Development of a cross-section based streamflow routing package for MODFLOW
Introduction
Understanding hydrological processes that occur in the stream-aquifer systems has significant importance to integrated water management, especially where streams and aquifers are hydrologically connected (Rassam, 2011, Sophocleous, 2002, Winter, 1995). The dynamics of stream-aquifer water exchanges are also important to the functioning of aquatic ecosystems (Hancock et al., 2005), pollutant and nutrient transport (Findlay, 1995, Jones and Holmes, 1996), as well as the quality and quantity of water supply to domestic, agricultural and recreational purposes (Winter, 1995). MODFLOW is the most widely used numerical model to simulate stream and groundwater interactions. However, it is challenging to represent stream-aquifer interactions for wide channel streams or for high-resolution grids, especially when the heterogeneity of streambed properties across the channel and stream morphology need to be taken into account.
The spatial variability of water and solute fluxes at the groundwater-surface water interface, the hyporheic zone, and their impacts on the stream and riparian ecosystems were studied extensively through experimental measurements in previous studies and continue as an area of active investigation (Biksey and Gross, 2001, Binley et al., 2013, Faulkner et al., 2012, Rosenberry and Pitlick, 2009, Wondzell et al., 2009). Such variability is largely attributed to streambed heterogeneity which is universal because of the streamflow dynamics and movement of streambed sediments, as well as streambed chemical reactions and invertebrate activities (Dong et al., 2012). Chen et al. (2008) discovered high spatial heterogeneity in streambed materials with depth, along the channel and cross the channel in the Platte River, Nebraska. Lu et al. (2011) also demonstrated that distribution of both horizontal and vertical hydraulic conductivities is significantly spatially dependent. Heterogeneity in permeability of streambed materials can form many flowpath connections between streams and aquifers, for both water and solutes (Conant et al., 2004), and consequently provide habitats of various scales which are important to different species (Brunke and Gonser, 1997). It has been demonstrated that neglecting streambed heterogeneity could cause significant errors in modeling results of stream-aquifer interactions (Irvine et al., 2012).
Brunner et al. (2010) pointed out the limitation in horizontal discretization of rivers could result in biased estimation of stream-aquifer exchange flux when the river width and cell size are mismatched. Mehl and Hill (2010) also examined the effect of the MODFLOW grid size on the simulation of stream-aquifer interactions. In the streamflow depletion analysis, on the other hand, Chen et al. (2008) showed that the depletion rate induced by groundwater pumping can be affected by the grid cell sizes. They all proved that refining the model grid can improve simulation accuracy. These findings underline the need for models to represent the geometry and properties both across and along stream channels with grids of high resolutions.
Because of the importance of streambed heterogeneity, and, particularly the need for simulating stream-aquifer interactions in wide channels or with refined grids, the Cross-Section based streamflow Routing (CSR) package was developed for MODFLOW. The objectives for developing the CSR package were: (1) to simulate stream-aquifer interaction and streamflow routing for the wide channel streams or with high-resolution grids; (2) to represent the heterogeneity in streambed properties not only along stream channels but also across them; (3) to physically represent the streambed geometry of high complexity. In this paper, we describe the concepts and algorithms used in the CSR package. A hypothetical stream-aquifer system is used to test the capabilities of the package. We also discuss the capabilities, assumptions and limitations, as well as potential improvement of the package.
Section snippets
Literature review on river and stream packages of MODFLOW
MODFLOW (McDonald and Harbaugh, 1988) was written in FORTRAN 77 and divided specific hydrologic features and calculation functions into separate modules. The modular structure, providing ease of modifying and adding features, is still being used in the current version. MODFLOW numerically solves the following finite-difference equation that describes the three-dimensional groundwater flow:
Overview of the model
The CSR package is developed to incorporate the capability of simulating streamflows and stream-aquifer interactions where the stream width is larger than the MODFLOW grid cell size. In the CSR package, the stream components are represented by different conceptualized geometry elements. Instead of a segment of polyline employed in most other stream packages to represent a stream segment, CSR uses a four-point polygon (quadrilateral) that is composed of an upstream cross-section and a downstream
Application of the CSR package
To test the capability of the CSR package, the proposed approach is applied to simulate the groundwater flow and stream-aquifer interaction of a hypothetical stream-aquifer system (Fig. 5). In the hypothetical model, the stream network is composed of a diversion channel, a tributary stream channel and a main stream channel. The CS numbers start with 1, 2 and 3 for the main, diversion and tributary channels, respectively. The diversion channel diverts stream water from the main channel to the
Advantages
The CSR package uses a cross-section based scheme to represent streams, and couples the Muskingum routing method and a rapid intersection area algorithm with MODFLOW, therefore providing some unique advantages in stream-aquifer simulation. These advantages can be concluded as follows: (1) By using a cross-section based scheme, the stream geometries are physically represented, which overcomes the limitation of other stream packages in representing channels with a width spanning over multiple
Conclusions
It is very challenging to simulate the streamflow and stream-aquifer routing for wide channels with previous stream packages. In this study, a cross-section based stream routing package is developed for MODFLOW. This approach uses a cross-section based routing scheme, thus being able to effectively represent the variability in the streambed properties such as morphology, permeability and roughness. Accordingly, heterogeneity of the stream-aquifer interactions can be simulated along and across
Acknowledgments
We thank the editor and anonymous reviewers for their valuable comments. The research was supported by Lower Platte North Natural Resources District of Nebraska, US National Science Foundation award ID0535255 and China Natural Science Foundation (project no. 41072183 and 41101015).
References (56)
Hydrologic connections of a stream-aquifer-vegetation zone in south-central Platte River valley, Nebraska
J. Hydrol.
(2007)- et al.
Electrical and hydraulic vertical variability in channel sediments and its effects on streamflow depletion due to groundwater extraction
J. Hydrol.
(2008) - et al.
A PCE groundwater plume discharging to a river: influence of the streambed and near-river zone on contaminant distributions
J. Contam. Hydrol.
(2004) - et al.
Comparison of vertical hydraulic conductivity in a streambed-point bar system of a gaining stream
J. Hydrol
(2012) - et al.
Hyporheic flow patterns in relation to large river floodplain attributes
J. Hydrol.
(2012) - et al.
Heterogeneous or homogeneous? Implications of simplifying heterogeneous streambeds in models of losing streams
J. Hydrology
(2012) - et al.
Surface-subsurface interactions in stream ecosystems
Trends Ecol. Evol.
(1996) A FORTRAN 77 program to calculate areas of intersection between a set of grid blocks and polygons
Comput. Geosci.
(1988)- et al.
Grid-size dependence of Cauchy boundary conditions used to simulate stream-aquifer interactions
Adv. Water Res.
(2010) - et al.
Modelling stream-aquifer seepage in an alluvial aquifer: an improved loosing-stream package for MODFLOW
J. Hydrol.
(2002)