Floyds Fork Nutrient TMDL Technical Advisory Committee (TAC) Meeting
MINUTES
/april 24, 2013
/9:00 AM
/floyds fork wwtp, Louisville KY
Meeting called by / KY Division of Water / Type of meeting / TAC MeetingGeneral Information: ,
Attendees / Facilitator: Paulette Akers
EPA: Chris Thomas; Tim Wool
TetraTech: Brian Watson, Madhu Smith
Note Taker (KDOW): Jenny Howard Owen, Paulette Akers
Atendees: Paulette Akers, Brooke Shireman, Jenny Howard Owen, Hui Chen, Amy Siewert, Ann Fredenburg, Abigail Rains (KDOW); Kevin Gibson (OCEA), Tara Brinkmoeller (Home Bldg Assoc.), Joshua Edwards (City of Shepherdsville), Richard Preston (KyCGA), Wayne Long (UK), Roger Recktenwald (KACo); Joe Cain (KyFB), Jay McCants (Agri Business Assoc.), Junfeng Zhu (KGS), Karen Schaffer (URS), Beth Stuber (Oldham Co.), Scott Smith (SMG/Ky Home Bldg), ShriVani Seripada (SMG), Dave Maples (KY Cattlemens), Maury Cox (KY Dairy Dev Council), Robert Klingenfus ( Dairy Dev Council), Dave Derrick (Derrick Eng), Tim Corrigan (GLI), Dale Salmon (Mt. Wash MS4), Laurent Rawlings (HBAK), Ward Wilson (KWA), Brad Lee (UK), Ben Albritton (KWRRI)
brian watson / This is our last meeting as an EPA contractor. Our contract stops May 14th. We will be a stakeholder in the future. The deliverables are the watershed model and instream model, a skeleton of the TMDL report. It will be filled out by DOW. We have provided 7 days of training, more than the 5 days we originally contracted for. Since our last meeting, there has been exciting and interesting information come out. Let’s go over what has been changing since TAC #3. Most of the time has been spent on the watershed model not the in-stream model. We have showed you the scenarios and we have revised 5 of those based on the revised information. Slide presentation.
In TAC #3 the subcommittees were formed and provided updated info for septics, sinkhole coverage. We have received a couple of questions and are currently revising language to make more clear.
Since TAC #4, we got the requested scenarios and ran them.
Since TAC #5, we updated the pastureland loading rates. The rates of cropland and pastureland in the table in revision 5. We came up with a way to determine fertilizer rates.
Cropland – the way it is calculated already accounted for manure application. Took manure out of cropland application. Numbers in report are loading rates from UK. Pastureland – 4% receiving manure, it looks skewed. In model was lower. The original numbers got moved down and Dr. Chad Lee sent an email saying that he agrees and understands how the values for cropland and pastureland were derived. They are now lower than revision 5, but have support from UK since they understand the numbers.
richard preston / The table had numbers that were printed that were incorrect. Dr. Chad Lee and I had a discussion yesterday. We need to be reasonable, but it is possible that about 15.5 pounds/year. That number seems fairly reasonable. Dr. Chad Lee was lower, but don’t want to interfere with the calibration and this seems reasonable.
brian watson / Sent email to corn growers yesterday that they agreed with values.
richard preston / Got an email yesterday from Dr. Chad Lee. Had passed info on since he didn’t feel he was an authority on pasture.
brian watson / Reading report it was a valid concern.
Cox / He said he hadn’t seen the revised report since last week.
richard preston / I think it has been resolved.
brian watson / Appeared to be the biggest concern at the last meeting. Adjust calibration a bit. Have started updating the modeling report for clarification and have run updated subset of scenarios. TMDL will probably refer to there documents. Any questions about evolution?
scott smith / When redistributed load do you know where it will go?
brian watson / We distributed loads to each station in Chapter 5. It increases the point source and MS4 load, but the load from NPS goes down a little bit. Around a 15% change. Didn’t shift as much as some.
karen schaeffer / You used hydrologic response to get ranges?
brian watson / Used to create ranges and picked percentiles within ranges for urban land uses.
karen schaeffer / The data collected by MSD from single land use drainage ditches with lots of data. From the MS4 land use, did you exclude ag? Can I get an MS4 boundary for Jefferson County?
Abby Rains / For Jefferson County it is the whole county, but that is because of the CSO consent decree here.
karen schaeffer / What land use applies?
brian watson / That is 2 separate things. Runoff from land use is exclusive of MS4. MS4 in the table is only urban land use.
tim wool / Table 5-7. For other land use types within the MS4 boundary, this table only uses high, medium and low density urban.
karen schaeffer / Maybe that should be labeled urban MS4 contribution.
tim wool / If it surrounds the farmland, then ag still contributes to the MS4.
richard preston / Is 11 pounds the atmospheric? Table 3-24.
brian watson / Percentile from MSD. It is atmospheric plus application.
scott smith / What is the atmospheric?
tim wool / National atmospheric deposition network has maps.
scott smith / I have friends in Chesapeake Bay that told me that the numbers they started with was much lower than what it ended up being.
brian watson / It could be higher or lower.
richard preston / Atmospheric could be a big component in Floyds Fork. Rates from 7-17. Need to look at it for the future.
brian watson / I can go to WASP or scenario revision. How should we present the results? Thought was to look at watershed as a whole. Have the ability to look at a watershed as a whole. Have the ability to look at a lot of different scenarios. 19 compiled originally. 5 redone. When showed last time, it was hard to see small changes. New maps show changes that occurred to baseline. Left is TN, right is TP. Green meets current endpoints, orange exceeds.
Scott smith / Implementation targets sampled every year to see if meet?
Brian watson / That is a question for DOW.
Laurent Rawlings / Is the exceedances for the 3 years of a permit or any 3 years?
brian watson / Any rolling 3 year time frame.
karen schaeffer / Targets exceeded more than maxes were.
brian watson / Most of the time if the target is violated, the max is violated. I can get a better sense for you later.
Karen schaeffer / Sure
richard preston / Is this where we are right now? Scenario 0 = Baseline?
brian watson / Just modeled to endpoints
richard preston / How does this match impairments?
Brian watson / Very well. The data itself matches very well.
Ann fredenburg / Went with segments of interest because some have not been listed or delisted yet with the data that was most recently collected.
chris thomas / Same information in 2 different ways. Listing is biology, but this is based on targets.
tim wool / It correlates well with monitoring information
richard preston / Maybe you are finding places that are impaired that aren’t listed.
tim wool / That is a circular argument. Nothing will be listed based on the model.
brian watson / Scenario 1 slide – 100% forested. If it doesn’t meet endpoints, there might be an error. Now, everything meets the targets.
tim wool / Everything of land use changed to upland forest. If there was red, it may point to a model problem or a target problem.
teena halbig / I recommended a wider buffer.
brian watson / This isn’t a buffer, but changing the entire land use to forest.
tim wool / This is an approximation of natural load.
laurent rawlings / If Garden of Eden doesn’t meet, then totally wrong. Is the Garden the same everywhere?
brian watson / We calibrated forest for this watershed.
laurent rawlings / Does this still take into account where we are?
tim wool / Yes.
pat dominik / Compared to baseline this would take everything to green?
brian watson / Yes
scott smith / On calibration is there a red line in forested area?
Kligenfus / How were samples taken?
brian watson / Collected data at variety of conditions. Model has output every day for 10 years.
junfeng zhu / Point source removed?
brian watson / This is only forest. No people.
Scenario #2 – point sources removed. Still urban land use, just load and flow removed. Many segments in baseline are just barely not meeting.
pat dominik / Best TAC meeting. Very interesting. Is this only max exceeded?
brian watson / There was processing of means and max value. Shown as red if exceeds either or both.
tim wool / If compared numbers, the model can go out to 32 decimals. If the model says off by 0.003, we probably wouldn’t worry about this. If we showed numbers it would be a hang up.
jenfeng zhu / When we removed the point source, what amount of load is removed?
brian watson / All of the point source load.
Scenario #4 – SSO removed. There is no impact to baseline if remove from the calibrated model.
wARD wilson / In Beargrass Creek, the same scenario showed nothing. The exceedance is very short in the time of the geo mean.
tim wool / Impacts of a single entity. They are causing an increase in the nutrients, just not the sole problem.
richard preston / In karst area. The SSO may be stored underground.
tim wool / These are pipes and are discrete. We know the flow and have monitoring data. They bypass during the storm. We have MSD discharge reports.
karen schaeffer / In TMDL they will be required to be removed.
ann fredenburg / SSOs are Illegal source
karen schaeffer / MSD is working to fix as part of a consent decree, but will continue as part of a TMDL.
scott smith / Shows that it wouldn’t be a big cost benefit to fix.
chris thomas / Not saying that. But overflow is really a bigger deal regarding the pathogens. This TMDL is for nutrients.
tim wool / If we determine we need a x% reduction, we can count the removal of the SSOs as that amount of reduction.
Net positive when remove point sources
brian watson / Meets target, not just incremental positive change.
tim wool / We know if we remove load it is always a positive change.
Brian watson / Scenario #5 – all facilities at permit limits
Tim wool / Says that some not hitting permit limit. Some are lower and looks like one is exceeding.
Brian watson / Scenario #18 – TN, several segments meet.
Summary slide – red means at least one segment meets the target, green means all meet. At the current end point, some combination can meet current targets
teena halbig / Means if a point source it would meet boatable.
brian watson / Yes. This says there would need to be a reduction from land use as well.
karen schaeffer / What about table differences?
brian watson / That is the 7 stations and the scenario is more than 200 segments. This is watershed wide.
jengfeng zhu / Curry’s Fork. It used to always show orange.
brian watson / Indicates a load from land.
tim wool / Anthropogenic since meets in all forest.
jengfeng zhu / Point source scenario, if remove 20% of load from point source does it meet? Can you run with removal of NPS?
tim wool / We could run scenarios without NPS but no water
jenfeng zhu / Springs point or nonpoint?
brian watson / Part of the natural balance
jengfeng zhu / If you change the concentration of N in the spring water would it make a difference?
brian watson / Yes, but springs are part of the natural condition.
shriVani / I have a question on point sources. Is this design flow or actual flow in Table 3-7?
brian watson / We will review the table label. For calibration, we used actual if it was available.
richard preston / Still interested in the interaction of surface and groundwater. Maybe we can talk privately.
teena halbig / I would like to hear that.
brian watson / When it rains, after interception, we supply an infiltration rate. What is left runs off. The groundwater is in 2 categories, shallow and deep. We lose water to the deep; it may never come back to the surface. This is in the watershed model. For springs, we used the KGS well monitoring data. May be increasing concentration, but the mass from the springs is small due to dilution.
richard preston / In karst, it is directly connected. Is this in the model? Where N from springs coming from?
brian watson / Surrounding land use. Sinkholes have a higher infiltration rate. Some may be lost, but most in spring.
ann fredenburg / Different loadings from spring versus groundwater?
brian watson / yes, dependent on land use and well data. Contribution info from KGS.
karen schaeffer / Error bar on range?
brian watson / Appendix C is sensitivity, based on loading rates. Hasn’t been updated.
karen schaeffer / Can go out to 30 decimals.
tim wool / Model a tool. Can’t do error analysis because 10 year model.
karen schaeffer / Can you estimate?
tim wool / No. We can get error on one set of conditions, but not on complete run. That’s why we do sensitivity analysis.
scott smith / Can you run through adjustments to default concentration?
Brian watson / Adjusted when didn’t meet measured in-stream concentrations. Only a few changes and described in report. No changes in those between Rev 5 and Rev 6.
Scott smith / Will there be additional locations where default are used?
brian watson / We will show pastureland and new calculation based on in-stream data, When look at specific.
teena halbig / Can you tell the number of trees that would need to be planted?
brian watson / No
shrivani / Did you account for leakage?
brian watson / No. If we had the data, but if they knew of a leak, they would fix.
scott smith / Chesapeake Bay said that info had been underestimated.
brian watson / You could see that in the USGS data. Some flow started coming into the system.
Laurent Rawlings / Forest. This is just a scenario to check.