[ncl-talk] linint2_point

Sat Jul 18 11:45:40 MDT 2020

Dear Ehsan,

First, apologies to NCL users for the back and forth as this seems much
more of an understanding into interpolation techniques than a typical NCL
user problem.

The statement "That is why more commonly higher resolution grids would be
interpolated to lower resolution.” was not made in isolation. It has to be
read in context.

Let me explain further: If you want to quantitatively evaluate some
datasets (as you are trying to do in your paper), one will ideally want to
regrid the different grids to a common grid.

In your case as you mention, 3 gridded datasets with 0.1, 0.5 and 0.75
resolution respectively. For a quantitative analysis of these datasets, one
will be required to interpolate to the lowest resolution grid, i.e 0.75.
However if this is not done, then one will have to explain why another
approach was preferred (as mentioned in my earlier mail, certain analysis
may be sensitive to interpolation and using the wrong technique may lead to
misleading results).

On the reviewer question, you will have to draw that conclusion from these
discussions. I will however mention that you should discuss how higher
resolution grids may have more information (detail) than one with lower
resolution. For instance higher resolution dataset will capture certain
spatial distribution of say precipitation which may not be necessarily
captured in lower resolution data.

I will entreat you to spend some time to read Dennis’s overview on
regridding here:
https://climatedataguide.ucar.edu/climate-data-tools-and-analysis/regridding-overview

Hope this helps.

Cheers
Kwesi

On 18 July 2020 at 13:11:49, Ehsan Taghizadeh via ncl-talk (
ncl-talk at mailman.ucar.edu) wrote:

Dears Ian and Kwesi,
I really appreciate for your helpful comments. I learnt much from them.
However, about Kwesi's comment, may I ask why someone wanna interpolate
higher resolution to the lower resolution, as you said "That is why more
commonly higher resolution grids would be interpolated to lower resolution".

May I ask more help?
My work included different grid data with different spatial resolution;
0.1, 0.5 and 0.75 degree. I think I should reply to the referee related to
these spatial resolution differences which included in my work, maybe the
influence of these differences in the interpolation.
The referee asked: "Please discuss also the spatial scale resolution
differences in products and interpolation implications."

*Sincerely,*
*Ehsan*

On Saturday, July 18, 2020, 12:00:41 PM GMT+4:30, Kwesi A. Quagraine <
starskykwesi at gmail.com> wrote:

Dear Ehsan,

Follow from Ian’s explanation; as an example, lower resolution grids have
less spatial variance than higher resolution grids and interpolating the
lower resolution to a higher resolution does not change this. Thus taking a
lower resolution grid to a higher resolution grid does not add information.
The interpolated data will have low spatial and temporal variance.
That is why more commonly higher resolution grids would be interpolated to
lower resolution.

This information has to be explicitly stated in your paper; because there
are some analysis that are sensitive to interpolation and readers have to
bear that in mind when comparing results.

Hope this helps.

Regards
Kwesi

On 17 July 2020 at 18:48:39, Ian Harris (ENV - Staff) via ncl-talk (
ncl-talk at mailman.ucar.edu) wrote:

Hi,

I'd say they want you to examine the resolution of your input and output
data, in case the interpolation could be judged, er, 'optimistic'. For
instance, if I had data on a 5° grid, I could easily interpolate it onto a
0.5° grid, but there are implications because you're not inventing new
data, you're still stuck with the 5° data but on a finer grid.

Cheers

Harry
------------------------------
*From:* ncl-talk <ncl-talk-bounces at mailman.ucar.edu> on behalf of Ehsan
Taghizadeh via ncl-talk <ncl-talk at mailman.ucar.edu>
*Sent:* 17 July 2020 17:12
*To:* Ncl-talk <ncl-talk at ucar.edu>
*Subject:* [ncl-talk] linint2_point

*Warning:* This email is from outside the UEA system. Do not click on links
or attachments unless you expect them from the sender and know the content
is safe.
Dear NCL group,
I hope you're doing great.
I have a general question and I hope I could have your help, as usual.
In one paper I wrote "The interpolation was done using the bilinear method
in NCL (NCAR Command Language)". Actually, I interpolated some grid
precipitation products on gauged stations using "linint2_points" from NCL.
The referee commented "With how many points? Please discuss also the
spatial scale resolution differences in products and interpolation
implications."
I wanna reply "As mentioned in the paper, we used NCL in the most
programming. NCL has a function named linint2_points which interpolates
from a rectilinear grid to an unstructured grid using bilinear
interpolation based on the four closest grid points to a particular (x, y)
coordinate pair." However, I don't know how reply to the part of discussing
about different spatial resolution. May I ask some help about that, please?

*Sincerely,*
*Ehsan Taghizadeh*
*PhD student of Meteorology,*
*Institute of Geophysics, University of Tehran, Iran*
_______________________________________________
ncl-talk mailing list
ncl-talk at mailman.ucar.edu
List instructions, subscriber options, unsubscribe:
https://mailman.ucar.edu/mailman/listinfo/ncl-talk
------------
*Try not to become a man of success but rather a man of value - Albert
Einstein*

Kwesi A. Quagraine
Department of Physics
School of Physical Sciences
College of Agriculture and Natural Sciences
University of Cape Coast
Cape Coast, Ghana

Alt. Email: kwesi at csag.uct.ac.za
Web: http://www.recycleupghana.org/
Office: +27 21 650 1563
Skype: quagraine_cwasi
_______________________________________________
ncl-talk mailing list
ncl-talk at mailman.ucar.edu
List instructions, subscriber options, unsubscribe:
https://mailman.ucar.edu/mailman/listinfo/ncl-talk

------------
*Try not to become a man of success but rather a man of value - Albert
Einstein*

Kwesi A. Quagraine
Department of Physics
School of Physical Sciences
College of Agriculture and Natural Sciences
University of Cape Coast
Cape Coast, Ghana

Alt. Email: kwesi at csag.uct.ac.za
Web: http://www.recycleupghana.org/
Office: +27 21 650 1563
Skype: quagraine_cwasi
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.ucar.edu/pipermail/ncl-talk/attachments/20200718/7eebe321/attachment-0001.html>