Understanding and being able to model the transport of water and solutes in the unsaturated zone is important for many reasons, including groundwater recharge, agricultural management, and not least pollution remediation. Different approaches have been made to model the process, including both physical and stochastic considerations. While well predicting overall transport features, these approaches are limited with respect to the reproduction of (1) high variability, originating from the strongly heterogeneous character of field soils, and (2) scaling properties, statistical process symmetries across scales. A new concept that can potentially improve process description in the respects is random cascade modeling. A cascade process redistributes a quantity into successively smaller spatial units, intuitively similarly to infiltration in (field) soils. Cascade processes have been successfully employed to describe other highly variable geophysical and hydrological processes, and their relevance for subsurface transport has been supported in recent studies. We intend to investigate the applicability of random cascade models to solute transport dynamics by comparing the results of detailed laboratory experiments with high-resolution field observations. Based on the data analyses, numerical modeling will be performed, including development of methodologies for up- and down-scaling of solute transport characteristics.

For more details, see the full Project Description.

Much of 2002 was spent designing and constructing a two-dimensional so-called Hele-Shaw cell for flow visualisation, including sophisticated equipment for photographic documentation of the flow dynamics as represented by infiltrating dye. Image analysis and calibration of the photos (from colour to concentration) were performed. During 2002-2003, a range of infiltration experiments took place, producing a large data base which has been analysed during late 2003 and early 2004. A paper on these analyses - including experimental setup, general properties of the data, and more elaborate description in terms of time moments - were submitted in 2004 and published in early 2005 (Persson et al., 2005). 

Status (October 2005): Scaling analyses and random cascade modelling of dye infiltration pattern dynamics started during autumn 2004 and was completed in early 2005. The results have shown that the dye patterns exhibit a very accurate scaling behaviour in terms of e.g. power spectrum and so-called breakdown coefficients (BDC). A BDC-based random cascade model was formulated and applied to reproduce the observed patterns, and an example is shown in Figure 1. The full results are reported in Olsson and Persson (2005).

Figure 1. Observed (OBS) and random cascade model (RCM) simulated dye infiltration patterns.

Duration: 2002-2004

Funded by: VR (Swedish Research Council)