Quantifying Stochastic Ucertainty in Detection Time of Human-Caused Climate Signals
In this study, LLNL authors and colleagues use Large Ensembles (LEs) to quantify the stochastic uncertainty in fingerprint detection time (td) arising from natural internal variability. They determine whether anthropogenic fingerprint detection time in satellite data is within the stochastic uncertainty in td estimated from model LEs. To address this question, they analyze LEs, satellite data, and a multi-model ensemble to study signal-to-noise properties of atmospheric temperature.
Climate observations comprise one sequence of natural internal variability and the response to external forcings. Large initial condition ensembles (LEs) performed with a single climate model provide many different sequences of internal variability and forced response. LEs allow analysts to quantify random uncertainty in the time required to detect forced “fingerprint” patterns. For tropospheric temperature, the consistency between fingerprint detection times in satellite data and in 2 different LEs depends primarily on the size of the simulated warming in response to greenhouse gas increases and the simulated cooling caused by anthropogenic aerosols. Consistency is closest to a model with high sensitivity and large aerosol-driven cooling. Assessing whether this result is physically reasonable will require reducing currently large aerosol forcing uncertainties.
The team found that: 1) Median detection time td in LEs performed with two climate models is ca. 1995 for stratospheric temperature and ca. 2000 for tropospheric temperature; 2) The stochastic uncertainty in td ranges from 8 to 15 years in the troposphere and from 1 to 3 years in the stratosphere; 3) The consistency between fingerprint detection times in satellite data and in two LEs depends on model climate sensitivity and aerosol forcing; 4) The tropospheric warming in the LE performed with a high-sensitivity, high aerosol forcing model is closer to observed warming; and finally 5) The differences in td between three satellite data sets can be larger than the stochastic uncertainty in td.