Modern Arctic tundra ecosystems are characterized by infrequent fires, but during warm, dry conditions tundra fires can burn thousands of hectares, with important implications for ecosystem processes including vegetation change and carbon cycling. Despite this evidence, global-change scientists and land managers lack critical information concerning long-term relationships among climate, fire, and tundra vegetation. This knowledge gap limits the ability to anticipate the response of the tundra fire regime to ongoing and predicted climate warming and potential feedbacks with the Earth system. Here we present four fire- and vegetation-history records from tundra ecosystems in the Noatak National Preserve, Alaska. Quantitative analyses of fossil pollen assemblages and peaks in macroscopic charcoal from lake sediments are used to reconstruct the timing and nature of vegetation and fire-regime changes since 5000 years before present (yr BP). Comparisons to paleoclimate records facilitate interpretations in the context of regional moisture changes

Results/Conclusions

Fire was a persistent process in the study region over the late-Holocene. Since 2000 yr BP, when regional moisture gradually decreased towards present, we estimate a site-specific (i.e. within a ≈ 1-km radius) median fire-return interval (mFRI) of 120 yr (95% CI 105-150). FRIs changed through time, however, coincident with millennial-scale changes in moisture. Higher FRIs during a period of steadily increasing moisture ca. 4000-2000 yr BP (p = 0.049; mFRI 165 yr [135-240]) suggest a direct link between climate and tundra burning. From 5000-4000 yr BP, regional moisture was again characterized by increasingly dry conditions; FRIs during this period (mFRI = 150 yr [90-195]) were shorter than those during the moist period (p = 0.092) but similar to FRIs over the past 2000 yr (p = 0.333). Links between fire-regime and vegetation shifts were complex and suggest a minor or secondary role of vegetation in influencing late-Holocene tundra fire regimes. Our results (1) indicate that tundra ecosystems can burn more frequently than observed in most modern tundra, and (2) suggest that tundra burning will increase in response to ongoing and future decreases in relative moisture in the Arctic.