Climate impacts of including plant-physiological CO2 effects in regional climate models.

Many countries rely on regional climate model (RCM) simulations to quantify the projectedimpacts of climate change and design their national adaptation plans. The accuracy andreliability of RCM simulations is thus of fundamental importance and, in particular, it iscrucial to include all relevant climate processes in RCMs. One of the processes notconsidered in most RCMs is the response of plants to elevated atmospheric CO2concentrations (the plant physiological CO2 effect); a process which is however includedin most global climate models (GCMs). Here we show that due to this missingprocess, RCMs underestimate the future increase of extreme temperatures compared toGCMs. The plant physiological CO2 effect takes into account that at higher CO2 concentrationsplants increase their water use efficiency, i.e. the fraction of CO2 uptake to transpiration. As aconsequence, plants reduce the openings of their leaf stomata at elevated CO2 concentrations.This causes a reduction of plant transpiration, resulting in a decrease of latent heat flux and anincrease of sensible heat flux. Ultimately, the plant physiological CO2 effect can thusalter near-surface air temperatures. Using a state-of-the-art regional climate model(COSMO-CLM2), we perform two model simulations: one that includes the plantphysiological CO2 effect and one where plants are insensitive to the atmospheric CO2increase. The simulations cover the European domain and the time span 1971-2099. In thesimulation that includes the plant physiological CO2 effect, evapotranspiration in central andnorthern Europe at the end of the 21st century is significantly lower than in the simulationwith insensitive plants (calculated as changes with respect to 1971-2000). In contrast, theevapotranspiration difference between both simulations is small in southern Europe,suggesting that other factors, such as soil moisture limitations, prevent a strongimpact of the plant physiological CO2 effect on evapotranspiration in this region. Theresults obtained with COSMO-CLM2 are in line with the fact that RCMs fromthe EURO-CORDEX multi-model ensemble also exhibit lower evapotranspirationcompared to GCMs in central and northern Europe at the end of the 21st century (withrespect to 1971-2000). Additionally, the results from COSMO-CLM2 agree verywell with estimations of the plant physiological CO2 effect from dedicated CMIP5experiments. Including the plant physiological CO2 effects in COSMO-CLM2 also affectsnear-surface air temperatures. In central and northern Europe, the annual maximumtemperature (TXx) is about 1 K higher in the COSMO-CLM2 simulation with plantphysiological CO2 effects compared to the simulation with insensitive plants at theend of the 21st century. This result can partially explain the discrepancy betweenthe TXx evolution in GCMs and RCMs in central and northern Europe where atthe end of the 21st century GCMs exhibit a TXx increase that is about 2 K higherthan the TXx increase in RCMs. According to the COSMO-CLM2 simulations,the plant physiological CO2 effect accounts for about 50 % of the TXx differencebetween GCMs and RCMs. Neglecting the plant physiological CO2 effect can thusresult in a substantial underestimation of future temperature increases in RCMs. [ABSTRACT FROM AUTHOR]