I have a time series of snow and temperature and I want to analyze changes over time for a specific period of time: 15 April - 15 June. In the perpetual calendar April 15 is the day 105 and in a leap year is 106. How should I handle/correct for leap years? Should I subtract one day on leap years? Is there a bias in the data because of leap years?

thanks for you answers!

  • 3
    $\begingroup$ Figure out what physical characteristic you think matters to your data -- the equinox? -- and choose the time period that is constant w.r.t. that physical characteristic. $\endgroup$ – cphlewis Sep 8 '17 at 2:22
  • 2
    $\begingroup$ Day of year is appropriate for some data. Month by month is appropriate for others. Really depends on what you are trying to do with the data. $\endgroup$ – f.thorpe Sep 8 '17 at 2:36
  • 1
    $\begingroup$ Try both, see if it makes a difference. If yes, ask yourself if you are analyzing the right thing in the first place $\endgroup$ – Christoph Sep 8 '17 at 4:00
  • $\begingroup$ Leap year highlights the issue, but it's actually a "bias" every year because the Spring Equinox can occur anytime between the 20th and 23rd. Thus, April 15th can be between the 27th and 30th day of spring (if I did the math) right, and you'd expect the 30th day of spring to be a little bit warmer and a little less snowy than the 27th (on average). That's a much bigger issue than leap day. $\endgroup$ – user967 Sep 8 '17 at 15:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.