The Water Package: Eleven Numbers You Control, Six That Wait for the Prospect

The power package has a hole at its center. What decides the deal is the carrying capacity of one specific delivery point, and that number cannot exist until the company hands the utility a load profile.

Water carries the same hole in a narrower form. The screening number is fully documentable before anyone calls: withdrawal authority, rated and permitted treatment capacity, maximum-day production, main sizes and pressure zones, storage, pump configuration, standby power, multi-year source chemistry. What waits on the prospect is the serving determination, meaning the utility's final engineering answer against a stated demand profile. Count what a site selector's engineer wants to see and you get seventeen fields. Eleven sit in records your utility already keeps: permits, operating reports, sanitary surveys, GIS, asset registers, lab data. Six do not. No single mandated report assembles the eleven, either, so somebody has to perform the reconciliation.

The deliverable to work backward from: a water package your office forwards within 48 hours of an inbound inquiry, without placing a call, that an engineering consultant treats as verified rather than aspirational.

Two habits keep that package from existing. The first is compiling nothing on the theory that water questions get answered when asked. They do get answered when asked. That is why cities are cut before the asking. The second is chasing a will-serve letter for an undefined load. No competent utility issues one, and what does get issued binds less than the director assumes.

Which lane the system is in

No federal dataset assigns facility-level water demand to NAICS codes. USGS names food, paper, chemicals, refined petroleum and primary metals as the heavy industrial users, but the national figures count self-supplied withdrawals rather than municipal deliveries, and the agency's own caution is that plant coefficients do not transfer between facilities. DOE's estimating method is project arithmetic: units per day, gallons per unit, less the recycled fraction.

The bins below are assembled from sourced facility records. They establish which lane you are in. Nothing more.

Above 1 MGD. Semiconductor fabrication, large meat and poultry first processing, pulp and paper, large wet-textile, major chemical complexes. The NIST environmental record for TSMC Arizona projected 4.755 MGD at Phase 1 and 17.29 MGD across three phases. The Micron New York record projected 7.85 MGD for a single fab and 48 MGD at four-fab buildout. EPA's 2023 meat and poultry dataset puts median process wastewater at 1.21 MGD for poultry first processors above 200 million pounds a year. EPA's technical record for pulp and paper puts a typical virgin bleached-chemical mill at 4,000 to 12,000 gallons per ton of pulp. That last one is an older benchmark. Use it for order of magnitude, not current averages.

250,000 gpd to 1 MGD. Fluid milk and cheese, larger breweries and distilleries, medium wet-textile, mid-scale protein, wet food and beverage generally.

25,000 to 250,000 gpd. Small and midsize breweries, poultry further processing, batch food, some dairy and textile, plastics, machining, chemical blending. Add any nominally dry plant carrying an evaporative cooling load.

Below 25,000 gpd. Warehousing, dry assembly, light electronics, small machining, closed-loop injection molding. EPA's Lean and Water material puts industrial domestic use at 20 to 35 gallons per employee per day. Five hundred employees with no process demand comes in under 20,000 gpd.

The poultry data is the reason sector names are worthless in a pitch. Median flow runs 0.0112 MGD in the band below five million pounds annually and 1.21 MGD in the top band. One NAICS code, two orders of magnitude between the ends of it. Cooling produces the same spread from the other direction. DOE reports towers commonly operating at two to four cycles of concentration where six or more is achievable, and the move from three to six cuts makeup water by 20%. A data center's choice between closed-loop and evaporative cooling relocates its water question by an order of magnitude before anyone has said the word megawatt.

Now sort your own system. A city genuinely in the wet-process conversation has particular things on the ground, and they are not the things on the brochure. Surplus firm capacity, meaning what remains with the largest critical unit out of service, not surplus rated capacity. A looped main of stated diameter at the park boundary, not a dead end. Elevated storage stated as a volume in gallons rather than shown as a tank in a photograph. Two independent sources, or a firm wholesale backup with a contract volume and an expiration date.

If your utility cannot demonstrate 250,000 gpd of bounded maximum-day headroom deliverable to a named parcel, pitching wet process spends credibility you will want on the next call. If it can demonstrate 25,000 to 250,000 gpd at the parcel, you hold a defensible lane: dry fabrication, assembly, light electronics, distribution, small brewing, lower-throughput food.

Volume is one screen. Source chemistry is a second and independent one. A system with 1 MGD of headroom and 400 mg/L chloride is not in the lane its headroom suggests. That gets its own section below.

The sentence that ends the call

We have a 12 MGD plant and we're only using 7, so there's 5 MGD available.

The 12 is probably rated capacity, which is design throughput under stated operating assumptions and carries no legal authorization at all. Permitted capacity is what the state allows. Virginia defines it as the limiting hydraulic capability of the entire system after accounting for withdrawal, treatment, finished storage, delivery and distribution. North Carolina defines plant permitted capacity as what may be treated on any given day. Permitted frequently sits below rated.

Above both sits the source: withdrawal authority, wholesale contract, or documented safe yield. A 12 MGD plant on an 8 MGD withdrawal permit is an 8 MGD system. Below both sits firm capacity, what remains with the largest critical unit down. EPA's sanitary survey guidance treats largest-unit-out as a planning convention and notes that state requirements vary. Six wells is not six wells of firm supply.

The 7 is almost certainly average day, which is a rate-setting number. What governs a new industrial connection is maximum day. Virginia's design regulation requires community waterworks to meet maximum-day demand while accounting for industrial and fire demand, holding at least 20 psi residual under the greater of peak-hour or maximum-day-plus-fire-flow.

Then subtract committed-but-unconnected allocation, the capacity assigned to approved projects whose meters have not turned on. That figure lives in the utility's development ledger. It appears in no state compliance filing.

The planning triggers come next, and they vary by state and get misquoted constantly. Virginia requires a written capacity plan once production hits 80% of permitted capacity for three consecutive months. Florida requires a capacity analysis within six months of maximum-day production first exceeding 75% of permitted maximum day. North Carolina's planning form asks whether production exceeded 80% and 90% for five consecutive days. None of the three is a national reserve rule, and none bars a new connection automatically. The director who says "we can't go above 80%" has imported a Virginia trigger into a state that never adopted it.

Then the parcel. System headroom means nothing if the site sits at the end of an eight-inch dead end in a low-pressure zone. This is the finding from the power teardown: system-level capacity is never delivery-point capacity, and the authoritative number belongs to the utility rather than the EDO. The hydraulic model settles it.

The defensible screening figure is the minimum of:

• source yield
• withdrawal authority
• firm treatment capacity
• firm pumping
• effective storage and delivery
• parcel-level hydraulic capacity

Less current maximum day, less committed unconnected demand, less any adopted planning margin. Every term carries a date and an operating condition.

Three systems, one brochure vocabulary

Constructed illustrations. The fields mirror the state filing structures above.

Riverbend's answer, written the way the utility would sign it and the director would forward it:

Wellfield is the ordinary case. Nothing on its brochure is false. The plant is rated at 6 MGD, current use is 3 MGD, the eight-inch main is in the ground, and the gauge reads 60 psi. All four statements survive isolation and none of them survives an engineer.

Junction is the case directors underuse, and it is where redundancy claims usually collapse. An emergency interconnection is not additive volume. It is a supplier's discretionary obligation under stated conditions, and no wholesale contract I have read grants simultaneous entitlement to routine and emergency supply. Document it regardless: purpose, pipe size, valve condition and position, activation authority, supplier-side restrictions, date last exercised. Present it as available capacity and you have lost the engineer for the remainder of the call. Backup wells and standby generators take the same discipline. An asset with no current pump test or load-test record does not enter firm capacity, because that is how the utility will treat it.

A bounded lane, stated with its limits, is a complete answer. Rome, Georgia's Enterprise Corner is documented that way: roughly three miles of 16-inch water main went to bid while a separate $4.5 million sewer extension waited on a tenant to define lift-station requirements. There is no partial-credit column on a screening matrix. Water without sewer scores full marks for dry process and zero for wet.

Where the numbers live

No single mandated filing contains all of this, and the belief that one does produces a specific failure: forwarding a Consumer Confidence Report and considering the question answered. The CCR reports detected regulated contaminants and source identification. It carries no rated capacity, no maximum-day production, no storage volume, no pump ratings, no pressure, and almost none of the chemistry an industrial user cares about.

Here is the request list your utility can work from:

• Withdrawal authority and conditions — state permit, water right, reservoir contract, or wholesale agreement, with daily, seasonal, drought, and expiration terms
• Safe or firm yield — approved yield study, well pump test, reservoir study, state supply plan
• Rated capacity — approved engineering design report and construction permit
• Permitted capacity — state waterworks operating permit
• Average-day and maximum-day production — SCADA and production logs
• Storage, pumps, standby power — sanitary survey file, asset register, pump curves, generator test logs
• Main size, material, pressure zone, parcel distance — utility GIS, record drawings, hydraulic model
• Static and residual pressure — model run or dated field test
• Process-relevant chemistry — the utility's laboratory information system

North Carolina's Local Water Supply Plan assembles more of this in one filing than any other state instrument I have reviewed: distribution materials and size ranges, finished storage, emergency interconnections, industrial use, monthly average and maximum-day withdrawals, wholesale contract volumes and expirations, plant permitted capacity, and production excursions above 80% and 90%. If your state maintains an equivalent, start there.

The capital improvement plan is the second document to learn to read, and it is where you can outrun your own utility. A project in a CIP occupies one of three states:

1. Listed — no appropriation.
2. Funded — appropriated but unscheduled.
3. Bid — design complete, easements acquired, bids opened against a budget.

Only the third supports a date. Rome's water extension reached it. An apparent low bid of $3.47 million against a $3.4 million budget establishes cost and procurement status, each with a date attached. It does not establish completion. The city had not set a construction timetable when it opened bids. Bid state supports a defensible cost and sequence, and saying that limit out loud is what keeps the rest of the package credible. Read the CIP against the current rate study and you can state an extension sequence, design through easement acquisition through bid through construction, with a responsible party at each step instead of a guess at each step.

None of this arrives pre-assembled as a site-selector answer. The utility has to reconcile it into a dated, parcel-level statement with a named signer and a refresh trigger, and that reconciliation is the one piece of the package an EDO cannot produce alone. Request it as standing work product on an annual update cycle. Request it as an emergency favor during a live RFI and you will find out on day two that your utility engineer is on vacation.

Drinking-water compliance is a floor

Source chemistry is the most recoverable unclaimed asset in a tier-3 system, and it goes unclaimed because directors read a clean compliance record as settling the industrial question.

EPA's secondary standards set chloride at 250 mg/L, iron at 0.3, manganese at 0.05, TDS at 500. They are non-enforceable aesthetic guidelines. They are not specifications for boiler feedwater, cooling towers, stainless process equipment, or branded beverage.

The distance between those guidelines and industrial specifications is wide. A DOE record carries a boiler manufacturer's feedwater recommendation of 0 to 5 ppm hardness, which means ordinary hard municipal water requires softening before steam generation. DOE's steam system guide shows allowable boiler-water silica falling from 100 mg/L at 150 psig to 3 mg/L at 1,500 psig, which is why source silica has no universal cutoff. It means something only when divided into an equipment limit at planned cycles. For stainless-intensive processes, the UK water industry standards board gives an ambient guide near 200 mg/L chloride for Type 304 and near 1,000 for Type 316, with corrosion possible below either figure under heat, deposits, or stagnation. Water that clears the taste guideline can still be a materials problem.

The EDO action here is narrow. Ask for a three-year raw and finished profile: hardness, alkalinity, silica, chloride, sulfate, TDS, conductivity, TOC, iron, manganese, color, pH, and monthly source temperature. It is an export from records the lab already keeps. Whether that chemistry fits a tenant's equipment and treatment economics is always the tenant's determination. But a city with soft, low-silica, low-chloride water and a stable seasonal temperature curve holds a real operating-cost advantage in steam- and cooling-intensive processes, and it is almost certainly not publishing the number, because nobody ever told the utility the number was worth publishing.

Fire flow and process flow are two screens

Fire flow is a high instantaneous rate held for a prescribed duration at a specified residual pressure. The 2024 International Fire Code defines it as water available at 20 psi residual, calculated from building area and construction type. Appendix B is not mandatory unless your jurisdiction adopts it, and its table starts at 1,500 gpm for two hours in the smallest nonresidential cases. Process demand is something else entirely: a daily and peak-hour operating load pulling on source, treatment, pumping and storage at once.

A parcel can pass one and fail the other in either direction. Storage and network pressure will carry a two-hour fire event at a site with no 24-hour production headroom. A system with ample daily production will fail the fire screen at an undersized dead end.

The pre-prospect artifact is a hydrant flow test under NFPA 291 practice, recording static, residual and pitot pressure, observed flow, and projected flow at 20 psi. Validity is set locally. Greensboro treats a test as good for one year; San Bernardino County uses six months. Find out which applies and put the expiration on the face of the document. The building's required fire flow stays project-specific. Construction type, area, sprinkler design, commodity, storage arrangement, the fire official and the insurer all move it.

What certification buys

Select Tennessee publishes the most detailed public water field list of any program worth benchmarking against: minimum 150,000 gpd, infrastructure within 500 feet or a utility-approved engineered extension, plus a utility letter covering line size, nearest tank distance and volume, static and residual pressure, excess treatment capacity, reserve capacity, extension cost, schedule, right-of-way and funding. Use it as a template outside Tennessee. Do not quote the 150,000 gpd as an industry threshold. It is a certification rule.

The other three programs reviewed publish no numeric water minimum whatsoever. California's Business Ready Sites requires utilities within 500 feet or a formal extension plan, with a will-serve letter carrying timelines. FAST NY Track C requires documented utility engagement, an upgrade plan and cost estimates. Virginia's December 2025 guidelines require preliminary engineering at Tier 3 and, at Tier 4, infrastructure in place or plans a Virginia PE deems deliverable within 12 to 18 months. Not one of the four reserves a gallon of water for anyone. They document, which is the standing limit on certified-site checklists generally: certification is a documentation threshold, never a guarantee of service.

Those are 2024–2025 program standards. What they jointly establish is that the water question gets asked and answered at the certification stage, before a prospect exists. The next step is my inference rather than survey data: a consultant screening a long list does not reconstruct your utility's capacity arithmetic for you. A missing number reads as a missing capability.

Know the instrument names before the call

Bainbridge Island publishes the clearest public ladder I have found. An information response confirms service-area location and feasibility and produces no formal document. A non-binding commitment letter states that capacity is available as of issuance, expressly contingent on capacity still being available later. A binding commitment letter reserves capacity once participation fees are paid and permitting conditions are met, with payment due in 30 days, after which the commitment does not expire.

That ladder resolves the will-serve problem. A pre-prospect availability letter is credible precisely because it does not pretend to reserve anything. It states present conditions, assumptions, signer and refresh date.

On extension cost, the authorized paths are a short list, and your utility's policy already names which of them apply:

• The CPUC's model Rule 15 developer advance, refunded as later connectors attach.
• Cary, North Carolina's four routes: developer-funded with no reimbursement; oversize reimbursement for facilities sized to serve other tracts; subsequent-user reimbursement from later development fees for up to ten years; town-funded capital extension after a verified petition.
• Special assessment districts of the kind Auburn, Nebraska's code establishes, with cost assessed against abutting property to the extent of special benefit.

Which instrument applies, and what the allocation comes to, requires the project. Which instruments are authorized in your jurisdiction, with specimen terms and current fee schedules, does not.

The package, and its two lists

The 48-hour package is seven files behind a dated cover page:

• The utility availability letter
• A capacity worksheet showing the arithmetic term by term
• A GIS excerpt marking the connection point
• Dated model or hydrant results
• A three-year chemistry table
• The relevant CIP excerpt
• The current rate and fee schedule

What those files can contain with no prospect in sight:

1. Source type and withdrawal authority with conditions and expiration
2. Rated, permitted and firm capacity
3. Recent average-day and maximum-day production
4. Applicable state capacity triggers
5. Storage inventory and pump configuration with standby power and test records
6. Main locations, sizes and pressure zones with parcel distance
7. Current model or hydrant results with a validity date
8. Multi-year raw and finished chemistry with seasonal temperature
9. CIP status and preliminary extension routes
10. Authorized funding instruments and standard fees
11. An indicative availability letter with signer and refresh date

What genuinely requires the prospect:

1. Exact daily, maximum-day, peak-hour, seasonal and construction-phase demand
2. The utility's final engineering determination against that profile
3. Required fire flow for the actual building and its insurer's criteria
4. Process chemistry fit, pretreatment train and recovery economics
5. Final extension cost allocation and construction responsibility
6. A binding capacity reservation

Eleven and six. That ratio is why water is the most recoverable documentation gap in a tier-3 portfolio, and why a director who has not closed it has no good account of the delay.

One boundary restated, because it is where forwarded packages come back. All of the above concerns supply. Whether the receiving treatment plant can accept a given tenant's discharge, its BOD, TSS, fats and oils, or categorical pollutants, is a separate screen with a different controlling utility, turning on quantified plant headroom, local limits and a named pretreatment path. Nothing in your water records speaks to it. A food plant can clear the source, the main and the pressure and die on discharge authority. A dry fabricator needs fire and domestic water and never triggers the discharge question at all.

Two packages, two controlling authorities, whether or not they report to the same utility director. Both fall inside the same 48-hour window.