[ncl-talk] computing the model SOI index
Dave Allured - NOAA Affiliate
dave.allured at noaa.gov
Wed Apr 14 09:48:15 MDT 2021
Okay, you want to stay with decimal coded, as in YYYYMM. Your input times
are YYYYMMDD.fffff where the fraction is non-zero. This is a case of
simple integer division and rounding down, except be careful of precision
with this many digits. The original time values are type double, so that
should take care of the precision issue.
Remove YYYY = cd_calendar(TIME,-2)/100. Then do this. Note that
converting double to integer will truncate (set to zero) the fractional
part of the result. Integer division will do the same thing.
YYYYMM = toint (TIME / 100) ; divide in double precision, then
truncate
YYYY = YYYYMM / 100 ; integer division and truncate
To check the results, compare adjacent values. For YYYYMM and sequential
months, the difference should always be the numbers 1 or 89, e.g. 185101 -
185012.
On Wed, Apr 14, 2021 at 1:31 AM Sri nandini <bax8609 at uni-hamburg.de> wrote:
> Hello,
>
> The output is taken from ECMWF output from the atmospheric component of
> the MPI model.
>
> They were orginally in grib format (hence the daily component), and
> converted to .nc with cdo.
>
> At the moment they are appearing as YYYYMMDD and i would like to convert
> the time calendar to YYYYMM so instead of 18500131.99444444 i wish to
> convert to 185001.
>
> Can someone help me with this?
>
> Sri
> On 13.04.21 18:18, Dave Allured - NOAA Affiliate wrote:
>
> Sri, there is something wrong with the time coordinates in fslp->time.
> You show values of [18500131.99444444..20051231.99444444], but units of
> "hours since 1850-01-31 23:15:00". Time coordinates should either be
> offsets relative to the base time in the units string, or else they should
> be decimal coded numbers. It looks like this file mixes the two
> approaches, resulting in your error. Relative offsets is the conventional
> approach, but decimal coded may also be used with care. Just don't mix the
> two.
>
> Do you prefer to fix the input file, or try to use the decimal time values
> the way they are written now? Please keep the conversation on the mailing
> list.
>
>
> On Tue, Apr 13, 2021 at 9:43 AM Sri nandini via ncl-talk <
> ncl-talk at mailman.ucar.edu> wrote:
>
>> Hello dear ncl-users,
>>
>> I am trying to compute the model SOI index to draw a timeseries plot. I
>> have studies the examples on the ncl page for indices_soi_1.ncl and
>> indices_soi_2.ncl , and made my script as per the methods here. But each
>> time an error occurs when reading the time: fatal:An error occurred
>> reading YYYY
>>
>> Variable: TIME
>> Type: double
>> Total Size: 14976 bytes
>> 1872 values
>> Number of Dimensions: 1
>> Dimensions and sizes: [time | 1872]
>> Coordinates:
>> time: [18500131.99444444..20051231.99444444]
>> Number Of Attributes: 4
>> standard_name : time
>> units : hours since 1850-01-31 23:15:00
>> calendar : standard
>> axis : T
>>
>> Variable: YYYY
>> Type: integer
>> Total Size: 7488 bytes
>> 1872 values
>> Number of Dimensions: 1
>> Dimensions and sizes: [1872]
>>
>> Could someone reply to me offline, or help point me to a query or
>> ncl-talk which already solved this or i can also transfer my datafile
>> via ftp? Below is my script.
>>
>> ; =============================================================
>> latT = -17.6 ; Tahiti
>> lonT = 210.75
>> latD = -12.5 ; Darwin
>> lonD = 130.83
>>
>> PLOT = True
>> pltType = "png" ; send graphics to PNG file
>> pltDir = "./" ; dir to which plots are sent
>>
>> pltName = "indices_soi"
>> yrStrt = 1950 ; manually specify for convenience
>> yrLast = 2010 ; 20th century ends 2010
>>
>> clStrt = 1950 ; reference climatology for SOI
>> clLast = 1979
>> ; ======================End User
>> Input=======================================
>> nmos = 12
>> nyrs = yrLast-yrStrt+1
>>
>> fslp = addfile("hist_1850-2005_ens_1-100.slp.nc", "r")
>> slp = fslp->slp
>> printVarSummary(slp)
>> TIME = fslp->time
>> printVarSummary(TIME)
>> ;YYYYMM = cd_calendar(TIME, -1) ;error here
>>
>> YYYY = cd_calendar(TIME,-2)/100
>>
>> tStrt = ind(YYYY .eq. (yrStrt*100 + 1)) ; indices 20th century
>> tLast = ind(YYYY .eq. (yrLast*100 + 12)) ; relative to entire
>> dataset
>>
>> yyyymm = YYYY(tStrt:tLast)
>> tStrtP = ind(yyyymm .eq. (yrStrtP*100+ 1)) ; relative to
>> yrStrt, yrLast
>> tLastP = ind(yyyymm .eq. (yrLastP*100+ 12))
>>
>> T = short2flt(fslp->slp(1632:1871,:,{latT},{lonT}))
>> D = short2flt(fslp->slp(1632:1871,:,{latD},{lonD}))
>> printVarSummary(T)
>> printVarSummary(D)
>> T&time = yyyymm ; convenience
>> D&time = yyyymm
>>
>> tClmStrt = ind(yyyymm.eq.(clStrt*100 + 1)) ; index for clStrt
>> tClmLast = ind(yyyymm.eq.(clLast*100 + 12)) ; index for clStrt
>> printVarSummary(tClmStrt)
>> printVarSummary(tClmLast)
>> TClm = new ( nmos, typeof(T), T at _FillValue) ; monthly
>> climatologies
>> DClm = new ( nmos, typeof(D), D at _FillValue)
>>
>> do nmo=0,nmos-1
>> ;do n=0,99 ; reference clm for each
>> month and each ensemble member?
>> TClm(nmo) = avg(T(tClmStrt+nmo:tClmLast:nmos))
>> DClm(nmo) = avg(D(tClmStrt+nmo:tClmLast:nmos))
>> end do
>> end do
>> printVarSummary(TClm)
>> printVarSummary(DClm)
>> TAnom = T
>> DAnom = D
>> do nmo=0,nmos-1 ; anomalies
>> reference clim
>> TAnom(nmo::nmos) = T(nmo::nmos) - TClm(nmo)
>> DAnom(nmo::nmos) = D(nmo::nmos) - DClm(nmo)
>> end do
>> printVarSummary(TAnom)
>> printVarSummary(DAnom)
>>
>> TAnomStd = stddev(TAnom(tClmStrt:tClmLast)) ; stddev of
>> anomalies over clStrt & clLast
>> DAnomStd = stddev(DAnom(tClmStrt:tClmLast))
>> printVarSummary(TAnomStd)
>> printVarSummary(DAnomStd) ; signal and noise
>> soi_signal = (TAnom/TAnomStd) - (DAnom/DAnomStd) ; (ntim)
>> copy_VarCoords(TAnom, soi_signal)
>> soi_signal at long_name = "SOI: 20th Century Reanalysis:
>> "+yrStrt+"-"+yrLast
>> printVarSummary(soi_signal)
>>
>> ;=======================================
>> ; lag-0 correlation
>> ;=======================================
>>
>> rslp = escorc(soi_signal,slp(lat|:,lon|:,ens:|,time|:))
>> copy_VarCoords(slp(0,:,:,:), rslp)
>> rslp at long_name = "Correlation: SOI-SLP: "+yrStrt+"-"+yrLast
>> printVarSummary(rslp)
>>
>> Best
>>
>> Sri
>>
>> --
>> Dr. Sri, Nandini-Weiss
>> Research Associate
>>
>> Universität Hamburg
>> Center for Earth System Research and Sustainability (CEN)
>> Cluster of Excellence 'Climate, Climatic Change, and Society' (CLICCS)
>>
>> Bundesstrasse 53, 20146 Hamburg
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.ucar.edu/pipermail/ncl-talk/attachments/20210414/9c2c43a0/attachment.html>
More information about the ncl-talk
mailing list