Re: [ferret_users] Seasonal_reg regriding questions

[Steve Hankin <steven.c.hankin@xxxxxxxx>]

On 12/19/2012 1:57 PM, Oots, Penny C. (LARC-E301)[SCIENCE SYSTEMS AND APPLICATIONS, INC] wrote:

Ansley,

Is there a method for modifying the seasonal_reg axis to change it to
1-Dec to 1-Mar
1-Mar to 1-Jun
1-Jun to 1-Sep
1-Sep to 1-Dec

Ie shift the seasons to be DecJanFeb, MarAprMay, JunJulAug, Sep,Oct,Nov?

It is a pretty straightforward matter to define any arbitrary list of time cells that you like on a climatological axis.  Have a look at this section of the Users' Guide
    ../../documentation/users-guide/commands-reference/DEFINE.html#_VPINDEXENTRY_1310

The second example down shows a (somewhat complicated) example of a climatological axis that divides the climatological year into two custom time chunks.  (look for "The following example shows the computation of a custom climatological average.")  The same procedure could be extended in a straightforward way to divide the year into the 4 chunks that you have defined.

    - Steve

I believe this is what was being asked by another user also (see
../fu_2002/msg00378.html)
.

Penny Oots

On 12/5/12 2:14 PM, "Ansley Manke" <ansley.b.manke@xxxxxxxx> wrote:

Hi Penny,
Coordinate values are at the middle of the grid cell, so

15-Feb runs from 1-Jan to 1-Apr
15-May runs from 1-Apr to 1-Jul
15-Aug runs from 1-Jul to 1-Oct
15-Feb runs from 1-Jan to 31-Dec

The command

yes? show axis/t seasonal_reg

lists the coordinates, the box sizes the date of the lower limit of the
axis and the time step, or time coordinate value.  The definition of
"month" is an ambiguous one, so in order for the grid cells to be evenly
spaced, things aren't quite as clean as what I listed above.  The axis
has four, equally spaced cells, each 1/4 year long.

(Coordinate axes can actually lie anywhere in a grid cell, IF one
specifies the edges of the cells as well as the coordinates, with the
DEFINE AXIS/EDGES or DEFINE AXIS/BOUNDS qualifier and those edges are
then saved in netCDF files. But that's not done on the climatological
axes in the Ferret dataset climatological_axes.cdf)



On 12/5/2012 9:50 AM, Oots, Penny C. (LARC-E301)[SCIENCE SYSTEMS AND
APPLICATIONS, INC] wrote:

I created a netcdf file by regridding a threads aggregation of monthly
data with
time values of integer number of months since 2006-06-01 (i.e. Time
values are
7.0 to 66.0) to a seasonal_reg axis.   The resulting time axis has the
following
values (in hours) that correspond to Feb-15, May-15, Aug-15, and Nov 15.

SEASONAL_REG[4]
1095.7275, 3287.1825, 5478.6375, 7670.0925


yes? show axis seasonal_reg
  name       axis              # pts   start                end
  SEASONAL_REG TIME              4mr   15-FEB 15:43         15-NOV 14:05
T0 =3D 01-JAN-0000 00:00:00
    Axis span (to cell edges) =3D 8765.82 (modulo length =3D axis span)

Question 1:  Does Feb 15 represent data from Feb 15 to May-15 or from
Jan-15 to Apr-15?  (Is Feb 15 the mid-point or the beginning point?)

And does Nov 15 represent data from Nov 15 to Feb 15 or from Oct 15 to
Dec 15?

And is there a Ferret command that I should have found to display this
info?

Question 2:  Why is the 15th of the month used instead of the first of
the month?

Penny Oots