Nutrient Criteria Stakeholders Notes

March 21, 2006

Governor’s Office Building

Jefferson City, MO

1-4 p.m.

In Attendance: Bob Bacon, Environmental Resource Coalition; Karen Bataille, MDC; Dorris Bender, City of Independence; Georganne Bowman, DNR-WPP; Robert Brundage, Newman, Comley & Ruth PC; David Casaletto, Table Rock Lake WQ; Cindy DiStefano, Peter Goode, Washington University; MDC; Dan Obrecht, UMC; Mark Osborn, DNR-WPP; Buffy Santel, St Louis MSD; Peter Scharf, UMC; Phil Schroeder, DNR-WPP; Tony Thorpe, LMVP; John Waitman, City of Springfield; Gary Welker, US-EPA; Tracey Winter, DNR-KCRO.

Mark Osborn: one way we can approach the rule is to classify lakes on size, and hydrology, use etc, Dan will discuss some options for this type of classification

Prioritization and Grouping of Lakes

Dan Obrecht: (PowerPoint presentation) How do we set nutrient criteria for more than 400 lakes that vary in size, watershed, hydrology and use?

Use a flow chart type of division. In this presentation there are two sets of numbers of which to be aware. The first is the 147 MU tested reservoirs, and the second is the 457 classified lakes.

Group 1 Big River remnant lakes (oxbows and blew holes)

• Set the criteria knowing that these lakes are nutrient rich due
• Shallow depth
• Difficult to define watershed
• MU has sampled 12 of the 15-25 classified oxbows
• Median values – TP 186, TN 1340

Group 2 Large reservoirs (>2000 acres)

• Set individual criteria for each reservoir
• Set multiple criteria to account for longitudinal gradient
• MU has sampled all the large reservoirs – however there is a difference on the classified list (those classified as L2 reservoirs) to the MU list.

Group 3 Small Reservoirs (<20 acres in size)

• Set criteria realizing that these reservoirs probably have a short residence time and internal loading issues
• Very little data on this group
• MU has data on 8 out of 118 classified lakes
• Median values – TP =24, TN=590
• Bias in the sampling (6 of 8 are in the Ozarks with 55% forest)

Group 4 Shallow lakes (< 1 meter deep)

• 0 in MU data set and Unknown Classified meet this criteria
• There isn’t sufficient data to make a group based on depth

Group 5 Reservoirs with Municipal and permitted point source pollution

• The NC should be set with TMDL considerations
• Only 4 are sampled by MU (out of 16?)

Group 6 Urban watersheds (> 50% urbanized)

• Most of the data is in the KC area
• MU data on 9 lakes (unknown classified)
• Bias in the sampling – chain of lakes
• Each lake becomes a settling basin for the next lake

Group 7 Regional breakdown (all other reservoirs)

• 200-300 classified reservoirs (no data available)
• 102 reservoirs in the MU data set
• Ozarks (n=37) Plains(n=65)

Group 8 Protected reservoirs (currently have very little impacts)

• Oligotrophic
• TP is < 22ug/l on the long-term average
• Not all reservoirs have low TP, TN and CHL

Group 9 Drinking Water Reservoirs (79 total)

• MU data on 5 in Ozarks, and 33 in Plains
• Not any difference between drinking water reservoirs and other regional lakes.

Hydrology and Phosphorus

TP values correlate to flushing index (R-squared of 0.51) in reservoirs located in the Plains Region. While a relation exists, there is still a fair amount of scatter in the data which limits the usefulness of the regression. We can divide the data into groups with pseudo-similar flushing indexes (the cut-points presented were actually based on having approximately the same number of reservoirs in each group) and calculate the range of TP values for each group. We can then develop a target TP value for each group based on a selected percentile of the ranges (25 th percentile was used in presentation). If we run a regression on the 25 th percentiles against the average flushing index for the reservoirs within the individual groups we get a strong correlation (R-squared = 0.93). Using this regression equation we can set criteria for reservoirs based on flushing index. A reservoir that has a high flushing index (or short residence time) would be allowed to have a higher TP level than one that has a low flushing index (or long residence time).

In the Ozark Region TP and flushing index show a weak relation (R-squared = 0.15) and the above method for setting target TP levels has limited value. There was only a small difference in the target TP values (21 VS. 14 micrograms per liter) for Ozark reservoirs with flushing indexes of 4 and .33 (residence time of 3 month and 3 year, respectively).

Hydrologic flushing index for each reservoir was estimated using regional runoff coefficient, watershed area and reservoir volume. (Matt Knowlton)

GROUP RESOLUTIONS ON LAKE GROUPINGS

Groupings are a good approach with the exemptions of point sources and urban watersheds. The <20 acre lakes and the shallow lakes should use the residence time/hydrology to set the nutrient criteria value.

Remove all lakes <10 acres from the criteria at this time (EPA guidance). More data is needed, but this group contains some Public Use areas, and we don’t want to dismiss them.

Discussion of rule –

Monitoring component – The latest version of the rule will include the current data collection procedures for lake data. This will enable stakeholders the opportunity to discuss what sampling and data requirements are necessary to list/delist lakes and reservoirs in the future. (large reservoirs containing more than one sampling location will be considered separately)

Sampling design –

Spatial - all MU data is currently collected from the Dam site. This may be the best area for comparability. All the watershed inputs and lake processes have occurred when the water reaches the dam. Large reservoirs often have more than one site. This will make compliance issues more difficult. Is the reservoir out of compliance when one arm/tributary or one area is higher than the nutrient criteria value set for that reservoir? Also, we need to consider the most critical point as the waterbody changes from river, to transitional to lake water. One option is to compare by hydrology and watershed area. Perhaps define the worst case scenario and sample more frequently/more critically in those situations.

Temporal – The lowest variability has been July to August. This is normally the driest time of the year. Springtime measurements will document an increase in nutrients with rainfall events, and runoff from winter pollution (animal waste, salt, etc). The Lakes of Missouri Volunteers sample during the growing season (April though October) It is important to sample when the protection of aquatic life is most critical. Fish mortality most often occurs August through September. Don’t sample 24-48 hours after a precipitation event.

Frequency – Statistical data has shown that 3-5 years of samples is needed to determine trends in the lake. Long-term average needs to include X samples for X years. Perhaps values more than 1 degree away from the standard deviation would warrant additional monitoring/investigation.

Implementation - Critical Sampling period needs to be determined. Advisory levels > Trigger amounts > Response variable > protection of aquatic life.

Tier 1 – lake shows obvious signs of water quality degredation

Tier 2 - lake is thriving, set the criteria value here. Use anti-degredation to maintain this level. (causual > response issue)

1. Georganne – review document from Washington and report back to group
2. Dorris Bender – provide comments letter from Independence with reference to Arizona’s rule.
3. Dan – compare bootheel reservoirs to Big Rivers lakes to determine if they belong in the same group
4. Dan – compare the growing season data to summer data to determine sampling regime.