PROJECT REPORT
Title: Comparison of climate models used in the IPCC Assessment Reports Four (AR4) and Five
(AR5) for the Southwest U.S.
Principal Investigator: Dan Cayan, Scripps Institution of Oceanography/UC San Diego and USGS
Collaborators and Partners: Suraj Polade, Scripps Institution of Oceanography; Mike Dettinger,
USGS and Scripps Institution of Oceanography; Alexander Gershunov, Scripps Institution of
Oceanography; David Pierce, Scripps Institution of Oceanography, Mary Tyree, Scripps
Institution of Oceanography
Project Summary: Evaluation of climate model simulations is an important step in the
investigation of trade-offs among potential resource management options. The models can
help inform decisions by providing information about projected regional changes in climate,
their potential effects on ecological and social systems, and their drivers.
In this project Cayan et al. sought to compare AR4 and AR5 models to determine to what extent
AR5 models represent an improvement in our understanding of climate change in the
Southwest United States. They focused, in particular, on the performance of the climate
models in simulating natural modes of variability and their teleconnections.
Their initial results indicate that the AR4 and AR5 models projections are generally similar
across the Southwest. The patterns of drying and warming seen in AR4 simulations are also
apparent in AR5 simulations. AR5 models appear to be more effective than AR4 is in simulating
key short-term Pacific climate modes and their teleconnections to North America, which should
provide a better representation of short-term climate variability over the Southwest.
The project team is continuing to evaluate the performance of the historical simulations in
capturing observed winter Pacific sea-surface temperatures (SSTs) and precipitation over the
Southwest and contiguous United States. However, their preliminary findings indicate that most
of the models reproduce the basic features of the natural climate variability patterns, including
their teleconnections. Importantly, the linkage between winter Pacific SST anomaly patterns and
North America (including the Southwest) precipitation anomalies were reasonably similar in
AR5 models and in observations, with similar structures shown by the spatial patterns of the
linked SST and precipitation. This linkage appears to be better captured in AR5 than in AR4. This
is a useful advance because much of the precipitation in the Southwest occurs in winter and the
precipitation variability is dominated by interannual-decadal, not long period changes. In turn,
this should help to inform many ecological processes as well as numerous resource
management issues, because precipitation variability is a key driver.