GitLab

Hello @dyb,

This deck went a bit faster. I have managed to map the following fields:

• Ship name
• Original values and units for: AT, SLP, SST

Incomplete fields that need some feedback from you:

• Platform sub-type: This code table was not documented in the online documentation. I used the same approach as for CLIWOC and identified the different type of sailing ships, grouped them into categories and assigned a code from the CDM for those type of ships that are in the CDM. However, more codes in the CDM are needed and a deep review of the categories that assigned to each sailing ship. A spread sheet with the categories and platform subtypes is located under the ICOADS_Documentation/structure folder

• Dates:

I was looking to a subset of data for deck 721: 152-721_1857-08_subset.imma The data from the c99 attachment looks like this:

The bottom part from c99_data corresponds to Dates and times of observations. The blue highlighted part corresponds to the Ship name STARLIGHT.

If we include the dates from ICOADS.core dates seem to be wrong:

"report_timestamp": {
        "sections": "core",
        "elements": ["YR","MO","DY","HR"],
        "transform": "datetime_imma1"

But if I include the attachment timestamp the dates are the following and correct:

"report_timestamp": {
        "sections": ["c99_data", "c99_data", "c99_data", "c99_data", "c99_data"],
        "elements": ["year", "month", "day", "hour", "hour_ind"],
        "transform": "datetime_to_cdm_time"

This improves:

There is no barometer height information for this deck.

A first draft of the CDM-mapper for this deck can be found in the branch deck721. I will push this to github once we have a better idea of the CDM codes for the platform subtypes and a feedback on the dates. Do we want to keep the attachment information instead of ICOADS.

Latitude and longitude seems fine form the CORE though!