# Q&A “On GCMs, Weather, and Climate”

Dan Hughes has sent in a question on climate science (thanks Dan!).

The title is  “GCMs, weather, and climate”, and following is his question.

Hello Professor Pielke,

I have a candidate for Question for the Day.

GCMs do not resolve weather, either spatially or temporally.  Even if sufficient spatial and temporal resolution were used at run time, GCMs are not applied in the same manner as NWP models and codes are applied.  All NWP calculations immediately begin to derivate from reality, and after a relatively short period of time ( a very few days ) the calculated numbers exhibit almost no fidelity to the real world.  The NWP models and codes, knowing that the forecast window into future time is severely limited, inject updated measured information into the calculations.  Obviously, GCMs applied to calculations of future states of the Earth’s climate systems cannot be applied in this manner.

If the above is a correct, but rough, description of the situation, how can the variability seen in the numbers calculated by GCMs be assigned to be weather.  If the numbers are weather, then it is clear that it cannot be the correct weather.  If climate is taken to be the average of weather, how can the average of the weather calculated by GCMs be expected to have any fidelity to actual future states of the Earth’s climate systems.

Kindly let me know if I have presented an incorrect description.

Thank you for taking time to consider this candidate question.

Dan

There is a fundamental difference in how scientists who have prompted the 2007 IPCC WG1 report view climate modeling and how other climate scientists view this modeling. The IPCC perspective is that numerical weather prediction is an initial value problem while  climate prediction is a boundary value problem in which levels of atmospheric CO2 and aerosols are the primary “boundary forcing”.  With this perspective, they claim that changes in the statistics of weather (and other climate features) can be skillfully predicted.

However, our research has shown this is a seriously flawed view as climate prediction is really an initial value problem. It even more complicated than weather prediction since there are more variables that need to be initialized accurately (e.g. ocean temperatures and salinity; land ice depth and area, vegetation type, amount and distribution, etc).  Moreover, there are feedbacks between components of the climate system (e.g. see Figure in NRC, 2005), which become important on time periods of seasons, years and decades.

The need for treating climate as an initial value problem has been documented in a number of publications, including

Pielke, R.A., 1998: Climate prediction as an initial value problem. Bull. Amer. Meteor. Soc., 79, 2743-2746.

Pielke Sr., R.A., G.E. Liston, J.L. Eastman, L. Lu, and M. Coughenour, 1999: Seasonal weather prediction as an initial value problem. J. Geophys. Res., 104, 19463-19479.

Rial, J., R.A. Pielke Sr., M. Beniston, M. Claussen, J. Canadell, P. Cox, H. Held, N. de Noblet-Ducoudre, R. Prinn, J. Reynolds, and J.D. Salas, 2004: Nonlinearities, feedbacks and critical thresholds within the Earth’s climate system. Climatic Change, 65, 11-38.

Indeed, the broader climate community is starting to come around to this view, as illustrated in the concept of “seamless climate prediction“, which I have discussed on my weblog (e.g. see).

They have a difficult challenge. On time scales longer than perhaps a single season, there is no demonstrated skill in regional predictions.