Hysteresis and reversibility of agroecological droughts in response to carbon dioxide removal

Agroecological droughts are expected to increase with climate change, becoming one of the greatest threats to ecosystems and human society. To mitigate climate change and the growing risk of agroecological droughts, carbon dioxide removal (CDR) is increasingly recognized as unavoidable. However, it remains unclear whether the increase of agroecological drought due to atmospheric CO2 emissions will be symmetrically reversed by an equivalent atmospheric CDR. Here we investigate this question by utilizing an idealized atmospheric CO2 emission and removal experiment from the CDR Model Intercomparison Project, involving eight Earth system models, and develop a new methodology to quantify climate hysteresis and reversibility. We find that drought increases in hotspot regions cannot be symmetrically reversed by an equivalent CDR: drought severity under the CDR pathway is 65% ± 30% greater than under the emission pathway; drought frequency increases are only partially reversed by 73% ± 18% when CO2 emissions are balanced by equivalent CDR. Drought hysteresis and irreversibility are most pronounced in the Mediterranean, northern Central America, west and east southern Africa and southern Australia. Our findings imply irreversible drought impacts associated with CDR, highlighting the need for planning long-term drought adaptations. Using an idealized multi-model experiment and a new hysteresis quantification method, this study shows that equivalent carbon dioxide removal fails to symmetrically reverse CO2-emissions-induced agroecological droughts, revealing irreversible impacts in hotspots in the Mediterranean, northern Central America, southern Africa and southern Australia, necessitating urgent adaptation planning.