Ingeniería de sistemas aplicada al acuífero kárstico de Fuenmayor (San Julián de Banzo, Huesca) y al glaciar Collins (isla King George, Antártida)

Resumen

System engineering, defined broadly as the art and science of creating whole solutions to complex problems, is applied herein to two natural systems, viz, a karst aquifer system and a glacier system, from an exclusively hydrological perspective. Identification techniques, developed to represent typical engineering artificial systems through linear and nonlinear models, can be applied in the study of natural systems with coupling effects between climate and hydrosphere. Methods evolve to meet new identification fields where strategies are needed to find the right model adapted to the peculiarities of the system. In this regard, special tools, as those based on wavelet transform, are been considered in preparing time series such as smoothing of signals, spectral analysis, coherence levels and outliers detection, among others. Under this approach, an application to be mentioned, among those treated in this thesis, is the analytical determination of seasonal effective core (SEC) through the study of wavelet coherence between air temperature and the discharge of the glacier, which provides a set of acceptably consistent sampling periods, from which glacial system models are defined. The study is specifically aimed to estimate the mutual influence between precipitation and discharge of Fuenmayor karst aquifer, in San Julian de Banzo, Huesca, Spain. In the same way, it deals with the implications of the air temperature in the glacier melting, manifested in the stream drainage of Collins glacier, King George island, Antarctica. These parametric and non-parametric identification processes try to find models that best represent the internal dynamics of the system by iterative testing, where models are systematically verified versus sampling data, with an efficiency criterion given. The best obtained solution, in the assessment of results drawn from dealt cases, is found among the model structures in blocks. This thesis means a formal statement of system identification methodology of engineering in the context of natural systems, which improves the results in many cases of karst hydrology that often use occasional \textit{ad hoc} methods of statistics. Likewise, approaches proposed in cases of glaciology with wavelet analysis and data-driven models scarcely appear in the literature, and reveal essential information when it is not possible to clarify the whole of physics governing the system. Remarkable results are stemmed from the characterization of Fuenmayor spring response and its correlation with precipitation, under the assumption of a linear system to be complemented with identification methods based on non-linear techniques. On the other hand, the model implementation for Collin glacier, obtained by black-box identification, can reveal certain instability at the boundaries of active periods in the discharge, and consequently, the current trend variability of the global climate change.