comments on response 1/13/04

The following e-mail was sent to Greg Bartow, San Francisco Public Utilities Commission, and Patrick Sweetland, Daly City Department of Water and Wastewater Management on January 13, 2004 in preparation for a meeting with the Regional Water Quality Control Board the following Thursday. These comments are in response to a commentary prepared by Greg and Patrick in turn responding to a RWQCB review of the original Vista Grande Canal monitorning plan.

>>>>>>>>>>>>>>>>>>>>

Greg & Patrick:

With regard to Dr. Cooper's letter: I agree with the overview that this general approach may well be a very good idea. And, as Dr. Cooper states, this "appears to be a good first step." However, appearances are not everything; the devil is usually in the details.

There is no way Dr. Cooper can conclude, based upon the plan as submitted to the RWQCB, that this Pilot Project has been well designed as there is no data provided in the monitoring plan upon which such a claim can be made. There is no indication in the plan as to how much water will be introduced, of what quality, over what time period, and with what testing protocol to evaluate the impact. In fact, the draft plan I have, dated 12/4/03, states that "Phase 2 of the pilot project is currently under design, and involves using a vegetated area . . . as a stormwater treatment wetland." In short, at that date no design was presented as no design existed.

Greg has stated that material will be provided in an operations plan; to date I haven't seen that plan, and my guess is that Dr. Cooper has not either. "Pilot Project" is itself not a scientific term. While it implies relatively low volumes in the initial stages those volumes need to be specified in order to permit evaluation of feasibility and to draw conclusions about the quality of the design.

With regard to beneficial uses: According to the RWQCB list of activities by class, either boating or fishing is considered to be contact recreation. With a quick search I couldn't find the list, but I think that it's fishing. I agree that seems strange, but that's the way the list was the last time I looked. You say, however, that "Therefore, the bacteriological criteria applicable to body contact recreation utilized for this pilot project . . . is highly conservative." I haven't seen those criteria; perhaps they are further along in the response; again I hope that they will be available when we meet.

We've been around the horn with respect to the question of potability a few times. The fact remains, however, that one such designated beneficial use is as an emergency supply of domestic, i.e., drinkable, water. "It's bad now so it doesn't matter much if we make it worse" isn't, in my view, a very good argument.

With regard to the purpose of this demonstration: Near the middle of your response you state, "The intent of the Plan is to provide preliminary screening information to determine if future, more detailed, evaluation is warranted. The intent of the Plan is not to provide an exhaustive analysis of all aspects of the pilot wetland treatment system." I'm afraid I don't know what this means. What information is needed to "determine if future, more detailed, evaluation is warranted?" What results would indicate that a next step should be taken, what would that step be, and how would that step be shaped, if at all, by results at this phase? On the other hand, what measurements would be appropriate for an exhaustive analysis that will not be included in this phase, will such an analysis be conducted before a full-scale project is undertaken (whatever that means!), and if so when? It seems to me that you are proposing an adaptive management plan without taking the trouble to detail the phases and the conditions for proceeding. That is not a complete plan.

In the same paragraph you complain that you don't think these requirements have been generally applied. "Mommy, Jimmy didn't have to do that." That is not a mature response to an effort to seek a solution that adequately protects public health and safety.

With regard to Phase 1 monitors: It would be nice if the CH2M Hill report were made available. The fact that this report was not prepared until this month indicates a high level of last-minute catch up. I've been taken to task for observing that this project has been in the planning process for at least a year, yet here is year-old data finally being analyzed in an organized way.

With regard to field testing: The response mentions testing with Quanti-trays, an approach that I believe requires two days to get results. As I described in my commentary of October 21, if the system operates at capacity for two days, and the coliform content is at the 80 per centile level as demonstrated in Phase 1 testing, then the blended quantitiy of E-coli in South Lake may be well over 3 times that prescribed as a limit for non-contact recreation. We either need faster test results or much reduced flows, or both.

A note on the sampling limits: The limits described for coliform sensitivity in Table 1 indicate the range of possible outcomes for undiluted samples. Tests should be conducted at various dilutions in order to increase the range of these results.

With regard to the testing protocol: The layout of the demonstration area makes close observation very difficult. Immediately next to the release area is a section of thick wild blackberry brambles. This is followed by an extensive growth of tules that continues several yards into the lake. Consequently neither the flow of water across the ground and into the lake, nor the distribution of water when it reaches the lake, can be closely observed.

It is then no surprise that the release of relatively small amounts of colored water onto relatively dry land would not produce an observable result outside the tule bed. However, it seems to me that this places an even greater burden on the number and location of samples to be taken. The contours of the land in this area are not likely to be uniform; as a consequence the flow rate of water into the lake will be non-uniform as well, depending on the local contour. Either sufficient dye testing should be conducted in order to identify the entry points or a random selection of sample points along the shoreline, at least seven, should be selected in order to assure some confidence that actual off-flow is being detected.

Similarly, it is this feature that argues for multi-depth testing, not the existence or lack of stratification in the lake itself. Considerable mixing is likely to occur in the tule area, and surface grabs alone outside this area may not be sensitive to this result.

If, as it seems, the tests will be taken outside the tule area then the time factor needs also to be taken into account. That is, testing should continue periodically both to evaluate delays in coliform reaching the open lake water as well as to evaluate die-off rates. The original plan indicated that "the primary goal (of Phase 2) is to demonstrate coliform die-off and nutrient removal." Again, this suggests the need for an extended time-based testing protocol.

Finally, it seems that in the past reliance has been placed on single tests. It is the nature of coliform that it comes in bundles, or colonies, and is not uniformly distributed. Again, in order to have meaningful results a number of samples, at least seven, should be taken at each sample point, with the geometric mean calculated to determine the best estimate of coliform density.

With regard to discharge rates and duration: The preliminary plan stated that "treated Vista Grande stormwater will be diverted to Lake Merced during four of the six sampling episodes." Your table, however, indicates quantities and durations for all six storm events. What was intended for the two non-diversion events that is now being neglected?

You observe that total discharge, even with the largest flow volume, will be less than 1% of total lake volume; the implication is that this isn't very much. However, the observed volume of E-coli during the Phase 1 testing demonstrated an 80 per centile level approximately 465 times your acceptable limit. That suggests that the entire lake might contain E-coli bacteria in quantities large enough to cause serious concern if the 1% volume is implemented. I think it would be appropriate, when you make seemingly dismissive comments such as this, that you follow up with an assessment of the potential consequences rather than leaving it to the reader to figure out that a 1% addition can lead to real problems.

Again, your statement of measurements and consequences continues to be vague. If limits are exceeded we'll either stop or slow down. If they are exceeded by how much for how long do you stop? If you slow down how much do you slow down? More structure in the design is needed.

Finally, if your results from last year are any guide, and your limits to be enforced, then I think we can stop right now. You suggest that an upper limit of E-coli of 400 MPN / mL will be observed. Last year, during Phase 1 testing, you observed a mean 'Beach Area' count of E-coli of 1,754; the geometric mean was 682.4. Of six readings taken, 3 were in the range 1,500 to 2,000, and one was more than 6,500.

Early in December I took three additional off-shore readings, near the Impound Bridge, with E-coli readings of 400, 410 and 630, all at or in excess of your limit. It had not been raining recently, and so the off-shore runoff associated with a storm event was probably not a factor. Mid-lake readings taken a day earlier suggested normal results with counts of 53 and 65 in South Lake.

In short, you may have shot yourselves in the foot with your own crieteria as normal readings may well exceed your limits. That suggests that you return to the protocol of measuring at most four of six storms, and perhaps take a reading from one storm with no diversion prior to each incremental test.

Again, you limits do not indicate if they are the mean, or geometric mean, of a number of tests, if there is a maximum observed reading among a sample group, or other testing protocols normal for coliform monitoring.

With regard to bio-retention capacity: Highly variable is better than no estimate at all. We know the area to be affected. We know the flora types that occupy that area. There must be some data that indicates how well wild blackberry does as a nutrient absorber. What other plants are present that might help out? Given the area and the best possible plant mix how long would the water introduced need to dwell in that area in order to remove observed quantities of coliform? Given that these tests will be conducted during and immediately following storm events, given that the ground will then be saturated or near-saturated with rain water, and given that the ground is not level but is sloped toward the lake, what retention time can be expected for water introduced during the test? Certainly some estimate can be made of each of these factors in order to demonstrate that a successful test is at least a feasible outcome.

This is particularly important as you are proposing to conduct this test with stormwater treated only with solids removal, an approach not expected to reduce coliform content. Other bio-filtration projects around the state generally rely on either secondary or tertiary treated recycled water augmented with ultra-violet treatment to remove contaminants before that water is introduced to a settling pond. I believe that in Arcata water is introduced into a settling pond before that water is in turn transferred to the open wetland. None of these precautions are included in the proposed demonstration.

You have neglected to mention that the RWQCB also asked you to comment of flow attenuation as well as pollutant removal. Again, given that the ground will be at or near saturation during these tests, and that this ground slopes toward the lake, some indication that residence time on shore will be sufficient to permit bio-filtering to take place is appropriate. Isn't it likely that some approach to flow attenuation will be required? I don't think that this question can be ignored.