Studie van de zoutindringing langs de Ma-rivier in Vietnam onder historische en toekomstige klimaatcondities

Water resources managers have been facing challenges posed by future climatic changes, especially in coastal regions of the world. The Ma catchment in Vietnam falls into this category given that agriculture is the key economic activity and it depends on water in the coastal rivers. Increased temperatures, changes in regional precipitation regimes, and potential increased sea level would have a great impact on salt intrusions and thus a serious threat to the regional agriculture. However, the ability of water resources managers tackling this problem in this catchment in particular and in this developing country in general are still limited due to the poor facilities, data and human resources. The goal of this research is to support the coastal decision-makers in the Ma catchment by providing information on potential future hydrology and sea level conditions under changing climatic conditions. To this end, the study investigates the threat of saltwater intrusion in the basin under conditions influenced by future climatic change. The first part of the dissertation presents an analytical salt intrusion model and its adaptations in a poor data environment. This model is suitable for a wide range of scenario investigations in support of water engineering and management in developing countries. With data provided by a network of fixed measurement stations only, the study sets up an analytic model to simulate salinity distribution along a multi-channel estuary at high water slack. A new approach is introduced to calibrate the parameters of a parsimonious salt intrusion model, which are typically extracted from estuary geometry measurements. Compared to the values derived from a field survey, the calibrated parameter values are in high agreement. Furthermore, by assuming a linear relation between inverses of the individual flows entering the estuary and inverses of the sum of flows gauged further upstream, the individual flows could be assessed. Evaluation on the model simulations shows that the model explains salinity distribution along the Ma Estuary reasonably well. This performance demonstrates the predictive power of the model and of the proposed parameter/input estimation approach in poor data environment. The second part of the study discusses a comprehensive comparison between two saltwater intrusion modelling approaches including 1) the analytical model as presented in the first part and 2) a hydrodynamic MIKE11 model, which is a prevailing tool for salt intrusion simulation. They are developed for the multi-channel Ma Estuary system in Vietnam. The analytical salinity intrusion model is applied for specific steady state conditions, whereas the hydrodynamic model simulates the spatiotemporal salinity variations. Both models show acceptable simulation performances at gauging stations along the Ma and Len Estuary rivers for the calibration and validation periods. For the Lach Truong river, the two modelling approaches show the poorest results, which are explained by the employment of alternative model boundaries. The dissertation also analyses the analytical and MIKE11 model simulations under changing conditions of upstream discharges and downstream water level. The two models have common equations for the river hydraulics and dispersion. However, the hydrodynamic model solves these equations dynamically for each river section, whereas the analytical one is based on assumptions of steady state conditions and of exponentially varying geometry. Because the latter assumption is valid for the Ma River, the result of both models show high agreement in the results under the current and scenario conditions. For the Len Estuary, abrupt changes in the cross-sectional area along the river lead to the disagreement in the simulation results between the two models. Furthermore, the data scarcity poses limitations to model parameter calibrations and explains partly the simulation differences. The third part of the study investigates the climate change impact on low flows in the Ma catchment in Vietnam for the mid and late 21st century. The considered climate model ensemble includes 15 CMIP5 climate model runs, with 8 RCP4.5 and 7 RCP8.5 runs. The projected meteorological input series for the hydrological model NAM is obtained using a quantile perturbation method. Results indicate that the low flow extremes may increase up to 50 % for the late 21st century, whereas the changes are less clear for the mid-century. These outcomes contradict the results of previous research for the same study area based on older generation (CMIP3) climate models, showing a decrease of the low flow extremes up to 5 % for the late 21st century. This could be attributed to the different behaviour of CMIP3 versus CMIP5 climate models, with higher autumn precipitation for the new CMIP5 models. Furthermore, from a water resources management perspective, the uncertainty in the obtained impact results need to be taken into account. In the unlikely event that the low flow extremes would decrease, the decrease would most likely be limited to 25 % or 15 % for the mid and late 21st century respectively. The final research objective sought to analyze the impact of anthropogenic climate change on the future salt intrusions for the two coastal rivers in the Ma catchment. An -SVR model and a linear regression model are further developed and integrated with the analytical models constructed in the second part of the study, hence the hybrid salt intrusion modelling tools are formed for the scenario analysis. Considering the future flow information downscaled in the third part of the study, the hybrid tools project the climate change impact on salt intrusions in the two rivers for the mid- and late century. Both RCP4.5 and 8.5 projections show minor changes in the salt intrusion length of 1 psu and 4 psu for these branches for the mid-century. For the late century, both salt intrusion lengths would penetrate about 1.7 km further upstream in the Ma River, whereas the increase in the intrusion length of 1 psu and 4 psu for the Len River is 3.7 and 5.2 km, respectively. The dissertation also investigates the impact of salt intrusion on the regional irrigation in the most extremes cases, when the upstream flow is the lowest and the salt intrusions reach their maxima. In most cases, the results show minor impact except for the RCP8.5 scenarios of the late century. In this event, nearly 8000 ha irrigated by water from the Ma River encounter the high salinity concentrations (higher than 1 psu) that are too high used for paddy rice, For the Len River, this is the case for about 7000ha. The procedures presented in this research are applicable to other areas as well, especially for areas with low data availability. They allow a better new knowledge of the climate variability and salt intrusion changes in river catchments. This knowledge allows coastal water managers to better account for the natural and anthropogenic climate variations in their basins, hence to achieve a better water engineering decision-making.status: publishe