At http://www.iers.org/SharedDocs/Publikationen/EN/IERS/Publications/ar/ar2012/ar2012_352.pdf?__blob=publicationFile&v=1 page 80 it is stated: "The LOD for the combination are derived directly from the UT1–UTC data. The analytical expression for the first derivative of a cubic spline passing through the UT1–UTC data is used to estimate the LOD at the epoch of the UT1–UTC data."

At http://www.physics.arizona.edu/~restrepo/475A/Notes/sourcea-/node35.html#fg.spline bottom, left figure, in the part of domain -1 to -0.7 and near -0.5 tangent lines to original curve and to the curve made by cubic spline, make great angles. It means great difference in trigonometric tangents that are first derivatives.

In general for some parts of the domain of any function, first derivatives of the function obtained by cubic spline from the original, are severely deferent than first derivatives corresponding to original function.

Cubic spline although is good for find values of dependent variable, is bad for find first derivatives.

My question: Is IERS/Rapid Service/Prediction Centre wrong?

  • There are various ways to do a cubic spline. It is not necessarily the case that the first derivative is very different from the underlying function at a given point of interest. But the document does not seem to explain the algorithm very well (or at least, I could not tell exactly what it was they do). – David K Oct 29 '15 at 15:59

Your Answer

 

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.