Three NPS Units Recorded Zero Visits During the 2025 Government Shutdown — All Three Are Staffing-Gated, Not Access-Gated

USE CASE. NPS regional planners and unit superintendents preparing FY27 budget appeals need per-park quantitative shutdown impact attribution to support arguments for staffing restoration funding and for shutdown-resilience capital improvements (automated entry counters, unstaffed kiosks). Congressional Interior Appropriations staff need to know which categories of parks are most damaged by shutdowns to weight closure cost-benefit during future continuing-resolution debates. Trade press covering the shutdown reports system-wide decline numbers but not per-park structural attribution. The non-obvious normalization here is share-of-annual rather than absolute count: a park might have 'low Oct-Nov 2025 visits' simply because Oct-Nov is its off-season, but a park whose Oct-Nov *share* collapsed against its own 5-year baseline is a park where the shutdown changed the visitor pattern. RESULT. Of 273 qualifying NPS units, 215 (78.8%) had a lower Oct+Nov 2025 share of annual visits than their 2020-2024 mean, and 58 had an elevated share. The system-wide deviation was -0.9 percentage points (Oct+Nov 2025 share = 14.13% vs baseline 15.03%), meaning the system as a whole barely shifted. The shutdown impact was concentrated on a specific subset. Three NPS units had ZERO Oct+Nov 2025 visits despite baselines of 15-24% of annual: Pennsylvania Avenue NHS DC (PAAV, baseline 16.94%, 100% collapse), Governors Island NM NY (GOIS, baseline 23.92%, 100% collapse), and The White House and President's Park DC (WHHO, baseline 15.00%, 100% collapse). 17 more units recorded between 0.79% and 3.76% of annual visits in Oct+Nov 2025 with baselines of 5.69% to 23.39% — a >78% collapse in shutdown-window share. The top 5 after the three zeros: Wright Brothers NMem NC (WRBR, 0.79% vs 14.27%, -94.5%), Pipe Spring NM AZ (PIPE, 1.06% vs 11.42%, -90.7%), Minuteman Missile NHS SD (MIMI, 1.00% vs 9.17%, -89.1%), Jewel Cave NM SD (JECA, 0.77% vs 5.69%, -86.5%), Ocmulgee Mounds NHP GA (OCMU, 2.65% vs 19.40%, -86.3%). STRUCTURAL READING. Every one of the top 20 collapsed-share units shares a common operational pattern: visitation is GATED BY FEDERAL STAFF rather than gated by physical access. Pennsylvania Avenue, the White House, and Governors Island are NPS units where the 'visit count' is generated by NPS interpretive staff on duty (PAAV is essentially the DC sidewalk corridor, recorded only when ranger-led programs are running; WHHO visits are tour groups; GOIS requires the National Park Service to operate ferries and the visitor center). Wright Brothers, Pipe Spring, Minuteman Missile, Jewel Cave, and Ocmulgee Mounds all have a single staffed visitor center as the only entry point — when the staff are furloughed, the gate closes, the parking lot empties, and the count goes to zero. By contrast, parks like Great Smoky Mountains, Blue Ridge Parkway, Natchez Trace Parkway, and Golden Gate NRA have through-roads and unstaffed entry points; vehicles continue to enter during a shutdown and the count keeps incrementing even with no rangers on duty. The shutdown's recorded-visit impact is therefore not 'how many people stopped going to national parks because of the shutdown' but 'how many parks have a recording method that depends on federal staffing'. The 20-unit subset identified here is the actionable category: these are the parks where shutdown-resilient automated visitor counting infrastructure (loop detectors, IR counters, gate sensors) would be the highest-leverage capital investment, both for accurate visitation data during future shutdowns and for FY27 budget defense. CAVEATS. (1) The 2.7% system-wide YoY decline matches the official NPS press release; the shutdown attribution per park is the new contribution. (2) Three more parks (RECA, etc.) appeared in earlier ranks of the iter 76 raw YoY analysis but were excluded from the shutdown-attribution analysis for missing baseline years. (3) STON (which had a separate 87% annual decline) appears in the elevated-shutdown-share list (+247%) — its annual visits were almost entirely concentrated in Oct+Nov, suggesting a different operational story (likely a long-term closure that reopened only briefly) rather than a shutdown signature.

When the federal government shut down in October and November 2025, news coverage said 'the national parks lost visitors.' The NPS later confirmed that 2025 visits were down 2.7% from 2024 — about 9 million fewer visits across the system. But the average 2.7% number hides something more specific: most parks barely budged, while a small group of parks completely zeroed out. I downloaded the NPS monthly visitation file and asked: which parks lost the most of their normal October-November visitation in 2025, controlling for what each park's October-November share normally is in non-shutdown years? The answer is striking. Three parks recorded ZERO visits in October and November 2025: Pennsylvania Avenue in DC, Governors Island in New York Harbor, and the White House. In a normal year, October and November account for 15-24% of these parks' annual visits. In 2025 they got nothing. Seventeen more parks had their October-November visitation collapse by 78% to 95%, including Wright Brothers National Memorial in Kitty Hawk, Pipe Spring in Arizona, Minuteman Missile in South Dakota, and Jewel Cave. By contrast, Great Smoky Mountains, Blue Ridge Parkway, the Natchez Trace, and Golden Gate Recreation Area kept counting visitors normally through the shutdown. What's the difference? Every park on the collapsed list shares one feature: visitation is GATED BY FEDERAL STAFF rather than by a physical entrance. Pennsylvania Avenue and the White House aren't really 'parks' in the gates-and-trails sense — they are interpretive sites where the visit count is generated by ranger-led tours, and when the rangers are furloughed, no count is generated. Governors Island only counts visits when the NPS-operated ferry runs. Wright Brothers, Pipe Spring, and the others have a single visitor center as the only gate; when the staff goes home, the parking lot empties and the count goes to zero. Every park that kept counting through the shutdown has a through-road or an unstaffed entry point where vehicles can enter without anyone checking. This matters for two reasons. First, when NPS prepares budget defenses for the next fiscal year, it should categorize its units as 'staffing-gated' or 'access-gated' and request shutdown-resilience funding (automated visitor counters, loop detectors, IR sensors at gates) for the staffing-gated parks specifically — those are the units that lose all of their data during a furlough. Second, the 'national parks lost 9 million visits' narrative is somewhat misleading: most of those 9 million visits weren't lost to people deciding not to go camping, they were lost to the recording method going dark for ~20 specific parks where federal staffing is the only way visits get counted at all.

PAAV Pennsylvania Avenue NHS 0.00% vs 16.94% (-100%) · GOIS Governors Island NM 0.00% vs 23.92% (-100%) · WHHO White House and President's Park 0.00% vs 15.00% (-100%) · WRBR Wright Brothers NMem 0.79% vs 14.27% (-94.5%) · PIPE Pipe Spring NM 1.06% vs 11.42% (-90.7%) · MIMI Minuteman Missile NHS 1.00% vs 9.17% (-89.1%) · JECA Jewel Cave NM 0.77% vs 5.69% (-86.5%) · OCMU Ocmulgee Mounds NHP 2.65% vs 19.40% (-86.3%) · SAIR Saugus Iron Works NHS 2.11% vs 14.35% (-85.3%) · WACA Walnut Canyon NM 3.36% vs 20.73% (-83.8%) · FOMC Fort McHenry NM 2.60% vs 15.80% (-83.5%) · FOFR Fort Frederica NM 2.60% vs 15.60% (-83.4%) · LIHO Lincoln Home NHS 2.41% vs 14.20% (-83.0%) · WUPA Wupatki NM 2.80% vs 15.63% (-82.1%) · STEA Steamtown NHS 3.33% vs 17.50% (-81.0%) · DEPO Dwight D. Eisenhower NHS 1.25% vs 6.54% (-80.8%) · ARHO Arlington House 3.04% vs 15.00% (-79.7%) · HAGR Hagerman Fossil Beds NM 4.83% vs 23.39% (-79.3%) · FOTH Fort Stanwix NM 3.76% vs 17.41% (-78.4%) · GEGR George Rogers Clark NHP 3.61% vs 16.51% (-78.1%)

2,905,440 monthly statistic rows in Main_Data.csv · 273 qualifying parks (full 2025 monthly + ≥3 complete 2020-2024 baseline years + ≥50K 2025 visits) · 215 parks with declined Oct+Nov share / 58 with elevated share · system-wide Oct+Nov 2025 share 14.13% vs baseline 15.03% (-0.90 pp) · 3 parks at 0.00% Oct+Nov 2025 (PAAV, GOIS, WHHO) · 20 parks with >78% collapse in shutdown-window share · structural pattern: every collapsed park is staffing-gated rather than access-gated