1948 - Spin-up ------------- Our observational database started in Jan 1948. To start the analyses, Bob Kistler did the following (1) computed 10-year monthly climatologies of the sigma and surface conditions from the earlier computed reanalyses and used these as the "previous" analyses. (A necessary input for the assimilation system.) (2) ran the analyses from January 1, 1948 for 5 days with no quality control. (3) after 5 days, used these analyses to restart the analysis system with quality control turned on. This procedure has the advantage that it will produce a first guess that has a reasonable approximation to the slowly changing features such as some of the long waves. It will also start the analyses a reasonable approximation of the transient KE and a reasonable soil moisture.