SearcharxivSearch

arXiv subjects

Sybren Drijfhout

Publications and source records attributed to Sybren Drijfhout.

2 recordsLinked to original sources

State transitions in land-vegetation systems emerge at Paris Agreement warming levels in CMIP6

Using an automatic detection workflow applied to the Coupled Model Intercomparison Project Phase 6 (CMIP6) ensemble under future emission scenarios, we identify 47 abrupt and more gradually developing state transitions in the land-vegetation component of the Earth system, classified into 9 categories. Over the Amazon, we find state transitions via vegetation dieback alongside greening cases; the contrast between them is traced primarily to differences in precipitation: models showing dieback experience either a larger absolute decline in precipitation, or one that translates more efficiently into soil moisture loss, particularly near the surface, while in greening models the CO$_2$ fertilization effect wins out where the soil moisture loss remains weaker. The precipitation decline in dieback-prone models appears driven through a weakening of moist convection. Across these Amazon cases, models with dynamic vegetation undergo dieback, whereas greening is confined to models with prescribed vegetation distributions. African cases include greening over eastern-central Africa and the Congo basin, and an abrupt soil-moisture drying also over the Congo. At high latitudes, boreal forest expands, while permafrost thaws once the regional above-zero temperatures persist for more than half the year. Additional categories cover transitions to a reduced snow-cover state over northeastern North America, increased vegetation biomass near the Tibetan Plateau and southeastern Asia, and increased leaf-area index over the northeast Northern America. Of particular concern, a global warming of 2$^\circ$C or below, within reach of the Paris Agreement targets, is already enough to trigger the onset of the majority of the identified categories in CMIP6.

physics.ao-ph

Western Europe is warming much faster than expected

The warming trend of the last decades is now so strong that it is discernible in local temperature observations. This opens the possibility to compare the trend to the warming predicted by comprehensive climate models (GCMs), which up to now could not be verified directly to observations on a local scale, because the signal-to-noise ratio was too low. The observed temperature trend in western Europe over the last decades appears much stronger than simulated by state-of-the-art GCMs. The difference is very unlikely due to random fluctuations, either in fast weather processes or in decadal climate fluctuations. In winter and spring, changes in atmospheric circulation are important; in spring and summer changes in soil moisture and cloud cover. A misrepresentation of the North Atlantic Current affects trends along the coast. Many of these processes continue to affect trends in projections for the 21st century. This implies that climate predictions for western Europe probably underestimate the effects of anthropogenic climate change.

physics.ao-ph