Fidelity of CMIP5 simulations in representing the dynamics of the large scale meteorology associated with California hot spells

Abstract:

Extreme hot spell events over the California Central Valley (CCV) are studied using their associated Large Scale Meteorological Patterns (LSMPs). The LSMPs include a ridge trough ridge spanning the north Pacific (Grotjahn and Faure, 2008; Grotjahn 2011, 2013, 2014). Climate models easily resolve the LSMPs even if they cannot resolve California’s complex topography and regional circulations. So, global models representation of CCV hot spells includes simulation of LSMP dynamics.

We compared hot spell dynamics in NCAR’s CCSM4 data with NCEP-NCAR Reanalysis (NNR) data. Normalized daily maximum temperature anomaly (using grid point’s long term daily mean, LTDM, and LTDM standard deviation) find 24 hot spell events in 1970-1999 CCSM4 simulations, similar to 22 events from NNR during 1981-2010.

CCSM4 biases are as follows. The lower troposphere temperature anomaly LSMP is similar in NNR and CCSM4, though further inland in CCSM4 data. But, the strong anomalous sinking occurs further inland in NNR than in CCSM4 data, though CCSM4 has a stronger maximum sinking offshore (where NNR also is maximum). NNR and CCSM4 have offshore low level flow that is much stronger in CCSM4.  The mid to upper level ridge trough ridge LSMP spanning the north Pacific is much weaker in CCSM4 data, the much weaker trough upstream is a crucial bias. The wave activity flux (WAF) convergence upstream of the CCV before onset is thus much weaker in CCSM4, so CCSM4 relies more on local heat generation (consistent with stronger sinking and horizontal flow locally) to build the ridge along the west coast. Air parcel trajectories sink strongly just prior to onset, but prior to sinking, 2/3 of the NNR events have paths coming from far to the west, the remainder hover near the CCV before sinking. CCSM4 trajectories do not come from the west, most hover near the CCV or come from the south before sinking. CCSM4 is missing the dynamics responsible for a majority of the heat wave events observed.

Such thermodynamic and dynamic analyses will be applied to other CMIP5 (Taylor et al., 2012) climate simulations based on the availability of the data needed for the analyses.  CCV hot spells in future simulations will also be included, time permitting.