Search

Arctic sea ice extent (SIE) is declining at an accelerating rate with a wide range of ecological consequences. However, determining sea ice effects on tundra vegetation remains a challenge. In this study, we examined the universality or lack thereof in tundra shrub growth responses to changes in SIE and summer climate across the Pan-Arctic, taking advantage of 23 tundra shrub-ring chronologies from 19 widely distributed sites (56°N to 83°N). We show a clear divergence in shrub growth responses to SIE that began in the mid-1990s, with 39% of the chronologies showing declines and 57% showing increases in radial growth (decreasers and increasers, respectively). Structural equation models revealed that declining SIE was associated with rising air temperature and precipitation for increasers and with increasingly dry conditions for decreasers. Decreasers tended to be from areas of the Arctic with lower summer precipitation and their growth decline was related to decreases in the standardized precipitation evapotranspiration index. Our findings suggest that moisture limitation, associated with declining SIE, might inhibit the positive effects of warming on shrub growth over a considerable part of the terrestrial Arctic, thereby complicating predictions of vegetation change and future tundra productivity.

Summary: Heterogeneity has been observed in the responses of Arctic shrubs to climate variability over recent decades, which may reflect landscape‐scale variability in belowground resources. At a northern fringe of tall shrub expansion (Yuribei, Yamal Peninsula, Russia), we sought to determine the mechanisms relating nitrogen (N) limitation to shrub growth over decadal time.We analysed the ratio of 15N to 14N isotopes in wood rings of 10 Salix lanata individuals (399 measurements) to reconstruct annual point‐based bioavailable N between 1980 and 2013. We applied a model‐fitting/model‐selection approach with a suite of competing ecological models to assess the most‐likely mechanisms that explain each shrub's individual time‐series.Shrub δ15N time‐series indicated declining (seven shrubs), increasing (two shrubs) and no trend (one shrub) in N availability. The most appropriate model for all shrubs included N‐dependent growth of linear rather than saturating form. Inclusion of plant–soil feedbacks better explained ring width and δ15N for eight of 10 individuals.Although N trajectories were individualistic, common mechanisms of varying strength confirmed the N‐dependency of shrub growth. The linear mechanism may reflect intense scavenging of scarce N; the importance of plant–soil feedbacks suggests that shrubs subvert the microbial bottleneck by actively controlling their environment. See also the Commentary on this article by Street & Caldararu, 233: 585–587. [ABSTRACT FROM AUTHOR]