Peak season carbon exchange shifts from a sink to a source following 50+ years of herbivore exclusion in an Arctic tundra ecosystem

To date, the majority of our knowledge regarding the impacts of herbivory on arctic ecosystem function has been restricted to short‐term (<5 years) exclusion or manipulation experiments. Our understanding of long‐term responses of sustained herbivory and/or herbivore exclusion on arctic tundra ecosystem function is severely limited. Recent evidence suggests lemming population outbreaks, which have historically been common in tundra regions, have become less frequent and in some cases disappeared. Here, we evaluate how 50+ years of lemming absence have impacted ecosystem carbon and energy exchange in an Arctic tundra region near Barrow, Alaska, using some of the oldest herbivore exclosures in the Arctic. We show that sustained lemming absence dramatically altered peak growing season ecosystem function (i.e. thaw depth, water‐table, NDVI, albedo and land atmosphere CO₂ and CH₄ exchange) in all land cover types (i.e. wet, moist and dry) investigated but was most profound in wet tundra, where the long‐term absence of lemmings altered the largest carbon sink during peak growing season into a source to the atmosphere. Landscape‐level analysis suggests the long‐term absence of lemmings may increase peak growing season albedo by +6% and decrease net ecosystem exchange, CH₄ flux and the combined radiative forcing potential by −96, −79 and −104%, respectively. Identified long‐term functional responses to herbivory differed from many short‐term exclusion studies, as we found higher rates of carbon uptake where herbivores are present, associated with a shift in plant community composition (i.e. graminoid to moss), while short‐term studies typically report a reduction in carbon uptake where herbivores are present, associated with defoliation events that decrease above‐ground biomass/photosynthesis. Synthesis. This work highlights the long‐term impact of herbivory on ecosystem function, which suggests (i) the measured response of arctic tundra to herbivory may be dependent on the duration of experimental observation, and (ii) if negative trends in lemming populations persist in arctic tundra regions, the historically strong carbon sink capacity may be severely reduced.