<p>Ecohydrological models are extensively used to evaluate land use, land management and climate change impacts on hydrology and in-stream water quality conditions. The scale at which these models operate influences the complexity of processes incorporated within the models. For instance, a large scale hydrological model such as Soil and Water Assessment Tool (SWAT) that runs on a daily scale may ignore the sub-daily scale in-stream processes. The key processes affecting in-stream solute transport such as advection, dispersion and transient storage (dead zone) exchange can have considerable effect on the predicted stream solute concentrations, especially for localized studies. To represent realistic field conditions, it is therefore required to modify the in-stream water quality algorithms of SWAT by including these additional processes. Existing reach-scale solute transport models like OTIS (One-dimensional Transport with Inflow and Storage) considers these processes but excludes the actual biochemical reactions occurring in the stream and models nutrient uptake using an empirical first-order decay equation. Alternatively, comprehensive stream water quality models like QUAL2E (The Enhanced Stream Water Quality Model) incorporates actual biochemical reactions but neglects the transient storage exchange component which is crucial is predicting the peak and timing of solute concentrations. In this study, these two popular models (OTIS and QUAL2E) are merged to integrate all essential solute transport processes into a single in-stream water quality model known as ‘Enhanced OTIS model’. A generalized model with an improved graphical user interface was developed on MATLAB platform that performed reasonably well for both experimental data and previously published data (R²=0.76). To incorporate this model into large-scale hydrological models, it was necessary to find an alternative to estimate transient storage parameters, which are otherwise derived through calibration using experimental tracer tests. Through a meta-analysis approach, simple regression models were therefore developed for dispersion coefficient (D), storage zone area (A_s) and storage exchange coefficient (α) by relating them to easily obtainable hydraulic characteristics such as discharge, velocity, flow width and flow depth. For experimental data from two study sites, breakthrough curves and storage potential of conservative tracers were predicted with good accuracy (R²>0.5) by using the new regression equations. These equations were hence recommended as a tool for obtaining preliminary and approximate estimates of D, A_s and α when reach-specific calibration is unfeasible. </p> <p> </p> <p>The existing water quality module in SWAT was replaced with the newly developed ‘Enhanced OTIS model’ along with the regression equations for storage parameters. Water quality predictions using the modified SWAT model (Mir-SWAT) for a study catchment in Germany showed that the improvements in process representation yields better results for dissolved oxygen (DO), phosphate and Chlorophyll-a. While the existing model simulated extreme low values of DO, Mir-SWAT improved these values with a 0.11 increase in R² value between modeled and measured values. No major improvement was observed for nitrate loads but modeled phosphate peak loads were reduced to be much closer to measured values with Mir-SWAT model. A qualitative analysis on Chl-<i>a</i> concentrations also indicated that average and maximum monthly Chl-<i>a</i> values were better predicted with Mir-SWAT when compared to SWAT model, especially for winter months. The newly developed in-stream water quality model is expected to act as a stand alone model or coupled with larger models to improve the representation of solute transport processes and nutrient uptake in these models. The improvements made to SWAT model will increase the model confidence and widen its extent of applicability to short-term and localized studies that require understanding of fine-scale solute transport dynamics. </p>

<p>The aim of this thesis was to test the capability of Artificial Neural Networks (ANN) to estimate respiratory mechanics during mechanical ventilation (MV). ANNs are universal function approximators and can extract information from complex signals. </p><p>We evaluated, in an animal model of acute lung injury, whether ANN can assess respiratory system resistance (R_RS) and compliance (C_RS) using the tracings of pressure at airways opening (P_AW), inspiratory flow (V’) and tidal volume, during an end-inspiratory hold maneuver (EIHM). We concluded that ANN can estimate C_RS and R_RS during an EIHM. We also concluded that the use of tracings obtained by non-biological models in the learning process has the potential of substituting biological recordings.</p><p>We investigated whether ANN can extract C_RS using tracings of P_AW and V’, without any intervention of an inspiratory hold maneuver during continuous MV. We concluded that C_RS can be estimated by ANN during volume control MV, without the need to stop inspiratory flow.</p><p>We tested whether ANN, fed by inspiratory P_AWand V’, are able to measure static total positive end-expiratory pressure (PEEP_tot,stat) during ongoing MV. In an animal model we generated dynamic pulmonary hyperinflation by shortening expiratory time. Different levels of external PEEP (PEEP_APP) were applied. Results showed that ANN can estimate PEEP_tot,stat reliably, without any influence from the level of PEEP_APP.</p><p>We finally compared the robustness of ANN and multi-linear fitting (MLF) methods in extracting C_RS when facing signals corrupted by perturbations. We observed that during the application of random noise, ANN and MLF maintain a stable performance, although in these conditions MLF may show better results. ANN have more stable performance and yield a more robust estimation of C_RS than MLF in conditions of transient sensor disconnection.</p><p>We consider ANN to be an interesting technique for the assessment of respiratory mechanics.</p>

The aim of this thesis was to test the capability of Artificial Neural Networks (ANN) to estimate respiratory mechanics during mechanical ventilation (MV). ANNs are universal function approximators and can extract information from complex signals. We evaluated, in an animal model of acute lung injury, whether ANN can assess respiratory system resistance (RRS) and compliance (CRS) using the tracings of pressure at airways opening (PAW), inspiratory flow (V’) and tidal volume, during an end-inspiratory hold maneuver (EIHM). We concluded that ANN can estimate CRS and RRS during an EIHM. We also concluded that the use of tracings obtained by non-biological models in the learning process has the potential of substituting biological recordings. We investigated whether ANN can extract CRS using tracings of PAW and V’, without any intervention of an inspiratory hold maneuver during continuous MV. We concluded that CRS can be estimated by ANN during volume control MV, without the need to stop inspiratory flow. We tested whether ANN, fed by inspiratory PAW and V’, are able to measure static total positive end-expiratory pressure (PEEPtot,stat) during ongoing MV. In an animal model we generated dynamic pulmonary hyperinflation by shortening expiratory time. Different levels of external PEEP (PEEPAPP) were applied. Results showed that ANN can estimate PEEPtot,stat reliably, without any influence from the level of PEEPAPP. We finally compared the robustness of ANN and multi-linear fitting (MLF) methods in extracting CRS when facing signals corrupted by perturbations. We observed that during the application of random noise, ANN and MLF maintain a stable performance, although in these conditions MLF may show better results. ANN have more stable performance and yield a more robust estimation of CRS than MLF in conditions of transient sensor disconnection. We consider ANN to be an interesting technique for the assessment of respiratory mechanics.