Extensive burning of Arctic tundra landscapes in recent years has contradicted the conventional view that fire is a rare, spatially limited disturbance in tundra. These fires have been identified as harbingers of climate change, despite our limited understanding of Arctic fire frequency and climatic controls. Much of this understanding relies upon sedimentary charcoal-based fire reconstructions, yet the controls on sedimentary charcoal abundance need to be rigorously tested in tundra ecoregions. We compile published charcoal-based tundra fire reconstructions in Alaska and compare them with historical, limnological, and spatial data to: (1) evaluate the spatial and temporal characteristics of tundra fires, (2) identify the primary controls of charcoal accumulation in lake sediments, (3) assess the accuracy of peak analysis in identifying local (<1 km) fire events, and (4) constrain charcoal source area. We find that tundra fires are spatially and/or temporally dichotomous in the historical (observational) period (1940-2015 CE), a characteristic which is mimicked by charcoal accumulation rates. We show that, despite methodological differences in published studies, analyses of charcoal peaks accurately identified local (<1 km) fire events in the historical period. Further, we find that lake properties (e.g. depth and surface area) have weak influences on charcoal accumulation in the tundra and that charcoal source areas in the Alaskan tundra are much broader than in forest and grassland ecosystems. We conclude that the unique nature of tundra fire regimes allows for the reliable use of total charcoal accumulation as a fire proxy.