[NARCCAP-discuss] HRM3-GFDL future time data

[lstack at syntectic.com]

Kelli,
Following on Seth's reply: you can strip February 29th dates out of Excel.
Here's an inelegant method that gets the job done:
• Load your NARCCAP data into an Excel spreadsheet
• Insert two columns
• In the first, create a date series that starts from 1/1/2038 and ends on
12/31/2070 (select a range that is a couple of dozen rows longer than the
length of your data, then fill-series and increment by 1-day increments)
• In the second column that you inserted, select a range that is as long
as the just-created date series. Create an =if() formula that searches
your just-created date series for instances of Month = February and Day =
29. The =if() function will return a true/false flag.
• Select the two columns of date series and 2/29 flags. Sort by the flag.
This moves all 2/29 dates to the beginning or end of the range (depending
on the flags and sort order that you used). Delete the offending dates,
and the column of flags.

If my recipe isn't clear, feel free to email me directly. Do you work with
Phil Mote's team at OSU? Do you know Josh Foster, in Phil's group? I'm up
the road from you in Portland.

Cheers
Latham

-----------------------------

Date: Tue, 29 Apr 2014 16:03:37 -0700
From: "Walters, Kelli Michelle" <waltersk at onid.oregonstate.edu>
Subject: [NARCCAP-discuss] HRM3-GFDL future time data
To: narccap-discuss at mailman.ucar.edu
Message-ID:
	<CAPWvrO-AGji_j384UK3s0+ghMPa3z3=TqDnVunH0Ed8VJk8AuQ at mail.gmail.com>
Content-Type: text/plain; charset=UTF-8

Hi Seth and NARCCAP users,

I am using the HRM3-GFDL future data (precip) and noticed that the future
dataset seems to start on 12/15/2037 instead of 01/01/2038 like all the
rest of the models. The other models I am using state that the time starts
counting from 1/1/2038, but the HRM3-GFDL model says that it starts at
01/01/1968 (like for the current data). When I add the time value to that
start date, the data begins at 12/15/2037.

If I bump the data to begin on 01/01/2038 like the other models, then it
ends on 12/23/2070 like the other models as well, which leads me to
believe
the starting date is just off.

Does anyone know if the date information was somehow shifted or should it
really be starting on 12/15/2037? Or am I maybe calculating something
wrong
with the dates? I am just using excel to add the date value (ex. 25551) to
the starting date.

Any help or information is appreciated.
Thank you,
Kelli

--
*Kelli Walters, *EIT
M.S. Candidate
Civil Engineering | Water Resources
Oregon State University
Office: Owen 233
Email: waltersk at onid.oregonstate.edu
Cell: 503.869.5118


------------------------------

Message: 2
Date: Tue, 29 Apr 2014 19:56:37 -0600
From: Seth McGinnis <mcginnis at ucar.edu>
Subject: Re: [NARCCAP-discuss] HRM3-GFDL future time data
To: narccap-discuss at mailman.ucar.edu
Message-ID: <53605855.8060004 at ucar.edu>
Content-Type: text/plain; charset=ISO-8859-1

Hi Kelli,

The CCSM, CGCM3, and GFDL GCMs use a 365-day calendar, which doesn't
have leap years.  There are 17 leap years between 1968 and 2038, so
Excel's calculation of the date (which uses a standard calendar that
*does* have leap years) is off by 17 days.

The starting date is really 2038-01-01.  Note that the ending dates are
actually 2070-12-31, because there are 8 leap years between 2038 and
2070, so the time span is actually 8 days shorter than you'd expect if
you used a standard calendar.

(The situation with the HadCM3-driven runs is even worse, because HadCM3
uses a 360-day calendar.)

Details and more information, including some important notes about
ragged end points and the spin-up period, are available on our website:

http://narccap.ucar.edu/about/time-periods.html

Cheers,

--Seth


On 4/29/2014 5:03 PM, Walters, Kelli Michelle wrote:
> Hi Seth and NARCCAP users,
>
> I am using the HRM3-GFDL future data (precip) and noticed that the
> future
> dataset seems to start on 12/15/2037 instead of 01/01/2038 like all the
> rest of the models. The other models I am using state that the time
> starts
> counting from 1/1/2038, but the HRM3-GFDL model says that it starts at
> 01/01/1968 (like for the current data). When I add the time value to
> that
> start date, the data begins at 12/15/2037.
>
> If I bump the data to begin on 01/01/2038 like the other models, then it
> ends on 12/23/2070 like the other models as well, which leads me to
> believe
> the starting date is just off.
>
> Does anyone know if the date information was somehow shifted or should
> it
> really be starting on 12/15/2037? Or am I maybe calculating something
> wrong
> with the dates? I am just using excel to add the date value (ex. 25551)
> to
> the starting date.
>
> Any help or information is appreciated.
> Thank you,
> Kelli
>
> --
> *Kelli Walters, *EIT
> M.S. Candidate
> Civil Engineering | Water Resources
> Oregon State University
> Office: Owen 233
> Email: waltersk at onid.oregonstate.edu
> Cell: 503.869.5118
> _______________________________________________
> narccap-discuss mailing list
> narccap-discuss at mailman.ucar.edu
> http://mailman.ucar.edu/mailman/listinfo/narccap-discuss
>


------------------------------

_______________________________________________
narccap-discuss mailing list
narccap-discuss at mailman.ucar.edu
http://mailman.ucar.edu/mailman/listinfo/narccap-discuss


End of narccap-discuss Digest, Vol 30, Issue 1
**********************************************



Latham Stack
Managing Scientist
Syntectic International LLC
syntectic.com
503-901-1939
lstack at syntectic.com