Global sensitivity analysis techniques for probabilistic ground water modeling

Identifying key factors of the seawater intrusion model of Dagu river basin, Jiaozhou Bay

Seawater intrusion is a complex groundwater - seawater interaction process, and it is influenced by many factors from ground surface to underground, from groundwater to seawater. Generally, for seawater intrusion model, some model parameters and boundary conditions are always specified by model users' personal experiences or literature's reference value. The defective model would damage the groundwater management for controlling and preventing seawater intrusion when making decisions are based on this model. In order to improve the reliability of seawater intrusion model, the influences of model inputs on output should be identified prior at optimizing model inputs. Dagu river basin, Jiaozhou Bay is one of the most serious areas of seawater intrusion in China, and it is chosen as the study area in this study. The seawater intrusion model of Dagu river basin is built based on a general program SEAWAT4. The key influence factors of model output are analyzed by two sensitivity analysis methods, i.e., stepwise regression and mutual entropy. The results demonstrated that the most important influence factors which have largest sensitivities to groundwater Cl^- concentration are the precipitation rate and groundwater pumping in agriculture area. In addition, the hydraulic conductivity of zone 1 has a non-negligible influence on seawater intrusion process. Stepwise regression analysis is capable of identifying most important influence factor, and it can't handle complicated nonlinear input-output relationship. Mutual entropy analysis is reliable for identifying the influence factors for complex seawater intrusion model.

Evaluating Groundwater Interbasin Flow Using Multiple Models and Multiple Types of Data

The groundwater interbasin flow, Q_y , from the north of Yucca Flat into Yucca Flat simulated using the Death Valley Regional Flow System (DVRFS) model greatly exceeds assessments obtained using other approaches. This study aimed to understand the reasons for the overestimation and to examine whether the Q_y estimate can be reduced. The two problems were tackled from the angle of model uncertainty by considering six models revised from the DVRFS model with different recharge components and hydrogeological frameworks. The two problems were also tackled from the angle of parametric uncertainty for each model by first conducting Morris sensitivity analysis to identify important parameters and then conducting Monte Carlo simulations for the important parameters. The uncertainty analysis is general and suitable for tackling similar problems; the Morris sensitivity analysis has been utilized to date in only a limited number of regional groundwater modeling. The simulated Q_y values were evaluated by using three kinds of calibration data (i.e., hydraulic head observations, discharge estimates, and constant-head boundary flow estimates). The evaluation results indicate that, within the current DVRFS modeling framework, the Q_y estimate can only be reduced to about half of the original estimate without severely deteriorating the goodness-of-fit to the calibration data. The evaluation results also indicate that it is necessary to develop a new hydrogeological framework to produce new flow patterns in the DVRFS model. The issues of hydrogeology and boundary flow are being addressed in a new version of the DVRFS model planned for release by the U.S. Geological Survey.

Pollution risk assessment based on QUAL2E-UNCAS simulations of a tropical river in Northern India

This paper exemplifies the application of U.S. Environmental Protection Agency's water quality model, QUAL2E-UNCAS in assessing the pollution risk of a tropical river. The rivers selected for study were Hindon (main river) and Kali (its tributary) flowing through Uttar Pradesh district of Northern India. The model application to the two rivers revealed poor water quality in terms of dissolved oxygen (DO), biochemical oxygen demand (BOD), and ammonia concentrations. Monte Carlo simulations were performed on two different data sets that were confirming to marked seasonal variations. The Monte Carlo simulation (MCS) derived 95% confidence level for these parameters strengthened the fact that all point sources were exploiting the assimilative capacity of the two rivers. In order to ascertain probabilistically the risk at which two rivers were falling short of desired water quality, probability curves based on effluent standards and available water quality were prepared. On mapping the two curves, it was found that at 95% probability, Hindon River was flowing with 53 to 100% less of desired DO, up to 100% more of minimum BOD, and probability with which ammonia concentration would not be more than the desired concentration was found to fall downstream. The Kali headwaters showed better quality during low river temperature but worsened downstream with up to 100% violation in all the above observed parameters. It is expected that similar studies wherein the dependable levels with which a polluted river can be understood to fall short of desired water quality can prove to be useful in ascertaining the efficacy of effluent standards and/or follow-up of pollution control measures.

Identification of Critical Molecular Components in a Multiscale Cancer Model Based on the Integration of Monte Carlo, Resampling, and ANOVA

To date, parameters defining biological properties in multiscale disease models are commonly obtained from a variety of sources. It is thus important to examine the influence of parameter perturbations on system behavior, rather than to limit the model to a specific set of parameters. Such sensitivity analysis can be used to investigate how changes in input parameters affect model outputs. However, multiscale cancer models require special attention because they generally take longer to run than does a series of signaling pathway analysis tasks. In this article, we propose a global sensitivity analysis method based on the integration of Monte Carlo, resampling, and analysis of variance. This method provides solutions to (1) how to render the large number of parameter variation combinations computationally manageable, and (2) how to effectively quantify the sampling distribution of the sensitivity index to address the inherent computational intensity issue. We exemplify the feasibility of this method using a two-dimensional molecular-microscopic agent-based model previously developed for simulating non-small cell lung cancer; in this model, an epidermal growth factor (EGF)-induced, EGF receptor-mediated signaling pathway was implemented at the molecular level. Here, the cross-scale effects of molecular parameters on two tumor growth evaluation measures, i.e., tumor volume and expansion rate, at the microscopic level are assessed. Analysis finds that ERK, a downstream molecule of the EGF receptor signaling pathway, has the most important impact on regulating both measures. The potential to apply this method to therapeutic target discovery is discussed.