Window Washing Guide
GUIDE / WATER ATLAS / PENNSYLVANIA
WATER ATLAS · NORTHEAST

Water hardness in Pennsylvania

TYPICAL RANGE
110–380 mg/L
6.4–22.2 grains/gal
BANDS SPANNED
Moderately Hard → Extremely Hard
DOMINANT SOURCE
Blended surface and groundwater

Moderately hard surface water in Philadelphia and Pittsburgh; substantially harder limestone-valley groundwater in the western Philadelphia suburbs and the Lancaster County belt; pre-1900 housing density that dictates the working calendar across most of the state.

HOW IT BREAKS DOWNmoderate (cities) to very hard (limestone valleys and wells)

Philadelphia draws from the Schuylkill and Delaware rivers via three treatment plants and runs moderately hard at 130-160 mg/L. Pittsburgh on the Allegheny and Monongahela rivers runs slightly softer at 110-140. The limestone valleys of Chester County and the western Main Line — the Great Valley running from Phoenixville through Malvern and Paoli — produce well water and some municipal supplies that run 250-340 mg/L, which is a separate working problem from the city tap. The Lancaster County aquifer carries similar hardness and adds a manganese-discoloration issue on long-dwell sprinkler overspray. The northern tier (Williamsport, Scranton, Wilkes-Barre) is more variable and depends heavily on whether the municipal supply is surface or groundwater. The deciduous calendar — sweet gum balls in late October, London plane bark and seed in November, oak pollen in April — is the through-line that ties the trade across the whole state together.

A blended system means hardness moves with the blend. Two addresses on the same utility can read differently, and the same address can read differently across a year, depending on which source is carrying the load that season.

What that means for the glass

Pennsylvania spans 4 hardness bands, which means there is no single answer for the whole state — the method changes as you move across it. Both ends are below.

Moderately HardSOFT END

61–120 mg/L

Tap water is still fine for most residential glass, with two exceptions worth knowing. Dark glass shows a faint mineral haze that lighter glass hides, and glass that is hot enough to flash the water off before you pull the squeegee will spot regardless of how soft the supply is. On both, a distilled final rinse costs about a dollar a window and removes the variable entirely. Everywhere else at this level, the water is not what is wrong.

HardMIDDLE

121–180 mg/L

This is the band where the water starts writing on the glass. Wash with tap — the surfactant holds the minerals in suspension while you work, so the wash pass is not the problem — then rinse with distilled and pull that. The rinse is the whole intervention: it replaces the mineral-bearing water sitting on the glass with water that has nothing in it to leave. This single change resolves most of the "I cleaned it and it still looks bad" complaints in this range, and it does not require buying a system.

Very HardMIDDLE

181–250 mg/L

A distilled rinse stops being an improvement and becomes the method. Tap water left to dry on glass at this concentration deposits a visible film within minutes, and the film is cumulative: each cleaning that ends in tap water adds a layer that the next cleaning has to get through first. Glass on a sprinkler line or under a runoff drip needs a maintenance interval, not just a better wash — the deposit is arriving faster than a cleaning schedule built around dust would predict.

Extremely HardHARD END

251+ mg/L

The hardest water in North America — Phoenix, Las Vegas, West Texas, much of the limestone Midwest. At this concentration calcium carbonate precipitates out of any water that touches the glass and dries, so what you are looking at is essentially limestone, growing one molecular layer at a time. Tap water cannot be part of the final pass under any circumstance, including "just a quick rinse." Above roughly 300 mg/L, buying distilled by the gallon stops making economic sense against a DI filter or a pure-water pole system, and a residential owner cleaning their own glass twice a year is usually better served by a service that already owns one.

By city

Hardest first — the order that matters, because the hard end is where the method has to change. Each figure is a service-area typical for the named utility. Within Pennsylvania the spread runs from Pittsburgh at 125 mg/L to Lancaster at 260 — a difference big enough that the same bottle of cleaner behaves differently in each.

CITY / UTILITYBANDMG/L
LancasterCity of Lancaster Bureau of Water · blended surface and groundwaterExtremely Hard26015.2 gpg
AllentownLehigh County Authority / City of Allentown · blended surface and groundwaterVery Hard22012.9 gpg
ReadingReading Area Water Authority · blended surface and groundwaterVery Hard19511.4 gpg
PhiladelphiaPhiladelphia Water Department · surface waterHard1458.5 gpg
ErieErie Water Works · surface waterHard1357.9 gpg
PittsburghPittsburgh Water and Sewer Authority · surface waterHard1257.3 gpg

This page is about Pennsylvania's water and nothing else. For how cleaning actually works in Pennsylvania across the year — climate, seasonal timing, the local contaminants, the housing stock — see the Pennsylvania cleaning guide.

For a figure at your own address rather than your city, the Hard Water Scorer takes a ZIP code. If you run a pure-water system, the TDS diagnostic reads the other end of the same problem.

Nearby in the Northeast

New York
35–350 mg/L
New Jersey
80–240 mg/L
Massachusetts
18–220 mg/L

Where these numbers come from

USGS national hardness survey data and utility Consumer Confidence Reports, 2023 vintage. City figures are service-area typicals for the named utility, not readings from any one tap. State ranges are the lowest and highest typical municipal values across the state, so they bracket the populated area rather than describing an average resident. Where a state blends sources seasonally, the range is wider than any single address will ever see.

Private wells are outside all of it. A well is whatever the rock under it says, and the only number that describes one is a test of that tap. Have a reading that disagrees with this table? Send it to us — a number that contradicts the table is worth more than one that confirms it.