Lau, W. K.-M, Y. Sud, and J.-H Kim, 1995a: Intercomparison
of hydrologic processes in global climate models. NASA Technical Memorandum
104517, NASA Goddard Space Flight Center, Greenbelt, MD, 161 pp.
Preliminary results of the Atmospheric Model Intercomparison Project
(AMIP) diagnostic subproject on "Intercomparison of Hydrologic Processes
in Global Models" are presented. The objective of the subproject is to
evaluate the ability of atmospheric general circulation models (GCM) in
simulating the global hydrologic cycle and to explore means of validating
GCM precipitation and hydrologic processes with space-based and ground-based
observations. Based on this evaluation, we hope to identify generic and/or
specific strengths and weaknesses of the participating climate models which
will help to formulate a strategy for model improvement. In this report,
we address the intercomparison of precipitation (P), evaporation (E), and
surface hydrologic forcing (P-E) for 23 AMIP GCMs including relevant observations,
over a variety of spatial and temporal scales. The intercomparison includes
global and hemispheric means, latitudinal profiles, selected areal means
for the tropics and extratropics, ocean and land, respectively. In addition,
we have computed anomaly pattern correlations among models and observations
for different seasons, harmonic analysis for annual and semiannual cycles,
and rain-rate frequency distribution. We also compare the joint influence
of temperature and precipitation on local climate using the Koeppen climate
classification scheme. Results show that the models collectively portray
an Earthlike climate with respect to the observed land-only global mean
surface temperature (=14.8°C ) and precipitation (=2.4 mm/day ) to
within 10%. The model consensus indicates a cold-wet bias of about 1.5
degrees C and 0.5 mm/day. Most of the models conserve atmospheric water
up to about 5%. While most models show a rain-rate intensity distribution
similar to that of the observed, almost all models underestimate the frequency
of occurrence of light rain (<1 mm/day), suggesting some fundamental
problems in the treatment of nonconvective rainfall in models. The major
areas of deficiencies in global rainfall distribution are in the eastern
Pacific Intertropical Convergence Zone (ITCZ), the South Pacific Convergence
Zone (SPCZ), and the Asian monsoon. The main discrepancy in the ITCZ and
the SPCZ is in the rainfall amount. In the Asian monsoon region, the problem
seems to be more severe. None of the models are able to reproduce the East
Asian monsoon rainbelt. Perhaps, not by happenstance, the disparities among
observed estimates are also large in the above regions. The model depiction
of the hydrologic cycle tends to be more realistic where there is a strong
annual cycle and where local moisture balance is the key operating mechanism,
i.e., over large interior land mass in the extratropics. In regions of
strong dynamical control (P-E >>0), i.e., over the tropical western Pacific,
the monsoon region and the ITCZ, the differences among models tend to be
large. In the interannual time scales, all models show enhanced rainfall
prediction over the tropics because of sea surface temperature (SST) forcing,
i.e., during El Nino Southern Oscillation (ENSO). However, the models do
not show any useful skill for rainfall prediction in the extratropics from
tropical SST forcing.