H31M-01
A Framework to Compare Lumped and Distributed Hydrological Models of Climate-Land Surface-Groundwater Dynamics
We have developed a framework to compare predictions of lumped and distributed hydrological models. Our work is motivated by the need to rigorously identify differences in the parameterization of atmosphere-land- groundwater dynamics across a range of scales. The framework is based on the prediction of the components of the long-term water balance, which can be approximated by the value of fluxes (runoff, evapotranspiration) and state variables (catchment storage) in steady-state conditions. We define the river basin as our control volume for which the mass conservation equation holds. In particular, we focus on the equilibrium catchment storage, discharge-storage relations and ratio of surface to subsurface runoff for different climate conditions. We develop an example comparing two well known hydrological models: the ARNO model, as an example of a lumped- conceptual model, and the tRIBS model, as an instance of a physically based distributed model. We create idealized scenarios in which the parameterizations of both models are consistent. We demonstrate that these two representations of the hydrological system do not yield equal steady-state behavior, and that the incompatibilities are not related to particular model parameters, but to the functional model structure. In particular, we provide an interpretation for the structural differences as a consequence of lateral redistribution of flows in the shallow saturated zone. We also apply the comparison framework to the VIC-2L and VIC-3L models. Finally, we compare the models in a more realistic situation where we incorporate a complex topography overlying an impervious bedrock layer. We argue that the proposed framework can provide insight into the functional behavior of a range of different climate-land surface-groundwater models, with implications toward the need for new observations and theory to accurately capture the dynamics of natural basins.
H31M-02
Simulated Soil Moisture Climatology in North America: Does the Water Table Matter?
We demonstrate the link between two terrestrial water reservoirs: the root-zone soil moisture and the groundwater. Since soil moisture strongly influences land-atmosphere fluxes, its link to the groundwater may affect the spatial-temporal variability of these fluxes. Here we simulate the climatologic water table depth at 30 arc-second resolution as constrained by US Geological Survey site observations. Then we use this water table climatology as the lower boundary for the soil, and VIC (Variable Infiltration Capacity) simulated landsurface flux climatology as the upper boundary, to calculate the soil moisture climatology (SMC) at 14 depths (down to 4 m). Comparisons with VIC, NARR (North America Regional Reanalysis) and observations suggest that, first, SMC is wetter than VIC, despite identical landsurface flux; second, while climate is the dominant signature in NARR and VIC, the water table manifest itself in SMC, with wet soil over shallow water table; third, while soils in VIC and NARR get drier with depth, soils in SMC get wetter in regions of shallow water table; and lastly, SMC has the highest root-zone (top-2 m) total soil water storage. These differences may have implications to climate modeling.
H31M-03
Regional Groundwater Recharge and Groundwater Evapotranspiration in Illinois
Several recent land surface modeling studies have found that the incorporation of groundwater representation can enhance evapotranspiration due to the additional moisture provided by the aquifer. However, there is a lack of regional-scale field evidence to support this finding. The role of shallow unconfined aquifers in supplying water for the evapotranspiration (groundwater evaporation) is investigated in this study based on a comprehensive 22- year (1984-2005) monthly hydroclimatic dataset in Illinois. State-average monthly groundwater recharge and the separation of streamflow components (surface runoff and groundwater runoff) were estimated by soil water balance computations and compared with the previous estimates. The seasonal and interannual variability of groundwater recharge estimates were investigated. It was found that during the summer, the capillary flux from the shallow water table to the root zone soil moisture helps to maintain a high rate of evapotranspiration (~120 mm/month), and its magnitude increases throughout the summer and reaches the maximum of 12 mm/month in August. Neglecting this mechanism in regional water budget studies or land surface modeling may lead to the underestimation of soil moisture and summer evapotranspiration.
H31M-04
Relocation of Southwestern US Terrestrial Moisture through Atmospheric Pathways
Evapotranspiration accounts for the majority of the water budget in arid and semiarid regions such as the Southwestern U.S. Using a Lagrangian analytical model forced with North American Regional Reanalysis (NARR) data, we are able to identify the regions where this evapotranspired moisture is advected and eventually falls as precipitation. Our work shows that moisture originating from evapotranspiration in the Southwestern US contributes to precipitation throughout North America, particularly the Central Plains and the Midwestern US. Evapotranspiration and the subsequent export of moisture are particularly important during the North American monsoon season, when the combination of increased soil moisture, energy availability and atmospheric demand promotes the transference of water to the atmosphere via direct evaporation from the soil or transpiration by vegetation. Moisture evapotranspired in the four corners region of the Southwest contributes to an average 15% of the total summer precipitation in regions of Wyoming, Nebraska and Kansas. We find that during the 2000- 2003 drought period, the decrease in precipitation and evapotranspiration within the four corners region significantly decreased the amount of recycled precipitation of downwind regions. Due to the uncertainty in NARR evapotranspiration, observations and independent model outputs are used as validation. Our study illustrates the importance of moisture transport through atmospheric pathways, a process that links the surface hydrologic stores and atmospheric moisture fluxes.
H31M-05
A Hydrologic Study of High Creek Fen: Groundwater Dynamics and Sources
High Creek Fen is a groundwater-fed wetland located in South Park, Colorado. Although the ecology and hydrology of the fen have been studied, the groundwater sources to the fen have not been identified. In addition, the spatial and temporal variation in the physical and chemical properties of groundwater at High Creek Fen are not well understood. This research has become a priority because land use changes in the South Park are exacerbating groundwater resource issues. Hydrologic modifications to High Creek Fen could have dramatic ecological effects that threaten globally rare plants and invertebrates. In addition, the Nature Conservancy is interested in identifying additional preservation lands to protect the hydrologic and ecologic integrity of High Creek fen. To characterize the physical properties of groundwater, a regional groundwater contour map was created to establish the regional groundwater flowpaths. Next, the local physical dynamics of groundwater was characterized using hydraulic head and groundwater level measurements at bi-weekly intervals from June 2007 to October 2007. These measurements were paired with stream discharge measurements at High Creek, the outlet of the fen, in order to determine the relationship between spatial and temporal groundwater dynamics and the surface water discharge. Water samples were collected from possible groundwater sources for chemical and isotopic characterization. These potential sources include groundwater in the piezometers throughout the fen, and surface water in the fen and in High Creek. This data supports the existence of multiple, relatively shallow, groundwater sources to High Creek Fen. Additionally, the data highlights the spatial and temporal variation of physical and chemical groundwater dynamics at High Creek Fen. Since it is a groundwater-fed system, understanding the hydrology of High Creek Fen is dependent on our understanding of groundwater dynamics. This research provides the foundation for studying other aspects of this unique and diverse ecosystem.
H31M-06
Correlation of groundwater withdrawal and fluctuation of stream stage for steam depletion
Stream depletions have two components which are due to groundwater withdrawal from a nearby pumping well and due to stream stage fluctuation itself. Previous studies often focus on each component separately. This study examines and compares two components together for various hydraulic settings and pumping schemes. It presents generalized solutions for stream depletions with and without streambeds. Since this study concerns seasonal stream depletion rates, it assumes that stream stage fluctuation is solely a function of time and a cosine function of time is chosen as an example. Three pumping schedules are 1) pumping for two months during dry season with a rate Q; 2) pumping for four months during the dry season with a rate of Q/2; and 3) constant pumping through out a year with a rate of Q/6. The values of Q are chosen to be 1000m3/d and 5000m3/d to simulate a small irrigation or a municipal well. One of the primary characteristics affecting stream depletion is the period of the stream stage fluctuation, in addition to the pumping rate and the hydraulic conductance of the streambed. For a pumping rate of Q=5000 m3/d and a period of one year, percentages of the stream depletion rate due to stream stage fluctuations to the total stream depletion rate range from 1.5% to 15% for a hypothetical case investigated. If reducing the stream fluctuation period to 30 days and keeping other setting the same, that percentages then range from 5.4 % to 38.5%. The amplitude of the stream stage fluctuation was found to have less important influence on the depletion rates.
H31M-07
Geodetic Monitoring Of A Karst Aquifer In The Larzac Region, South Of France
In this study we attempt to understand the hydrological behavior of a karst aquifer on the Larzac Plateau (South of France) using geodetic techniques. In this Jurassic plateau mainly formed of dolomite, the uppermost weathered zone, or epikarst, acts as a potential perched aquifer. Water then percolates down the infiltration zone to the saturated or vadose zone, which represents the main water reservoir and is connected to the outlet. On this studied karst aquifer (110 km2), water input is exclusively rainfall and draining occurs at the Durzon perennial spring in a karstic valley. Because the entire Larzac Plateau is drained by a few karstic springs, it constitutes a significant water resource. It is therefore important to understand the physical links between pluviometry, infiltration through the karst and the spring discharge. Our basis assumption is that the transient water storage in the epikarst and in the infiltration zone has a Newtonian effect on the gravity signal and also deforms the earth surface. The karst aquifer is hence being monitored since January 2006 with monthly absolute gravity measurements at three sites, differential relative gravity measurements at the surface and at 60 m depth, continuous high resolution tiltmeters at two sites, continuous surface and underground rainfall recordings, pressure head variations in bore holes and underground natural caves and hourly Durzon spring flow. Absolute gravity measurements corrected for regional scale gravity using water storage models exhibit an important seasonal trend (100-150 nm/s2 of amplitude). Heavy rainfall events are both seen by FG5 absolute gravimetry and tiltmeters. Global mass balance considerations linking rainfall, evapotranspiration and spring discharge provide a framework for the understanding of observed gravity variations. We find that water storage is not spatially uniform on the karst system and that it can be explained according to a geomorphologic model accounting for the degree of karstification of the studied area. Significant deformation in tilt (10-6 rad) is closely correlated to precipitation. During each event, tilt azimuth remains constant and tilt amplitude rises to a maximum and then slowly declines. The tilt azimuth points in the spring's direction, as if water convergence at the spring during high flow periods deformed the plateau reversibly. However, large observed tilt amplitude is not currently explained in the framework of an elastic behavior of the plateau loaded by a distributed water flow through the aquifer
H31M-08
The Amazon Water Budget as Diagnosed by ERA40
The ERA40 surface water budget is evaluated by comparing to observed precipitation (P), streamflow/runoff (R), and evapotranspiration (ET) for the period of 1980-2001 in the Amazon River basin. ERA40 precipitation shows an overall agreement with observations in annual and interannual scales, although the seasonal cycle shows a clear precipitation low bias during the wet season. Rainfall distribution on sub-basin scale disagrees with observations by as much as 30%. ERA40 rain rates compare to those estimated by TRMM at the 2.5o spatial resolution indicating interception loss may not account for the 25% underestimation of runoff (R) in annual scale, although it remains unclear whether the ERA40 rain rate agrees with observation at the scales of convective rainfall. ET in ERA40 appears to be higher than observations by as much as 30%. The residual of precipitation and runoff (P-R) in ERA40 generally agrees with observations. Thus the large scale surface water budget could provide a realistic annual ET, but soil moisture nudging appears to cause overestimation of annual ET. The seasonal scale analysis however, shows that soil moisture nudging is necessary to provide realistic ET during the dry season. This is because ERA40 provides only 45 mm surplus of P-R relative to ET during the wet season, whereas the deficit in the dry season is almost four times greater. This low bias in wet season soil moisture recharge may be caused by the underestimation of wet season rainfall in ERA40.