Are these description criteria intended for L2 to L4? Some things may be harder or less meaningful to do at certain levels.
Page 9, Section 4.2.4. Temporal resolution. What does "95% of ocean" mean in this description of daily coverage? Is this relative to the sea mask used averaged over the entire year or period of record? Is it the satellite footprint?
Page 9. 4.2.5 Timeliness of new data. Perhaps this should contain more flexibility. For example, our "preliminary" 1-day latency is deleted after 15 days and the "final" version is produced and put in LTSRF. However, the reprocessed time series that uses harmonized higher quality inputs extends only up to 2006. So the product is most likely lower quality after 2006.
Pag 9. 4.2.7 Valid data fraction. In the example: 14% of total what? L3 or 4 ocean gridpoints on x-km grid? L2 pixels in original satellite projection? On an annual average basis?
Page 12. Section 18.104.22.168 Will the TAG identify or provide the buoy dataset to use? There are different ideas on what high quality obs are. Any chance for a lower quality but "consistent" ship dataset so we can test for temporal consistency? Also essential to have some validation data for the period with volcanic eruptions of Pinatubo and El Chichon.
Page 13. Sec22.214.171.124., Will the TAG provide a "blessed" Argo dataset in one location?