Terrestrial carbon sink observed from space: variation of growth rates and seasonal cycle amplitudes in response to interannual surface temperature variability

摘要

The terrestrial biosphere is currently acting as a net carbon sink on theglobal scale, exhibiting significant interannual variability in strength. Toreliably predict the future strength of the land sink and its role inatmospheric CO growth, the underlying biogeochemical processes andtheir response to a changing climate need to be well understood. Inparticular, better knowledge of the impact of key climate variables such astemperature or precipitation on the biospheric carbon reservoir is essential.It is demonstrated using nearly a decade of SCIAMACHY (SCanning ImagingAbsorption spectroMeter for Atmospheric CHartographY) nadir measurements thatyears with higher temperatures during the growing season can be robustlyassociated with larger growth rates in atmospheric CO and smallerseasonal cycle amplitudes for northern mid-latitudes. We find linearrelationships between warming and CO growth as well as seasonalcycle amplitude at the 98% significance level. This suggests that theterrestrial carbon sink is less efficient at higher temperatures during theanalysed time period. Unless the biosphere has the ability to adapt itscarbon storage under warming conditions in the longer term, such atemperature response entails the risk of potential future sink saturation viaa positive carbon-climate feedback.Quantitatively, the covariation between the annual CO growth ratesderived from SCIAMACHY data and warm season surface temperature anomalyamounts to 1.25 ± 0.32 ppm yr K for the NorthernHemisphere, where the bulk of the terrestrial carbon sink is located. Incomparison, this relationship is less pronounced in the Southern Hemisphere.The covariation of the seasonal cycle amplitudes retrieved from satellitemeasurements and temperature anomaly is −1.30 ± 0.31 ppm Kfor the north temperate zone. These estimates are consistent with those fromthe CarbonTracker data assimilated CO data product, indicating thatthe temperature dependence of the model surface fluxes is realistic.