If you interrupted the model at a critical point to make the backup, then this may have caused the error. Once post processing has started it\'s best to not touch them.

You only get back what you saved, so if the backup was from before the new model arrived, then you\'ll lose it by restoring the backup.

The way around this, is to make a new backup, using a different save name. (And make sure that you FULLY exit from BOINC first, not just from the manager.)
Next, clear the working area by deleting everything, and restore the backup.
Run the backup and see what happens.
If it\'s no better, then just let it report to the server. (Or you can wait until the next step first.)

Once you\'ve finished with the old model, exit from BOINC again, clear the area, and restore the second backup.
Continue with the new model.

edit
And as a follow up to your other post about no trickles, you won\'t see these in the Transfers tab; they only show as the message that Mo posted.
You can look back through the messages to find when trickles were sent, and also see a list of them on the server page for each model.

Ed, your BOINC Manager messages should be like this:

28/03/2009 16:53:27|climateprediction.net|Sending scheduler request: To send trickle-up message. Requesting 0 seconds of work, reporting 0 completed tasks
28/03/2009 16:53:32|climateprediction.net|Scheduler request succeeded: got 0 new tasks

Each time a trickle(s) is sent a message should say so. The type of model you have sends a trickle roughly every 8 model months.

This is the model you are running, and all appears well with recent trickles, and well into the third phase. It looks to be running along fine.