Stacked hybridization to enhance the performance of artificial neural networks (ANN) for prediction of water quality index in the Bagh river basin, India.

Water quality assessment is paramount for environmental monitoring and resource management, particularly in regions experiencing rapid urbanization and industrialization. This study introduces Artificial Neural Networks (ANN) and its hybrid machine learning models, namely ANN-RF (Random Forest), ANN-SVM (Support Vector Machine), ANN-RSS (Random Subspace), ANN-M5P (M5 Pruned), and ANN-AR (Additive Regression) for water quality assessment in the rapidly urbanizing and industrializing Bagh River Basin, India. The Relief algorithm was employed to select the most influential water quality input parameters, including Nitrate (NO3-), Magnesium (Mg2+), Sulphate (SO42-), Calcium (Ca2+), and Potassium (K+). The comparative analysis of developed ANN and its hybrid models was carried out using statistical indicators (i.e., Nash-Sutcliffe Efficiency (NSE), Pearson Correlation Coefficient (PCC), Coefficient of Determination (R2), Mean Absolute Error (MAE), Root Mean Square Error (RMSE), Relative Root Square Error (RRSE), Relative Absolute Error (RAE), and Mean Bias Error (MBE)) and graphical representations (i.e., Taylor diagram). Results indicate that the integration of support vector machine (SVM) with ANN significantly improves performance, yielding impressive statistical indicators: NSE (0.879), R2 (0.904), MAE (22.349), and MBE (12.548). The methodology outlined in this study can serve as a template for enhancing the predictive capabilities of ANN models in various other environmental and ecological applications, contributing to sustainable development and safeguarding natural resources.

Assessing the protective effect of cutoff walls on groundwater pumping against saltwater upconing in coastal aquifers.

Subsurface physical barriers are amongst the most effective methods to mitigate seawater intrusion in coastal aquifers. The main objective of this study was to examine the impact of cutoff walls on saltwater upconing using laboratory and numerical modelling experiments. Physical experiments were first completed to reproduce the saltwater upconing process in a laboratory-scale coastal aquifer model incorporating an impermeable cutoff wall. Numerical modelling was used for validation purposes and to perform additional simulations to explore the protective effect of cutoff walls against saltwater upconing. The results suggest that the cutoff wall did not substantially delay the saltwater upconing mechanism in the investigated configurations. Laboratory and numerical observations showed the existence of some residual saline water, which remained on the upper part of the aquifer on the seaward side of the wall following the retreat of the saltwater. The protective effect of cutoff walls was noticeably sensitive to the design parameters. Specifically, cutoff walls installed close to the pumping well enabled the implementation of higher pumping rates, therefore a more optimal use of the freshwater, especially for deeper wells. The results highlighted that the penetration depth of the cutoff walls may not necessarily need to exceed the depth of the pumping well to ensure effectiveness, which is of great importance from construction and economic perspectives.

The effectiveness of cutoff walls to control saltwater intrusion in multi-layered coastal aquifers: Experimental and numerical study.

The objective of this study was to examine the performance of cutoff walls in controlling saltwater intrusion in stratified heterogeneous coastal aquifers. Numerical and laboratory experiments were completed in laboratory-scale aquifer where the effectiveness of cutoff walls was assessed in three different configurations, including a homogeneous scenario, a stratified aquifer with high K-low K-high K pattern (case HLH) and another stratified aquifer with low K-high K-low K pattern (case LHL). The results show that the cutoff wall was effective in reducing the saltwater wedge in all the investigated cases of layered-aquifers with toe length reduction of up to 43%. The wall exhibited more wedge reduction in shallower than steeper hydraulic gradients. However, the soil stratification appeared to lessen the overall performance of the wall compared to the homogeneous case. The aquifer stratification disrupted the flow dynamics, and thus affected the freshwater velocity at the wall opening to various degrees, depending on the layering pattern. The presence of an interlayer of low k (case HLH) inhibited the downward movement of the freshwater towards the wall opening, and thus decreasing the repulsion ability of the wall. Moreover, the presence of an underlying low permeability layer (case LHL) was found to obstruct the freshwater flow in the lower part of the aquifer, thereby slowing down the velocity through the wall opening. Numerical analysis of other layering patterns of monotonically increasing/decreasing permeability from top to bottom showed that the cutoff wall remained effective in repulsing the seawater wedge.

A lysimeter experimental study and numerical characterisation of the leaching of incinerator bottom ash waste.

Incinerator bottom ash (IBA) is a residual produced from incinerating municipal solid waste. In the past, IBA presented a big waste disposal problem; however, various recycling approaches have been adopted in recent years to mitigate this problem, as well as to provide a useful alternative to using primary aggregate resources. The use of IBA as an alternative to conventional aggregates in different civil engineering construction applications helps to conserve premium grade aggregate supplies; however, when IBA is in contact with water in the field, as a consequence of precipitation events or changes in water table, elements, such as salts and heavy metals, may be released to the soil and ground water. In this work, IBA waste was mixed with limestone aggregate to produce a blend with acceptable mechanical properties and minimum environmental risks for use as road foundation. The study focused on evaluating potential environmental impacts of some constituents, including sulphate, chloride, sodium, copper, zinc and lead in IBA blends using a lysimeter as a large scale leaching tool. Moreover, a specific scenario simulating field conditions was adopted in the lysimeter to assess the potential impact of changing conditions, such as IBA content in the blend, liquid to solid ratio (L/S) and pH value, on long-term release of heavy metals and salts. Then, numerical modelling was used to predict the release of the aforementioned constituents from IBA based on initial measurement of intrinsic material properties and the kinetic desorption process concept. Experimental results showed that zinc and lead were released in very low concentrations but sodium and sulphate were in high concentrations. The control limestone only blend also demonstrated low release concentrations of constituents in comparison to IBA blends, where constituent concentrations increased with increase in IBA content. Experimental results were compared with numerical results obtained using a non-equilibrium desorption model. Good agreement was found between the two sets of data.

Field scale DNAPLs transport under nonequilibrium sorption conditions.

The purpose of this work is to study the desorption of dense nonaqueous phase liquids (DNAPLs), TCE in particular, from solid particles in field scale heterogeneous aquifers upon their remediation. A computer program, capable of simulating the fate and transport of NAPLs in porous media, has been developed to work under nonequilibrium sorption conditions. The model has been applied to a field scale site at Hill Air Force Base, Utah, which has been contaminated by DNAPLs. The simulated domain was 155 ft (47.25 m) long, 60 ft (18.29 m) wide, and 15.5 ft (4.72 m) thick. This thickness represents only the saturated zone of the aquifer. Changes in permeability, grain size distribution, and sorptive properties throughout the site have been incorporated into the model. Immediately after the aquifer cleanup, the DNAPL concentration in the aqueous phase was assumed to be zero, and this was considered the start-off time for the simulation. Results show that, with an increase in time, the TCE diffused out of the solid particles, forming a plume. The rate of contaminant diffusion was observed to be very fast at the start, followed by a very slow stage, with a number of years required for substantial desorption of the contaminant from the solid particles. There were local variations in contaminant concentration in the fluid phase across the site due to aquifer heterogeneity. A comparison between numerical results and water samples taken from the site after the end of the cleanup operation is also presented.

Effects of contamination age on phenanthrene desorption from aquifers.

The length of contact time of contaminant with aquifer solids has a significant impact on the subsequent desorption process that normally occurs after the aquifer cleanup. In the present study, a diffusion model was built and validated with experimental results. Phenanthrene was brought in contact with sand and soil particles for different lengths of time. The distributions of solid-phase concentration within intraparticle pores were determined numerically and then used as the initial conditions for the subsequent desorption process. The purpose was to investigate the effect of contaminants' age on desorption time. Results revealed that a significant fraction of contaminants desorbs very slowly, even for freshly contaminated materials. For example, it has taken years for phenanthrene to release from freshly contaminated Jurassic limestone and Triassic limestone. The desorption time for aged contaminated material was found to be longer than for freshly contaminated material, at a more or less constant ratio with respect to the degree of aging.