Integrating Real-Time Chemical Sensors into Understanding of Groundwater Contributions to Surface Water in a Model Urban Observatory

PIs: C. Welty, S. Kaushal, P. Groffman, L. Band, A. J. Miller, M. McGuire, R. Maxwell, K. Belt, J. Duncan

Funding Source: National Science Foundation (8/1/10 - 7/31/13)

Project Web Site:


Intellectual Merit
The purpose of this proposal is to build onto the efforts developed in the first round of the WATERS Test Bed program to quantify the significance of groundwater in the urban water cycle.  We will deploy nitrate analyzers and electrical conductivity sensors in Baltimore watersheds and conduct mathematical modeling to address the following questions: (1) How can sources, timing and fluxes of solutes from groundwater to surface water vary as a function land use (ultraurban, suburban, exurban, forest) and stream position (headwater vs downstream)? (2) How can transport time scales and subsurface flowpaths vary with flow regime (base flow vs storms) and antecedent conditions? (3) How can information from high frequency sensor deployment across a range of hydrologic conditions be used to “fill in the gaps” from our current weekly long-term monitoring to explain interannual changes in residence times and flushing of solutes?  (4) How well can a physically-based watershed flow and transport model represent solute transport behavior across a range of time scales?

A long-term goal of hydrologic modeling and observational activities in the Baltimore region is to establish an “end-to-end system” of field-deployed sensors and sensor networks feeding real-time data into hydrologic and water quality models to enable prediction of water and chemical fluxes in streams and aquifers.  A principal objective is to understand how the urban landscape and infrastructure partitions water in all components of hydrologic cycle at multiple scales. This understanding is critical to quantifying biogeochemical cycles, and to aid in understanding transport pathways of contaminants to major tributaries and the Chesapeake Bay. Existing infrastructure in the Gwynns Falls watershed provides a robust backbone for quantifying fluxes of water, but has not yet been augmented to support high-resolution real-time collection of water-quality data. The proposed sensor deployment test program will also complement the 10-year weekly sampling of chemical species that has been carried out by the Baltimore Ecosystem Study LTER. An additional objective is to contribute to the national CUAHSI Hydrologic Information System effort by beta-testing new software and assimilating legacy regional environmental data into the HIS Observations Data Model.

Broader Impacts
Our work involves frequent communications with managers and agency personnel concerned with planning and implementing regulations affecting land use and water resources, including the Chesapeake Bay Program. The importance of water resource management to serve the public interest is a topic of growing importance to the State precisely because of stresses induced by patterns of growth, but with inadequate tools and insufficient data currently available to support decision-making. Results of this and related studies in the Baltimore region can provide information to assist local entities with stream restoration and other types of land preservation activities.

It is expected that data produced by this project will be used in dissertation research by a number of graduate students in our programs, whose expertise spans environmental engineering, hydrology, biogeochemistry, aquatic ecology, economics, and public policy.  The growth of our monitoring and modeling network will enable us to broaden the scope of training and research opportunities provided to students at all levels.

We have held discussions with the Florida (Santa Fe) Testbed (Wendy Graham, PI) about cross-testbed comparisons of sensors and models as part of the test-bed renewal program.  We have already used to advantage the N sensor testing that the Florida group carried out in Phase 1 to choose the Satlantic nitrate analyzer for this proposal; we both plan EC as a surrogate sensor – in our case for chloride.   We both plan to use PARFLOW and SLIM in these very different hydroclimatic/ geologic environments and have discussions about how well the model works at each of our sites. We believe that this kind of informal cross-site comparison can contribute to concepts of building a network of sites, and can also contribute to educational goals of involved post-doctoral associates and graduate students.