# Georgia Water Loss Audits 2016-2025: Northwest Georgia Findings

Oct 12, 2026 · @Elliot Pierce

## The short answer

In 2025, the nine public water systems in Walker, Catoosa, Dade and Chattooga counties lost about one of every three gallons they put into their pipes: 2,280 million gallons out of 6,899 million, or 33%. The statewide figure was 22%. (Sum of the nine 2025 audits; statewide sum of 241 audits. Source: *2025 Water Audits (8-27-26) FINAL.xlsx*.)

That is about 6.2 million gallons a day across the four counties. The local share lost has stayed near or above 30% in every year from 2016 to 2025.

- **Worst locally:** LaFayette. It lost 612 million gallons in 2025, 47% of its supply, and leaked more per customer connection than 94% of Georgia utilities.
- **Walker County Water & Sewer Authority** lost 46% of its supply in 2025. Its losses bounce between about 30% and 47% from year to year.
- **Best locally:** Catoosa Utility District, at 13% lost in 2025. But its 2025 audit also has a population figure that cannot be right, so its sudden improvement needs checking.
- **Data problems are part of the story.** Dade's 2019 audit is a copy of its 2018 audit. Several local audits carry population, pipe-length or cost figures that jump or never change. These are in the open questions.

Everything below separates what the audits **show** (fact, with file and row), what it **might mean** (labeled interpretation), and what to **ask next**. Spreadsheet companion: *ga\_water\_audits\_cleaned.xlsx* (cleaned data, change log, open questions).

All volumes are in million gallons (MG). One million gallons equals 1,000 of the "thousand gallons" units most water bills charge by. Dollar figures are as the utilities reported them, not adjusted for inflation.

## What's in the project

The project holds three documents that explain the rules and ten spreadsheets of results. Every Georgia water system serving 3,300 or more people must file a yearly "water loss audit" with the state Environmental Protection Division (EPD) by March 1 (Manual p. 7-8). The spreadsheets are EPD's statewide compilations of those audits, one per calendar year.

**Documents**

| Document | Who made it, when | What it covers | Key points |
| --- | --- | --- | --- |
| *Geogia Water Loss Control Manual V3 2024-02.pdf* (87 pages) | Georgia Association of Water Professionals committee, with EPD; Feb. 2024 | How to do, check and certify an audit; definitions | Audits follow the American Water Works Association method (p. 12). Each audit must be checked by a state-qualified auditor (p. 28). A data validity score under 50 makes loss-control spending risky (p. 27). Leakage index below 1.0 often means bad data (p. 35). Leakage index is not reliable for small or low-pressure systems (p. 76). |
| *Guidance for Demonstrating Progress in Improving Water Efficiency* (10 pages) | EPD, July 2019 | What a utility must show to renew or expand a water permit | Systems must show improving data scores and falling losses per connection (p. 2-4). EPD asks if a system has kept a data validity score of at least 50 (p. 9). |
| *Water Supply Efficiency Improvement EPD Approach* (15 slides) | Johanna Smith, EPD; GAWP conference, Nov. 19, 2020 | EPD's review of loss-control programs | Programs required since July 2016 (slide 11). EPD wants losses per connection, not percentages (slide 13). Permits now require yearly progress reports (slide 12). |

**Spreadsheets** (one row per utility per year; full column list in the workbook)

| File | Audit year | Utilities | Layout |
| --- | --- | --- | --- |
| *2016 Water Audit Compilation (11-15-18).xlsx* | 2016 | 230 | Older (v5) |
| *2017 Validated Water Audit Results.xlsx* | 2017 | 233 | Older (v5) |
| *2018 Water Audit Compiled Spreadsheet\_Final 7-31-19.xlsx* | 2018 | 243 | Older (v5) |
| *2019\_water\_audits\_compiled\_-\_all\_9-11-20.xlsx* | 2019 | 241 | Older (v5) |
| *2020 Water Audit - Posting Version (7-26-21).xlsx* | 2020 | 243 | Older (v5) |
| *2021 Water Audits Fianal Version (9-29-22).xlsx* | 2021 | 241 | Newer (v6) |
| *2022 Water Audits Final Version 9-25-2023.xlsx* | 2022 | 241 | Newer (v6) |
| *2023 Water Audits (9-24-2024)(10-8-2025) - UPDATED10062025.xlsx* | 2023 | 241 | Newer (v6) |
| *2024 Water Audits Final\_Jan2026.xlsx* | 2024 | 242 | Newer (v6) |
| *2025 Water Audits (8-27-26) FINAL.xlsx* | 2025 | 245 | Newer (v6) |

The most recent data is calendar year 2025, from a file dated Aug. 27, 2026. Each file covers all of Georgia, so all four focus counties are in every year.

**Problems with the files**

- Slide 7 of the presentation is a picture with no readable text; its content is unknown.
- The 2017 file has two empty extra tabs.
- Blank-looking cells hold codes like "Note 1" (not applicable) or "Note 3" (cannot be calculated). The meaning of "Note 2" changed: in 2016-2022 it meant "utility does not know"; in 2023-2025 it means "optional, no longer required."
- Each file ends with a block of footnotes inside the data area, and the 2022-2025 files have merged cells in a mid-table note.
- Two cells mix a number and a note in one cell ("7.745 Note A").
- Water system ID numbers are inconsistent: typos, stray spaces, and several rows filed under another utility's ID (details below).
- Two utilities reported 2024-2025 volumes about 1,000 times too large, and two reported 2025 volumes far too small.
- No file was unreadable. All numbers were stored as true numbers, not text.

The four counties' public systems in the audits are: Walker County Water & Sewer Authority, LaFayette and Chickamauga (Walker); Catoosa Utility District, Fort Oglethorpe and Ringgold (Catoosa); Dade County Water & Sewer Authority (Dade); and Chattooga County Water District and Summerville (Chattooga). Trenton, Trion, Rossville and Lookout Mountain do not appear as separate systems in any year. Ringgold has no 2017 audit in the 2017 file.

## How the data was cleaned

The ten files became one table of 2,400 utility-years covering 265 distinct water systems. Your original files were not changed; all work was done on copies. The workbook's Change Log sheet lists every step with a count.

The main steps, in plain terms:

1. **Kept only real data rows.** Dropped headers, blank rows and the footnote blocks (2,400 rows kept).
2. **Turned codes into blanks.** 5,301 cells held codes like "Note 1" or "N/A." They became blanks so the math works, and each original code is saved in a notes column. Two cells that mixed a number and a note kept the number.
3. **Fixed system ID numbers.** 65 IDs had spaces or odd prefixes ("CG," "CP," "CS" for "GA"). Three six-digit typos were fixed (Newnan, Fayetteville, Gwinnett). Fifteen rows filed under another utility's ID were moved back, using name and county as proof: Hiawassee, Leesburg, Milledgeville, Washington, Oconee County, Montezuma and an Oak Park subdivision row.
4. **Linked combined audits.** Some utilities file one audit for two systems (Cobb, Gwinnett with Lawrenceville in 2020-22, Barrow, Newton, Satilla, Moultrie). These 42 rows are linked to the main system and labeled, because a scope change can make a year look better or worse than it was.
5. **Standardized names.** County names were put in one format and six misspellings or mistakes fixed (258 cells, e.g. "Chattoga" to Chattooga). Each utility got one standard name from its latest audit; raw names are kept.
6. **Checked for duplicates.** No utility appears twice in a year after the ID fixes, and no row is an exact copy of another.
7. **Calculated the same measures the same way for every year** (defined in Terms explained): losses per connection per day, losses per mile of main, share of supply lost, dollar cost of losses, and gallons supplied per person. These match the figures already in the files to the decimal, which confirms the files are internally consistent.
8. **Flagged, but did not change, suspect values.** Flags sit in a column beside each row: negative losses, impossible volumes, leakage index below 1.0, data validity under 50, population jumps over 50%, combined audits, and audits identical to the prior year.
9. **Set aside six rows with impossible volumes** from statewide totals and rankings: Eatonton-Putnam and Hinesville (2024, 2025), Glen of Robin Hood and South Atlantic Utilities (2025). Their numbers were not rescaled, because the right fix is a guess.
10. **Blanked two population-based figures** where population was clearly wrong (Columbus 2020, entered as 2,299,000; Oak Park subdivision 2023, which carries the City of Douglas's population).

What was **not** fixed: suspicious local values that might be right. They are listed as open questions instead.

## Key findings across Georgia

**Georgia's water systems are losing a growing share of their water.** In 2025 the 241 usable audits reported 88,976 million gallons lost out of 401,064 million supplied: 22.2%, or about 244 million gallons a day. In 2016 the figure was 16.8%. (Sums of each year's file; workbook sheet *State by Year*.)

- **Not a change in who reports.** The 208 utilities with a usable audit in all ten years show the same climb, from 16.7% to 21.9% (sheet *Same 208 Utilities*).
- **Leaks are most of it.** Real losses, water that physically leaks out, were 80,329 million gallons in 2025, or 90% of all losses. The rest is "apparent" loss: water used but not billed because of bad meters, billing errors or theft.
- **The typical system got leakier.** The middle (median) utility leaked 47.7 gallons per connection per day in 2016 and 55.3 in 2025.
- **The reported dollar cost rose faster.** Utilities valued their 2025 losses at $149.3 million, up from $85.7 million in 2016. These are not inflation-adjusted, and utilities price losses in different ways (see Terms).
- **The data got better.** The median data validity score rose from 63.6 to 68.0. Systems scoring under 50 fell from 11.7% to 5.4%.
- **A few systems skew the average.** In 2025 the average leak rate was 86 gallons per connection per day; the median was 55. That gap means a small number of very leaky systems pull the average up, so medians are used for comparisons.
- **No single cause shows up statewide.** Across 2025 audits, leak rates barely track pressure (rank correlation -0.11), size (-0.22) or data score (-0.26). Correlation of 0 means no pattern; 1 means a perfect one.
- **The worst leakers statewide are mostly small or military systems.** The top five per connection are the Waycross-Ware Industrial Park, Moody Air Force Base, Dawson, Vienna and Fort Stewart. LaFayette is 15th worst of 241.

*Interpretation:* A higher share lost while data scores improve could mean leaks are truly growing, or that better meters and record-keeping now reveal losses that were hidden. The audits alone cannot tell which.

[embedded content: Sums of water losses and water supplied from the ten EPD audit files, 2016-2025. Local = the nine audited systems in Walker, Catoosa, Dade and Chattooga counties. Georgia excludes 6 rows with impossible volumes.\]

*Takeaway:* the four-county systems have lost 11 to 18 points more of their water than Georgia as a whole in every year since 2016.

## County-by-county findings

Seven of the nine local systems leaked more per customer connection in 2025 than the typical Georgia system. Latest data for all nine is calendar 2025 (*2025 Water Audits (8-27-26) FINAL.xlsx*; Excel row numbers given below). "Leaks per connection" means real losses in gallons per service connection per day, the size-adjusted measure EPD prefers (slide 13).

[embedded content: 2025 Water Audits (8-27-26) FINAL.xlsx, columns Annual Volume of Real Losses and Number of Active and Inactive Service Connections. Georgia median of 241 usable 2025 audits.\]

*Takeaway:* only Catoosa Utility District and Ringgold leaked less per connection than the typical Georgia system in 2025.

**2025 snapshot** (source: 2025 file; "lost" = all water losses)

| County | System | Connections | Supplied (MG) | Lost (MG) | Share lost | Leakage index | Data score | Reported cost of losses |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| Walker | LaFayette | 7,809 | 1,312.7 | 612.0 | 46.6% | 8.01 | 74 | $140,858 |
| Walker | Walker County Water & Sewer Authority | 13,679 | 828.5 | 379.9 | 45.9% | 3.12 | 69 | $258,929 |
| Walker | Chickamauga | 2,485 | 289.5 | 124.8 | 43.1% | 3.86 \[UNCERTAIN\] | 52 | $62,762 |
| Chattooga | Chattooga County Water District | 3,595 | 375.1 | 163.3 | 43.5% | 2.04 | 51 | $226,109 |
| Chattooga | Summerville | 5,300 | 841.2 | 314.2 | 37.4% | 5.18 | 67 | $399,853 |
| Catoosa | Fort Oglethorpe | 3,097 | 301.1 | 108.8 | 36.1% | 4.01 | 51 | $220,908 |
| Dade | Dade County Water & Sewer Authority | 8,886 | 921.5 | 320.3 | 34.8% | 2.53 | 75 | $316,557 |
| Catoosa | Ringgold | 1,420 | 192.6 | 26.9 | 13.9% | 1.80 \[UNCERTAIN\] | 60 | $80,283 |
| Catoosa | Catoosa Utility District | 22,668 | 1,837.4 | 230.3 | 12.5% | 0.83 \[UNCERTAIN\] | 73 | $204,266 |

[UNCERTAIN\] leakage index: Chickamauga and Ringgold are too small for the index to be valid (Manual p. 76); Catoosa Utility District's 0.83 is below the 1.0 floor the manual says often signals data error (p. 35).

By county, 2025 (all audited systems added together): Walker lost 45.9% of its supply, Chattooga 39.3%, Dade 34.8% and Catoosa 15.7% (workbook sheet *County Rollup*).

### Walker County

**LaFayette** (row 129)

- *Shows:* Lost 612.0 MG of 1,312.7 MG in 2025, 46.6%. Leaks of 211.8 gallons per connection per day were the 15th worst of 241 Georgia systems and the worst locally. Real losses rose from 375.4 MG in 2016 to 603.8 MG in 2025, up 61%. It bought 618.0 MG from other suppliers in 2025, nearly half its supply.
- *Shows:* The reported dollar cost of its losses fell from $571,715 in 2024 to $140,858 in 2025 because the production cost it entered fell from $824.98 to $192.95 per million gallons. Its reported total cost to run the water system is $226,074. Miles of main read 238.8 (2016-20), 638.8 (2021-23) and 272.6 (2024-25). \[CONFLICT\]
- *Might mean (interpretation):* LaFayette appears to be paying for outside water and then losing nearly as much as it buys. The low cost figures look like entry errors and likely understate what the losses cost ratepayers.

**Walker County Water & Sewer Authority** (row 235)

- *Shows:* Lost 379.9 MG of 828.5 MG in 2025, 45.9%. Up from 29.7% in 2024. The share lost has swung between 29.7% and 46.7% since 2016. Real losses were 369.6 MG in 2016 and 373.1 MG in 2025.
- *Shows:* Leaks per connection look lower since 2022 (124.5 in 2021, 74.7 in 2025). But that year the connection count jumped from 11,164 to 13,211 and the pressure entry fell from 115 to 75 psi. Miles of main is 429.4 in all ten audits; connections are 13,679 in each of 2023-2025.
- *Might mean (interpretation):* The authority loses about as many gallons as it did a decade ago. The apparent per-customer improvement comes mostly from changed inputs, not fewer gallons lost. Unchanging pipe and customer counts suggest some inputs are carried forward rather than re-measured.

**Chickamauga** (row 49)

- *Shows:* Lost 124.8 MG of 289.5 MG in 2025, 43.1%. That is a real improvement from 54% to 80% lost in 2016-2020. Real losses fell from 238.5 MG (2020) to 122.3 MG (2025).
- *Shows:* Average pressure of 145 psi (165 psi in 2016-18) is the highest of the nine. Its data score was 46 in 2017 and 2019-2022, below the EPD benchmark of 50; it is 52 now.
- *Might mean (interpretation):* Chickamauga cut its losses roughly in half after 2020, but still loses more than four of every ten gallons. Very high pressure tends to drive leaks (Manual p. 68).

### Catoosa County

**Catoosa Utility District** (row 42)

- *Shows:* Lost 230.3 MG of 1,837.4 MG in 2025, 12.5%, the best locally. Leaks of 26.3 gallons per connection per day beat 78% of Georgia systems.
- *Shows:* Real losses fell by half in one year, from 456.0 MG in 2024 to 217.7 MG in 2025. From 2016-2024 it lost 17.6% to 25.8% each year. The 2025 population is 7,800, against 60,000 in 2023-2024 and 22,668 connections. \[CONFLICT\]
- *Might mean (interpretation):* Either the district made a major leak fix in 2025 or the audit has input errors. The impossible population and the sub-1.0 leakage index both point toward checking the data.

**Fort Oglethorpe** (row 93)

- *Shows:* Buys all its water (301.1 MG in 2025). Lost 108.8 MG, 36.1%; leaks of 95.4 gallons per connection per day; leakage index 4.01.
- *Shows:* Its data score fell from 75.6-79.9 (2016-2022) to 41 in 2023 and 51 since. Its production cost entries were $1.53-$2.34 in 2016-2021 and $1,827-$2,250 in 2022-2025. Its 2023 retail rate is $0.10 per thousand gallons versus about $8 other years. \[CONFLICT\]
- *Might mean (interpretation):* The cost figures changed units, so Fort Oglethorpe's dollar losses cannot be compared across years. The data score drop suggests record-keeping slipped after 2022.

**Ringgold** (row 182)

- *Shows:* Lost 26.9 MG of 192.6 MG in 2025, 13.9%; 40.5 gallons per connection per day. No audit appears in the 2017 file.
- *Shows:* Apparent losses (meter and billing) jumped from 0.72 MG in 2024 to 5.87 MG in 2025. Production cost entries jumped from $291-$360 per million gallons (2016-2022) to $2,364-$2,619 (2023-2025).
- *Might mean (interpretation):* Ringgold's leaks are low, but its meter or billing losses rose eightfold in a year.

### Dade County

**Dade County Water & Sewer Authority** (row 66)

- *Shows:* Lost 320.3 MG of 921.5 MG in 2025, 34.8%. Leaks of 97.9 gallons per connection per day are up from 62.6 in 2018. Water supplied grew from 684.5 MG (2018) to 921.5 MG (2025), a 35% rise with 5% more connections.
- *Shows:* The 2019 audit repeats the 2018 audit exactly: every volume, cost and connection count matches (2018 file row 70, 2019 file row 70). Only population differs. \[CONFLICT\] Population is 36,400 in 2017-2018 and 16,251-18,200 in all other years.
- *Shows:* Apparent losses fell from 13.8 MG (2023) to 2.6 MG (2024) and 2.7 MG (2025). Its data score has stayed between 72 and 80, the most consistent locally.
- *Might mean (interpretation):* Dade has good-quality data but rising leaks. The 2019 copy means one year of Dade's record is not a real measurement.

### Chattooga County

**Summerville** (row 205)

- *Shows:* Lost 314.2 MG of 841.2 MG in 2025, 37.4%. Leaks of 146.8 gallons per connection per day are more than triple the 44.2 of 2016. Real losses grew from 83.4 MG to 283.9 MG over that span.
- *Shows:* Population drops from 11,651 to 4,500 in 2023-2025 while connections stay at 5,300. \[CONFLICT\] Connections (5,300) and miles of main (200) are identical in 2023, 2024 and 2025. A production cost of $2,913 per MG in 2023 ($279-$591 in other years) pushed that year's loss cost to $864,360.
- *Might mean (interpretation):* Summerville's leaks have grown faster than any other local system's over ten years.

**Chattooga County Water District** (row 47)

- *Shows:* Lost 163.3 MG of 375.1 MG in 2025, 43.5%; 110.6 gallons per connection per day; leakage index 2.04.
- *Shows:* Data score 51 in 2022-2025, 48 in 2021, 49.7 in 2017: at or below EPD's benchmark of 50. Miles of main is 300 in all ten audits; population is 7,800 every year since 2019. The 2022 retail rate is $16.12 per thousand gallons versus $5 to $8 in other years. \[CONFLICT\]
- *Might mean (interpretation):* The district's numbers are the least trustworthy locally, so its loss figures carry the most uncertainty.

## Trends and what changed most recently

Over ten years, leaks per connection rose at five local systems (Dade slowly), fell at two, and held about level at two. In the latest year, three systems broke sharply from their own pattern: Catoosa Utility District down, Walker County Water & Sewer Authority and Summerville up.

All nine have audits for 2016 through 2025 (Ringgold is missing 2017), so every system is compared over the same years. Figures are real losses in gallons per connection per day (workbook sheet *Local Summary*).

| System | Avg. of first 3 audits | Avg. of last 3 audits | Ten-year direction | 2024 | 2025 |
| --- | --- | --- | --- | --- | --- |
| LaFayette | 113.5 | 178.7 | Rising (+9.4 a year) | 177.6 | 211.8 |
| Summerville | 56.4 | 121.2 | Rising (+9.9 a year) | 93.4 | 146.8 |
| Chattooga County Water District | 86.9 | 111.0 | Rising (+3.8 a year) | 115.7 | 110.6 |
| Ringgold | 12.6 | 39.8 | Rising (+4.0 a year) | 27.7 | 40.5 |
| Dade County Water & Sewer Authority | 70.3 | 85.7 | Rising slowly (+2.7 a year) | 94.8 | 97.9 |
| Fort Oglethorpe | 96.6 | 81.0 | About level (-1.0 a year) | 102.3 | 95.4 |
| Catoosa Utility District | 45.8 | 42.6 | About level (-0.6 a year) | 56.0 | 26.3 |
| Walker County Water & Sewer Authority | 89.0 | 67.0 | Falling (-3.8 a year)\* | 50.6 | 74.7 |
| Chickamauga | 270.1 | 128.0 | Falling (-20.0 a year) | 146.4 | 134.8 |

\*Walker County Water & Sewer Authority's decline comes from a changed connection count and pressure entry in 2022, not fewer gallons lost (see county findings).

"Direction" is a straight-line trend through all audits. It is called rising or falling only if it moves more than 3 gallons per connection per day each year.

**What changed most recently (2024 to 2025)**

- **Catoosa Utility District:** leaks per connection fell 53%, from 56.0 to 26.3, the largest one-year improvement locally. The same audit has the population error noted above.
- **Summerville:** up 57%, from 93.4 to 146.8, its worst year on record.
- **Walker County Water & Sewer Authority:** up 48%, from 50.6 to 74.7, after a good 2024.
- **LaFayette:** up 19% to its worst year on record.
- **Four-county total:** water lost was 2,266 MG in 2024 and 2,280 MG in 2025, essentially flat. Across all ten years the nine systems lost a combined 21,183 MG.

## What to ask or investigate next

The most useful records are each utility's **Water Loss Control Program** and its yearly **progress reports** to EPD. Both are required under Rule 391-3-33 (slides 11-12), and EPD reviews them when a utility seeks a permit renewal (Guidance p. 9). They should name the goals and leak-repair steps behind these numbers.

**Records to request** (Georgia Open Records Act)

- From EPD's Water Supply Program: the full audit workbooks (with the Comments and Notes tabs) and auditor certification forms for all nine systems, 2016-2025. The compiled files omit the auditor's notes, which explain odd entries (Manual p. 8, 57).
- From EPD: each system's Water Loss Control Program, annual progress reports, and any permit review that judged "demonstration of progress" (Guidance p. 9).
- From each utility: meter testing results, leak repair logs and main-break counts for 2023-2025; the basis for pressure, pipe-length and connection counts.

**Questions for officials**

- **LaFayette:** Who sells you the 618 million gallons you bought in 2025, and at what price? Why did the production cost you report fall from $824.98 to $192.95 per million gallons in one year? Which miles-of-main figure is right: 238.8, 638.8 or 272.6? Is the $226,074 operating cost the full cost of running the water system?
- **Walker County Water & Sewer Authority:** Why does the share lost swing between 30% and 47% from year to year? Were miles of main (429.4) and connections (13,679 in 2023-25) re-measured each year? What explains the 2022 change in connections and pressure?
- **Catoosa Utility District:** What did the district do in 2025 that cut leaks in half? Why is population listed as 7,800? Has the auditor reviewed the leakage index below 1.0, which the manual says often signals data error?
- **Dade County Water & Sewer Authority:** Why is the 2019 audit identical to 2018? Why did apparent losses drop from 13.8 to 2.6 million gallons in 2024? Why has supply grown 35% since 2018 with 5% more connections?
- **Summerville:** Why did population drop from 11,651 to 4,500 in the 2023 audit? What drove leaks per connection up 57% in 2025? Were connections and pipe miles re-counted after 2023?
- **Chattooga County Water District:** What is the plan to raise a data score that has sat at or below 51 since 2021? Is the 300 miles of main a measured figure?
- **Fort Oglethorpe:** Why did the data score fall from about 80 to 41 in 2023? Which unit did the production cost use before 2022? What happened in 2023, the one year losses dipped to 21%?
- **Chickamauga:** What changed after 2020 to halve losses? Is there a plan to lower pressures of 145 psi or more?
- **Ringgold:** Why is there no 2017 audit in EPD's compilation? What caused apparent losses to rise eightfold in 2025?
- **EPD:** Has any of these systems been asked to explain its numbers during a permit review? Does EPD check for audits copied from the prior year?

## Also noticed

- **EPD's own published files contain clear errors.** Eatonton-Putnam and Hinesville reported 2024 and 2025 volumes about 1,000 times too large (Hinesville 2025 shows 1,612,799 "million gallons" supplied for 28,099 people). Columbus's 2020 population is 2,299,000. These sit in the files EPD posts publicly.
- **The 2023 file was revised after release.** Its name shows dates of Sept. 24, 2024 and Oct. 8, 2025. Anyone who downloaded it earlier may hold different numbers.
- **Percentages are not EPD's preferred yardstick.** The manual and EPD's slides say to judge leaks per connection, not by percent lost (Manual p. 33, 75; slide 13). This report gives both because percent is easier to grasp, but the rankings use per-connection figures.
- **Dollar losses are hard to compare.** 140 statewide audits priced leaks at the retail rate instead of production cost, which makes them look costlier. Locally that includes Fort Oglethorpe (2016-18, 2022), Summerville (2016-18), Chattooga County Water District (2016) and Ringgold (2016-19).
- **Small systems get a pass on the leakage index.** The index is not valid for 807 of 2,400 audit-years statewide because the system is small or low-pressure (Manual p. 76). Locally that includes Chickamauga and Ringgold.
- **Population figures are weak everywhere.** 109 audits statewide show population jumping more than 50% from the prior year. EPD estimates population at 2.6 people per connection for residential systems (Manual p. 8), so population should not be used to judge these utilities.
- **Not everyone in the four counties is covered.** Only systems serving 3,300 or more must file (Manual p. 7). Customers of smaller systems, or of utilities based outside these counties, do not appear here.

## Terms explained

| Term | What it means | Why a ratepayer should care |
| --- | --- | --- |
| Water loss audit | A yearly accounting of where every gallon a utility pumps or buys ends up, filed with EPD by March 1 (Manual p. 7-8). | It is the public's main record of how much treated water is wasted. |
| Water supplied | Water a utility makes or buys, minus water it sells to other utilities. | It is the starting total that losses are measured against. |
| Water losses | Water supplied minus everything customers and the utility were authorized to use. | Someone paid to treat and pump that water; nobody paid for it at the tap. |
| Real losses | Water that physically leaks from mains, service lines and tanks before reaching a meter (Manual p. 75). | Leaks waste the cost of treating and pumping water and can force new supply spending. |
| Apparent losses | Water used but not billed, from slow meters, billing errors or theft (Manual p. 73). | That is lost revenue, which the paying customers make up. |
| Non-revenue water | All water that earns no money: losses plus authorized free uses like firefighting (Manual p. 75). | Shows the full volume a system gives away or loses. |
| Gallons per connection per day | Real losses divided by the number of service connections and by 365. | Lets a big and small system be compared fairly; EPD's preferred measure (slide 13). |
| Infrastructure leakage index | The ratio of actual leaks to the lowest level technically possible. A 1.0 is near perfect; 4 means four times the minimum (Manual p. 35, 74). | A quick read on how much leak reduction is possible. Not valid for small or low-pressure systems. |
| Data validity score | A 1-to-100 grade of how trustworthy the audit's own numbers are (Manual p. 74). EPD looks for 50 or more (Guidance p. 9). | Low scores mean the loss figures could be well off in either direction. |
| Production cost (VPC) | The cost of treating and pumping, or buying, one million gallons (Manual p. 24). | Used to put a dollar value on leaks. |
| Retail unit cost (CRUC) | What customers pay per thousand gallons (Manual p. 23). | Used to value apparent losses, and sometimes leaks. |
| Cost of losses | Apparent losses times the retail rate plus real losses times production cost, as reported. | A rough dollar figure for what losses cost a system each year. |
| Median | The middle value when all are lined up. | Unlike an average, a few extreme systems cannot pull it. |
| Million gallons (MG) | The unit used for all audit volumes; equal to 1,000 of the thousand-gallon units on most bills. | Keeps the scale of these figures clear. |
| Water Loss Control Program | A utility's written plan, required since 2016, to cut losses and improve its data (Manual p. 59). | It is the document that says what officials promised to do. |