3 ms·

A good example of this is revisions EDIT: of the interpretations :ENDEDIT of historical data: https://www.science.org/content/article/world-1-3%C2%B0c-or-1-5%C

A good example of this is revisions EDIT: of the interpretations :ENDEDIT of historical data: https://www.science.org/content/article/world-1-3%C2%B0c-or-1-5%C2%B0c-warmer-historical-ship-logs-hold-answers https://www.science.org/content/article/world-1-3%C2%B0c-or-... Apparently the Japanese continued using wooden buckets longer than other navies, resulting in Pacific data being skewed.

defrost 1y ago
Closer reading should tell you that no one, no data scientist at least, is rewriting raw historical data - the temperatures recorded in ships logs by various ships are preserved unaltered. What is discussed there is part of all data interpretation, the transfer functions between raw measures of <something> and the inferred values of <interest>.
- WillAdams 1y ago
  Thank you for that, I have edited my (hasty/poorly-worded) comment.
  - defrost 1y ago
    No drama, as first worded it irked my experience in multichannel geophysical processing; every data source has potential biases, tweaks, scaling, shifts and normalizing applied to get them to best fit and reveal a larger picture. eg: https://www.ga.gov.au/bigobj/GA13928.pdf https://www.ga.gov.au/bigobj/GA13928.pdf as an example of stitching 40 years of survey data collected with a range of different instruments and procedures. In the climate domain the same care is taken to preserve raw data from historic methods, the challenge is syncing data streams that have covariant features (such as matching years against tree patterns from differing logs from a variety of species), in this case logs from a variety of ways of sampling sea water temperature. Glad you took my comment well, no offense was intended :)
    - jasonjayr 1y ago
      This is an excellent example too -- all too often science reporting will just put in the headline "Updated historical temperatures make threat of global warming (less|more) serious!" -- and folks latch on to "how can history change?!?!" and immediatly go to conspiracy theories, rather than trying to understand the subtlty of what's actually going on

defrost 1y ago

Closer reading should tell you that no one, no data scientist at least, is rewriting raw historical data - the temperatures recorded in ships logs by various ships are preserved unaltered. What is discussed there is part of all data interpretation, the transfer functions between raw measures of <something> and the inferred values of <interest>.

WillAdams 1y ago
Thank you for that, I have edited my (hasty/poorly-worded) comment.
- defrost 1y ago
  No drama, as first worded it irked my experience in multichannel geophysical processing; every data source has potential biases, tweaks, scaling, shifts and normalizing applied to get them to best fit and reveal a larger picture. eg: https://www.ga.gov.au/bigobj/GA13928.pdf https://www.ga.gov.au/bigobj/GA13928.pdf as an example of stitching 40 years of survey data collected with a range of different instruments and procedures. In the climate domain the same care is taken to preserve raw data from historic methods, the challenge is syncing data streams that have covariant features (such as matching years against tree patterns from differing logs from a variety of species), in this case logs from a variety of ways of sampling sea water temperature. Glad you took my comment well, no offense was intended :)
  - jasonjayr 1y ago
    This is an excellent example too -- all too often science reporting will just put in the headline "Updated historical temperatures make threat of global warming (less|more) serious!" -- and folks latch on to "how can history change?!?!" and immediatly go to conspiracy theories, rather than trying to understand the subtlty of what's actually going on