Premium
Surface air relative humidities spuriously exceeding 100% in CMIP5 model output and their impact on future projections
Author(s) -
Ruosteenoja Kimmo,
Jylhä Kirsti,
Räisänen Jouni,
Mäkelä Antti
Publication year - 2017
Publication title -
journal of geophysical research: atmospheres
Language(s) - English
Resource type - Journals
eISSN - 2169-8996
pISSN - 2169-897X
DOI - 10.1002/2017jd026909
Subject(s) - coupled model intercomparison project , supersaturation , relative humidity , environmental science , climatology , spurious relationship , climate change , truncation (statistics) , humidity , atmospheric sciences , surface air temperature , climate model , meteorology , thermodynamics , mathematics , geology , physics , statistics , oceanography
In 17 out of the 29 Phase 5 of Coupled Model Intercomparison Project (CMIP5) climate models examined in this work, near‐surface air relative humidity (RH) frequently exceeded 100% with respect to ice in polar areas in winter. The degree of supersaturation varied considerably across the models, and the same evidently applies to the causes of the phenomenon. Consultations with the modeling groups revealed three categories of explanations for supersaturation occurrence: specification of RH with respect to ice rather than liquid water; inconsistencies in the determination of specific humidity and air temperature for the near‐surface level; and the nonlinearity of saturated specific humidity as a function of temperature. Modeled global warming tended to reduce the artificial supersaturations, inducing a spurious negative trend in the future RH change. For example, over East Antarctica under Representative Concentration Pathway 8.5, the multimodel mean RH would decrease by about 10% by the end of the ongoing century. Truncation of overly high RHs to a maximum value of 100% cut the RH response close to zero. In Siberia and northern North America, truncation even reversed the sign of the response. The institutes responsible for the CMIP6 model experiments should be aware of the supersaturation issue, and the algorithms used to produce near‐surface RH should be developed to eliminate the problem before publishing the RH output data.