(...CONTINUED FROM PREVIOUS POST)
The following is essentially what I sent to the manufacturer and talks about my analysis of the data. I would have preferred to have discussed this in private, but don't see much choice but to publish this [EDIT] since I haven't heard from them in a while [END-EDIT] and I'd like to continue to get feedback, more data, and do studies to confirm or refute my hypothesis. This was not my first E-mail to them, but it was the most recent where I used actual calculated HOCl values instead of a proxy estimate (FC/(CYA+1)), plus I thought more about what might be going on.
-----------------------------------------------------------------------
PINELLAS COUNTY STUDY DATA WITH HOCl
Attached is the spreadsheet of the Pinellas County 1992 commercial pools study where I have calculated the disinfecting chlorine (hypochlorous acid, HOCl) concentration in ppm Cl2 units (as all chlorine is measured). The spreadsheet has the 486 pools in the study sorted from highest to lowest HOCl concentration. Though the conclusions in the study are largely based on overall correlation analysis, I don’t think that this sort of analysis is particularly useful for this kind of data because the correlations aren’t strong enough to be used with reliability and safety. The mechanisms for chlorine inactivation of pathogens and inhibition of algae growth are more like step functions so if the data doesn’t show that kind of behavior, then the model is wrong and there are missing parameters that are not getting measured. For the Pinellas County data, this seems to be true both with the study assertion that FC is the sole parameter to be used and with my assertion that the calculated HOCl from pH, FC and CYA should be used. Neither approach is valid for this study data, so something is missing.
LOOKING AT GROUPS, NOT CORRELATION
To see this more clearly, let’s take a look at a percentage grouping analysis which would normally make any sort of “step function” much more readily apparent. The groupings were done using geometric (logarithmic) sizes since it is well known that chlorine inactivates mostly in a linear fashion (i.e. CT) so that doubling the chlorine should double the rate of inactivation. Something approaching a step function is to be expected because the amount of chlorine that inactivates roughly half of the pathogens (or algae) in the time it takes for the pathogen to double in quantity (i.e. the generation time) is a “critical” amount of chlorine. Below that amount, the number of pathogens is expected to increase while above that amount it is expected to decrease. So an “S” shaped curve is expected (actually, an upside-down “S”, so more like a “Z”, if chlorine increases on the X-axis and pathogen counts increase on the Y-axis).
The following table shows the percentage of failures (HPC>500, TCOLI,FCOLI>0, NCOLI>200, BK,YL,GN,PK<>”N”) in each grouping that has “Count” data points.
Count .. FC ...... HPC ....... TCOLI .... FCOLI .... NCOLI ... PSEUD .. TSTAPH . FSTREP .... BK ......... YL ........ GN ....... PK
49 ... < 0.1 .... 48.98% ... 36.73% . 20.41% .. 57.14% .. 6.12% ... 6.12% .. 2.04% .. 48.98% . 14.29% .. 0.00% .. 0.00%
. 4 .. 0.1-0.19 . 25.00% .... 0.00% ... 0.00% .. 75.00% .. 0.00% ... 0.00% .. 0.00% .. 75.00% ... 0.00% .. 0.00% .. 0.00%
14 .. 0.2-0.39 . 14.29% .... 7.14% ... 7.14% .. 14.29% .. 7.14% .. 28.57% .. 0.00% .. 42.86% . 14.29% .. 7.14% .. 0.00%
35 .. 0.4-0.79 .. 8.57% ... 11.43% ... 5.71% .. 20.00% .. 2.86% .. 11.43% .. 0.00% .. 34.29% ... 5.71% .. 2.86% .. 2.86%
44 .. 0.8-1.59 .. 9.09% .... 6.82% ... 4.55% .... 9.09% .. 2.27% .... 2.27% .. 0.00% .. 43.18% ... 6.82% .. 4.55% .. 0.00%
111 . 1.6-3.19 .. 4.50% ... 6.31% ... 4.50% .... 9.01% ... 0.00% .... 7.21% .. 0.00% .. 41.44% ... 7.21% .. 0.90% .. 0.90%
189 . 3.2-6.39 .. 4.76% ... 5.82% ... 3.70% ... 12.17% .. 0.00% .... 3.17% .. 0.53% .. 34.39% ... 5.29% .. 0.00% .. 0.00%
29 .. 6.4-12.79 . 0.00% ... 6.90% ... 3.45% .... 0.00% ... 0.00% .... 3.45% .. 0.00% .. 20.69% ... 3.45% .. 0.00% .. 0.00%
11 ... > 12.8 .... 0.00% ... 9.09% ... 0.00% .... 0.00% ... 0.00% .... 0.00% .. 0.00% .. 18.18% ... 0.00% .. 0.00% .. 0.00%
486 . Overall .... 9.88% ... 9.67% ... 5.76% ... 15.84% ... 1.23% .... 5.56% .. 0.41% .. 37.65% ... 6.79% .. 1.03% .. 0.41%
Now let’s look at the same table using the sorted table based on HOCl.
Count ... FC ............ HPC ...... TCOLI .... FCOLI ..... NCOLI .... PSEUD .. TSTAPH .. FSTREP .... BK ......... YL ......... GN ......... PK
58 .... < 0.001 ...... 43.10% .. 31.03% .. 17.24% .. 51.72% ... 5.17% ... 5.17% ... 1.72% .. 50.00% .. 12.07% ... 1.72% ... 0.00%
10 . 0.001-0.0019 . 30.00% .. 20.00% .. 20.00% .. 50.00% ... 0.00% .. 10.00% ... 0.00% .. 70.00% .. 10.00% . 10.00% ... 0.00%
22 . 0.002-0.0039 ... 9.09% .... 4.55% ... 4.55% .... 9.09% ... 0.00% .. 13.64% ... 0.00% .. 59.09% .... 9.09% ... 9.09% ... 4.55%
44 . 0.004-0.0079 ... 2.27% .. 13.64% ... 6.82% ... 13.64% .. 4.55% .... 9.09% ... 0.00% .. 36.36% .... 4.55% ... 0.00% ... 0.00%
92 . 0.008-0.0159 ... 7.61% .... 6.52% ... 3.26% .... 9.78% ... 0.00% .... 2.17% ... 0.00% .. 39.13% ... 8.70% ... 1.09% ... 1.09%
94 . 0.016-0.0319 ... 4.26% .... 5.32% ... 2.13% ... 12.77% .. 0.00% .... 6.38% ... 1.06% .. 24.47% ... 8.51% ... 0.00% ... 0.00%
55 . 0.032-0.0639 ... 3.64% .... 3.64% ... 3.64% ..... 3.64% .. 0.00% .... 0.00% ... 0.00% .. 32.73% ... 1.82% ... 0.00% ... 0.00%
28 . 0.064-0.1279 ... 0.00% .... 0.00% ... 0.00% ..... 7.14% .. 0.00% .... 0.00% ... 0.00% .. 25.00% ... 3.57% ... 0.00% ... 0.00%
83 .... >0.128 ......... 4.82% .... 8.43% ... 6.02% ... 10.84% .. 1.20% .... 9.64% ... 0.00% .. 40.96% ... 3.61% ... 0.00% ... 0.00%
486 ... Overall ......... 9.88% .... 9.67% ... 5.76% .. 15.84% ... 1.23% .... 5.56% ... 0.41% .. 37.65% ... 6.79% ... 1.03% ... 0.41%
174 ... < 0.011 ...... 20.69% .. 18.39% .. 10.34% .. 27.59% .. 2.87% .... 6.90% ... 0.57%
312 .. >= 0.011 ...... 3.85% .... 4.81% ... 3.21% .... 9.29% ... 0.32% .... 4.81% ... 0.32%
312 ... < 0.03 .................................................. .................................................. .......... 39.10% .. 8.97% .. 1.60% .. 0.64%
174 .. >= 0.03 .................................................. .................................................. ......... 35.06% .. 2.87% .. 0.00% .. 0.00%
You can readily see something that looks like a bit of a step function in the first one or two categories, but this step function does NOT go to 0 above a certain chlorine level (except FC>5.0). The earlier Pinellas County research study of about 1500 pools over 8 years (1973-81) stated the following:
They showed that as long as a minimum of 1 ppm of free available chlorine was present, algal growth was controlled and coliform bacteria were absent, even for pools containing up to 800 ppm of cyanuric acid. In addition, they showed that pools maintained the desired minimum of 1 ppm of free available chlorine more readily when cyanuric acid was present.
The 1992 study said the following:
The results of this study verified the results of the 1973-81 Pinellas County study.
Of course, that is simply not true since the first study claimed that 1 ppm FC was sufficient when clearly it is not. In fact, the 1992 study also concludes:
The free chlorine standard of 1.0 - 5.0 ppm of the Florida swimming pool code increases the probability that a swimming pool will be bacteriologically satisfactory for swimming. Hence, this standard is a more effective than the 1.0 - 3.0 ppm free chlorine standard employed by some of the other state regulatory agencies.
The main problem, of course, is that a free chlorine standard of 1.0 – 5.0 ppm says nothing about what FC level to actually use. Should it be 1.0? Should it be 5.0? Something in between?
Also, pool #318 had an HPC of 31000 and an NCOLI of 320 (and 2 TSTAPH) and yet had 5 ppm of Free Chlorine with a pH of 7.2 and no Cyanuric Acid. Just looking at HPC alone, there were 5 pools with an HPC>500 and FC >= 5.0; and a total of 9 pools (i.e. 4 more pools) with an HPC>500 and FC >= 4.0. How can one possibly have a standard or claim that FC alone makes a pool safe when so many pools are bacteriologically unsatisfactory?
I also find it very odd that pools with no measurable FC have lots of bacterial counts, but no green algae. Could the bacteria be consuming nutrients needed for algae? The study showed that some pools with no FC and lots of bacteria still had nitrates — but we don’t know about phosphates. Anyway, I just thought it particularly strange. Could it be that there is some use of algaecides in these pools (and maybe other pools as well)?
(CONTINUED ON NEXT POST...)
Bookmarks