Observing System Simulation Experim

Observing System Simulation Experiments—CLARREO-like Datasets Show Ability to Discriminate between Climate Models Since the last SDT meeting, significant progress has
been made on faster Principal Component-based Radiative Transfer Models (PCRTM), which enable the development of CLARREO Observation System Simulation Experiments (OSSEs) using a diverse set of climate models found in the Coupled Model Intercomparison Project 5 (CMIP5) archive. Work
conducted by UC Berkeley and LaRC show PCRTM
achieved a 25-to-30-fold processing speed increase for
longwave (LW) and shortwave (SW) radiation cal-
culations over MODerate-resolution atmospheric
TRANsmission (MODTRAN) processing. This foun-
dational work has enabled a larger number of simula-
tions, permitting more-comprehensive examination
of the signatures of climate change and—most impor-
tantly—differentiation between climate models.
Recent results from UC Berkeley’s work, comparing
climate models with CLARREO-like datasets to deter-
mine how long of an observational record is needed
to detect changes in the climate system, suggest that a
10-year record of outgoing longwave radiation (OLR)
can differentiate between high- and low-sensitivity
climate models—see Figure 2. Additional work is
planned to determine how long of an observational
record is required to distinguish between low- and mid-
sensitivity models.
This topic generated a great deal of discussion on how
OSSEs can be used to inform the next NASA Earth
Science Decadal Survey (planned for 2017), i.e., focusing
on the advances in computational power and the ability
of models to get to finer and finer spatial resolution.