6
SUMMARY AND FUTURE SCOPE
SUMMARY AND FUTURE SCOPE
(a) Analysis of rainfall variations indicated that there was an average of six-year cyclic period for the wet year and nine-year cyclic period for the dry year over the 29 years of the analysis period (1982-2010).
(b) LULC change analysis for the year 1985-2005 indicated that there was no significant change in the LULC in the study region, which was mostly dominated by agricultural lands (90%).
(c) Spatio-temporal variations of evapotranspiration (ET) showed that there was a shift of one month in the cropping pattern from October to September during the study period (1982-2010).
(d) The seasonal groundwater estimation results showed that WTF method considered the fluctuations in groundwater levels and RIF method simply followed the rainfall patterns. But both the methods were not able to capture the wet and dry year impacts on groundwater recharge.
(e) Surface water hydrological model (SWAT) was calibrated and validated with the observed monthly discharge time series to get the monthly groundwater recharge at sub-basin scale. Groundwater delay and shallow groundwater storage were found to be the most sensitive calibration parameters.
(f) SWAT model also over-predicted the recharge and also not able to capture wet and dry year effects on groundwater recharge due to the under-prediction of simulated Actual Evapotranspiration (38% of observed AET).
(g) SWB method (in which high resolution remote sensing-based satellite datasets such as precipitation, actual evapotranspiration (ground truth validated) and soil moisture along with runoff obtained from calibrated SWAT model were considered) improved the monthly groundwater recharge estimations and also captured wet and dry year effects on groundwater recharge in the study zone.
6.1.2 Estimation of Fluid Flux and Hydraulic Conductivity
In this objective, heat was used as a natural tracer to estimate the fluid flux and bed hydraulic conductivity through the sediment-water interface based on laboratory experiments conducted in a sandbox. In order to understand the variation in fluid flux and hydraulic conductivity under different conditions, a series of laboratory experiments were conducted with four different scenarios (soil combinations, temperature gradients (gradient between hot water and normal water) and ponding depths). The following conclusions were drawn from this study
1. Vertical thermal profiles were measured in soil column for different conditions to estimate the fluid flux in different soil layers using four different analytical solutions i.e. Keery and Hatch Amplitude Ratio (AR) and Phase Difference (PD).
Among these four methods, fluid fluxes estimated using Keery AR method were observed to have a good agreement with measured seepage velocity. It might be due to the lack of consideration of dispersion term in Keery analytical method and violating the one-dimensional assumption.
2. Inter-comparison of four different scenarios show that fluid fluxes were observed to be higher values in Scenario 1 (Fine Sand I at top, Fine Sand II at middle and Medium Sand at bottom layers) than other three scenarios. It is resulted due to the placing of highly permeable media at bottom, moderately permeable soils at middle and less permeable media at the top layers.
3. The influence of ponding depth and temperature gradient on fluid flux were analyzed using ANOVA (TukeyKramer HSD test at P ≤ 0.05) by considering different cases i.e. varying temperature gradient while ponding depth remained constant and in other case, temperature was constant against varying ponding depth. This analysis revealed that ponding depth had a significant influence on fluid flux over temperature gradient in all the four scenarios considered.
4. Effective Hydraulic Conductivities (Keff) calculated using hydraulic conductivities (estimated from the fluid fluxes) of different soil layers in all the four scenarios followed Gaussian normal probability distribution and the influence of ponding depth and temperature gradient on Keff was observed to be not significant.
However, the influence of temperature gradient was noted to be slightly more compared to ponding depth.
6.1.3 Assessment of River-Aquifer Interaction
The objective aimed the assessment of spatio-temporal variation of river-aquifer interaction exchange flux of Kosi river. For this purpose, a fully distributed groundwater model (MODFLOW) was used along with remote sensing inputs (groundwater recharge and evapotranspiration), aquifer characteristics (hydraulic conductivity, porosity, and specific storage) and riverbed conductance. Groundwater recharge incorporated in the groundwater modelling was estimated using two different methods using SWAT modelling and SWB method. In order to understand the best accurate recharge estimation method for groundwater modelling in Kosi river basin, groundwater modelling was carried out for two
SUMMARY AND FUTURE SCOPE
cases, i) modelling with recharge estimated using SWAT model and ii) modelling with recharge estimated using SWB method. Also, the influence of riverbed conductance on river-aquifer interaction process was assessed by incorporating different riverbed conductance (calculated using Keff values of different soil combinations considered in the sandbox experiments) in the groundwater modelling. The major observations in the study lead to the following conclusions.
First, monthly groundwater levels simulated using SWAT recharge-based groundwater flow modelling (case I) did not correlate with observed seasonal groundwater levels and also was not able to capture the wet and dry year effects. However, simulated groundwater levels in the groundwater modelling with SWB based recharge were observed to have a better agreement with observed groundwater levels and also found to be captured the effects of wet and dry years. In addition, the model was able to capture the effects of wet and dry years on the exchange process. Second, three different interaction zones were identified from upstream (Kosi barrage) to downstream (conferencing point with Ganga river) in the study reach. It is observed that the river always loses water to the aquifer (as influent) in Zone I (80 km from upstream) and the river mostly gains water from the aquifer (as effluent) in Zone III (40 km above downstream). However, in the middle reaches (Zone II), the river has a combination of both influent and effluent natures. Third, at the upstream, the river always loses water to the aquifer and mostly gains water from the aquifer at downstream during wet and dry years which indicates significant influence of topography on river-aquifer exchange flux. Fourth, the results obtained in case II modelling with different riverbed conductance values of different scenarios (soil combinations) indicate that the river-aquifer interactions are the most sensitive to the riverbed conductance.
From this study, it can be concluded that use of satellite remote sensing inputs (recharge and evapotranspiration used in SWB method for estimating recharge) and estimated river bed conductance based on laboratory sandbox experiments improved the assessment and understanding river-aquifer interaction process in an alluvial River basin.