The other day, Gavin Schmidt stated that the amplification factor over land for tropospheric trends to surface trends for GISS models was only 0.95. A reader reported that an amplification factor of 1.1 had been reported in an article by Pielke Sr et al, relying on a pers comm. Some readers expressed frustration over the lack of documentation of a seemingly important number.
In our analysis of Santer et al 2008, Chad Herman did an enormous amount of work benchmarking an algorithm to extract synthetic satellite information and then extracting results for A1B models.
This collation has wider application than our article e.g. for verifying Schmidt’s assertion about amplification factor. While the underlying data is at PCMDI, huge data sets have to be downloaded and processed to extract the relatively small data sets that are of interest for downstream analysis.
I’ve made the data available as R-files at http://www.climateaudit.info/data/chadh/GLOBAL/sat/ and http://www.climateaudit.info/data/chadh/TROPICS/sat/. The objects are in files land.tab, ocean.tab, loti.tab but the name of the object in each case is “runs”. Each matrix is 3000x37x3. By month from (1850,1) to (2099,12). 57 A1B runs summarized in http://www.climateaudit.info/data/chadh/info.chad.tab. R-objects can be downloaded easily using download.file(…..,destfile, mode=”wb”); load(destfile). Each object is about 3 MB in size.
I’ve also made Chad’s collation to 17 tropospheric levels available in corresponding directories named “TA” instead of sat. The structure is the same but each matrix is 3000x57x17 and is about 17 MB in size.
I’ll show how to use these objects in a companion post examining Gavin Schmidt’s assertion about amplifcation factors.
B model runs showing synthetic satellite runs are fundamental to a variety of analyses.