Integrated Hydrologic Modeling of Lake Tahoe and Martis Valley Mountain Block
and Alluvial Systems, Nevada and California
Justin L. Huntington , Richard G. Niswonger , Seshadri Rajagopal , Yong Zhang , Murphy Gardner ,
Charles G. Morton , Donald M. Reeves , David McGraw , Greg M. Pohll
Desert Research Institute, [email protected], [email protected], [email protected],
[email protected], [email protected], [email protected], [email protected],
[email protected], Reno, NV, USA
U.S. Geological Survey, [email protected], Carson City, NV, USA
The U.S. Department of Interior has identified the Truckee River basin as highly likely for potential water
supply conflict in the future. A critical water supply to the Truckee River is outflow from Lake Tahoe, and
surface and groundwater contributions from the Martis Valley hydrographic area. This paper highlights
the development of an integrated surface water and groundwater model, GSFLOW, in the Lake Tahoe
and Martis Valley hydrographic areas to ultimately assess the effects of changing climate on surface and
groundwater resources, and to identify the hydrologic mechanisms responsible for observed and
simulated effects. Maintaining a balance between accurate representation of spatial features (e.g.,
geology, streams, and topography) and computational efficiency is a key objective for developing a
realistic and computationally efficient model that can adequately simulate important hydrologic processes,
including groundwater, stream, lake, and wetland flows and storages. Computational efficiency is required
in order to calibrate to a diverse set of observation data, including surface water flows, lake levels, and
groundwater head observations. Our work highlights how maintaining continuity between mountain block,
stream zones, and basin fill units through data driven and conceptual layering is efficient, and results in
accurate calibration to both surface water and groundwater observations. Additionally, we treat spring
and wetland areas as groundwater head observations equal to land surface elevation, which provides
constraints on groundwater heads over a much broader part of the model domain relative to well head
The U.S. Department of Interior (DOI) is currently initiating a historical and future water supply and
demands investigation in the Truckee River basin as part of the WaterSMART (Sustain and Manage
America’s Resource for Tomorrow) program. Snowmelt runoff from the Sierra Nevada is a multibillion
dollar resource that is critical to the region’s economy, forest health, aquatic ecosystems, and agriculture.
Snow melt runoff and storage water from Lake Tahoe and Martis Valley hydrographic areas is the primary
water supply to the Truckee River, Reno, NV, and surrounding agricultural areas. Significant shifts in the
timing of snowmelt and streamflow, and reductions in summer streamflow have recently been observed in
the Lake Tahoe basin and larger Sierra Nevada region (Coats, 2010; Kim and Jain, 2010). Groundwater
in the Martis Valley watershed has recently been identified as vulnerable to changing climate (Singleton
and Moran, 2010).
Groundwater will be pivotal for future water supplies and the health of groundwater dependent
ecosystems (GDEs), yet our understanding of climate change impacts on surface water and groundwater
(SW/GW) supply and exchange is very limited. The majority of water in the Truckee River basin
emanates from high-altitude mountain catchments, thus a better understanding of how climate change
affects hydrology in mountain catchments is essential for long-term water and biological resources
planning in the region. Hydrologic modeling capabilities of mountain catchments are not very well
developed. Most climate change hydrologic modeling studies of mountain catchments have relied on
simple bucket and linear reservoir representation of groundwater, while either ignoring or over simplifying
the effects of the unsaturated zone. These models calculate recharge independently of dynamic
groundwater levels and SW/GW interactions. Furthermore, the important interplay between snowmeltderived streamflow and SW/GW interactions are not simulated in a coupled manner, which is essential for
evaluating climate-change impacts on summertime flow (baseflow) and GDEs in snow-dominated regions
(Huntington and Niswonger, 2012).
Recent developments of integrated hydrologic models provide a means of simulating coupled hydrologic
processes in mountain catchments. Integrated hydrologic models can provide greater insight into climate
change effects on watershed hydrologic processes due to their ability to more realistically simulate
feedback between hydrologic processes that occur above and below land surface (Maxwell and Kollet,
2008; Ferguson and Maxwell, 2010; Sulis et al., 2011; Huntington and Niswonger, 2012). In this paper,
we present an application of the integrated hydrologic model, GSFLOW (Markstrom et al., 2008) to
simulate climate change effects on the hydrology in all the mountain catchments tributary to Lake Tahoe,
and Martis Valley hydrographic areas. The regional scale and hydrologic complexity of these catchments
pose difficult challenges for hydrologic modeling; however, we have constructed useful models through
innovative methods and automated calibration procedures. Century-long simulations of surface water
and ground water flow provide unique insights into the feedbacks between climate and hydrologic fluxes
that drain from the Sierra-Nevada to the Truckee River. For this paper, we focus on conceptual model
development and calibration of GSFLOW and present some preliminary results that demonstrate the
unique hydrologic conditions of these mountain catchments.
CONCEPTUAL AND NUMERICAL MODEL DEVELOPMENT
Maintaining a balance between adequate horizontal and vertical grid discretization, while honoring known
stream, wetland, and lake locations and elevations (i.e., heads), is particularly difficult in mountain
catchments due to complex topography and geology. Generally, the geology in the Sierra Nevada is
characterized by low-permeable mountain block overlain by thin, high-permeable alluvium and glacial
deposits near stream channels that gradually thicken in the down-valley direction. We developed gridblock representations of the alluvium and mountain block subsurface geology using deductive reasoning
and a combination of data-driven automated and manual interpolation to observed lithologies, cross
sectional and surficial geologic maps, and geophysical surveys, while explicitly considering known stream
and wetland locations. Key to the development of the model grid is the conditioning of the model gridscale digital elevation model to ensure proper location of streams and wetlands, and their sub-grid scale
geometries. This conceptual model, which merges data-driven hydrostratigraphic interpolation with
conceptual understanding of the surface water and groundwater systems, is useful for constructing
integrated models in data limited mountain block regions.
Lake Tahoe and Martis Valley basins are largely representative of typical topography, geology, climate,
and hydrology of the greater Sierra Nevada region. Important characteristics that are shared among the
upland watersheds of the region are the large topographic relief, high precipitation gradients with
significant winter snowfall, and relatively impermeable shallow bedrock that accentuates the dominance of
shallow groundwater-flow paths in the regional system. Because the alluvial aquifers are typically small
and have limited storage, they are likely to be more sensitive to climate fluctuations as compared to thick
valley aquifers. Mean annual precipitation over the model domains ranges from 380 to 1,650 mm, with 90
percent of the precipitation occurring between November and March. Monthly average extreme
temperatures range from 30 C in August to -10 C in January. Vegetation consists of subalpine and
conifer forest, with some deciduous riparian and meadows association. Mountain block geology is
primarily composed of granitic and volcanic rocks, overlain with glacial moraines and stream deposits in
low-elevation areas that primarily make up the alluvial aquifers, while soils are generally shallow and
derived from parent rock consisting of mostly sand and silts. Gridded datasets of elevation, geology,
vegetation, soils, and land use were used to discretize and parameterize GSFLOW. Precipitation and
temperature was distributed spatially across model domains (1,900–3,000 m above Mean Sea Level,
AMSL) using the Parameter-elevation Regression on Independent Slopes Model (PRISM) monthly
precipitation spatial distributions (Daly et al., 1994), and daily temperature and precipitation recorded at
the Mt. Rose, Squaw Valley, Tahoe City Cross, and Truckee #2 SNOTEL stations, and Tahoe City
cooperative-observer weather station.
Spatial hydrogeologic and stratigraphic data, primarily reported by Brown and Caldwell (2012) and Plume
et al. (2009), were used to develop the conceptual hydrogeologic framework model (HFM) and vertical
Figure 1. Total alluvial thickness (layers 2
plus 3) for Lake Tahoe and Martis Valley
models. Mountain block layers were
assigned zero alluvial thickness shown
as no data.
and horizontal model discretization for both study
areas. Well logs and geophysical data were used to
develop the HFM and define the layer thicknesses that
represent alluvium, and these data were interpolated
and mapped to the model grid using the geologic
modeling software, Leapfrog. HFM results from
Leapfrog were modified in GIS to simulate thin alluvial
stream deposits along streams in the mountain block,
as well as simulating gradual transitions of alluvial
layer thicknesses from the mountain front to the valley
floor while maintaining layer continuity (Figure 1).
Model cells were set to a 300×300 m spatial resolution
over the 1,310 km and 500 km model domains for
Lake Tahoe and Martis Valley, respectively. The HFM
was discretized vertically into five layers, and
horizontally into approximately 15,000 and 5,600 active
grid cells per layer, for a total of 73,100 and 28,000
active cells for Lake Tahoe and Martis Valley,
respectively. The HFM was divided into four basic
geologic units, including top soil, alluvium, weathered
bedrock, and less-weathered bedrock. Layer 1 (soil), 2
and 3 (alluvium), and 4 and 5 (mountain block) ranged
in thickness from 0-4m, 0-210m, and 60-120m,
respectively. Based on the steep topography near the
watershed divides, no-flow boundary conditions were
assigned along the edges of the model domain that
coincide with watershed divides.
MODEL CALIBRATION AND RESULTS
GSFLOW is typically calibrated using a 3-step process, where PRMS is calibrated independent of
MODFLOW-NWT, and MODFLOW-NWT is calibrated for a steady state stress period, and lastly, PRMS
and MODFLOW-NWT are calibrated jointly for transient daily stress periods in GSFLOW. In this work,
we calibrated PRMS for a 29-year period by matching observed streamflows. PRMS was calibrated in a
stepwise automated framework utilizing the Differential Evolution Adaptive Metropolis (DREAM) algorithm
(Vrugt et al., 2008), where solar radiation, precipitation, evapotranspiration, and parameters controlling
the shape of the hydrograph are calibrated independently, in that order. Goodness of fit between the
simulated and observed streamflow was assessed using the Nash-Sutcliffe statistic (N-S = 0.75) at the
monthly timestep (Figure 3).
For the second step, MODFLOW-NWT (Niswonger et al., 2011) was
calibrated independent of PRMS using a steady-state stress period,
including representation of stream flow (SFR2), Lakes (LAK7), and
unsaturated-zone flow (UZF1). Mean annual PRISM precipitation was
scaled to represent the mean annual streamflow (i.e., sum of recharge,
interflow, and overland runoff), and utilized as net infiltration for UZF1
and MODFLOW-NWT. Calibration of UZF1 and MODFLOW-NWT was
performed by adjusting the precipitation scaling factor and layer specific
homogeneous aquifer hydraulic conductivity values until there was a
good correspondence between the simulated steady-state flows in
streams, lake levels and lake outflows,
Figure 2. Observed
groundwater heads, and the locations of
and simulated mean
major discharge and wetland areas.
Wetland areas were also used to calibrate
for Martis Valley basin
the model by comparing surface elevations
to simulated head. Goodness of fit between
observed and simulated heads was assessed using the RMSE
(Figure 3). Results of the calibration generally show excellent
agreement between simulated and observed heads for Martis
Valley, where the RMSE is 8.6m, and normalized RMSE,
NRMSE (RMSE/total head loss), is 2.8%. A small NRMSE
indicates that model errors are only a small part of the overall
model response (Anderson and Woessner, 1992). The
calibrated spatial distribution of depth to water (DTW) is very
intuitive, where there is shallow DTW near streams, valleys,
and meadows, with DTW increasing in mountain block and
high elevation areas (Figure 4). Groundwater heads are
shown for layer 4 and are above land surface in some valley
Figure 3. Observed and simulated
floor mountain transition zones, while heads in layer 2 in these
steady state heads at wells and
areas are only slightly above land surface and discharging as
wetland areas for Martis Valley.
groundwater ET and stream seepage. Preliminary steady state
simulations for Lake Tahoe also reveal very interesting but intuitive results. Figure 5 illustrates the spatial
distribution of steady state net flux (recharge – discharge) for Lake Tahoe and tributary watersheds,
where significant groundwater discharge is occurring along stream valleys and around the lake rim at
major transition areas of changing topography and hydraulic head gradients. Analyses of model results
indicate that the preliminary calibration and model results of PRMS and MODFLOW-NWT for both Martis
Valley and Lake Tahoe are fairly robust and accurate. In addition, our preliminary calibrated water budget
compares well to precipitation, ET, and recharge percentages derived from recent watershed modeling,
chloride mass balance, and Darcian flux estimates of recharge in adjacent watersheds with similar
geology, vegetation, and precipitation magnitudes (Maurer and Berger, 1997; Jeton and Maurer, 2007).
Figure 5. Lake Tahoe
net recharge and
by lakes, streams,
and diffuse recharge
for each model cell.
and negative values
Figure 4. Martis Valley simulated
depth to groundwater using
layer 4 (mountain block) heads.
Here we present preliminary results from two integrated hydrologic models of complex mountain basins.
Upland catchments represented in these models are very important because the majority of available
water in the Truckee River basin emanates from these catchments as snowmelt runoff. Because these
models can simulate the interactions among all the major co-varying hydrologic processes, including
snowmelt, runoff, evapotranspiration, and SW-GW interactions, they are a big step forward in terms of
simulation capabilities for assessing the effects of climate on water resources. As described in this paper,
efficient and accurate representation of hydrogeologic features within the model is paramount for
developing a robust model that can be calibrated using automated procedures that ensure objective
model performance relative to diverse sets of observation data.
This research was financially supported by the U.S. Bureau of Reclamation and U.S. Department of
Energy. We thank the U.S. Geological Survey, Office of Groundwater, and Nevada Water Science
Center for their committed technical support of this project.
Anderson, M. P., and W. W. Woessner, 1992. Applied Groundwater Modeling: Simulation of Flow and
Advective Transport, 2nd edition,Academic, San Diego, Calif.
Brown and Caldwell Environmental Consultants (2012. Draft Martis Valley Groundwater Management
Plan, November, 19, 2012.
Coats, R., 2010. Climate change in the Tahoe basin: regional trends, impacts and drivers, Clim. Change,
Daly, C., R. P. Neilson, and D. L. Phillips, 1994. A statistical-topographic model for mapping climatological
precipitation over mountainous terrain, J. Appl. Meteor., 33, 140–158.
Ferguson, I. M., and R. M. Maxwell, 2010. Role of groundwater in watershed response and land surface
feedbacks under climate change, Water Resour. Res., 46, W00F02,
Huntington, J. L., and R. G. Niswonger, 2012. Role of surface-water and groundwater interactions on
projected summertime streamflow in snow dominated regions: An integrated modeling approach,
Water Resour. Res., 48, W11524, doi:10.1029/2012WR012319.
Jeton, A.E., and D. K. Maurer, 2007. Precipitation and runoff simulations of the Carson Range and Pine
Nut Mountains, and updated estimates of ground-water inflow and the ground-water budget for
basin-fill aquifers of Carson Valley, Douglas County, Nevada, and Alpine County, California, U.S.
Geological Survey Scientific Investigations Report 2007–5205, 56 p.
Kim, J. S., and S. Jain, 2010. High-resolution streamflow trend analysis applicable to annual decision
calendars: a western United States case study. Clim. Change, 102, 699-707.
Markstrom, S. L., R. G. Niswonger, R. S. Regan, D. E. Prudic, and P. M. Barlow, 2008. GSFLOWCoupled Ground-water and Surface-water FLOW model based on the integration of the
Precipitation-Runoff Modeling System (PRMS) and the Modular Ground-Water Flow Model
(MODFLOW-2005), U.S. Geological Survey Techniques and Methods 6-D1, 240 p.
Maxwell, R. M., and S. J. Kollet, 2008. Interdependence of groundwater dynamics and land energy
feedbacks under climate change, Nature Geoscience, 1, 665–669.
Maurer, D. K. and Berger, D. L., 1997. Subsurface flow and water yield from watersheds tributary to Eagle
Valley hydrographic area, west-central Nevada, U.S. Geological Survey Water-Resources
Investigations Report: 97-4191
Niswonger, R.G., S. Panday, and I. Motomu, 2011. MODFLOW-NWT, A Newton formulation for
MODFLOW-2005: U.S. Geological Survey Techniques and Methods 6–A37, 44 p.
Plume, R. W., M. L. Tumbusch, and T.L. Welborn, 2009. Hydrogeology of the Lake Tahoe Basin,
California and Nevada. U.S. Geological Survey Scientific Investigations Map, 3063, 1 sheet.
Singleton, M.J., Moran, J.E., 2010. Dissolved noble gas and isotopic tracers reveal vulnerability of
groundwater in a small, high elevation catchment to predicted climate change, Water Resour.
Res. 46, W00F06. doi:10.1029/2009WR008718.
Sulis, M., C. Paniconi, C. Rivard, R. Harvey, and D. Chaumont, 2011. Assessment of climate change
impacts at the catchment scale with a detailed hydrological model of surface-subsurface
interactions and comparison with a land surface model, Water Resour. Res., 47, W01513,
Vrugt, J.A., C.J.F. ter Braak, M.P. Clark, J.M. Hyman, and B.A. Robinson, 2008. Treatment of input
uncertainty in hydrologic modeling: Doing hydrology backward with Markov chain Monte Carlo
simulation, Water Resources Research, 44, W00B09, doi:10.1029/2007WR006720.