For the purpose of describing the prospects for including in-situ data in obs4MIPs, we distinguish between three broad classes of in-situ data:
With its first community-wide data call, obs4MIPs has made an important step of moving beyond satellite-only data products to include gridded in-situ data. A number of gridded in-situ products have been proposed and are in the process of being included in obs4MIPs. Examples include SOCAT and a several products from the Met Office Hadley Centre (e.g., HadCRUT4).
Within CMIP/CFMIP, some high frequency site-specific model output is archived. Currently, discussions are underway to adapt the obs4MIPs data specifications to this precedent, and thus enable site-specific in-situ to be included in obs4MIPs. There are important technical issues to consider, e.g., whether or not data from different sites should be stored in the same file. A prototype was developed early on in obs4MIPs with ARM sites using the original obs4MIPs data specifications, but the datasets were not consistent with the project focus at the time. Anyone interested in contributing to this discussion of how best to include site-specific data in obs4MIPs should contact obs4MIPsemail@example.com.
Lagrangian data (e.g. ocean drifters) may ultimately be an important addition to obs4MIPs, but currently resources are not available to address the technical aspects of how best to fit such data within the CMIP data structures.
Another example of a class of data (obviously not in-situ, but having similar technical considerations) to be incorporated into obs4MIPs will be satellite “swath” data. Again, resource limitations have pushed this to the back burner.