Tuesday, August 5, 2008 - 8:40 AM

COS 16-3: Latitudinal variation in nitrogen loading to coastal North Pacific lakes as inferred from nitrogen stable isotopes

Gordon Holtgrieve1, Daniel Schindler1, Daniel Selbie2, Peter Leavitt3, Bruce Finney4, and Irene Gregory-Eaves2. (1) University of Washington, (2) McGill University, (3) University of Regina, (4) Idaho State University

Background/Question/Methods

The transport of materials from terrestrial to aquatic environments is integral to the global cycling of nutrients as well as important for supporting productivity in lentic, lotic, and nearshore marine environments. In lakes, quantifying the magnitude of these linkages is an arduous task involving either long-term monitoring/mass-balance approaches or expensive stable isotope labeling experiments. As a result, there are currently few estimates of nutrient loading to lakes from which large-scale comparisons within or across geographic regions can be made. In coastal regions of the North Pacific, anadromous salmon return to lakes and streams annually to spawn and eventually die, and in so doing supply large quantities of isotopically enriched nitrogen (N) to these systems. By knowing the magnitude of this flux it is possible to utilize migrating salmon as a large-scale natural tracer experiment where watershed N flux can be determined by pool dilution of N isotopes. We used surface sediment δ15N values and records of salmon returns to determine watershed N flux to 90 lakes from coastal British Columbia to western Alaska
Results/Conclusions
Sediment δ15N was positively correlated with salmon density reflecting marine-derived N input to lakes. However, the response between salmon density and sediment δ15N enrichment differed among regions owing to differences in hydrologic regime and associated variation in watershed N loading. In lakes with spawning salmon, the proportion of total N in sediments derived from salmon ranged from 1 to 70% and was negatively related to precipitation and relative watershed size. Using maximum likelihood estimation, we tested the significance of factors controlling watershed N flux to lakes including geographic region, precipitation, watershed morphometry, and vegetation. The strongest models of N loading include positive effects of precipitation, vegetation (% tree cover), and the watershed-to-lake-area ratio as predictors. This work highlights a novel method for determining the transfer of N from terrestrial to aquatic ecosystems and indicates wide variation in N loading rates from watersheds to aquatic ecosystems across latitudes in coastal North America.