How do we know what we know?

1. 1We start with every record attached to the road section that is still active. Records withdrawn by their source, merged into another record or replaced by a newer version are left out.
2. 2We keep the ones that count as Confirmed service. RMWB statements are tracked separately, as the last officially reported date.
3. 3Each record is dated by when the work happened: its end time, else its start time, else the day it happened. The day an update was published is never used as the day of the work.
4. 4The newest one is the last confirmed service, shown only as precisely as its source allows.
5. 5If nothing qualifies since tracking started on Oct 24, 2026, the answer is Unknown.

The last officially reported date works the same way with one difference. When the RMWB names a street without saying when the work happened, we date it by the day the RMWB published it, and the page says the work date wasn't given.

A street page leads with the days since its last plow record, and so does the map's road card. That counts from the last confirmed service, or from the last RMWB report when nothing is confirmed, the same order the map's Last plowed view uses. Its badge says which kind of record it is. When an editor has found records that disagree, Conflicting evidence replaces both.

A record keeps the precision its source gave it: an exact time, an hour, a day, a week, a month or a season. We show a time only when we know it.

Daily counts put a record on every day from its first day to its last. A record that spans more than 14 days is left out of daily counts, because it can't say which day the work happened. The counts are rebuilt every 20 minutes.

Times are Fort McMurray time, written MT. Alberta stays on UTC-6 for good from Nov 1, 2026, so the clocks don't fall back this year. We store every time in UTC and convert it for display. Some computers still carry the old fall-back rule, and our code corrects for it, so no time after Nov 1 is shown an hour early.

An RMWB update and a work order can describe the same plowing run. We keep each as its own record and link the two, and the pair counts once. A record that covers only part of a road section is labelled partial, and partial coverage never adds kilometres.

A road section is the unit we track: one stretch of the RMWB's road map with its own ID, such as RLC1234. A street is all the sections that share a name, and kilometres add up section lengths. The three measures differ, so every share on this site says what it's out of.

We count the current road sections inside the Fort McMurray boundary on the RMWB's service-area map and leave out the 122 sections (66 km) of provincial highway (Highways 63 and 686), including the ramps of their interchanges. The RMWB says Alberta Transportation maintains Highway 63, including its interchanges, and doesn't say who maintains Highway 686 inside the city. Saprae Creek Trail, the former Highway 69, is municipal and counted.

42 road sections (14 km), such as resource roads and streets crossing a highway, don't say who maintains them in the RMWB's data. They stay in the count.

Kilometres count only records that cover a whole road section. An RMWB update that names a street doesn't say how much of it was done, so it adds road sections and no kilometres.

Environment Canada hasn't measured fresh snowfall at any Fort McMurray station since Nov 30, 2007. What it does measure is snow depth at the Fort McMurray CS station (climate ID 3062696), along with precipitation in millimetres of water and the temperature. So snowfall on this site is always an estimate, and it's labelled as one.

A day counts as a snowfall day when the snow depth rose by at least 2 cm from the day before. Snowfall days in a row, or with one day between them, make one snowfall, and the estimate is the sum of the daily rises. Snow settles as it piles up, so the rise usually understates how much fell. A blank depth reading counts as 0 cm right after a reading of 2 cm or less, and before the first reading of a season, because the readings on file start in the autumn and Environment Canada leaves the cell blank until snow lies on the ground. A rise counted from that start is never rated high confidence. A blank in the middle of a snowpack is a gap, and that day's rise is unknown.

Each estimate is checked against precipitation and temperature, and Environment Canada snowfall warnings issued around it are listed with it. Confidence is high only when every day has readings, every rise came with precipitation, the temperature fits and the estimate is at least 5 cm. It's low when readings are missing for more than half the days, or no rise came with precipitation, and medium otherwise.

The station is near the airport, about 13 km southeast of downtown and about 116 m higher, so the depth there can differ from the depth in town. An estimate of 5 cm or more counts as a snowfall for the map, the response times and the policy comparisons, and it's worded "estimate meets threshold".

See the snowfall history

For each snowfall estimated at 5 cm or more, we look for the first confirmed or RMWB-reported plowing or snow removal record on each road that starts after the snowfall began. The time is measured from the end of the snowfall, so a negative time means the work started while snow was still falling. When a record is known only to the day, the time is only as precise as that day. A road with no such record is Unknown.

We compare records with the written policy only where it gives a time window to compare against. Each comparison ends in one of six phrases, and none of them says anyone broke a rule. Statements on Promises vs. Pavement use five other words, listed there.

• Within the documented windowA rule with a time window applies, and a confirmed record shows the work inside it. A record known only to the day gets the benefit of the doubt.
• Officially reported completeThe RMWB reported the work done inside the window. That shows what the RMWB said, and no other record confirms it.
• Outside the documented window based on available evidenceThe window has ended, and a source that reports every street shows the work after it, or not at all. Records we don't have could still change this.
• Evidence incompleteThe window has passed without a record inside it, and no source reports every street, so the missing record proves nothing either way.
• No conclusion possibleThe window hasn't ended yet, or we can't tell when it started or which class of road this is. The RMWB hasn't published its road classes as data.
• Policy rule not applicableThe rule doesn't cover this road or this time: a provincial highway, a different class of road, a road outside the zones, or a schedule suspended by an Extreme Weather Response.

Only the RMWB's daily progress map or its plow records could report every street. Neither is connected today, so no comparison can end outside the window yet.

Policy OPE-150, in force since Oct 6, 2026, sets no hours-based deadline for Priority One, Two or Three roads. It sets an order instead, Priority One first, and calls for continuous plowing of Priority One routes in a Significant Snowfall or Severe Weather. Its time windows are in the zone program, where plowing all subzones takes two weeks and removing the snow takes four under normal weather conditions, and in a 24-hour limit for removing windrows of 15 cm or more from zone driveways and accessible parking stalls. Page 1 of the policy says its service levels "are not guarantees of service".

The policy doesn't say when the two-week clock starts. We start it at the first activation the RMWB publishes for the road's subzone, counting from the day before the snowfall began.

Residential streets are plowed through the winter maintenance zone program, from Oct 15 to Apr 15. There are five weekday zones with four subzones each, and subzones are worked every other week. Zone streets are plowed after 5 cm of snow, and while a subzone is active, parking there is banned from 7 a.m. to 6 p.m. on its weekday. During an Extreme Weather Response the zone schedule is suspended and rolling parking bans apply.

Zone activations come from the RMWB's notices. The zone outlines come from an RMWB map layer dated 2018 and last edited in 2021, so a road's zone label carries a warning.

The policy sorts roads into Priority One, Two and Three routes, plus residential streets in the zone program. The RMWB publishes its 2026 priority maps only as PDFs, so a road's priority reads Not published until a class is imported with a citation. We never guess a road's priority from how the map draws it.

A storm report covers one snowfall estimated at 5 cm or more. Only reports an editor has published appear, and the figures beside the text come from the same records as the rest of the site, so they change when a record is corrected.

Last plowed

Each road shows its newest confirmed record this winter, up to the day you pick, or else its newest RMWB report. Solid lines are confirmed and dashed lines are RMWB reports. Brighter means more recent, in five steps from within 24 hours to more than 14 days ago. Any recorded winter maintenance counts, including sanding and snow removal.

After snowfall

Pick a snowfall estimated at 5 cm or more. Each road shows whether a confirmed or RMWB-reported record of work starts after that snowfall began.

Residents

Moderated resident reports, one dot each, on the road point the report snapped to. The pin a resident dropped is never shown.

One day

Records of work on the day you pick, with that day's resident reports. A record that covers only part of a road is drawn fainter.

RMWB vs residents

Roads the RMWB reported serviced this winter, next to resident reports on the same roads. A disagreement shows only after it has been recorded; the map never guesses one.

Zones

The winter maintenance zone outlines, each with its status now, or scheduled when it starts within 7 days. The outlines come from a 2018 RMWB map, so they may not match this winter exactly.

A road with no record is drawn dotted grey, and solid grey lines are provincial highways, which this site doesn't count.

For each day and each source, coverage is the share of road sections in the count with at least one record from that source that day. Daily counts leave out records that span more than 14 days. Which sources are working right now is on the data sources page.

Unknown roads are dotted grey on the map and never get a recency colour. Nobody has published evidence for them, so we can't say whether they were plowed.

• No operational records are public. Section 63 of policy OPE-150 says the RMWB keeps records of when crews worked, with equipment, operators and GPS where available. Those would give us Confirmed service.
• The RMWB hasn't published the daily progress map that section 60(c) of the policy calls for. We last checked Oct 25, 2026, 10:41 p.m. MT.
• Snowfall is estimated from snow depth at one sensor near the airport, so it can differ from what fell in your neighbourhood.
• The zone outlines come from 2018 data, last edited in 2021.
• Road priority classes are published only as PDF maps.
• The RMWB publishes no licence for its map data. We credit © Regional Municipality of Wood Buffalo.
• A resident report describes one place at one moment, and not every road gets reported.
• Alberta 511 needs a key and covers provincial highways only, so we don't use it.
• A missing record never means a road wasn't plowed.

These are the definitions our code uses, written as database queries with the column names from our database schema.

-- Road sections every count is out of
select count(*) as road_sections, sum(length_m) / 1000 as km
from core.road_segment
where retired_at is null
  and not is_synthetic
  and community_code = 'FMM'
  and road_authority <> 'provincial';

-- Last confirmed service for one road section this winter
select max(coalesce(e.ended_at, e.started_at,
                    ymm_local_day_start(e.occurred_local_date))) as last_confirmed
from core.maintenance_event e
join core.maintenance_event_segment es on es.event_id = e.id
where es.road_segment_id = :segment_id
  and e.status = 'active'
  and not e.is_synthetic
  and e.verification_status <> 'retracted'
  and e.evidence_level in ('official_operational_record',
                           'independently_verified_observation')
  and coalesce(ymm_local_date(e.ended_at), e.occurred_local_date,
               ymm_local_date(e.started_at)) between date '2026-10-24' and date '2027-09-30';