This folder contains the following correspondence:

   Reply received from Paula Kehoe, March 9, 2005, proposing a 'business-as-usual' monitoring plan, and offering to arrange a meeting to discuss this plan

Paula:

I receive quarterly reports on water quality in Lake Merced from David Dingman, PUC's Water Quality Chemist. Unfortunately, David has a full plate, and it usually takes another three months for him to get the report prepared and distributed.

The most recent survey, conducted in October 2004, produced a very alarming result. Coliform counts were way up, especially in North and East Lakes, and E-coli, the bad stuff, recorded levels at or near the maximum permissible for contact recreation in all lakes, and twice that maximum in North Lake. I realize that the measurement technique is subject to large errors. However, the simultaneous increase in all of these measurements is unlikely to be attributable to measurement error.

In short, then, we have a potentially serious problem in that E-coli levels can and do periodically exceed safe limits for recreational use, and clearly a quarterly monitor, reported three months later, is not sufficient to assure public health and safety.

I suggested to Greg Bartow that this sampling frequency should be increased, and offered to try to develop volunteer support for gathering samples on a more frequent basis. I am pasting the exchange with Greg below. Put in simple terms, Greg's response comes down to this:

1) Volunteers are incompetent and unreliable, and are unlikely to provide any useful service, at least over the long run.

2) Greg doesn't feel any overriding responsibility for conditions at Lake Merced, and I should take my request to David Dingman, Susan Leal, or whomever else I can think of.

I think that both the City's health department and the Regional Water Quality Control Board should be advised of the potential health risks associated with the presence of E-coli in Lake Merced. I would rather that the PUC provide this advice, together with your proposal as to what monitoring program is needed and how that will be implemented. If necessary, however, we will advise these regulatory agencies of our concern, and request that they work with the PUC to develop an appropriate monitoring program.

Also, I hope that the PUC will designate a representative who does have a comprehensive responsibility for water conditions at Lake Merced. The recent findings at the Pacific Rod and Gun Club, the proposed study of the Vista Grande watershed, possible use of recycled water as a source for maintaining lake level, and this new finding of coliform levels exceeding safe standards, together these issues require and deserve comprehensive overview. We are now uncertain as to how that overview and coordination will be provided.

John

---

John- with regards to your email below, I have met with staff at the SFPUC to discuss your concerns.

Issues pertaining to monitoring and reporting, the SFPUC will commit to the following:

Limnological monitoring surveys will be conducted in Lake Merced every two months for a total of 6 completed surveys in each calendar year. Specifically in the 2005 calendar year surveys will be collected three times from January through June then surveys will be conducted monthly from August 2005 through February 2006 during the planned period of drinking water discharge to raise Lake levels. Bacteria samples will be collected in duplicate [versus single samples] to better estimate bacteria density and evaluate long-term trends. Also a hardcopy of the final data summary for each survey will be made available and sent to you within 30 days of the completion of the field effort.

Additionally, based on the review of bacteria data prepared by Dr. Michael Casteel from 2002 to the present the following conclusions can be drawn:

Significant differences exist between summer and winter bacteria levels in the lake over the three year period but no difference is seen when wet [winter] levels are compared to each other or dry [summer] levels compared.

The levels of bacteria in the Lake are not elevated compared to other urban waterways and are in fact are quite good given natural inputs from waterfowl and other wildlife sources.

Bacteria samples should be taken in duplicate to allow the geometric means values to be calculated to determine site variance.
The small risk of infection for boaters and fisherman at Lake Merced can virtually be eliminated if people follow some common sense recommendations as prepared by Dr. June Weintraub [epidemiologist at the SFDPH]. The recommendations include washing hands before eating food and showering if total immersion occurs. It should be noted that Lake Merced does not contain potable water and full contact recreation is not allowed. (See attachment)

Our water quality staff would welcome the opportunity to meet with you in person to discuss these issues in greater detail.
Please let me know and I will schedule a meeting.

If you have any additional concerns or comments, please contact me. Thanks PK


Paula:

Thank you very much for your prompt response to my e-mail of February 23.

I am disappointed, however, by the lack of urgency the proposed program represents in terms of addressing water quality and public health issues. Also, the analysis provided leaves much to be desired in terms of statistical validity.

I will comment on your suggested program in parts:

John- with regards to your email below, I have met with staff at the SFPUC to discuss your concerns.

Issues pertaining to monitoring and reporting, the SFPUC will commit to the following:

Limnological monitoring surveys will be conducted in Lake Merced every two months for a total of 6 completed surveys in each calendar year. Specifically in the 2005 calendar year surveys will be collected three times from January through June then surveys will be conducted monthly from August 2005 through February 2006 during the planned period of drinking water discharge to raise Lake levels.

Friends of Lake Merced, along with several other environmental groups, has opposed relying on Tuolumne River water for maintaining lake level. The “planned period” to which you refer has not, to my knowledge, been exposed to the light of public comment. Who did the planning? What is the target lake level for February 2006? What consideration has been given alternate water sources for use during this period?

Bacteria samples will be collected in duplicate [versus single samples] to better estimate bacteria density and evaluate long-term trends. Also a hardcopy of the final data summary for each survey will be made available and sent to you within 30 days of the completion of the field effort.

This proposal represents inadequate concern for public health. Testing at 60-day intervals when an out-of-limit reading has been found does not meet standard protocol for rapid follow-up testing when potentially dangerous conditions have been identified. Actual readings will be available within 24 hours of the sample collection, at the most 48 hours. Samples held longer than that without evaluation have timed out, lost their validity. Why wait a full month before releasing that information? Is there any reason that the information can't be released to the public within 2 business days? If so, what is that reason?

And why send the report to me. I neither boat nor fish, and do not have good connections to those groups. Shouldn't some alert be provided the lake’s users?

Additionally, based on the review of bacteria data prepared by Dr. Michael Casteel from 2002 to the present the following conclusions can be drawn:

Significant differences exist between summer and winter bacteria levels in the lake over the three year period but no difference is seen when wet [winter] levels are compared to each other or dry [summer] levels compared.

The data does not support this observation. The second highest recorded reading of E-coliform bacteria in North Lake occurred in August 2004 at 135 MPN. A year earlier, in September the level was just 9 MPN, and by December had risen to just 22 MPN. East Lake, which is directly joined to North Lake, shows a similar pattern. With regard to the potential of increased pollution resulting from early rains, in 2003 there was 9.2” of rain during the period between tests, in 2004 just 3.1”. No rain fell during the three months prior to the earlier test in either year.

However, as you say, data has been collected for just a little more than three years. That is certainly not sufficient history to establish either seasonal patterns or wet-year/dry-year comparisons, let alone to attempt to evaluate both. However, the fact that E-coli levels exceed maximum accepted limits for the two most recent readings indicates the possibility of a trend that deserves careful monitoring.

The levels of bacteria in the Lake are not elevated compared to other urban waterways and are in fact are quite good given natural inputs from waterfowl and other wildlife sources.

Again, no data is provided to support this observation.

Bacteria samples should be taken in duplicate to allow the geometric means values to be calculated to determine site variance.

Two points per site will not provide a reliable estimate of the geometric means. A minimum of five readings is needed to secure a reliable estimate of either the mean (geometric or arithmetic) or the spread (variance) of the data. It is likely that the second reading will better serve as insurance against accepting outliers than as a means to taking an average. Similarly, five years of testing is the minimum requirement to identify significant seasonal patterns, and at least four or five years each of wet and dry years to evaluate group-to-group differences.

The small risk of infection for boaters and fisherman at Lake Merced can virtually be eliminated if people follow some common sense recommendations as prepared by Dr. June Weintraub [epidemiologist at the SFDPH]. The recommendations include washing hands before eating food and showering if total immersion occurs. It should be noted that Lake Merced does not contain potable water and full contact recreation is not allowed. (See attachment)

Swimming is not allowed. Contact recreation is an existing beneficial use. There is nothing in the Regional Water Quality Control Board’s (RWQCB) definition of contact recreation that distinguishes between “full” and “partial” contact recreation. See http://www.lakemerced.org/Plans/Water/Beneficial_uses_of_Lake_Merced.html, or visit the RWQCB site if you prefer.

With regard to potable water, it is certainly true that most people would not like to drink Lake Merced water fresh out of the tap. However, the lake is supposed to contain potable water, and did contain potable water before the PUC took it over. Should the PUC now be rewarded for letting the quality of the water decline? Next year we could in a similar way state that Lake Merced does not contain water that is suitable for contact recreation. In fact, in both August and October that statement could be made quite accurately. Solution: remove the beneficial use; far cheaper than cleaning up the water.
In what ways has June Weintraub’s memo been published? Has any attempt been made to contact the users of the lake with this advice since the unprecedented high counts of E-coli bacteria have been recorded?

Our water quality staff would welcome the opportunity to meet with you in person to discuss these issues in greater detail. Please let me know and I will schedule a meeting.

If you have any additional concerns or comments, please contact me. Thanks PK

John

---

Paula:
 
I do appreciate your invitation to meet with a group at the Public Utilities Commission to discuss the water quality monitoring program at Lake Merced.  I'm sorry that our calendars don't quite mesh for the rest of the month.
 
At the same time, however, I don't think that we are close enough to being on the same page to warrant the full assemblage you suggested.  There are a number of issues that can and should be addressed before such a meeting can be productive.  I think that I raised them all in my earlier response to your e-mail, but let me restate them here, slightly rearranged to facilitate follow-up.  I won't check my notes, so let's see if I come up with the same list.
 
First let me say that the proposed monitoring program is quite adequate so long as normal conditions of water quality prevail.  However, recent readings of fecal coliform (E-coli) bacteria are not consistent with historical performance, and may, and according to advice I received from the PUC do, exceed acceptable limits for contact recreation.  The main question is, What is an appropriate response when conditions are encountered that exceed limits set for public health and safety?
 
Parenthetically, the number I was provided is 104 MPN/ml (Most Probable Number per millileter) for E-coli as the acceptable limit.  A review of published standards for fresh water lakes (http://www.dhs.ca.gov/ps/ddwem/beaches/freshwater.htm) suggests that this standard may apply to enterococci bacteria, a parameter not measured with the existing monitors, and may be too stringent a test for E-coli.  However, until a new standard has been accepted I will continue to work with the one provided.
 
I think that the issues fall into three major categories:
 
1) Testing protocol when a threat is identified:
 
Bi-monthly monitoring has been proposed.  That is certainly adequate during normal periods.  However, the literature generally agrees that monitoring less often than weekly is unlikely to be sufficient when out-of-standard conditions have been identified.  I was referred to one study, done in Los Angeles but which I haven't yet been able to find, that concluded that testing less than weekly is so unreliable as to obviate the value of the tests.
 
I had suggested that volunteers, perhaps recruited from the boat clubs, might assist in grabbing these samples.  That suggestion was first ignored and then dismissed out of hand.  Perhaps it should be revisited.
 
2) Procedure for pubic notice
 
There are two issues  here.  First, the proposal was that public notice would be provided within 30 days of conducting the test.  Again, this is certainly adequate when normal conditions prevail.  However, it takes just two days at most to complete the E-coli monitor and determine the result.  Putting that finding on the shelf for nearly a month after out-of-standard conditions have been identified is completely unacceptable.  There is no reason that this information should be held in confidence while the results of other tests are processed.
 
Second, the proposal is to advise me of the test results.  Once again, when normal conditions prevail I see no need for aggressive public posting.  Sending me the data to put on the Lake Merced web site seems to me an adequate step.  However, I do not represent the public, nor do I boat or fish.  When an out-of-standard condition is encountered this is not adequate pubic notice.  Provision needs to be made to get the information to the people most affected.
 
Parenthetically, while June Weintraub's advisory may be quite useful, and as someone I think observed suggests common-sense precautions. I do not believe that this advise has been circulated broadly to the lake's users.  The organized groups, boat clubs and fisherpersons, should have access to this information.

3) Testing protocol under normal conditions
 
Finally, the proposal suggests that during normal conditions 2 readings of E-coli count will be taken, and a geometric average of those readings calculated.  There are two problems with this approach.
 
First, during normal conditions two readings are probably sufficient.  However, the main purpose here is to control against spurious, and usually excessively high readings.  If two readings are taken that are similar in value it will not matter much what averaging technique is used; the results will be about the same.  If the two readings are markedly different it is probable that one of them is simply wrong.  Averaging a good number and a bad number does not produce a good result, no matter how much finesse is applied to selecting the averaging technique.  It seems that someone has read a good book without much real understanding of the subject.
 
Second, when conditions are not normal two readings are not enough.  At least five readings should be taken, and either the mean of the logs or the geometric mean used to reduce the influence of higher values.  Again, however, values that are far outside the range of the remaining readings are quite likely simply to be wrong.
 
There were other issues in the proposal, generally dealing with the analysis conducted.  By and large there was not sufficient data to support the conclusions drawn, in one instance there was no data at all, and in one instance the data contradicted the findings.  Since there was little connection between these analyses and the proposed monitoring program we can chose to let those problems alone.  However, if someone would like to discuss them I would be happy to do so.
 
I hope that someone can take a look at these issues in the next couple of weeks, and that progress can be made before our meeting.  I will call you when I get back to town.
 
Thanks again.
 
John