DRAFT MEETING NOTES
Lake Koocanusa Technical Subcommittee Notes
October 25-26, 2016
Cranbrook, BC
Meeting Objectives
· To review and discuss 2016 research and monitoring activities on Lake Koocanusa;
· To Present Karen’s Conceptual Se Model.
· Identify and prioritize data needs for 2017.
Technical Subcommittee Meeting Highlights
· The presentation of a Selenium Conceptual Model Framework by Karen Jenni
· Three key areas of research (further study) needed to move forward with the proposal of a protective selenium number for Lake Koocanusa (actions 2-4).
Actions
Action / Who / TargetCompletion Date /
1 / Add a monitoring node in Canada (directly south of the Elk input) and determine monitoring needs at that node. Compile available information (fish tissue/macro/phyto and zooplankton and water chemistry) that could be used in model. / Sokal/Naftz / December 2016
2 / Kd study for 2017, with a node added in Canada. Study will collect suspended sediment (via large volume water samples and sediment traps) and dissolved selenium at three nodes (2 in US, 1 in Canada) and zooplankton for selenium concentrations. / Naftz/Mavencamp/Sokal/Presser / December 2016
3 / Whole body or muscle to egg/ovary conversion factors. How do conversion factors change pre/post spawning , what is the variability w/in fish communities, what knowledge is necessary for us to be comfortable using conversion factors in the Se model? / Skorupa/ Miller / February, 2017
4 / Food Web for Key fish Species identified by Technical Subcommittee. / Mavencamp/Selch / February, 2017
5 / Joe Skorupa to send a summary to Committee Chairs on Bird Expedition 2016 for sharing out with the group / Skorupa/Mavencamp/Sokal / Jan 2017
Goal - discuss 2017 sampling priorities to fill out the Se conceptual model. Karen would like to discuss the pieces of the model we need to customize (species, locations, timing).
State of the Lake document - a compilation of data. A snapshot. Identify data gaps for future work, and how to move forward with ongoing monitoring.
Jon - what would KNC's role be in the subcommittee and in future sampling? KNC has input into future sampling programs.
Mike - Jesse and Katrina were participating as observers.
Jon - KNC has lots of local knowledge with respect to sampling efforts
Jesse has converted Teck’s data into excel documents.
Draft information sharing document for internal use?
Locations for modelling
· Do we need to model multiple locations? Or just one big "pond"
o David N - err on the side of too much data. Is monthly frequency (on the US side) enough?
§ Three Nodes: south of elk including a mixing zone where the elk comes into the reservoir - riverine
§ International border – transition area
§ Forebay – most lake-like
o Greg - Forebay and international border sites are historic sampling locations and have long-term data set- including selenium, but selenium data with a lower reporting limit LRL ~1 only since ~2012 (same as for Total recoverable selenium)
o Marco - have seen differences between sites historically at sites. April - October sampling period - why was that chosen? Safety concern?
o Karen - for conceptual model, don't need to go above Elk River.
o Downstream of the Elk, during low water still acts like a river. The "backwater" area.
o Mike - need 3 stations: river like, full lake, transition area – at a minimum for model
o Theresa - need matched data sets to run the model. Is it possible to get matched data sets at these sites?
o Karen - fewer modelling locations the better. But monitoring should occur at more stations.
o Terri - were Kd values, etc. different up and down the reservoir?
o Theresa - the kds are acting like a reservoir in the north part of the U.S.
o Carla - Dec - April is more riverine …. Rest of the year is more reservoir like.
o Joe S - any value in a node below the dam in a true river to provide context? Even though we are focusing on the reservoir.
o Karen - it might help us understand loadings and how much Se is trapped. But maybe not as a node.
o Turbidity rarely comes through the dam.
· Data Needs:
Data GAP (Lana) - particulate data in Canada
Data need in Canada: suspended sediment matched to dissolved selenium in the water column for calculation of Kd’s that can be compared to the other two nodes.
Ideally the Kd’s would be matched with the other data (food web data) taken at the same locations and the same times.
GAP (Lana) - invertebrate data in US
US side does not have invertebrate Selenium concentration data – does have zooplankton, selenium data from 2016- data still pending a full QC. Community analyses have been done on US side, though data is largely from the 80’s (e.g., Chisholm paper).
Does Canada have selenium concentration data on individual portions of the food web i.e., zooplankton, aquatic insects, terrestrial insects, etc.
· Timing
· Build models for different times at the nodes that represent dam conditions
· Key biological piece is relationship of concentrations to spawning ?
· Food chain … is the diet work (expensive) actually worth it? Suggestion - use data from mid 80s and do a sensitivity analyses. And some opportunistic sampling when are taking other lethal samples and use it to validate the assumptions from the previous study.
· Need some Se concentrations in macro inverts - this would take a lot of effort and money
· How much do we need to know to use literature TTFs? Just need the Kd to predict Se levels in all other levels. Then just need observed values to validate the prediction from the model.
· Bird eggs reflect Se conditions 2-5 days before egg laid. (vs. fish that can be months to years). Fish and bird egg Se can match up really well when they feed on the same food source as they build their eggs.
· Kd
· Most sensitive part of model.
· Have suspended data from 2015 to 2016 from 2 stations in the US, International border and Forebay. Could calculate Kds for the Canadian site using sediment grab samples and dissolved Se concentrations (would not be directly comparable to the suspended sediment/dissolved Kds) but DEQ has bottom sediment data from the two US stations that could be used for comparison.
· Very large Kd difference observed using the two methods.
· Theresa - bottom sediment represent long-term system, suspended sediment an active shorter term system
· TTFs
· Can we use literature values? We have an indication of food webs from studies in the 80’s
· We can validate tissue Se using data we have already (validate invertebrate TTFs w/ Teck’s data)
· Toxicity (to ovary from whole body)
· Use literature values from USEPA Se criterion document, as we don't have any site specific data.
· Need to "de-homogenize" info to make it site specific. "un-pool" the data and use the species specific data (available in EPA docs)
· Joe S - even if we have a target ovary concentration we want to protect, we have to translate to whole body. These conversion factors are species and site specific.
· What percentile of the Conversion factor do we use (EPA used a median, is this sufficiently protective? Joe S)
· Where are we the most uncomfortable, if we populate the model with the data we have now?
· Kd
o Lana, David - think Kd is important to focus on.
o What's the size distribution, DOC? Help explain the variability in the Kds (David N)
· Whole-body to ovary conversion factors that are species and site specific (Joe S). Need to measure Se in multiple tissue compartments in fish captured in the reservoir.
o Need to know the range of variability within the fish community. Then can choose a factor to protect a certain proportion of the communities/individuals.
o How conversion factor changes with pre-, post- spawning, etc. It is published.
· Theresa - concerned about the constructing the actual food web, for the species that we pick. Get fish people to help us out.
· Trevor - have been looking at food habit study - time consuming to replicate, but could focus on specific studies.
· David - is the food study a good research study for someone from U of Mont. Will be expensive
· Erin - is helpful to have specific research questions that can be shared at U of Mont, and people could write grants to get the money
· Jon - how does life history affect model? Maybe there are other things we can use as better surrogates to determine egg levels in Se (e.g. bird eggs to fish eggs relationships)
· Trevor - separate out daphnia? They seem to eat a lot of these.
· What species should be considered for the model?
· Considerations
o Present in lake
o bioaccumulate
o sensitive - do we want range or just most sensitive
o social/cultural importance
o exposure pathways
o Site fidelity
o Abundant and well distributed
· Top 3-5 fish species
o Lana - WCT, LNSC - link to ER work; burbot - cultural importance a benthic, peamouth/shiner or pikeminnow - different place in food web, abundant, kookanee - we have data, different piece of food web,
o Joe S - peamouth, red shiner, burbot, bull trout, mountain whitefish
o Theresa - small fish, long or large scale suckers, bull trout
o Trevor - cutthroat is a good one, but not many in the reservoir, rainbow trout* could be another sensitivity, burbot, peamouth (high accumulator- but is less sensitive), kokanee had the lowest variation for Se levels- very low diet variability, mainly eat zooplankton (Dalbey et al., 1998)
o Jon - peamouth, kokanee (also get the ovary samples), WCT travel lots but are an important species, mountain whitefish also display high variability in selenium tissue concentrations. Yellow perch and kokanne are introduced, which may make them easier to sample.
· Wrap up
o Had 3 areas of uncertainty to address next year. 2 have to do with fish.
o 3-4 modeling nodes
o Shrunk the list of fish a little
· Who is taking Kd samples? Vertical profiles? Large volume or traps? What are our priorities with Kds
o Dave - Kd - need to be more efficient. Use continuous centrifuges in the field (a capital investment). Would allow us to sample more depths. Let lake profile guide where we sample that month (i.e. where turbidity is high). Sediment traps keep them further off the bottom to avoid contamination from bottom sediments. Large volume better than sediment traps. When? Target before during and after runoff, then decrease frequency. If we collect more mass, can also look at grain size, carbon, etc.
· Will have follow-up discussions (conference calls) on the 3 areas that will lead to field work. Need field season plans end of Jan.
· Three Field season Plans:
o Kd (USGS/DEQ/MOE) Will Canada do their own site or have USGS do it?
o Food Webs (DEQ/FWP)
o Conversion Factors (DEQ/FWP)
Next face to face meeting - tail end of MRC. (Tentatively in February 2017)