in the final results. The activities were developed with the theoretical scope of environmental systems modeling and based on geoprocessing and remote sensing techniques. Results gathered show that the sensitivity of the parameters is specific to each type of watershed that was modeled, C and P factors being the most sensitives for Monjolo Grande river basin (sandy soil), and C and LS factors were the most sensitives for Jacutinga river basin (clay soil). The uncertainties are more prominent in the areas where the soil is predominantly sandy. There was a significant correlation between the level of uncertainty and the results from the models with geomorphological characteristic, especially in concave hillslope areas.
Descrição
Palavras-chave
Modelagem geoespacial, Geomorfometria, Modelos preditivos, Análise de sensibilidade, Geotecnologia, Geospatial modeling, Geomorphometry, Predictive models, Sensitivity analysis, Geotechnology