Wildfires and fire seasons are commonly rated largely on the simple metric of area burned (more hectares: bad). A seemingly paradoxical narrative frames large fire seasons as a symptom of a forest health problem (too much fire), while simultaneously stating that fire-dependent forests lack sufficient fire to maintain system resilience (too little fire). One key to resolving this paradox is placing contemporary fire years in the context of historical fire regimes, considering not only total fire area but also burn severity distributions. Historical regimes can also inform forest restoration efforts by illuminating the pace and scale at which fires historically maintained (i.e., 'treated') fire-resilient landscapes. Here we ask, for a broad extent of the inland Pacific Northwest (eastern Oregon and Washington, USA), a region predominated by drier forest types and recently experiencing record-breaking fire years: 1) How much annual fire area would have been needed to support historical fire regimes?; and 2) How do contemporary fire years (1985-2020) compare to historical fire amounts and severities?

To meet historical fire frequencies for each forest type, annual area burned would have averaged at least 224,000-291,000 ha per year regionally prior to the 20th century (notwithstanding interannual variability) - presumably arising from both indigenous-cultural and lightning-ignited fires. Drier forests would account for 82-88% of annual burn area. In contrast, despite the seemingly fiery contemporary era, contemporary fire years average just ∼49,100 ha·yr−1 (∼17-22% of historical), with only one year approximating historical area burned. Contemporary years average well under historical rates for virtually all severity classes across dry and moist (but not cold) forests. Annualized fire deficits relative to historical rates are especially conspicuous for low-severity area in dry forests (on average missing 127,000-161,000 ha·yr−1 regionally) and moderate-severity area in both dry (missing 34,000-44,000 ha·yr−1) and moist (missing 9000-12,000 ha·yr−1) forests. Ten-year moving averages in burn area are increasing in recent years, but remain below historical levels. Trends are similar across states and major land ownerships.

With current forest restoration efforts occurring at a fraction of historical fire rates, our findings highlight that successful restoration and maintenance will require a) increasing active treatment rates, and b) incorporating managed wildfire to attain substantially more treated area. As such, beneficial fire years may be those not with less, but rather more, area burned - with characteristic severity and patch distributions, minimal clearly-negative impacts (e.g. loss of life and property), and contribution to restoration/maintenance objectives.