.|  Baltimore Ecosystem Study
BES Long Term Stream and Watershed Study Update Overview

The watershed approach is perhaps the most powerful and useful approach in the field of Ecosystem Ecology (Likens 1992). Development of the ability to quantify inputs and outputs of water, energy, nutrients and carbon to hydrologically defined drainage basins has provided a basis for evaluating whole ecosystem function and response to disturbance and environmental change. The watershed approach also has been fundamental to the development of theory about patterns of ecosystem development in time and space and about the regulation and maintenance of ecosystem function via mechanisms of resistance and resilience (Odum 1969, Holling 1973, Bormann and Likens 1979). Watershed-based approaches have been applied to fundamental studies of forest, grassland, wetland, desert, agricultural and other ecosystem types.
The watershed approach has been a central component of the Baltimore Ecosystem Study (BES) since its inception. We hypothesized that this approach would allow us to establish a basis for comparing the relatively novel urban, suburban and exurban ecosystems that we were studying with the more well studied less human-dominated systems in the LTER network. Would water and nutrient retention by biogeochemical processes in these novel systems be similar to freshly disturbed, highly leaky forest ecosystems or to more mature, aggrading forests dominated by young, actively growing vegetation? The watershed approach also provided an opportunity to develop theories to guide analysis of the role of humans as components of ecosystems. Would human fluxes of water and nutrients be independent of and/or overwhelm natural fluxes? Are there coherent feedback relationships between environmental changes and human actions that function to maintain a quasi "steady-state" in urban watersheds? Finally, how does the built environment and human social processes co-evolve with abiotic and biotic components as components of developing ecosystems to function as adaptive processes that underlie sustainability?
Establishment of long-term watershed studies in Baltimore

Figure 1. Watersheds of Baltimore City and County.

Baltimore City and County are ideally suited for watershed ecosystem studies. Three principal watersheds span the urban region: the Gwynns Falls, Herring Run and Jones Falls (Figure 1). Watershed studies in BES have focused on the Gwynns Falls, which has headwaters in suburban Baltimore County, traverses older (1950s) suburban areas, enters the northwest corner of Baltimore City, and drains into Baltimore harbor just south and west of the Inner Harbor (Figure 2) (Doheny 1999). We established four longitudinal long-term sampling stations on this stream: (1) Glyndon,in the headwaters; (2) Gwynnbrook, approximately 25% downstream; (3) Villa Nova, at the mid-point of the watershed; and (4) Carroll Park, near the confluence with the harbor (above tide). We also established an array of additional sampling stations to provide land-use contrasts in similar-sized watersheds including: (1) Pond Branch, a forested reference site; (2) Baisman Run, an exurban site with low-density residential development served by septic systems; (3) McDonogh, an agricultural watershed; and (4) a series of urban watersheds (Dead Run, Gwynns Run, Rognel Heights Storm Sewer Outfall, Maiden Choice Run).
Continuous data on stream stage and stream discharge are collected by the USGS. USGS protocols for data collection and processing are well developed and widely used, allowing for comparison of the data collected at our sites with watersheds across the world. Data from most of the gages utilized by BES are available in "real time" (e.g., http://waterdata.usgs.gov/usa/nwis/uv?01589352).

Figure 2. Land use (top) and long-term sampling stations in and near the Gwynns Falls watershed.

Weekly "grab" sampling of stream water for water quality analyses began at most sites in Fall 1998. Automated samplers have been added at several sites to provide flow proportional sampling along with the weekly grab sampling. Weekly analyses include nitrate, phosphate, total nitrogen, total phosphorus, chloride and sulfate, turbidity, temperature, dissolved oxygen and pH. Cations, dissolved organic carbon and nitrogen, E. coli and contaminants such as metals and pharmaceuticals have been measured on selected samples. Data are made publicly available through the BES web site within one year of collection.
The long-term "signal" in Baltimore watersheds

Figure 3. Nitrate concentrations in streams draining forested, suburban and agricultural watersheds in the Baltimore metropolitan area sampled weekly from October 1998 through April 2010.

The signature long-term watershed graph for BES shows nitrate concentrations in streams draining forested reference, suburban and agricultural watersheds (Figure 3). Nitrate is the most common and mobile form of reactive nitrogen in the environment and is a drinking water pollutant and a prime agent of eutrophication in coastal waters such as Chesapeake Bay. The long-term estimates of nitrate (and total nitrogen) export are compared with inputs from the atmosphere, fertilizer, food and other sources used to compute watershed input/output budgets and to calculate watershed nitrogen retention.
The long-term nitrate signals from the forested and agricultural watersheds are as expected. Forests are known to have conservative (highly retentive) nitrogen cycles, leading to low concentrations of nitrate in streams draining forested watersheds. Indeed, the nitrate concentrations in the BES forested reference watershed are quite similar to those observed at the forested Hubbard Brook, Coweeta, Harvard Forest and Andrews LTER sites. We note that, similar to some of the Coweeta watersheds, peak nitrate concentrations and loads in our forested reference watershed are in the summer, in contrast to the snowmelt-dominated Hubbard Brook site. The nitrate signal from the agricultural watershed is also not a surprise. Agricultural watersheds often have high nitrogen inputs from fertilizer and resulting high nitrate concentrations in streams. The concentrations of nitrate observed in the BES agricultural stream are similar to those observed in other agricultural streams that have been studied throughout the eastern U.S. In this small catchment, concentrations and loads are very sensitive to local agricultural practice, and the large rise in concentrations is probably related to manure application and management.
The long-term forested reference and agricultural watershed nitrate signals provide an established context, and are a platform for detailed, process-level research for the urban and suburban watersheds that are the focus of BES. Why do suburban watersheds have nitrate concentrations that are clearly higher than concentrations in forested watersheds and lower than agricultural watersheds? What mix of nitrogen sources (e.g., fertilizers, atmospheric deposition, human waste, animal waste) and sinks (e.g., plant uptake, wetland denitrification) and hydrologic processes produce these patterns? Can we develop models capable of depicting these patterns and predicting nitrate concentrations in the future that can then be verified by continued monitoring? Just as in other LTER sites, long-term watershed-scale monitoring is useful for propelling the progress of our analysis of urban watershed ecosystems.

Figure 4. Watershed nitrogen retention versus precipitation for forested (Pond Branch), suburban (Glyndon, Villa Nova, Baisman, Gwynnbrook), urban (Carroll Park, Dead Run, Rognel Heights) and agricultural (McDonogh) watersheds in the Baltimore metropolitan area from 1998 - 2011. From Bettez et al. (2013).

One of the most surprising and important results of the long term watershed studies in BES is that nitrogen retention is much higher than we expected. We expected that watersheds with as little as 10 or 20% impervious surface would be so hydrologically and biogeochemically disrupted that the capacity of plants and soils to retain nitrogen would be greatly reduced. We predicted that this disruption would reduce nitrogen retention in the watersheds and that more than 50% of the nitrogen that enters the watersheds from the atmosphere and fertilizer would be exported in the stream. Certainly, our urban, suburban and agricultural watersheds have lower retention (60 - 80%) than the forested watershed (~95%), but these results suggest that vast amounts of anthropogenically-derived N are being processed, stored and retained in urban, suburban and exurban watersheds (Figure 4) (Groffman et al. 2004, Kaushal et al. 2008, Bettez et al. 2013). This result raises a series of basic science questions about watershed hydrology and biogeochemistry as well as a series of social science questions about human activities that influence the anthropogenic fluxes, and has significant implications for management and planning efforts to improve the environmental performance of these watersheds.
We observed marked variation in nitrogen retention both within and among watersheds that appears to be driven by precipitation (climate) and land use (urbanization) (Figure 4) (Kaushal et al. 2008, Bettez et al. 2013). While retention was consistently high in the forested reference watershed (> 95%), retention in urban and agricultural watersheds ranged from 33 - 95% between 1999 and 2010. There were marked decreases in retention with increased precipitation and urbanization. If hydroclimatic variability increases with climate change, our results suggest that this will lead to marked increases in the temporal range of nitrate export from these watersheds, with the potential for very high loading rates from urbanized areas during wet years.
Literature Cited
Bettez, N. D., P. M. Groffman, J. M. Duncan, L. E. Band, J. O'Neil-Dunne, S. S. Kaushal, and K. T. Belt. 2013. Effects of urbanization and climate on nitrogen retention in coastal watersheds. Ecosystems Submitted.
Bormann, F. H. and G. E. Likens. 1979. Pattern and Process in a Forested Ecosystem. Springer-Verlag, New York.
Doheny, E. J. 1999. Index of hydrologic characteristics and data resources for the Gwynns Falls watershed, Baltimore County and Baltimore City, Maryland. USGS Report OFR 99-213.U.S. Geological Survey, Denver, CO. 24 pp.
Groffman, P. M., N. L. Law, K. T. Belt, L. E. Band, and G. T. Fisher. 2004. Nitrogen fluxes and retention in urban watershed ecosystems. Ecosystems 7:393-403.
Holling, C. S. 1973. Resilience and stability of ecological systems. Annual Review of Ecology and Systematics 4:1-23.
Kaushal, S. S., P. M. Groffman, L. E. Band, C. A. Shields, R. P. Morgan, M. A. Palmer, K. T. Belt, C. M. Swan, S. E. G. Findlay, and G. T. Fisher. 2008. Interaction between urbanization and climate variability amplifies watershed nitrate export in Maryland. Environmental Science & Technology 42:5872-5878.
Likens, G. E. 1992. The Ecosystem Approach: Its Use and Abuse. The Ecology Institute, Oldendork-Luhe, Germany.
Odum, E. P. 1969. Fundamentals of Ecology, 3rd Edition. Philadelphia, W.B. Saunders & Company.
This research was supported by funding from the NSF Long-term Ecological Research (LTER) Program. This material is based upon work supported by the National Science Foundation under Grant No. 1637661. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.