MSc Dissertations

Advisors: Yangxiao Zhou, PhD, Associate Professor of Hydrogeology, and Shreedhar Maskey, PhD, Associate Professor of Hydrology & Water Resources, Water Resources and Ecosystems Department, IHE Delft, Netherlands

Abstract

The Toluca Valley is located in central Mexico's Upper Lerma River basin, where pollution of surface water and growing demand from population and industry put pressure on decreasing groundwater storage. The valley's topography (slope, drainage networks) and climatological features (precipitation and runoff potential) have implications for soil-water conservation and water harvesting (terracing and infiltration ponds). But before the practical experiment, a combination of existing surface and groundwater models can simulate these interventions and present the effects for future landscape modification purposes. The conjunctive use scenario of existing reservoir storage and groundwater pumping facilities is attempted to be improved by applying terraces and ponds in the Soil Water Assessment Tool (SWAT) and MODFLOW (GMS – Groundwater Modelling System).

At first, reduction in curve number and fixed practice factor are applied in the management operation database for hydrological response units, where the land use is agriculture and the land slope is within 5-47%. For the final analysis, a SWAT terrace model is selected with a 30% reduction in curve numbers. The SWAT recharge data is applied to the MODFLOW reference model to create a MODFLOW terrace model. Infiltration ponds (at the foothills) are applied in the MODFLOW terrace model to sink the runoff coming down from hills.

Compared to the reference scenario, adding terraces to SWAT directly lowers the average annual runoff by more than 10%. Indirect effects on the other components (lateral flow, evapotranspiration, and percolation) of the water balance result in an increase of less than 1% of the reference parameters. The combined effect on these components is the same as the runoff reduction. For all hydrological components, a difference between the reference model and terrace model only appears during the wet season. Over time, the baseflow volume for terraces increases.

Seasonal differences in storage changes depend on recharge (SWAT), showing an increase in cumulative storage in the MODFLOW terrace model. In MODFLOW, the changes in groundwater discharge to the river are following the identical behaviour changes of SWAT baseflow. But using both SWAT lateral flow and baseflow together helps to lessen the lag effect when compared to groundwater contribution to rivers by MODFLOW.

The pond water infiltration primarily contributes to the inflows during dry months, significantly boosting storage and baseflows. During the last decade of the simulation time, the cumulative storage changes of the terrace+ponds model show near-zero storage loss with a higher increase in aquifer water than the terrace model, indicating the achievement of sustainability in water resources. The baseflow estimation from MODFLOW is very high than SWAT due to the combined effect of terraces and ponds.

With these results, it's clear that terracing lowers peak flow and increases lateral flow and percolation, ponds improve long-term storage changes, and overall baseflow rises throughout the year because of all these effects working together. Terracing and infiltration ponds are significantly helpful in enhancing the conjunctive use scenario of the Toluca valley.

Advisor: Mohammad Mozaffar Hossain, PhD, Professor and Shyamal Karmakar, PhD, Professor of Environmental Science, IFESCU, Bangladesh

Abstract

Developing and testing hydrological models in data-poor regions like Bangladesh is vital in supporting water management facilities. In this study, Quantum-GIS was used to integrate the Soil and Water Assessment Tool (SWAT) to simulate streamflow and sediment transport in Halda Basin using three different digital elevation models (DEMs): ASTER, ALOS, and SRTM. This study reveals that though ASTER DEM delineates more accurate watershed sub-basins, SRTM delineates a more accurate stream network than ALOS and ASTER. Surface runoff curve number, baseflow, available water capacity of the soil layer, saturated hydraulic conductivity, threshold depth of water in the shallow aquifer required for return flow, groundwater evaporation coefficient, and deep aquifer percolation fraction are the sensitive parameters for the developed hydrological model. The model simulation was done for the years 1982-2017 (with a 3-year warm-up period), where the years 1985-2005 were the observation periods, 2006-2009 were used for calibration, and 2010-2012 were used for validation. For the calibrated models, calibration (R2 = 0.87 - 0.88, NS = 0.81 - 0.87) and validation (R2 = 0.95 - 0.97, NS = 0.84 - 0.93) using streamflow of Panchapukuria station suggested good agreement in the seasonal cycle of streamflow (except pre-monsoon and peak flows). Among the DEMs, SRTM shows an overestimation in streamflow and total sediment transport than ALOS and ASTER. The streamflow of the river is decreasing after the dam construction and showing the lowest amount in the pre-monsoon season. Sediment transport from the Halda to Karnaphuli river is 63745.30 to 93388.15 tonnes per year, following an increasing trend. This study improves understanding of the key processes of a catchment in a data-poor, monsoon-driven river basin and could serve as a baseline for scenario modelling (e.g., climate change) for future water management and policy framework.