Diagnosing the average spatio-temporal impact of convective systems – Part 1: A methodology for evaluating climate models 1Department of Earth and Space Sciences, Chalmers University of Technology, Gothenburg, Sweden
12 Dec 2013
2Department of Computer Science, Electrical and Space Engineering, Division of Space Technology, Luleå University of Technology, Kiruna, Sweden
3Program for Climate Model Diagnosis and Intercomparison, Lawrence Livermore National Laboratory, Livermore, California, USA
4Swedish Meteorological and Hydrological Institute, Norrköping, Sweden
5European Centre for Medium-Range Weather Forecasts, Reading, UK
Received: 21 March 2013 – Published in Atmos. Chem. Phys. Discuss.: 23 May 2013 Abstract. An earlier method to determine the mean response of upper-tropospheric
water to localised deep convective systems (DC systems) is improved and
applied to the EC-Earth climate model. Following Zelinka and
Hartmann (2009), several fields related to moist processes and
radiation from various satellites are composited with respect to the
local maxima in rain rate to determine their spatio-temporal evolution with deep
convection in the central Pacific Ocean. Major improvements to the
earlier study are the isolation of DC systems in time so as to prevent
multiple sampling of the same event, and a revised definition of the
mean background state that allows for better characterisation of the
Revised: 04 October 2013 – Accepted: 04 November 2013 – Published: 12 December 2013
The observed DC systems in this study propagate westward at
~4 m s−1. Both the upper-tropospheric
relative humidity and the outgoing longwave radiation are substantially
perturbed over a broad horizontal extent and for periods >30 h.
The cloud fraction anomaly is fairly constant with height but small maximum can
be seen around 200 hPa. The cloud ice water content anomaly is mostly
confined to pressures greater than 150 hPa and reaches its
maximum around 450 hPa, a few hours after the peak convection. Consistent
with the large increase in upper-tropospheric cloud ice water content,
albedo increases dramatically and persists about 30 h after peak convection.
Applying the compositing technique to EC-Earth allows an assessment of the
model representation of DC systems. The model captures the large-scale
responses, most notably for outgoing longwave radiation, but there are a
number of important differences. DC systems appear to propagate eastward in the
model, suggesting a strong link to Kelvin waves instead of equatorial Rossby waves.
The diurnal cycle in the model is more pronounced and appears to trigger new convection
further to the west each time. Finally, the modelled ice water content anomaly
peaks at pressures greater than 500 hPa and in the upper troposphere between
250 hPa and 500 hPa, there is less ice than the observations and it does
not persist as long after peak convection. The modelled upper-tropospheric cloud
fraction anomaly, however, is of a comparable magnitude and exhibits a similar
longevity as the observations.
Citation: Johnston, M. S., Eliasson, S., Eriksson, P., Forbes, R. M., Wyser, K., and Zelinka, M. D.: Diagnosing the average spatio-temporal impact of convective systems – Part 1: A methodology for evaluating climate models, Atmos. Chem. Phys., 13, 12043-12058, doi:10.5194/acp-13-12043-2013, 2013.