On Sunday April 25 2010, Kevin sent me the e-mail below which continues the constructive discussion we are having on Climate Science. It is reproduced below with his permission.
I am just back from travel: actually I was at a mtg in Atlanta with your son!
I saw Roy Spencer’s comment for the first time and it is not correct. The CERES data are indeed processed to give the reflected solar radiation and outgoing longwave radiation, which combine to give the net radiation. The biggest single change, which occurred abruptly, is a drop in OLR at the beginning of January 2008 and lasting throughout most of 2009 with only a brief return to “normal”. Values are between 0.5 and 1 W m-2 below the normal which is the mean for the record from 2000 through about 2005 (I think) and thus not normal in the sense of being radiatively balanced (the zero is not a true zero). On the other hand, the reflected solar is higher from 2000 through 2003, and runs up to about 0.5 W m-2 below the mean thereafter.
The abrupt nature of the drop in OLR made it look very suspicious to me and we looked at the daily values. The drop seems to occur very near the start of 2008. We have raised this with the CERES team who find that it is replicated in a separate measurement from AIRS data. It occurs during La Nina and is thus consistent with changes in high cloud (more cloud), which has some plausibility. We explored the cloud records from three sources but all disagree and the quality of cloud information is not yet good enough for this sort of thing. Fortunately cloud data (ISCCP) are being reprocessed. However, definition of cloud and changes in sensitivity of the instruments are difficult challenges. The more sensitive the instrument, the more cloud is found, even though it is very thin.
We hope that our paper is a stimulus to help improve the records of all aspects of this problem, the radiation , the ocean heat, and the way they are processed.
I am inviting Roy to respond to Kevin’s comment.