Petitioning City Manager, City of Falls Church Mr. Wyatt Shields and 8 others

Conduct an open, independent audit of the City of Falls Church water billings

A number of Falls Church City residents recently have received water bills that are substantially higher than normal.  Many have complained to City staff and officials and have been brushed off with unjustified and unsupported explanations.

This petition :

(1) summarizes the initial public comments;

(2) recounts the City's inadequate approach to the abnormal usage problem;

(3) rebuts the City's principal explanations; and

(4) outlines an approach by which the City can employ its own data to determine the extent of the problem and focus a more detailed investigation.

The City should use a small amount of the millions of dollars gained from its sale of its water utility to fund an independent and open audit of Falls Church’s water accounts to address citizens’ concerns.  This is required to restore City credibility on this issue.


Letter to
City Manager, City of Falls Church Mr. Wyatt Shields
through the City Clerk cc: Falls Church City Council Members
Vice Mayor, City of Falls Church David Snyder
and 6 others
City Council Member Nadr Baroukh
City Council Member Marybeth Connelly
City Council Member Phil Duncan
City Council Member Karen Oliver
City Council Member Dan Sze
Mayor, City of Falls Church David Tarter (Mayor, City of Falls Church)
300 Park Avenue, Suite 303 East
Falls Church, VA 22046

Water usage shown in the fall 2013 Falls Church City water bills was abnormally high for many residential customers. To date, your response has been inadequate.

This petition summarizes the public comments to date, recounts the City’s approach to the abnormal usage problem, and rebuts your principal explanations. It then recommends using a small amount of the $14 million gained from the City’s sale of its water utility to fund an independent and open audit of Falls Church’s water accounts to address citizens’ concerns. We feel that is required to restore City credibility on this issue.

Water Customer Comments--The Specifics

This is a resident by resident issue, not a more general one of whether system-wide usage was higher or lower than the previous year. Although you have said that overall fall 2013 water usage was only 3.4% higher this year, some customers saw significant, sometimes huge, proportionate increases in their billed water usage. An early January posting by a resident in the online community news letter, “Commonplace,” drew 31 comments; 20 people commented specifically on their water usage in 2012 and 2013.


• Of those 20, 18 reported significant increases in their fall 2013 usage compared to 2012.

• Thirteen of the 20 reported their billed water usage in 2013 and 2012. The 13 accounts had a combined increase in water usage of 66% (including the two accounts where water usage decreased), or about 95,000 gallons, compared to the fall of 2012.

• On an account by account basis, of the 13 who reported comparable volumes, the median usage increase was 57%. The highest three increases were 110%, 135%, and 600%. One of those was absent during three weeks of the fall quarter, and one consulted a plumber who found no leaks.

Since the original posting on “Commonplace,” other people have come forward in response to articles in the “Falls Church Times,” the “Falls Church News Press,” and a subsequent posting in “Commonplace.”

• Two more people posted on “Commonplace.” One said billed usage had doubled and one implied usage was about 37% over the summer quarter of 2013.

• In the “News Press,” one person commented that his water usage fell in the fall quarter of 2013 by 3000 gallons, but another said his had risen from 8000 gallons in the fall quarter of 2012 to 15,000 in 2013.

• In the “Falls Church Times,” one person said her usage was 3000 gallons higher in 2013 than the 2012 fall quarter. In the worst case apparent so far, a woman reported that her bill was for 2000 gallons in the summer quarter of 2013, but was 62,000 gallons in the fall quarter. Our follow up found that the bill is for a townhouse with virtually no yard and was partially vacant during both the summer and fall quarters. Inspections found no leaks.

See Appendix 1 for more detail.

The City’s Reaction to High Water Bill Complaints

In public comments, you and your staff have brushed off complaints of a significant number of Falls Church water system customers, and offered only generalities to explain the higher usage:

• You have said that the number of complaints is not unusual, but have given no figures to support that, even when a City Council member, in the January 13, 2014 Council meeting, specifically asked how many complaints you have received.

• People who called about high water usage initially were referred to Fairfax County, even though their meters were read by, and the bills prepared by, Falls Church City. The City later setup a hotline, but a caller is still faced with only the general explanations discussed below and in Appendices 2 and 3 to this petition. Your pubic statements infer that the hotline has solved most of the issues, but you have presented no quantitative information to support that contention.

• There is no evidence, either in your public comments or those of your staff, that any attempt has been made to determine how many accounts are either higher or lower than historic patterns, and by how much (See Appendix 4 for a description of how that might be accomplished).

• The explanations you have offered are questionable:

1) Traditionally, the City’s approach has been to blame leaking toilets when a resident complains about an unexpected surge in a water billing. Appendix 2 shows how this is not likely to explain more than a few hundred gallons of increased consumption in a quarter.

2) Perhaps resulting from the number of recent complaints, and that several residents stated publicly that plumbers confirmed they had no leaks, you have now embraced a weather-related argument. It is that residents simply watered their lawns more last fall than in 2012 due to a shortage of rainfall last August and September. However, August is not in the suspicious fall quarter, and rainfall was actually twice as high in the entire summer quarter of 2013 than in 2012. It was about equal in the fall quarters. In addition, you have not conducted any analysis to show the quantitative relationship between rainfall and City-wide water usage (see Appendix 3 for more discussion of this issue).

What You Should Do

Given the level of publicity, the City’s reputation is tarnished by its failure to get to the heart of this issue. Accordingly, the City should use a small amount of the $14 million cash gain from the sale of its water system to conduct an independent audit of its billings in 2012 and 2013, and to evaluate some of the arguments you have put forth to date. This should be an open audit accessible to concerned residents, with well publicized public progress briefings while it is being conducted.

Initial activities of the audit should include the following objectives:

1) Determine, on an account by account basis, the extent of the abnormally high and low billings in the fall quarter of 2013 compared to 2012; provide detailed statistical summaries in the format described in Appendix 4.

2) Establish whether 2013 did, or did not (as alleged), experience an unusual number of high and low billings compared to the previous years (examine all quarters).

3) Examine account histories, water office records and conduct staff interviews to determine if abnormal 2013 bills were catch-up bills due to measurement errors in previous billing cycles.

4) In addition to the fall quarters, examine winter quarter usage in both 2012 and 2013 to determine if there are anomalies that would skew sewer bills (which are based on winter usage). Determine if abnormally low billed usage in the fall quarter of 2013 corresponds to higher usage earlier in the year

5) Assess the accuracy of the meter reading process. Interview meter readers to determine if procedures or equipment malfunctions may have contributed to the problems seen by some residents.

6) Review City records to assess the results of previous meter replacements and equipment maintenance.

7) Conduct a simple correlation analysis of the relationship of local rainfall to residential water use to determine if City’s argument has any merit in the aggregate (go back 8 to 10 years).

8) Attempt to address the fall lawn watering argument by merging the City’s real estate files with water files by street address and determine if there is any relationship between increased 2013 water usage and lot size (e.g., a townhouse with a large increase in fall water usage is unlikely to be due to lawn watering).

The basic data for these analyses would be the City’s billing database, real estate files, and readily retrievable data from the National Weather Service. It is likely that two junior analysts working in concert with City staff could perform items (1) – (8) in less than three weeks. From this “scoping” of the situation should spring a more focused audit of the problems seen at the individual account level.

In addition, the City Manager’s Office should present information in a report to justify its public statements on this problem. That report should include:

• A quantitative examination of the logs of phone calls that led City Manager to say,

“...the trend has been that either a) the consumption is normal, but their bill is larger than normal because they had a past due balance from a prior bill, b) they remember they actually did water the lawn quite a bit back in the early fall, or c) they find that they had a leaking toilet or faucet down in the basement of which they weren’t aware”

• The names (with their permission) of all residents who called in and were satisfied with the above explanations, along with the names of those who were not satisfied.

A draft final report on this audit should be available to the Falls Council and interested former Falls Church water customers by May 1, 2014.


Concerned Residents in the Falls Church Area

Appendix 1.

Postings in the “Commonplace” Neighborhood Newsletter

Of 31, 20 substantively commented on their water billings (several comments were opinions, including two explanations by the City)

Of the 20 substantive comments, 18 reported that their water unexpectedly rose significantly in the fourth quarter of 2013 from a year earlier. Two were down slightly.

Of 13 water customers who reported volumetric usage for the fourth quarters of 2012 and 2013, the net total volume increased, year-over-year, by 66%, or 95,000 gallons (this includes the two customers whose usage went down).

The additional 95,000 gallons cost the 13 City residents about $342 in higher water charges. (Sewer charges are based on water usage, but for the preceding winter quarter. Our calculation assumes that whatever problem exists for the fall quarter of 2013 is unique to that quarter and does not affect readings for the preceding winter quarter.)

Six people reported trying to contact the City, of which 3 could not because either the City did not return the call or the customer could not find a working phone number or point of contact. (This has been partially remedied as a result of the publicity afforded this issue. The City has made available a City phone number to contact a temporary employee.)

Of the three people who reported successfully contacting the City about apparent high water usage, all were either told to check their toilets or that the meter had been read and was correct.

Three people who reported abnormally high apparent usage contacted, or had recently contacted plumbers to either confirm that they had no leaks, or had plumbers fix leaks in the recent past before the fourth quarter of 2013.

Originally, the City’s postings on “Commonplace” simply referred complaints and questions to Fairfax County, despite the fact that the City of Falls Church made the meter readings, computed the billings, and will receive the payment for water and sewer usage for the fourth quarter of 2013. They also again suggested the high apparent usage problems are associated with leaking toilets. A later posting included the lawn watering argument addressed in Appendix 3.

Appendix 2.

The Toilet Leak Explanation
For years, any inquiry to the City about excessive water bills has met the "leaky toilet" argument. It is possible, but unlikely, that a leaky toilet could waste a few thousand gallons per quarter. To account for an increase of 8 to 10 thousand gallons in a quarter, such as many residents apparently saw in the fall quarter of 2013, a toilet flapper valve would have to be trapped open for long time. Further, the presence of a leaky toilet does not, by itself, account for a sudden increase in quarterly consumption. Because flapper valve leaks are slow to develop, it is likely that any leak could have been present for a long period, and not be the cause of a sudden quarterly jump in water consumption.
A typical toilet leak is one that would drain water from the tank and cause a partial refill of the tank. It is easy to measure how much water is lost to a partial refill. Simply hold the flush handle down slightly to let the water leak into the bowl. Mark how far the water in the tank has gone down when the valve trips and the tank starts to refill. Measure and compute the cross section of the tank; then multiply by the amount the water dropped. Convert cubic inches to gallons (231 cubic inches in a gallon). You will find that the partial refill volume is about 1/4th gallon.
If a tank leaks that much every 15 minutes, it loses a gallon per hour, or about 2000 gallons per quarter. Most people would notice a toilet turning on and off every 15 minutes. Much more likely is a leak that would cause an infrequent refill, say, once per hour. At that level only 500 gallons per quarter would be lost.
Finally, almost all toilets leak a little around the flapper valve. The food coloring test for a leaky toilet proposed by the City staff is often recommended. However, it is highly sensitive to very small leaks, which are very common, and it does not provide a measure of how much is being lost. A better procedure is to fill the tank and then shut the water off to the toilet. Come back in a set period of time, say 2 to 4 hours, and measure how much the water has dropped in the tank. Use the procedure above to compute how much water is lost. Equate this to a quarterly day use rate.

Appendix 3
The Low Rainfall Explanation

Recently, the City has seized on another excuse for higher usage. It calculated that the year-over-year fall quarter residential water consumption increased 3.4% for 2013 compared to 2012. It then attributed this to increased lawn watering due to lower rainfall in August and September of 2013 compared to 2012. While this has intuitive appeal, it does little to explain why individual citizens’ billed water usage doubled or more. Nevertheless, part of the City’s explanation is now that citizens with high usage watered more and forgot about it.

We agree that rainfall (as measured by the National Weather Service at National Airport) was lower in August and September in 2013 than in 2012. However, it was also higher in July and November of 2013. In fact, when the summer and fall quarter totals for 2013 and 2012 are compared:

• For the summer quarter (June-August), rainfall in 2013 was the highest since 2006 and almost twice that of 2012.

• September 2013 rainfall was the second lowest in the preceding 7 years (September 2007 was lower). But October and November 2013 were higher than 2012, so that the total rainfall for the entire quarters of 2013 and 2012 were essentially equal.

• Last, it is worth noting that August is in the summer quarter, and the problem we have in this issue is anomalous water usage in the fall quarter. Rainfall, or lack thereof, in August would tend to impact the summer quarter’s water consumption.

The City staff seem to have made their observation about 2012 and 2013 and seized upon it as an excuse for higher water consumption. Logically, it would appear to contribute in general, but it begs for more detailed analysis. One suggestion would be to simply graph average quarterly residential water usage (total residential water usage divided by number of residential accounts) for the last 8-10 years and plot that versus quarterly rainfall. Specific comparison of the summer and fall quarters would determine if there really is any correlation between measured water consumption and rainfall, and what is that correlation factor. A refinement would be to provide this graph for the separate categories of townhomes and residential single family dwellings.

Appendix 4

Scoping the High Usage Issue

Any analysis of an issue should depend on data, not supposition. In reports of abnormally high water usage, the City has a problem that it has met with only supposition. This is particularly egregious because it has at its disposal several bodies of data that would shed light on the issue and focus a subsequent investigation. These data bases will not answer all questions on an account by account basis, but, properly sorted and examined, the data will provide a focus for further audit and investigation.

The principal database we would use contains water billing information. We have been told by former City staff that this database will be retained by the City of Falls Church; that is, control of the historic billing files will not transfer to Fairfax County. According to the information we have now, it has account level data that provides water usage quarter by quarter for at least the last 5 quarters, perhaps much longer. Further, accounts can be subdivided into commercial (maybe with several subcategories), residential single family detached homes and townhouses. Geographically, in addition to the account address, there is a designator denoting whether each account is inside or outside the City limits. It is likely that a long term history can be obtained for each account by resurrecting archived files.

The property assessment database is a secondary source of information. Therein, physical properties of residential homes, such as lot size and each home’s square footage are available.

Merging these two databases using street address as a common delineator would allow an analyst to provide information to evaluate hypotheses about water use versus property characteristics. For example, if it is true that the surge in fall quarter water usage is due to increased lawn and garden watering in 2013, then detached homes should tend to show proportionally higher increases than townhomes. Further, among single family homes, there should be a positive correlation between higher fall quarter water use and bare ground area, i.e., lot size minus home square footage. (This measure could be made better by subtracting out ground area covered by concrete, but we are not sure about the City’s progress in producing this measure.)

Initially, the City should conduct analyses to determine the scope and magnitude of the apparent abnormal use in the fall quarter of 2013. It should begin by calculating the percentage increase (or decrease) for each account in the fall quarter of 2013 compared to the summer quarter of 2013 (e.g., an account that used 10,000 gallons in the summer quarter and 12,000 in fall quarter would have increased by 20 percent). Usually, water usage goes down in the fall quarter, so accounts that increase are historic anomalies. These percentages should be sorted and compared to make observations about the character and scope of the high usage.

For example, the analyst could sort the account level data into those whose usage went up and those that decreased, and perform the following evaluations.

• Show how many accounts had increased usage in the fall and a distribution of their proportionate increases (how many rose by 0 to 10%, 10% to 20%,...., were over 100%, etc.).

• Provide the same display for the accounts that had decreased water usage.

• Examine the distributions to determine, as the City might allege, that there were a very small number of accounts with large proportionate increases in usage, and a large number that showed the expected decrease in the fall quarter from summer usage.

Then, the analyst might perform the same calculation and sort to compare the fall quarter of 2013 to the fall quarter of 2012. The resulting data distributions would show what percent of the total accounts had higher usage. This, in turn, would indicate whether the problem here is associated with a small number of accounts that have singular problems, or if there is a reason that applies more broadly. The latter might be the case if lower rainfall caused more watering (more analysis is necessary to make this link).

The question of whether 2013 patterns were the same or different from other years could be addressed by performing the same calculation and sort, comparing individually, the winter spring and summer quarters of 2013 and 2012 to each other. The data would answer the following questions:

• Are there seemingly anomalous increases or decreases in some accounts?

• Are there more anomalies in the fall 2013 quarter than the other quarters (this relates to the City Managers statements that we did not see anything unique in the fall 2013 billing cycle)?

• Do the accounts that have abnormally low usage in the fall of 2013 also have abnormally high usage in the winter quarter of 2013 (this would imply, but not prove, that there is some sort of account adjustment taking place).

• For the accounts that showed large increases in 2013 fall quarter versus 2012, is there a positive relationship between the amount or percentage of the increase and bare (unimproved) lot size? If so, one might suggest that the low rainfall argument might have some credibility.

As with any evolutionary analysis, we cannot say ahead of time where this data examination will lead. One should expect to see unexpected results that require additional analyses investigation. The more that can be accomplished by data analysis, the more efficient will be the overall audit. In general, the purpose of examining data freely available to the City is to focus the audit to improve its result and to avoid costly investigative dead ends.

Concerned Resident started this petition with a single signature, and now has 167 supporters. Start a petition today to change something you care about.