Background

Human sources now rival natural sources of fixed N to the biosphere, changing the biogeochemistry of both terrestrial and aquatic environments by adding biologically available N, primarily from fossil fuel combustion, agricultural fertilizer applications, and legume cultivation.  N pollution of the atmosphere is also one of the leading causes of acid precipitation.  The fate of much of the anthropogenic N load to the biosphere is uncertain because mass balance studies of major watersheds show that most of the added N is not exported to the oceans but is missing, having disappeared from our accounting somewhere within these watersheds.  Among the several possible fates of the missing N are uptake and denitrification in streams and rivers.  Despite this “missing N”, anthropogenic N pollution produces N loading to estuaries and coastal oceans that has resulted in increasingly severe eutrophication problems, such as harmful algal blooms on the Atlantic coast and oxygen depletion in the Gulf of Mexico.  Understanding the controls of transport of excess N from land to rivers and ultimately to the ocean is one of the major challenges for biogeochemistry and ecosystem ecology. 

Most of the inorganic N entering and transported by streams and rivers is accounted for as nitrate by many ecosystem scientists.  Unlike ammonium, which is considered to be biogeochemically “sticky”, nitrate has been thought to flow freely downstream to lakes and coastal ecosystems once it enters streams.  Recently, however, this view has been challenged.  Mass balance analyses for the Mississippi River drainage have shown that large quantities of total N and nitrate are lost as water travels through its tributary streams and rivers (Alexander et al. 2000).  Headwater streams clearly can alter the quantity and forms of N in water passing through them (Peterson et al. 2001).  Additionally, larger downstream fluvial networks contain other potential hotspots of N retention, including reservoirs, semi-impounded reaches (i.e., low-head navigation dams), floodplains, estuaries, and deltas.  The relative importance of these diverse environments in controlling the transport and fate of excess N is not well understood for any large river system, nor are the impacts of human alterations of these environments.

Existing research networks

            There are already several research networks dedicated to understanding the N problem, among which the LINX (Lotic Intersite Nitrogen eXperiment) group is most specifically dedicated to investigating the role that headwater streams play in N transformation and export.  LINX can serve as a model for future synthetic efforts throughout the LTER network.  This group first formally met in a workshop during the 1993 LTER All Scientists Meeting, during which they planned a proposal for a workshop (later funded by an NSF incubator grant) to meet at Coweeta Hydrologic Lab, NC in July 1995.  The objectives of the workshop were to conduct a preliminary stable isotope labeling project, and to plan for a full NSF proposal that would incorporate 6-week stream ¹⁵N-ammonium tracer experiments at multiple sites.  The proposal was submitted to NSF Ecosystems panel and funded from June 1996 to May 1999 for about $4 million (LINX I).  Research at 10 sites was supported and 8 out of 10 were LTER sites.  The project included numerous meetings to standardize methods and synthesize data, and has resulted in 25 publications to date, plus more than 62 presentations at scientific meetings.  At the end of the first LINX project, a second proposal was developed (LINX II) and subsequently funded by the NSF IRCEB program for $3,000,000 over a 5-year period (1/02 to 1/07).  This proposal expanded the scope of the research by funding stable isotope tracer measurements of ¹⁵N-nitrate in multiple small streams (9 total per site) in pristine, agricultural, and urban watersheds of 9 sites.  The LINX II proposal also included a modeling effort to extrapolate ¹⁵N tracer results from small streams to the watershed scale including networks of stream systems.  This group received further funding from the LTER office to hold a workshop in February 2003 at Sevilleta, NM to facilitate data management aspects of the project, growing out of an increase from data generated from 10 streams in LINX1 to 81 streams in LINX2. In addition, there is an on-going NSF-funded cross site comparison lead by Mark Williams (Niwot) is investigating concentrations, forms, analytical techniques and fluxes of DON at 10 LTER sites.

            Significant prior research on N cycling in lotic systems has also been directed at the largest rivers, with particular attention to the role of off-channel aquatic environments associated with navigation dams along the upper Mississippi, and of natural floodplains along unregulated very large rivers such as the Orinoco and Amazon in South America.  This research has often entailed a team effort but has not been previously organized into networks performing simultaneous but distinct studies in the same way that the LINX project has been structured, and has had no linkage to the smaller streams in the fluvial system.  The N biogeochemistry of estuaries and coastal oceans has also been studied in some detail, and these ecosystems are the focus of several new LTER sites, but traditionally there has been little linkage of this work to the watersheds that contribute N loading.  Recent LTER research is making that connection, however, as for example at Plum Island or the Santa Barbara Channel. 

Concurrently, synthesis efforts have been directed at very large-scale (regional to global) analyses of nitrogen transport from continents into the oceans (e.g., The International Nitrogen Initiative: A Joint Programme of SCOPE and IGBP).  Currently, the linkages between small-stream models and watershed-scale transport are not often made.  Furthermore, several models of N transport through fluvial systems exist, each with its own strengths and limitations (Alexander et al. 2003), but all suffer from inadequate knowledge of processes and mechanisms across the entire fluvial system, and hence tend to rely on correlative analyses and to focus on one end or the other of the spectrum of stream size. 

Future synthetic research

A large-scale, cross-site, synthetic effort to characterize the controls on nitrogen transport through streams and rivers would serve to unite disparate research efforts and produce a comprehensive picture of where and how the fluvial system can intercept and reduce excess N.  By originating in the LTER network, such an effort has a high probability of success because it builds on a core group (LINX) and the data and methods they have developed while conducting synthetic research on the topic for the past 7 years.  The LTER network is uniquely suited as a base for such a synthetic project, but the synthesis will obviously require expanding well beyond the traditional boundaries of existing LTER sites. 

Key Issues on N Cycling and Transport

Three workshops at the most recent LTER All-Scientists Meeting focused on issues important to this synthesis: 1) Development of coupled hydrological-biogeochemical models of materials transport at the landscape scale; 2) Methods of determining denitrification rates in lotic ecosystems; and 3) Exploring nitrogen dynamics in streams: Using models to scale up from headwaters reaches to stream networks.  Participants from all three groups have contributed to the ideas presented here for this synthesis activity.  The LTER-ASM workshops identified several key issues that need to be addressed to scale up existing information to large watersheds:

1)     Adequate hydrological modeling across spatial scales is necessary to quantify physical parameters influencing N transport, including the effects of river size, flood and drought, impoundments, surface-groundwater interactions, and channelization.

2)     Biological parameters have been or will shortly be relatively well characterized in headwater streams, but the ability to extrapolate the biological activities to large rivers needs to be tested with empirical data.

3)     Whole-stream estimates of ecosystem function will be required to account for N flux, and particularly the processes of denitrification, assimilative uptake, and mineralization.  Whole-stream methods need to be cross-validated to allow comparison of data across systems.

4)     Several different approaches have been proposed to model N transport but systematic comparisons of their performance are lacking. The approach that will best capture spatial and temporal scaling issues of both hydrological transport and nitrogen cycling needs to be determined.

The LINX stream nitrogen group as well as other LTER researchers working on hydrological and biogeochemical aspects of fluvial systems will form the central core group to get this synthesis project off the ground.  Scaling from studies of small streams up to entire fluvial systems will require greater modeling expertise and collaboration with sites conducting research on large rivers (e.g., the USGS Upper Midwest Environmental Sciences Center and participants in other large-river biogeochemistry groups).  Such an initiative would also require data from LTER sites that are not currently involved, particularly hydrological, land use, and nutrient data and the LTER network contains unique repositories of such data.

References:

Alexander RB, Smith RA, Schwarz GE (2000) Effect of stream channel size on the delivery of nitrogen to the Gulf of Mexico. Nature 403:758-61

Alexander, RB, Johnes PJ, Boyer EW, Smith RA A comparison of models for estimating the riverine export of nitrogen from large watersheds Biogeochemistry 57/58: 295–339, 2002

Peterson BJ, Wollheim W, Mulholland PJ, Webster JR, Meyer JL, Tank JL, Grimm NB, Bowden WB, Vallet HM, Hershey AE, McDowell WB, Dodds WK, Hamilton SK, Gregory S, D’Angelo DJ (2001) Stream processes alter the amount and form of nitrogen exported from small watersheds. Science 292:86-90