The Transition of Benthic Nutrient Sources after Planned Levee Breaches Adjacent to Upper Klamath and Agency Lakes, Oregon

Executive Summary

Four sampling trips were coordinated after planned levee breaches that hydrologically reconnected both Upper Klamath Lake and Agency Lake, Oregon, to adjacent wetlands. Sets of nonmetallic pore-water profilers were deployed during these trips in November 2007, June 2008, May 2009, and July 2009. Deployments temporally spanned the annual cyanophyte bloom of Aphanizomenon flos–aquae (AFA) and spatially involved three lake and four wetland sites. Profilers, typically deployed in triplicate at each lake or wetland site, provided high-resolution (centimeter-scale) estimates of the vertical concentration gradients for diffusive-flux determinations. Estimates based on molecular diffusion may underestimate benthic flux because solute transport across the sediment-water interface can be enhanced by processes including bioturbation, bioirrigation and groundwater advection. Water-column and benthic samples were also collected to help interpret spatial and temporal trends in diffusive-flux estimates. Data from these samples complement taxonomic and geochemical analyses of bottom-sediments taken from Upper Klamath Lake (UKL) in prior studies.

This ongoing study provides information necessary for developing process-interdependent solute-transport models for the watershed (that is, models integrating physical, geochemical, and biological processes) and supports efforts to evaluate remediation or load-allocation strategies. To augment studies funded by the U.S. Bureau of Reclamation (USBR), the Department of the Interior supported an additional full deployment of pore-water profilers in November 2007 and July 2009, immediately following the levee breaches and after the crash of the annual summer AFA bloom.

As observed consistently since 2006, benthic flux of 0.2-micron filtered, soluble reactive phosphorus (that is, biologically available phosphorus, primarily as orthophosphate; SRP) was consistently positive (that is, out of the sediment into the overlying water column) and ranged from a negligible value (-0.19±0.91 milligrams per square meter per day; mg m^-2 d^-1) within wetlands of the Upper Klamath National Wildlife Refuge to 74±48 mg m^-2 d^-1 at the newly restored wetland site removed from the levee breach (TNC1); both observed in May 2009 before the annual AFA bloom. When areally averaged (13 km² for the newly restored wetlands), an SRP flux to the overlying water column is determined of approximately 87,000 kilograms (kg) over the 3-month AFA bloom season that exceeds the magnitude of riverine inputs (42,000 kg for the season). Elevated SRP benthic flux at TNC1 relative to all other lake and wetland sites (including TNC2 near the breached levee) in 2009 suggests that the restored wetlands, at least chemically, remain in a transition period after engineered blasts on October 30, 2007, restored hydrologic connectivity between lake and wetland environments. As reported in previous lake studies, ammonium fluxes to the water column were consistently positive, with the exception of two measurements at the restored wetland sites (TNC1 and TNC2) immediately following the levee breaches in November 2007. The flux of ammonia, particularly at elevated pH in the overlying water column, has toxicological implications for endangered fish populations in both lake and wetland environments. For dissolved nitrate, with the exception of a single positive flux measurement at TNC1 in June 2008 (0.16±0.02 mg m^-2 d^-1), consistently negative (consumed by the sediment) or undetectable nitrate-flux values were observed (-21±12 mg m^-2 d^-1 to undetectable fluxes due to concentrations for dissolved nitrate <0.03 milligrams per liter (mg L^-1) in both porewaters and overlying waters near the sediment-water interface). Such negative fluxes for dissolved nitrate are typical of microbial transformations, such as dinitrification (dissimilatory nitrate reduction), that benthically consume nitrate from the water column. The diffusive-flux measurements reported herein serve as conservative (under-) estimates of benthic flux, because solute transport across the sediment-water interface can be enhanced by other processes, including bioturbation, bioirrigation, groundwater advection, and wind resuspension. They do, however, provide lower bounds to indicate the potential importance of this internal solute source.

In the water column, the average molar ratio for dissolved inorganic nitrogen to phosphorus was 3.4±2.9, well below the Redfield Ratio of 16. This condition favors nitrogen-fixing primary producers like AFA. Soluble reactive phosphorus (SRP) concentrations were lower than detection limits (<0.03 mg-P L^-1) at all lake and wetland sites in November 2007 following the planned levee breaches. As indicated in previous studies, SRP concentrations for 2009 sampling trips indicate higher concentrations at the end of the annual AFA bloom (July 2009) relative to before (May 2009) for all six lake and wetland sites. This observation suggests the possibility of another limiting factor or factors. Wetland samples from the newly restored wetland areas (sites TNC1 and TNC2) in July 2009 yielded the highest SRP concentrations (0.21 and 0.20 mg-P L^-1, respectively), providing additional evidence that the newly restored wetlands remain in a period of chemical transition. Dissolved-nitrate concentrations were consistently low, with detectable concentrations (>0.03 mg-N L^-1) only observed after the annual AFA bloom. As previously reported in studies of Upper Klamath Lake, concentrations for dissolved ammonium, silica, and organic carbon, like that of SRP, were higher after the annual AFA bloom compared to before at all six lake and wetland sites.

To our knowledge, the flux of trace elements across the sediment water interface has heretofore not been quantified and reported for any part of the Klamath River Basin. Using calculations based on Fick’s law, as for nutrients above, we present what is believed to be the first benthic-flux estimates for the lake and wetland sites for the biologically reactive metals iron (Fe), manganese (Mn), and copper (Cu). Other metals discussed in the water-column section below (cadmium, cobalt, copper, nickel, and zinc) did not exhibit statistically significant gradients, so fluxes were negligible and not tabulated.

Both Fe and Mn exhibited dramatic concentration differences between overlying waters and porewaters, with the deepest porewater sample at 10 cm often being two or more orders of magnitude higher than the overlying sample. Calculations based on these gradients yielded flux estimates for Fe ranging from -0.08±0.12 mg m^-2 d^-1 in the Wildlife Refuge in May 2009 to 61±76 mg m^-2 d^-1 at site TNC2, also in May 2009. The elevated error estimate about the TNC mean flux is a reflection of small-scale (within site) variability. The disparity between Fe fluxes in the Wildlife Refuge (WET) and sites within the newly restored wetlands (TNC1 and TNC2) suggests an ongoing transition for areas inundated in 2007 by the planned levee breaches. Dissolved-iron fluxes observed at wetland sites in this study (9.39±16.74 mg m^-2 d^-1, n=12) were more variable than for lake sites (5.45±3.81 mg m^-2 d^-1, n=10). The maximum diffusive Fe flux observed in this study (61±76 mg m^-2 d^-1) generates an areally averaged benthic flux of 789±985 kg d^-1 that is comparable in magnitude to the highest measurement of riverine inputs (257 kg d^-1) determined during spring high flow in May 2009. It also exceeds previous estimates of riverine-Fe inputs by orders of magnitude (areally averaged at 0.7±0.1 mg m^-2 d^-1 during spring high flow and 0.3±0.1 mg m^-2 d^-1 during summer low flow). Like iron, dissolved-manganese flux estimates were typically positive and ranged from -0.01±0.02 mg m^-2 d^-1 in the Wildlife Refuge in November 2007 to 8.78±4.91 mg m^-2 d^-1 at site WMR, near the Williamson River delta in July 2009. The maximum diffusive Mn flux observed in this study (8.78±4.91 mg m^-2 d^-1) generates an areally averaged benthic flux of 1,756±983 kg d^-1, which exceeds by more than an order of magnitude the highest measurement of riverine inputs (13 kg d^-1) determined during spring high flow in May 2009. In summary, benthic flux of iron and manganese may serve as a significant source of these essential micronutrients to the lake water column. However, for other trace metals, including those that could retard algal growth (for example, dissolved copper and zinc), water-column and porewater concentrations are low and relatively constant across the sediment-water interface.

Benthic invertebrates are extremely abundant in Upper Klamath Lake and the Williamson River delta. Densities per site within the delta ranged from a low of 178 individuals m^-2 approximately a week after the levee breach to 32,519 individuals m^-2 in July 2009. Although chironomids dominated delta sites sampled in 2009, oligochaetes (primarily tubificids) typically dominated the benthos at other sites during the study. Given the densities of invertebrates in the ecosystem and their behavioral activity at the sediment-water interface, it is likely they have a substantial influence on solute cycling via metabolic activity and bioturbation (fig. 2, benthic-invertebrate video).

In summary, both the lake and wetland benthic environments substantively contribute to both macro- and micronutrients in the water column. The wetland areas undergoing restoration and those being used for water storage function very differently relatively to the established Wildlife Refuge. It may therefore be prudent, in terms of developing long-term management strategies for water quality in the Upper Klamath Basin, to extend examination of these historically perturbed wetlands as they transition toward more traditional seasonal function with regard to nutrient (both metal and ligand) fluxes.

Potential Management Implications

Long-term biological goals of maintaining or restoring fisheries resources in Upper Klamath Lake and the downstream Klamath River Basin below are dependent on solute transport through benthic and pelagic food webs. The base of those food webs, from which trophic transfer begins, is dependent on the availability of nutrients to primary producers. In order to meet water-quality objectives for the lake (for example, to reduce the impact of nutrient cycling on endangered fish populations), an understanding of the processes governing nutrient transport and distribution is required to explain temporal and spatial trends in benthic and pelagic community composition. The structure and abundance of these communities are in turn linked to solute accumulation within the lake and wetlands as well as solute loads downstream. In hypereutrophic Upper Klamath Lake, predominance of the nitrogen-fixing cyanophyte Aphanizomenon flos–aquae (AFA) suggests limitation by a solute other than nitrogen. Hence, this study provides measurements of benthic sources of dissolved macro- and micronutrients that may be significant relative to potentially regulated allochthonous sources. Because wetland areas became hydrologically connected to both Upper Klamath and Agency Lakes after the engineered levee breaches of October 30, 2007, this study focuses on the benthic nutrient sources from these wetlands undergoing restoration after decades of agricultural use. The rate at which benthic fluxes from these restored wetlands converge to those observed for established wetlands or adjacent lake habitats may also be of management interest. In addition to phosphorus, we examine the consistency of significant ammonia fluxes from the sediments in these wetland areas, as was previously reported for the lake. The information provided herein is intended to help evaluate proposed management and remediation strategies for the basin and appropriate response times for those remediation efforts. Deployment of nonmetallic porewater profilers, designed and fabricated for previous studies of Upper Klamath Lake has been, with minimal modifications, applied to adjacent wetland environments. Benthic-flux measurements in the wetland areas begin to locate and quantify major nutrient sources for subsequent assessments of restoration activities in the lake and basin in general.

First posted March 26, 2010