When is bee swarm season in North America?

Swarm reports build from late March and peak in May. In 2026 the first 10% of the year's North American reports (counted to 20 September) were in by 31 March, the busiest week began on 11 May, half had arrived by 6 June and 90% by 1 September. May carried about a fifth of the year's reports. A smaller second rise follows in late summer, concentrated in the South.

Why do bee reports rise again in September?

Mostly because people find established colonies, not swarms. Across 2024 and 2025, three quarters of September and October reports were colonies living in walls, soffits and roof voids, against a third in April and May. What remains once those are set aside is a genuine late swarming period in the South: below about 37 degrees north, swarm reports roughly doubled from August to September, while the band from 37 to 42 degrees north stayed flat.

Did the 2026 swarm season start earlier than 2025?

Swarmed doesn't publish a year-over-year timing figure. When we tested shifts in report timing between 2025 and 2026, they tracked the growth of the reporting network more closely than local spring weather, so a days-earlier number would describe our growth rather than the bees.

Are there more or fewer bees in 2026?

Swarm report counts cannot answer that, because they track the growth of the reporting network rather than the number of swarms, which is why this report states dates and proportions rather than totals. On colony losses, US beekeepers lost 55.1% of colonies in 2023-24 and 55.6% in 2024-25, the two highest years since annual tracking began, and the 2025-26 survey published in June 2026 reported losses back down to 39.9% annual and 30.3% winter, a return to the level of the preceding decade.

← Bee Swarm Data

Swarmed Season Report · Published 21 August 2026 · Revised 29 September 2026

The 2026 North American Swarm Season

Q: When is bee swarm season in Canada?

Later than in the United States. American swarm reports peak between April and June, while Canadian reports peak in July, on a season that runs from the second half of May into September. A Canadian beekeeper following American swarm-season advice sets bait hives roughly six to eight weeks early. The figures in this report are continental and weighted toward the United States, where almost all of our reports originate; the Canadian base is 89 reports as of 26 August 2026, which is enough to date the national peak and not enough to date an individual province.

The season peaked in mid-May, and its autumn rise is mostly colonies already living in buildings

Swarmed's network watched the 2026 North American season build from late March, peak in the week of 11 May and pick up again in late summer. This report sets out the shape of that year: when reports came in, what the late-summer rise was made of, and where it was concentrated.

31 March

The first 10% of the year's reports were in

Counting reports from 1 January to 20 September 2026. Half had arrived by 6 June and 90% by 1 September.

21%

Of the year's reports came in May

The busiest month, with the peak week starting on 11 May. April and June each carried a little over 15%.

3 in 4

Autumn reports were established colonies

September and October, pooled across 2024 and 2025, against a third in April and May. Most of the autumn rise is colonies in buildings, not swarms.

Published by beeswarmed.org · 21 August 2026 · Based on Swarmed's North American swarm report record, collected since March 2024

If you have found a swarm while reading this, you can report it free of charge and we will alert the beekeepers registered near you, usually within minutes. Everything below is a record of how the season ran.

Line chart of weekly North American swarm reports through 2026, with the same weeks of 2025 for reference. The 2026 curve rises from late March, peaks in the week of 11 May, eases through July, and climbs again from August into September. — Swarm reports across North America, week by week. The 2026 season built from late March, peaked in mid-May, and picked up again in late summer, as 2025 also did. Dotted lines mark the points at which 10%, 50% and 90% of the year's reports to 20 September had come in. 2025 is shown over matching weeks for reference only: the gap between the two curves mostly reflects how much the reporting network grew, not how many more swarms there were.

Two record years of colony loss

Beekeepers entered the 2026 season after two consecutive record years of colony loss, 55.1% in 2023–24 and 55.6% in 2024–25, the highest since annual tracking began, which left unusually high demand for caught swarms as replacement stock. The 2025–26 survey, published in June 2026, put losses back down at 39.9% annual and 30.3% winter, a return to the level of the preceding decade.

“The 2024-25 Apiary Inspectors of America survey showed beekeepers lost 55.6% of their colonies over winter, the highest loss rate since the annual survey began in 2010. So this spring beekeepers are extremely eager to catch swarms to replenish their hives with strong local bees.”
Mateo Kaiser, Master Beekeeper, Swarmed

Swarm Season

Swarming is how honey bee colonies reproduce, and when a colony grows strong enough it splits, the old queen leaving with roughly half the workers to find a new home while the bees left behind raise a new queen, which in spring is generally a sign of a colony doing well rather than one in trouble.

Most swarms go unreported, which is the gap Swarmed's own network has been filling with georeferenced North American reports since March 2024, alongside archival records from iNaturalist reaching back to 1967 that extend the historical baseline.

Earlier seasons generally reflect colonies that built up strength faster, which in turn reflects winter survival, forage availability and spring weather, so the timing of reports gives beekeepers and researchers a ground-level view of what colonies are doing well before any of it reaches survey data.

How the season actually ran

The shape of a season carries more information than its start date. In 2026 the first 10% of reports to 20 September were in by 31 March, the halfway point fell on 6 June, and 90% had arrived by 1 September.

May was the busiest month, carrying about a fifth of the year's reports, and August ran ahead of July as the late-season rise began.

Share of 2026's North American reports, by month

Jan1.1%
Feb1.6%
Mar7.3%
Apr15.4%
May21%
Jun15.7%
Jul12.6%
Aug14.7%
Sep10.7% (partial)

Shares of reports from 1 January to 20 September 2026; September was still in progress.

The season has a second, smaller act

In both complete years on record, reports per beekeeper rose in September after falling through July and August, up 63% over August in 2024 and 38% in 2025 once established-colony removals are set aside, and pooling both years September's share of reports runs well above what the growth in registered beekeepers over that month would predict (p = 0.0007), so it is not simply more people joining.

What rises in autumn is not really the same thing as the spring peak, given that three quarters of autumn reports (75% in September and October) are established colonies in walls, soffits and roof voids rather than swarms, against 33% in April and May, which makes autumn mostly the season when people notice the colony that moved into their house back in spring.

What remains of the September rise once cutouts are removed sits at lower latitudes, with the 30–37°N band roughly doubling from August to September across 2024–2025 and the band south of 30°N doing much the same while the 37–42°N band stayed flat, which is the signature of late-season reproductive swarming in places whose autumn forage window runs long enough to support it, and matches the secondary swarming period Rutschmann et al. (2025) recorded in unmanaged colonies.

Two panels. Left: swarm reports per beekeeper from June to December as a share of each year's May peak, for 2024, 2025 and 2026, showing a September uptick in 2024 and 2025. Right: bar chart of the share of reports that were established colonies by month, at 33% in spring and 75% in September and October. — Left: swarm reports per registered beekeeper through the late season, shown as a share of each year's own May peak, with established-colony removals excluded. Right: the share of reports that were established colonies rather than swarms, by month, pooled across 2024 and 2025.

“By running a national swarm reporting network we're able to get more swarms to beekeepers, but we're also building up data on seasonal patterns that we couldn't see before, and share those with beekeepers and researchers.”
Mateo Kaiser, Master Beekeeper, Swarmed

What Swarmed does

When someone spots a honey bee swarm, they report it free of charge. We alert the beekeepers registered near them, usually within minutes, and a beekeeper who takes it on relocates the swarm safely instead of it being exterminated or moving into a wall. The service is free for the public and free for beekeepers.

Not sure what you are looking at? Learn how to identify a bee swarm and tell honey bees apart from wasps and bumble bees.

Beekeepers can sign up to receive swarm alerts in their area and help protect honey bees in their community. More than 12,000 beekeepers across North America, Europe and Australia are registered on the network.

Live reports, right now

Everything above is a closed-season retrospective on 2026, whereas the feed below is live, updating as reports come in, and forms no part of any figure on this page.

Swarm tracker

This page is the closed-season summary. The tracker behind it runs all year: every swarm reported to Swarmed carries a GPS coordinate, a date and an outcome, and the North American record runs from March 2024, which is three seasons. Report totals are deliberately absent from the findings above, because they track how fast the reporting network grew rather than how many swarms occurred. For the same reason, treat the live counts below as activity on the network, not as a census of bees.

Live reports : swarms as they are reported, across the whole network.
Swarm lists by state : season totals, how many went unclaimed, and peak timing for each state we publish.
Report a swarm : adds to the record and routes it to nearby beekeepers.

Among the counties with at least 10 swarm reports in both 2025 and 2026, these have their own local swarm pages: Ventura, Riverside, San Diego, Contra Costa, Los Angeles, Santa Clara, San Mateo and San Bernardino in California, Maricopa in Arizona, Clark in Nevada, and Broward, Miami-Dade and Palm Beach in Florida.

Methodology & Data Sources

Scope.Everything on this page covers North America, taken as latitude 15 to 72°N and longitude 170 to 50°W, so the roughly 11% of 2026's reports that came from Swarmed's European and Australian operations are excluded from every figure here, as are test entries, deleted records and reports identified as wasps rather than honey bees. Established-colony removals, which beekeepers call cutouts, are counted throughout except in the late-season section, where the caption says they are not.

No year-over-year timing claims.This report makes no claim about whether the 2026 season came earlier or later than 2025's. When we tested year-over-year shifts in report timing, they tracked the growth of the reporting network more closely than local spring weather: raw counts peaked two weeks earlier in 2026 than in 2025 while Swarmed's registered-user base grew from 1,917 to more than 12,000, and expressing volume per registered beekeeper makes that shift disappear and weakly reverse. Earlier versions of this page led with season-start shifts in days; they have been withdrawn, and for the same reason report totals appear nowhere in the findings above.

The dataset

Every report in Swarmed's own network carries a GPS coordinate, a date and an outcome, and the North American record runs from March 2024, which gives three seasons of tracking in 2024, 2025 and 2026, with archival iNaturalist records reaching back to 1967 extending the historical baseline. Report totals appear nowhere in the findings above, since they measure how quickly the reporting network grew rather than how many swarms occurred.

The 2025 reference curve

2025 appears here as a reference curve in the weekly chart and in the late-season comparison, and it had an odd autumn, with October to December report volume running 18 to 29 times its 2024 level while registered-user growth had already fallen back to roughly 1×, a pattern that belongs to a one-time outreach push rather than to bee activity, and one that had decayed back to 0.95× by July 2026. The weekly chart shows 2025 over matching weeks only.

The weather model

Swarmed scores daily weather conditions for each monitored region using five variables: accumulated growing degree days since 1 January, a standard measure of accumulated warmth above 10°C used in agricultural forecasting to track biological readiness in insects and plants; maximum daily temperature; peak solar radiation; the number of warm days in the preceding week; and whether rain occurred.

Surge events

When a colony has been confined by rain or cold for three or more consecutive days, swarming pressure accumulates. The first suitable flying day after such a period produces a spike in swarm activity. Swarmed detects these events automatically and sends alerts to beekeepers in affected regions. Surge conditions require daily solar radiation above 25 MJ/m², maximum temperature above 20°C, and no rain on the release day. About 18% of region-days in the weather record clear that radiation bar.

Habitat factors

Two locations with identical weather can still differ in swarm activity, so each is also scored on historical swarm density at the site and on proximity to active registered beehives within 3km. Mature tree canopy cover, taken from the Meta/WRI Global Canopy Height Model (Tolan et al., 2024), sits alongside those as a third factor and remains unvalidated, its apparent predictive power disappearing as soon as you restrict the comparison to locations that actually carry a measured value, which suggests coverage rather than canopy was doing the work. It is named here because it is in the model, not because we lean on it.

Trap placement

Swarm trap recommendations follow the nest site preferences established by Thomas Seeley's controlled experiments at Cornell and Appledore Island (summarised in Honeybee Democracy, Princeton University Press, 2010): 40-litre cavity volume, entrances of 12.5–75 cm², mounted about 4.5 metres up, southeast-facing orientation. The same figures drive the trap recommendations in the Swarmed dashboard.

Key references

Henneken, R., Helm, S., & Menzel, A. (2012). Meteorological influences on swarm emergence in honey bees (Hymenoptera: Apidae) as detected by crowdsourcing. Environmental Entomology, 41(6), 1462–1465. doi:10.1603/en12139
Rutschmann, B., Kohl, P. L., & Steffan-Dewenter, I. (2025). Swarming rate and timing of unmanaged honeybee colonies (Apis mellifera carnica) in a forest environment. Insectes Sociaux, 72(4), 505–512. doi:10.1007/s00040-025-01028-y
Tolan, J., et al. (2024). Very high resolution canopy height maps from RGB imagery using self-supervised vision transformer and convolutional decoder trained on aerial lidar. Remote Sensing of Environment, 300, 113888. doi:10.1016/j.rse.2023.113888
Seeley, T. D. (2010). Honeybee Democracy. Princeton University Press.
US Beekeeping Survey, 2024–25 and 2025–26. Apiary Inspectors of America / Auburn University / Oregon State University. 2025–26 results published 24 June 2026.

Data limitations

The 2024 season carries too few North American reports to serve as a comparison year, and 2025's late autumn is distorted by a one-time outreach push, which is one more reason this page makes no year-over-year timing claim. Late-season figures for 2026 stop at the 21 August cutoff. The week-by-week chart and the monthly shares were refreshed on 27 September 2026 and run to 20 September.

When is swarm season in Canada?

Later than in the United States, and by enough to matter. American swarm reports peak between April and June. Canadian swarm reports peak in July, with a season that runs from the second half of May into September and is still producing in August when most American beekeepers have finished. A Canadian beekeeper working from American swarm-season advice will set bait hives roughly six to eight weeks early.

The figures above are continental and weighted toward the United States, where almost all of our reports come from, so read them as the North American picture rather than the Canadian one. Our Canadian base is thin: 91 reports as of 2026-08-26. It is enough to date the peak and not enough to date a province, which is why there is no provincial breakdown here.

Province by province, including what each one requires of a registered beekeeper, is on beekeeping in Canada, with the two provinces we have real coverage in at Ontario and British Columbia.