mkoohafkan / cder

R Interface to the CDEC Web API
https://mkoohafkan.github.io/cder/
2 stars 0 forks source link

triple check CDEC time zone #5

Open mkoohafkan opened 4 years ago

mkoohafkan commented 4 years ago

The web services page now states

Dates, times, and output will always be Pacific Standard Time.

Which is different from the documentation that led to the decision for #3. Need to verify whether the timezone should be "GMT+8" or "US/Pacific".

mkoohafkan commented 4 years ago

Testing with 2020 data at AFO and HUN showed that timestamps jumped from 01:45 to 03:00, indicating that CDEC days includes daylight savings and the note on the web service page is incorrect. Submitted a ticket to CDEC to verify.

mkoohafkan commented 3 years ago

Testing with 2016 NSL data showed that timestamps did not jump from Nov 6 01:45 to 01:00 (PST to PST), indicating that CDEC does not account for daylight savings. Or worse, perhaps it varies from station to station...

mkoohafkan commented 3 years ago

Actually, it's even worse: looking at 2020 HUN data it appears that CDEC leaps forward on 2020-03-08 for PST -- > PDT, but does not fall back for PDT --> PST on 2020-11-01.

mkoohafkan commented 2 years ago

Looking at SGN yields the same result: CDEC leaps forward on 2020-03-08 for PST --> PDT, but does not fall back for PDT --> PST on 2020-11-01.

My best guess as to what is happening: CDEC is actually in US/Pacific (i.e., PST/PDT) but cannot handle the duplicate timestamp that occurs at fall back from PDT --> PST, which results in a lost hour of data every year. Therefore it is correct to use "US/Pacific" as the timestamp when downloading CDEC data.