A. G. Bobba

Application of first-order and Monte Carlo analysis in watershed water quality models

To achieve effective environmental control, it is important to develop methodologies for dealing with uncertainties in model simulation of pollution behaviour and effects. Several procedures have been proposed to quantify uncertainties in modelling studies. This paper utilizes the two methods that are widely applied, i.e. functional analysis and Monte Carlo Simulation.The first-order part of the functional analysis method provides a measure of uncertainties in dependent variables in terms of uncertainties in independent variables. The procedure is based on first-order terms in the Taylor series expansion of the dependent variable about its mean value with respect to one or more independent variables. The major assumption in this procedure is that all independent and dependent variables are the second moment variables (SMV), which means that the behaviour of any SMV is completely described by its mean and standard deviation. The mathematical simplicity of the procedure allows application by simple input-output models. Consequently, it has been applied to many environmental simulators, e.g. hydrological models, stream water quality models, lake water quality models and ground water pollution models.The Monte Carlo Simulation (MCS) method uses a large number of repeated trials or simulations with the values for stochastic inputs or uncertain variables selected at random from their assumed parent probability distributions to establish an expected range of model uncertainty.

Mathematical models for saltwater intrusion in coastal aquifers

Flow of freshwater and saltwater intrusion in coastal aquifers has drawn the attention of many investigators. Several laboratory, as well as mathematical models have been developed to study the pattern of flow of groundwater in coastal aquifers. Mathematical models have wider range of application and are the concern of this paper. Due to the complex nature of the problem, each of these mathematical models are based on certain simplifying assumptions and approximations. This paper presents a critical review of various methods of solution which have been proposed. The validity of the results abtained and the limitations of these models are also discussed.

APPLICATION OF A WATERSHED RUNOFF MODEL TO NORTH-EAST POND RIVER, NEWFOUNDLAND: TO STUDY WATER BALANCE AND HYDROLOGICAL CHARACTERISTICS OWING TO ATMOSPHERIC CHANGE

The hydrological sensitivities to long-term climate change of a watershed in Eastern Canada were analysed using a deterministic watershed runoff model developed to simulate watershed acidification. This model was modified to study atmospheric change effects in the watershed. Water balance modelling techniques, modified for assessing climate effects, were developed and tested for a watershed using atmospheric change scenarios from both state of the art general circulation models and a series of hypothetical scenarios. The model computed daily surface, inter- and groundwater hows from the watershed. The moisture, infiltration and recharge rate are also computed in the soil reservoirs. The thirty years of simulated data can be used to evaluate the effects of climatic change on soil moisture, recharge rate and surface and subsurface flow systems. The interaction between surface and subsurface water is discussed in relation to climate change. These hydrological results raise the possibility of major environmental and socioeconomic difficulties and have significant implications for future water resource planning and management

Application of uncertainty analysis to groundwater pollution modeling

Prediction and evaluation of pollution of the subsurface environment and planning remedial actions at existing sites may be useful for siting and designing new land-based waste treatment or disposal facilities. Most models used to make such predictions assume that the system behaves deterministically. A variety of factors, however, introduce uncertainty into the model predictions. The factors include model and pollution transport parameters and geometric uncertainty. The Monte Carlo technique is applied to evaluate the uncertainty, as illustrated by applying three analytical groundwater pollution transport models. The uncertainty analysis provides estimates of statistical reliability in model outputs of pollution concentration and arrival time. Examples are provided that demonstrate: (a) confidence limits around predicted values of concentration and arrival time can be obtained, (b) the selection of probability distributions for input parameters affects the output variables, and (c) the probability distribution of the output variables can be different from that of the input variables, even when all input parameters have the same probability distribution

Field validation of ‘SUTRA’ groundwater flow model to Lambton County, Ontario, Canada

A computer simulation model (SUTRA) is applied to the Fresh Water Aquifer of Lambton County, Ontario, Canada. The model is calibrated by matching computed hydraulic heads with observed hydraulic heads. The heads are compared using statistical methods. The model indicates that virtually all of the water flowing through the Fresh Water Aquifer discharges to the St Clair River and Lake Huron. Discharge of groundwater from the Canadian side of the Fresh Water Aquifer is calculated to be between 0.45 and 0.50 m3/sec for that portion of the river between Lake Huron and Stag Island.

Lead-210 sedimentation in Lake Ontario

The sedimentation rates and diffusive sediment mixing coefficients at several Lake Ontario locations have been derived from measurements of unsupported210Pb profiles in sediment cores. The values of mixing coefficients obtained in the present study are significantly lower than those obtained previously through an analysis of porosity profiles. The present estimates, however, are consistent with the rather well-preserved pollutant profiles at some of these locations. It is observed that the more realistic value of the mixing coefficient, obtained by inclusion of the sedimentation rate parameter, follows the sign opposite to that for the constant obtained by regression analysis of the porosity data. Further work is required to delineate this apparent relationship between two important physical characteristics of deposited sediments.Analysis of available suspended sediment data shows that Niagara River supplies about 1.8 million tonnes of sediment annually to Lake Ontario. This value is significantly lower than that (4.6 mt/yr) used previously in constructing sediment and pollutant budgets for Lake Ontario. From the presently derived sedimentation rate and suspended solid discharge estimates, an average value of 441 km2 (range 220-938 km2) is obtained for the minimum area of Lake Ontario over which the Niagara River-supplied fine sediment is deposited.

Application of four watershed acidification models to Batchawana Watershed, Canada

Four watershed acidification models (TMWAM, ETD, ILWAS, and RAINS) are reviewed and a comparison of model performance is presented for a common watershed. The models have been used to simulate the dynamics of water quantity and quality at Batchawana Watershed, Canada, a sub-basin of the Turkey Lakes Watershed. The computed results are compared with observed data for a four-year period (Jan. 1981-Dec. 1984). The models exhibit a significant range in the ability to simulate the daily, monthly and seasonal changes present in the observed data. Monthly watershed outflows and lake chemistry predictions are compared to observed data. pH and ANC are the only two chemical parameters common to all four models. Coefficient of efficiency (E), linear (r) and rank (R) correlation coefficients, and regression slope (s) are used to compare the goodness of fit of the simulated with the observed data. The ILWAS, TMWAM and RAINS models performed very well in predicting the monthly flows, with values of r and R of approximately 0.98. The ETD model also showed strong correlations with linear (r) and rank (R) correlation coefficients of 0.896 and 0.892, respectively. The results of the analyses showed that TMWAM provided the best simulation of pH (E=0.264, r=0.648), which is slightly better than ETD (E=0.240, r=0.549), and much better than ILWAS (E=-2.965, r=0.293), and RAINS (E=-4.004, r=0.473). ETD was found to be superior in predicting ANC (E=0.608, r=0.781) as compared to TMWAM (E=0.340, r=0.598), ILWAS (E=0.275, r=0.442), and RAINS (E=-1.048, r=0.356). The TMWAM model adequately simulated SO4 over the four-year period (E=0.423, r=0.682) but the ETD (E=-0.904, r=0.274), ILWAS (E=-4.314, r=0.488), and RAINS (E=-6.479, r=0.126) models all performed poorer than the benchmark model (mean observed value).

Determination of diffusion coefficients associated with the transport of210Pb radionuclides in lake bed sediments

An iterative least-squares optimization technique is utilized in conjunction with a one-dimensional representation of the mass transport equation to generate theoretical210Pb concentration/depth profiles beneath the water-sediment interface that are best-fit approximations to directly measured210Pb concentration/depth profiles at various locations within the Great Lakes system. The outputs of such an optimization analysis are the diffusion coefficientsDM (molecular) andDB (bioturbation) associated with the transport of210Pb radionuclides in lake bed sediments.For all stations studied, the estimated values ofDB are consistently larger than the estimated values ofDM, emphasizing the importance of accounting for the effects of bioturbation in the modelling of contaminant transport through lake bed sediments.

Interfacing a hydrological model with the RAISON expert system

The RAISON expert system allows for the interactive analysis of spatial data related to water resources investigations. A conceptual design for the interface between the RAISON expert system and hydrological model includes functions for the simulation of runoff. Design criteria include ease of use, minimal equipment requirements, a generic data base management system, and use of a micro computer. An application is demonstrated for Northeast Pond River watershed, Newfoundland, that performs to predict the runoff.

Estimation of mixing rates in western Lake Ontario sediments by finite element analysis of210Pb profiles

A finite element model is used to generate theoretical excess210Pb concentration/depth profiles that are “best-fit approximations” to directly measured profiles at five locations in western Lake Ontario. The best-fit approximations are derived using a minimum error analysis technique. The results indicate that, while mixing is prevalent at all the study locations, its influence is severe at locations closer to the mouth of the Niagara River. The nonoccurrence of the implicit effects of sediment mixing in measured pollutant profiles at these locations is pointed out and it is postulated that the general assumptions of constant flux of excess210Pb and constant sedimentation rate in such models lead to unrealistic estimates of mixing.