In March, 225,817 people visited Yosemite. That is 45 percent more than the March before it and the busiest March this park has seen since 2016. Nobody who lives here was surprised. The National Park Service announced in late December that the entrance reservation system would not return in 2026, and the visiting public heard it as an open gate. Through April, the park has logged 836,458 visits against 739,313 over the same months last year. Then Memorial Day arrived and gave us the preview: 90-minute entrance lines, the Curry Village lot full by 8:30 in the morning, cars parked on the meadow edges because there was nowhere else to put them.
I made my argument about whether this was a good idea after that weekend. This piece is not that argument. This is the practical question that lands in my inbox every week now: with no reservation system and visitation climbing, when should you actually come?
To answer it properly I went through the National Park Service's visitor use statistics: monthly visitation for the past decade, the year-to-date numbers for 2026, and the gate-by-gate traffic reporting the park publishes in its press releases. Then I built a month-by-month forecast for the rest of 2026. The numbers, the method, and the charts are all below, and so are the specific days I would pick.
Some history, because the data makes no sense without it. Yosemite has spent six years running an accidental experiment in demand management:
Notice the direction of that visitation column. Even with throttles in place, demand has climbed every single year since 2020. The reservation systems never reduced the desire to be here. They moved it, spread it, and metered it. What 2026 removes is the metering, in a year when the underlying demand is the highest it has been since the 2016 record of just over 5 million visits.
Here is the year so far, with my projection for the months ahead, plotted against what 2025 actually did.
The year-to-date math is simple and striking. Through April, 2026 is running 13 percent ahead of 2025. The growth is not evenly spread. January and February ran modestly ahead of last year. March, the first full month after the no-reservation announcement made national news, exploded: up 45 percent, although that comparison flatters 2026 a little because storms held down March 2025. April landed in line with a strong 2025 April. The lesson of the spring is that headlines move visitation, and the headline this year is that Yosemite is open, no ticket required.
My projection for the rest of the year puts 2026 at roughly 4.6 to 4.7 million total visits. If that holds, 2026 becomes the second-busiest year in the park's recorded history, behind only 2016 and ahead of 2019 and 2017. July alone projects to about 665,000 visits, which would be the biggest single month here in nearly a decade.
To see why I expect summer to set near-records, look at what the throttles did to the shape of the last two summers.
In 2024, with reservations covering 5 a.m. to 4 p.m. across most of the season, July topped out at 596,711. In 2025, with the requirement trimmed to 6 a.m. to 2 p.m. and only eight weeks of coverage, July rose to 628,400 and June to 607,410. Every loosening of the valve has produced more volume, immediately. The park's own August 2025 reporting showed the surge arriving on every road: visitation through the Highway 140 gate at El Portal and the Highway 41 gate at Wawona up about 8 percent year over year, and up more than 10 percent on Highway 120 from the west.
2026 removes the valve entirely in a year that was already pacing at record demand. That is the basis of the dashed line above, and it is why the people who study this park's traffic were not surprised by Memorial Day weekend. The only force left to cap a July Saturday is the road network itself: when the Valley loop saturates, the line simply backs up outside the gates, which is exactly what the two-to-three-hour afternoon delays the park now warns about look like in practice.
The method is deliberately simple, and I would rather show it to you than hide it. I took 2025's monthly actuals as the base. For each remaining month of 2026, I applied a growth rate anchored to what 2026 has already demonstrated: the 13 percent year-to-date pace, damped hard in June, July, and August. The damping matters. March can absorb a 45 percent jump because March has empty road. July cannot, because July was already operating near the ceiling of what the park's entrances, lots, and Valley loop can physically process in a day. For the peak months I assumed 6 to 7 percent growth, which is roughly what the system demonstrated it could swallow when the 2025 restrictions came off in mid-August. Shoulder months got 10 to 15 percent, because that is where the headroom is and where displaced summer visitors tend to land.
Honest caveats: the 2026 figures are preliminary and the Park Service revises them. A bad smoke season can erase an August. A government shutdown, a Merced River flood, a midsummer policy reversal would all bend the curve. This is a forecast built on five months of data, not a promise. But the direction is not in doubt, and the direction is up.
The forecast turns into a usable plan when you break it into days. The grid below is my crowd-pressure index for 2026: projected visits per day, split by weekday and weekend, scaled so that 100 equals a peak July weekend.
Read it month by month, with the second half of the year in mind:
Here is the part the monthly numbers hide: the difference between a miserable Yosemite day and a great one is mostly the hour you reach the gate. The park's traffic guidance is blunt about the shape of the day, and it matches what I see on Highway 140 every morning.
It is no accident that every reservation system this park has run gated the same hours. The 2024 system covered 5 a.m. to 4 p.m.; 2025 covered 6 a.m. to 2 p.m. The Park Service has effectively published its own definition of when Yosemite is over capacity, twice. With no system in 2026, you enforce it on yourself:
If I were planning a 2026 trip from outside the area, in order of preference:
Forecasts age badly in public, so let me say what would change this one. If the early-season chaos keeps producing national coverage, some fraction of casual visitors will self-select out and the summer numbers come in under my line. If smoke closes the air in August, that month craters. And if the gridlock gets bad enough, the park retains the authority to manage entries again; the Park Service has said it will manage 2026 with staffed intersections, parking management, and real-time traffic monitoring instead, but that toolkit has limits a busy Saturday will find quickly.
None of that changes the planning logic. Demand for this park rises every year the gate is open, the gate is wide open in 2026, and the only variables you control are the month, the day, and the hour. Choose all three on purpose. The park at 6:45 on a September morning is still the same park that made the record books, minus the line to get in.