ECMWF system-2 seasonal forecast fields
Some fields of the ECMWF seasonal forecast system-2 have been included to enable exploratory verification studies. For real work you'll need the raw data, which I am not allowed to redistribute. A lot of verification studies have also been done at the ECMWF. Some results are available at their member state server (restricted access).
At the moment I have the ensemble means of the following fields
The fields are not homogeneous. They consist of ensembles started at the first of the month with perturbed ocean analyses.
Each forecast month is treated as a separate "reality". This has non-obvious consequences. If you study the winter (DJF) temperature at lead time "+2" months, you will average the December data from the runs started on September 1, the January data from runs started around October 1, and the February data from runs started on November 1. If you want the equivalent of the ECMWF +2 forecast you have to select the "Oct" forecast
Some example analyses with step-by-step instructions are here.
Please contact me if you have any questions or would like to have other variables included.