Abstract:Accurate modeling of contamination in subsurface flow and water aquifers is crucial for agriculture and environmental protection. Here, we demonstrate a parallel method to quantify the propagation of the uncertainty in the dispersal of pollution in subsurface flow. Specifically, we consider the density-driven flow and estimate how uncertainty from permeability and porosity propagates to the solution. We take an Elder-like problem as a numerical benchmark and we use random fields to model the limited knowledge on the porosity and permeability. We construct a low-cost generalized polynomial chaos expansion (gPC) surrogate model, where the gPC coefficients are computed by projection on sparse and full tensor grids. We parallelize both the numerical solver for the deterministic problem based on the multigrid method, and the quadrature over the parametric space
What problem does this paper attempt to address?
The problems that this paper attempts to solve are: how to accurately simulate and quantify the uncertainty of contaminant dispersion in groundwater flow, especially in the case of density - driven flow. Specifically, the study focuses on how the uncertainties of permeability and porosity propagate into the solutions of pollution dispersion. These problems are crucial for agriculture and environmental protection, because groundwater pollution may cause serious impacts on ecosystems and human health.
### Core problems of the paper
1. **Pollution dispersion in density - driven flow**:
- Density - driven flow refers to the flow phenomenon caused by fluid density differences, such as saltwater intrusion or wastewater leakage. This kind of flow will cause contaminants to spread more quickly through convection.
2. **Impact of parameter uncertainty**:
- Parameters such as porosity and permeability in the groundwater system are difficult to measure accurately, and usually only the average values can be obtained. The uncertainties of these parameters will affect the prediction results of pollution dispersion.
3. **Selection of numerical methods**:
- Although the traditional Monte Carlo method can handle high - dimensional random variables, it has a slow convergence speed and high computational cost. Therefore, more efficient numerical methods are needed to quantify the propagation of uncertainty.
### Solutions
The paper proposes a technique based on the generalized polynomial chaos expansion (gPC) to estimate the propagation of uncertainty. The specific steps are as follows:
1. **Model establishment**:
- Use the Elder problem as a benchmark test case, which is a classic density - driven flow problem, describing the process of saltwater invading an aquifer from the upper boundary.
2. **Random field modeling**:
- Use random fields to represent the uncertainties of porosity and permeability, and these random fields reflect the complexity and heterogeneity of the geological medium.
3. **gPC expansion**:
- Represent the mass fraction \( c(t, x, \theta) \) as a series expansion of multivariate Legendre polynomials, that is:
\[
c(t, x, \theta) = \sum_{\beta \in J} c_\beta(t, x) \Psi_\beta(\theta)
\]
- Where \( \Psi_\beta(\theta) \) is the multivariate Legendre basis function, and \( c_\beta(t, x) \) is the gPC coefficient.
4. **Coefficient calculation**:
- Use the Clenshaw - Curtis quadrature rule to calculate the gPC coefficients:
\[
c_\beta(t, x) = \frac{1}{\langle \Psi_\beta, \Psi_\beta \rangle} \int_\Theta \hat{c}(t, x, \theta) \Psi_\beta(\theta) P(d\theta) \approx \frac{1}{\langle \Psi_\beta, \Psi_\beta \rangle} \sum_{i = 1}^{N_q} \Psi_\beta(\theta_i) \hat{c}(t, x, \theta_i) w_i
\]
5. **Parallel computing**:
- In order to improve computational efficiency, a two - level parallelization strategy is adopted. The first level is to run all scenarios in parallel on different computing nodes, and the second level is to use 2 - 32 cores for parallel computing within each scenario.
### Significance of application
Through this method, researchers can more accurately predict the migration paths of contaminants in groundwater and evaluate the uncertainties under different parameter settings. This is of great significance for applications such as optimizing experimental design, data assimilation, and Bayesian inference, and ultimately helps to better protect water resources and the environment.
In summary, this paper aims to solve the uncertainty problem of contaminant dispersion in groundwater flow through advanced numerical methods (such as gPC and parallel computing), thereby improving the prediction accuracy and reliability.