Does plant ecosystem thermoregulation occur? An extratropical assessment at different spatial and temporal scales.

To what degree plant ecosystems thermoregulate their canopy temperature (T c ) is critical to assess ecosystems' metabolisms and resilience with climate change, but remains controversial, with opinions from no to moderate thermoregulation capability. With global datasets of T c , air temperature (T a ), and other environmental and biotic variables from FLUXNET and satellites, we tested the 'limited homeothermy' hypothesis (indicated by T c & T a regression slope < 1 or T c  < T a around midday) across global extratropics, including temporal and spatial dimensions. Across daily to weekly and monthly timescales, over 80% of sites/ecosystems have slopes ≥1 or T c  > T a around midday, rejecting the above hypothesis. For those sites unsupporting the hypothesis, their T c -T a difference (ΔT) exhibits considerable seasonality that shows negative, partial correlations with leaf area index, implying a certain degree of thermoregulation capability. Spatially, site-mean ΔT exhibits larger variations than the slope indicator, suggesting ΔT is a more sensitive indicator for detecting thermoregulatory differences across biomes. Furthermore, this large spatial-wide ΔT variation (0-6°C) is primarily explained by environmental variables (38%) and secondarily by biotic factors (15%). These results demonstrate diverse thermoregulation patterns across global extratropics, with most ecosystems negating the 'limited homeothermy' hypothesis, but their thermoregulation still occurs, implying that slope < 1 or T c  < T a are not necessary conditions for plant thermoregulation.