Snow removal invoices: the photo shows a clear lot, the weather record shows whether it snowed

The December invoice from the snow contractor lists nine visits across your three buildings, each billed per push. Attached are eighteen photos: a parking lot with snow on it, then the same lot clear, for every visit. You approve it in March along with everything else from the winter, and by then nobody remembers which mornings it actually snowed.

This piece is about checking that invoice against two records you can get for free: the camera’s own clock inside each photo file, and the public record of how much snow fell. Neither one is proof, and we’ll be specific about why. It is not legal advice.

Why the “after” photo proves the least

A cleared lot looks the same whether it was plowed at 5 a.m., melted by noon, or never snowed on at all. The photo that carries the weight is the “before” one: snow on this lot, on this morning. And a picture of snow is the easiest photo to reuse, because one snowy lot looks very like the next, and like the same lot a week earlier.

We have written before that the repair-invoice fraud to expect first is not AI. It is the same photo billed twice. Snow contracts are where that is easiest: dozens of visits, near-identical scenes, billed in batches months later.

Record one: the camera’s clock

When a phone takes a photo, it writes the date and time its clock showed into the file, often with the time zone, alongside the phone’s make and model. Our checker reads that in your browser. A phone original comes back as “Camera Original”, with the device and the capture time printed, so you can set each photo beside the visit date on the invoice.

Two things to look for:

• Dates that don’t match the visit. A “before” photo for the 14th whose clock says the 3rd is a question for the contractor. It may have an innocent answer: a wrong photo attached, a phone clock set wrong. But it is a specific question.
• The same file used twice. Every saved report lists each file’s SHA-256, a fingerprint of its exact bytes. If you check the whole winter’s photos in one batch, any two identical files are marked “Identical file”, whatever they are named. The checker only compares files you put in front of it together, and only exact copies: a resized or re-saved copy gets a different fingerprint.

Phones can also record where a photo was taken. Our checker does not read or show location, and plenty of crews have it switched off, so don’t build your process around it.

Record two: whether it snowed

NOAA’s National Centers for Environmental Information publish daily snowfall and snow depth from thousands of US weather stations, many of them run by volunteers. There are two free ways in:

• Daily U.S. Snowfall and Snow Depth: pick a state, month and year, and get a table of each station’s snowfall for each day, in inches, downloadable as CSV. This is the quickest way to see a whole month at once.
• Climate Data Online: choose Daily Summaries, set the dates and search by ZIP code. NOAA’s own walkthrough says “(these data are free)”; the file arrives by email.

Then line the invoice up against it. Nine billed visits in a month when the nearby stations report two snowfalls is worth a conversation. Nine visits in a month with eight snowfalls probably isn’t.

Why a mismatch is a question, not proof

The weather record is easy to misread, and every one of these has an innocent explanation:

• The station is not your lot. NOAA’s guide warns that “not all stations record all variables; about half the stations only report precipitation.” The nearest one that measures snow may be miles away, and in the Mountain West a few hundred feet of elevation can mean the difference between six inches and rain.
• The day is not midnight to midnight. Volunteer observers report the 24 hours ending at an observation time that suits them, “most often 7:00 AM or 8:00 AM”, according to the Minnesota DNR’s explanation of the same data. Snow that fell at 4 a.m. and was plowed at 6 can sit in a report dated the day before or the day of, depending on the station. Look at the days on either side.
• Not every visit follows new snow. Wind drifts snow back over a cleared lot, and ice treatment can be needed after a thaw and refreeze with no new snow at all. Read the contract for what a billable visit is before you count.
• The record is not instant. NCEI checks station data before publishing it, which it says delays public availability, and early numbers can be revised. Check last month, not this morning.

So the useful finding is narrow: this visit was billed on a day with no snow reported nearby, and its “before” photo was taken on a different day. Two independent records disagreeing with the invoice, both of which the contractor can see and answer. Neither alone is an accusation.

Where AI comes into it

AI image tools can now produce a convincing photo of a snowy parking lot, or add snow to a real one. If the file says it was made or edited with AI, through Content Credentials or a tool’s own marker, the checker says “AI Generated” or “AI Edited” and names the tool when the file does. Many files say nothing. We have written about what the file records when a photo is AI-edited, and how often it records nothing.

A file with no camera record at all comes back “Can’t Verify”. That is not a fake. It is what happens to an honest photo sent by text or through WhatsApp or Messenger, which strip the record as they send. It means: ask for the original.

A routine for this winter

1. Put one sentence in the snow contract: each billed visit is supported by a before and an after photo of the property, sent as the original files (email attachment or upload, not a screenshot or a messaging app) with the invoice.
2. Ask for a wide shot with something fixed in frame: the building sign, a numbered light pole, a dumpster enclosure. A close-up of snow on asphalt could be anywhere.
3. Each month, run that month’s photos through the checker in one batch. Note any capture date that doesn’t match its visit, any “Identical file” pair, and any “Can’t Verify”. Ask about those while the crew still has the originals on their phones.
4. Download the month’s snowfall for your area and mark the billed visits against it, reading the days either side.
5. Save the report with the invoice. It is dated and lists every file’s SHA-256, so if the bill is disputed later, anyone can confirm the file you produce is the one you checked.

The limits, stated plainly

• “Camera Original” is not proof the lot was plowed, or that the photo shows your lot. It means a camera or phone recorded the file and the record is intact. Camera clocks can be set wrong, and camera data can be edited by someone who means to.
• The weather record describes a station, not your property, over a day that may not start at midnight. A billed visit on a dry day is a question, not a finding.
• Our checker gives no confidence score and does not judge a picture by looking at it. A file with no record is not a fake, and a file with no AI marker is not proven real.
• We have not seen any snow contractor’s billing practice and are not suggesting the industry bills falsely. Almost every invoice will check out; the point is to know which ones to ask about.
• NOAA pages were read in October 2026. NOAA has noted that some ordering services are intermittently slow while it moves to new systems.
• This is not legal advice.

---

ImposterShield reads the evidence inside image and video files. It runs in your browser and uploads nothing. If a file has been stripped, it says so. That is the product working, not failing.