Review of the Edwards Aquifer Habitat Conservation Plan Report 1 (2015) / Chapter Skim
Currently Skimming:
2 Hydrologic Modeling
2 Hydrologic Modeling
Pages 41-75
The Chapter Skim interface presents what we've algorithmically identified as the most significant single chunk of text within every page in the chapter.
Select key terms on the right to highlight them within pages of the chapter.
From page 41... ...
will affect water levels and spring flows. This chapter discusses the appropriateness of the hydrologic modeling strategies used for the Edwards Aquifer and it suggests additional analyses to quantify and, if possible, reduce uncertainty and improve defensibility.
|
From page 42... ...
First a conceptual model is developed that, for a fractured rock system like the Edwards Aquifer, should include (a) identification of the most important boundary conditions and features of heterogeneity in the system; (b)
|
From page 43... ...
This distinction becomes important later in the chapter as we review the updates made to the MODFLOW model of the Edwards Aquifer. GROUNDWATER FLOW MODELS OF THE EDWARDS AQUIFER Modeling of flow and transport in systems with conduit flow such as the Edwards Aquifer requires the use of complex models in order to account for its unique hydrogeology, which is characterized by significant heterogeneity in both porosity and permeability (see Box 2-1)
|
From page 44... ...
A fault is a fracture or fracture zone along which there has been displace ment of two blocks of the earth's crust or a geological formation. The effect of the fault zone on groundwater flow can be diverse; some faults enhance groundwater flow along the fault plane (and hence are preferential pathways)
|
From page 45... ...
Equivalent porous media model, using uniform aquifer parameters; (c) Equivalent porous media model in which highly fractured zones such as shown in (e)
|
From page 46... ...
to provide technical assistance and oversight. EAA'S 2014 MODFLOW Model The EAA's 2014 MODFLOW model is a single layer transient flow model to simulate heads and spring flows (HDR, 2011; EAA presentations to the NRC Committee)
|
From page 47... ...
EPM (as of May 2014) NOTE: EPM, equivalent porous media; K, hydraulic conductivity.
|
From page 48... ...
48 REVIEW OF THE EDWARDS AQUIFER HABITAT CONSERVATION PLAN FIGURE 2-3 This map of the Edwards Aquifer region shows the 2004 model domain used in the MODFLOW model as well as the expanded domain used for the finite element model that includes the contributing zone. SOURCE: Winterlee (2014a)
|
From page 49... ...
Spring flow at Comal Springs tended to have better fits than at San Marcos Springs; Well J-17 was better than Well J-27, and the well water levels had better fits than the spring flows. Linear correlation coefficients (R2)
|
From page 50... ...
(labeled the diffuse model)
|
From page 51... ...
(labeled the diffuse model)
|
From page 52... ...
showing predicted Comal Springs discharge under the different management scenarios shown in Figure 2-6. SOURCE: Winterlee (2014a)
|
From page 53... ...
The finite element model is more appropriate for aquifers with complex hydrostratigraphy because it is not limited by the rectangular model cells typically used in finite difference models such as MODFLOW. In addition, standard MODFLOW models represent the porous medium as having relatively slow laminar flow, whereas the flows in the conduit networks such as those found in the Edwards Aquifer may occur under turbulent conditions.
|
From page 54... ...
The current model uses the same equivalent porous medium (EPM) approach that was applied in the EAA 2014 MODFLOW model; that is, high transmissivity zones are being incorporated to model rapid flow movement in known conduits.
|
From page 55... ...
However, it can address the stated purpose of having another groundwater model, which was to make comparisons between the MODFLOW and FEFLOW model results and thereby reduce uncertainty. This purpose cannot be achieved under the current modeling strategy.
|
From page 56... ...
. Despite this, the Committee feels that the rapid aquifer response to recharge and management actions will not be adequately simulated without the use of conduit-like features, especially if the modeling effort moves to a shorter time step in the future (see subsequent discussion)
|
From page 57... ...
. The Puente method uses streamflow measurements upstream and downstream from the recharge zone and estimates tributary inflow to determine stream recharge on a monthly basis; this method also estimates base flow and recharge from direct precipitation.
|
From page 58... ...
. The recharge zone for the Edwards Aquifer consists of karst rocks that allow fast infiltration of precipitation and surface water through fractures and dissolution openings such as caves.
|
From page 59... ...
FIGURE 2-8 Comparison of recharge estimates from the Puente method and HSPF. SOURCE: Winterlee (2014d)
|
From page 60... ...
Furthermore, given that the calibration errors observed for the MODFLOW model are as large as the minimum required spring flows in the HCP, it is imperative to better understand model and input uncertainties and how they translate into uncertainties in the simulated management actions. Hence, this section describes ways for the EAA to formally consider uncertainty in their hydrologic models, both for MODFLOW and the finite element model.
|
From page 61... ...
To be most useful for EAA managers, they are presented in order from easiest-to-implement to most complex: sensitivity analysis, formal model validation, PEST predictive uncertainty analysis, and the ensemble method. Some of these methods can address multiple types of uncertainty, as will be discussed.
|
From page 62... ...
Formalizing Model Validation Validation refers to executing a calibrated model for a period of the data record not used in the calibration and comparing the results to measured data (e.g., groundwater levels and spring flows)
|
From page 63... ...
Having a validated model will allow the EAA to make better predictions of spring flow or groundwater-level responses for future scenarios. Box 2-2 describes how to determine confidence intervals based on validation-period residuals in order to quantify model uncertainty.
|
From page 64... ...
Plans to incorporate PEST into the current modeling efforts were mentioned by EAA scientists, but no results were available for the Committee to review. Ensemble Method Another source of uncertainty is associated with the assumptions in the conceptual model.
|
From page 65... ...
That is, an ensemble plot composed of multiple model outputs would show a range of spring flow responses to the same stress scenario. On the basis of this range of model outputs, error bars could be shown around the spring flow hydrographs that were simulated by the selected or preferred model version.
|
From page 66... ...
This method quantifies the reduction in model uncertainty that could be achieved by adding monitoring sites at specific locations. The utility PREDUNC, which is part of the suite of utilities available with PEST, is used to conduct the predictive uncertainty analysis.
|
From page 67... ...
First, it does not align with the finer time step of the ecological models to be developed in support of the HCP, most of which are daily. According to the HCP, the groundwater models will produce "reliable and defensible" spring flows and the ecological models will be "linked with the groundwater model" (EARIP, 2012; p.
|
From page 68... ...
is used with discrete flow features and under turbulent flow conditions. One of the difficulties presented by shortening the time step is that pumping rates for the hundreds of wells in the Edwards Aquifer region are only available on a monthly basis.
|
From page 69... ...
Future Code Selection Currently there are two models with two competing codes running for the Edwards Aquifer. As stated previously, the HCP-stated purpose for developing an alternate groundwater model was to "reduce uncertainty in the modeling results and increase the reliability and the defensibility of the projections of aquifer and spring flows." Compared to the proven methods for conducting sensitivity and uncertainty analysis discussed above, having two models is not generally useful for quantifying predictive uncertainty.
|
From page 70... ...
Whatever code is chosen, the conceptual model improvements described in the FEFLOW model section should be incorporated. CONCLUSIONS AND RECOMMENDATIONS The hydrologic modeling effort has shown continuous improvement in both the use of models and the incorporation of new data.
|
From page 71... ...
The hydrologic modeling should move toward making predictions on a daily time scale, e.g., by developing telescoping models of smaller regions. Unlike the monthly time step of the current modeling effort, a daily time step would better (1)
|
From page 72... ...
2012. Habitat Conservation Plan.
|
From page 73... ...
2004. Conceptualization and simulation of the Edwards Aquifer, San Antonio region, Texas: USGS Scientific Investigations Report 2004–5277, 143 pp.
|
From page 74... ...
2012. Origin and characteristics of discharge at San Marcos Springs based on hydrologic and geochemical data (2008–10)
|
From page 75... ...
Aquifer in the San Antonio region, Texas. Texas Water Development Board.
|
Key Terms
This material may be derived from roughly machine-read images, and so is provided only to facilitate research.
More
information on Chapter Skim is available.