Stochastic simulation of karst conduit networks
Karst aquifers have very high spatial heterogeneity. Essentially, they comprise a system of pipes (i.e., thenetwork of conduits) superimposed on rock porosity and on a network of stratigraphic surfaces and frac-tures. This heterogeneity strongly influences the hydraulic behavior of the karst and it...
| Autores: | , , , |
|---|---|
| Tipo de recurso: | artículo |
| Fecha de publicación: | 2011 |
| País: | España |
| Institución: | Consejo Superior de Investigaciones Científicas (CSIC) |
| Repositorio: | DIGITAL.CSIC. Repositorio Institucional del CSIC |
| OAI Identifier: | oai:digital.csic.es:10261/276808 |
| Acceso en línea: | http://hdl.handle.net/10261/276808 https://doi.org/10.1016/j.advwatres.2011.09.014 |
| Access Level: | acceso abierto |
| Palabra clave: | conduit geometry Network topology diffusion-limited aggregation inception horizon rose diagram Z-histogram |
| Sumario: | Karst aquifers have very high spatial heterogeneity. Essentially, they comprise a system of pipes (i.e., thenetwork of conduits) superimposed on rock porosity and on a network of stratigraphic surfaces and frac-tures. This heterogeneity strongly influences the hydraulic behavior of the karst and it must be repro-duced in any realistic numerical model of the karst system that is used as input to flow and transportmodeling. However, the directly observed karst conduits are only a small part of the complete karst con-duit system and knowledge of the complete conduit geometry and topology remains spatially limited anduncertain. Thus, there is a special interest in the stochastic simulation of networks of conduits that can becombined with fracture and rock porosity models to provide a realistic numerical model of the karst sys-tem. Furthermore, the simulated model may be of interestper seand other uses could be envisaged. Thepurpose of this paper is to present an efficient method for conditional and non-conditional stochasticsimulation of karst conduit networks. The method comprises two stages: generation of conduit geometryand generation of topology. The approach adopted is a combination of a resampling method for generat-ing conduit geometries from templates and a modified diffusion-limited aggregation method for gener-ating the network topology. The authors show that the 3D karst conduit networks generated by theproposed method are statistically similar to observed karst conduit networks or to a hypothesized net-work model. The statistical similarity is in the sense of reproducing the tortuosity index of conduits,the fractal dimension of the network, the direction rose of directions, the Z-histogram and Ripley’s K-function of the bifurcation points (which differs from a random allocation of those bifurcation points).The proposed method (1) is very flexible, (2) incorporates any experimental data (conditioning informa-tion) and (3) can easily be modified when implemented in a hydraulic inverse modeling procedure. Sev-eral synthetic examples are given to illustrate the methodology and real conduit network data are used togenerate simulated networks that mimic real geometries and topology |
|---|