Systematic land climate and evapotranspiration biases in CMIP5 simulations

Area of Science:

• Climate Science
• Earth System Science
• Atmospheric Science

Background:

• Land climate significantly influences human populations and inhabited areas.
• Accurate simulation of land surface processes is crucial for climate modeling.

Purpose of the Study:

• To evaluate the realism of simulated evapotranspiration (ET), precipitation, and temperature in the CMIP5 multimodel ensemble over continental regions.
• To identify and quantify systematic biases in these key climate variables.

Main Methods:

• Utilized a newly compiled synthesis dataset from the LandFlux-EVAL project for evapotranspiration (ET) evaluation.
• Compared CMIP5 model outputs against reference datasets for ET, precipitation, and temperature.
• Analyzed biases across continental areas and assessed their seasonal dependence.

Main Results:

• CMIP5 simulations exhibit systematic ET overestimation across most continental regions, averaging 0.17 mm/d globally.
• This ET overestimation is more pronounced in CMIP5 compared to the previous CMIP3 ensemble.
• Precipitation is also overestimated, and these ET biases are linked to systematic temperature biases, with seasonal variations observed.

Conclusions:

• The study highlights significant biases in simulated land climate variables within the CMIP5 ensemble.
• Evapotranspiration and precipitation overestimations likely contribute to temperature biases in climate models.
• These findings underscore the need for improving land surface modeling in climate projections.