Definition of biweekly aggregates

How the biweekly aggregates for t2m and tp are calculated is not very well described and explained. You can see what I did in the renku dataset notebook. In this notebook I show how I get from climetlab data to the biweekly datasets used for scoring. However, I mixed up the lead_time definition for tp.

Working on the changes:

• climetlab: https://github.com/ecmwf-lab/climetlab-s2s-ai-challenge/pull/24
• website: https://github.com/s2s-ai-challenge/s2s-ai-challenge.github.io/pull/25
• template: https://renkulab.io/gitlab/aaron.spring/s2s-ai-challenge-template/-/merge_requests/14