Seasonal Case Studies Reveal Significant Variance in Large-Scale Forcing Data
Xie, S., Lawrence Livermore National Laboratory
General Circulation and Single Column Models/Parameterizations
Cloud Modeling
Xie, S, R.T Cederwall, M. Zhang, and J.J. Yio, Comparison of SCM and CSRM forcing data derived from the ECMWF model and from objective analysis at the ARM SGP site, J. Geophys. Res., 108(D16), 4499, doi:10.1029/2003JD003541, 2003.
Observed (left) and ECMWF-derived (right) forcing fields of time-height distributions of the derived (top) vertical velocity, (middle) total advective tendency and temperature, and (bottom) total advective tendency of moisture during the selected strong precipitation period during summer 1997 at ARM's Southern Great Plains Site in Oklahoma.
One of the major objectives of the Department of Energy's Atmospheric Radiation Measurement (ARM) Program is to improve the cloud component of models used for predicting weather and modeling the Earth's climate. One method to test the cloud component of a climate model is to extract the cloud physics in a single column of the model and compare what it predicts when forced by the real atmosphere to actual cloud observations.
The results of the approximation to real clouds, referred to as a single column model (SCM), can also be compared with the results of a high-resolution, 3D model run over a limited area, known as a cloud system resolving model (CSRM), which must also be forced at its lateral boundaries with atmospheric data, such as the advection of temperature and moisture, and vertical velocity produced by the horizontal wind convergence. These data can be produced either from a sophisticated blend of high resolution data or from the initial conditions of weather forecasting models. The latter is less expensive because no special data are required and are always available. The question is whether these weather model analyses are sufficiently accurate depictions of the real atmosphere.
As described in the Journal of Geophysical Research (August 2003) ARM researchers conducted a series of case studies to assess large-scale forcing data derived from the European Center for Medium Range Weather Forecast (ECMWF) model. Since 1995, the ECMWF has provided continuous forcing sets for ARM''s three instrumented observation sites across the globe to support cloud research with the Single Column Model (SCM) and Cloud System Resolving Model (CSRM).
Researchers used observational data collected at ARM's Southern Great Plains (SGP) site in Oklahoma during the summer of 1997 and the spring and fall of 2000. These seasonal data allowed them to assess the ECMWF-derived forcings under different weather conditions. The SGP seasonal data were then processed with a proven objective analysis scheme, called the "constrained variation analysis approach" (developed by Zhang and Lin, 1997) to produce a large-scale data set for comparison against the ECMWF. In addition, the researchers conducted sensitivity testing to assess the uncertainty in the variational analysis forcings derived from the summer 1997 SGP data.
For the high precipitation conditions in the summer case, the forcing data derived from the ECMWF model differed substantially from those derived from the objective variational analysis. Though much better overall agreement resulted for the spring case under moderate precipitation conditions, weak correlations for precipitation-related terms (such as latent heating and moisture convergence) remained. Disagreement between the ECMWF model data and objective analysis was further reduced in the fall nonprecipitation case, however with notable discrepancies related to heat and moisture convergence. Based on these results, the use of ECMWF forcing data should be avoided during periods of strong convection (precipitation). This recommendation is further strengthened by results of the sensitivity testing, which showed less uncertainty in the variational analysis forcings than the error calculated for ECMWF-derived forcings.
Information gained from case studies such as this are reducing the uncertainty associated with cloud models and increasing the accuracy of cloud models used in climate research. Results from ARM research continues to produce valuable information that will improve our ability to simulate climate and predict the weather and help decision makers consider climate-related issues relative to economics, society, and policy.