[Velvet]

Originally Posted by tuccillo

Numerical Weather Prediction models include subroutines to compute the heating (or cooling) rates from longwave and shortwave radiation. The longwave codes account for the effects of CO2 and the CO2 level in an input (along with specific humidity, temperature, modeled clouds, etc.). I don't recall what value I used when I was doing model development at NCEP. CO2 is an important part of the radiative transfer calculations that are performed in these codes. A previous post that suggesting some sinister element to the inclusion of CO2 in the radiative transfer calculations is comical, at best. The ECMWF IFS model (in it's many versions) has had a small performance advantage over the NCEP GFS code (in it's many versions) for the last 40 years as measured by the 500mb anomaly correlation coefficient. Both models have improved over the years at about the same rate. However, it is a little bit like comparing apples to oranges. The IFS has a later data dump and I believe they dedicate more computing resources to the simulation than NCEP because NCEP has many more modeling requirements.

Just when you’re needed most, you’re back!