- 1 -

Roy Jenne
11 October 1991
rev. 4 Mar 1992



Daily World Meteorological Data

A.	Daily World Data (CAC)

	This is about the daily world station data prepared by CAC.  NCAR has the data for 
1979-on.  An update for July 1989 - present will soon arrive.  This daily data is derived 
from the 3-hourly world synoptic data that arrives at NMC in real-time by communications.  
The code forms have special groups for precipitation and max/min temperature.  These 
groups are included for some hours of the day.  If the groups are missing, CAC derives 
daily temperature data from the 3-hour data if enough data are present.

	There is now a project under formation that involves several centers including 
NCAR.  This project will result in updates, and has the hope of recovering data for missing 
months and will extend the archive back in time.

	This dataset of world daily station data has started to receive heavy use by the 
research community during the past few years.

1. 	Paul Dyke, Texas A & M, is under contract to use this data (1977-91) to 
produce a lot of data on CD-ROM for agricultural use.

		%  I gave CAC information about Paul's project.  There are now good
discussions between CAC - Paul - NCDC - NCAR.

		%  I told CAC that, in principle, there is data to calculate dailies back to
1967.  The various groups are interested in this idea.  We could help 
supply the inputs.

2.  Daily data for July 1989-on

	We don't have it.  NCDC is probably missing 70% of it.

	CAC is getting it together on an Exabyte tape and will send copies to 
NCAR, NCDC, and Paul Dyke.  Lloyd Thomas (x 8071) at CAC called Dennis 
Joseph about the format.  They talked about a VMS backup format (sounds 
potentially messy).  They will send the data in an ASCII format.

	CAC data through May 1991 is ready to put onto Exabyte tape at NMC.  
June-August 1991 will be ready by November 1; then CAC will send the tape.

3.  Monthly Data on the CAC Tape

	The tape has monthly data derived form the daily data.  It also has monthly 
data that are sent separately by stations in CLIMAT reports.  CAC also gets 
some paper data from places like the Sahel, and puts that data on the tape.

4.  CAC expert, Verne Patterson, retired from CAC a year or so ago, but he still
does consulting (once or so a week).

	%  He confirmed that daily data for 1977-78 was low quality.  He would like
to re-do it.  He is enthusiastic about going further back in time.

	%  Dave Miskus, CAC, is acting as the CAC coordinator on this (301) 763-
4670.  He worked closely with Dave on the project a few years ago.  Jim 
Laver is head of the Analysis and Information Branch (was Fred Finger).  
Miskus is in this branch and works closely with Jim.

5.  Work in Texas

	Paul Dyke, (Texas) has computers and grant money and can enlist about 
fifteen undergraduates from Texas A & M.  (telephone 817-770-6612)

	%  He told me that he has some of the software to prepare daily data from 
three-hour data, but needs help from Verne Patterson.  This is now in the 
bag, I think.

	%  This could fill in the four missing months and give us earlier data.  The 
missing months are July and Sept. 1984; June and Sept. 1986.

	%  Paul also has a special need for daily precipitation for East Africa.  NCAR
 has some special data of those countries.  However, we want the results 
of a merge project at a university.

6.  Sort of Data

	Both Dyke and NCDC are aiming for time series data by station.  This is 
good, but I told CAC we should also have a synoptic sort of the data.

7.  New Variables

	The new data, 1989-on, has some new variables: Dave Miskus is sending 
me a list.  It includes water vapor.  If we can recalculate the series, we should 
add a few new variables, not an infinite number.

8.  CD-ROM

	Paul Dyke was going to make some.  Now it sounds as if NCDC will do 
that job.

9.  Projects at NCAR

	Roland Madden and Dennis Shea are using these data.  Wilhelm (Willy) 
May, ext. 1356, is here from Germany until 01 May 1992 and needs the data as 
soon as possible.


B.	Sources of Daily Data (written December 1991)

	Surface synoptic reports are generally taken each three or six hours by stations 
around the world.  These reports are usually sent over GTS and received by NMC and 
other weather centers.  At certain times of day, the stations report max or min temperature, 
and precipitation.  NCAR has an archive of these basic 3-hour data from NMC for July 
1976-on, and from the USAF (DS 463.0) for 1967-1980.  In addition, NCAR has TD13 
from massive USAF projects that were done 20-30 years ago to key enter old surface 
synoptic data from many countries.  Daily data could be extracted from these three large 
datasets that would provide data for about 7500 global stations for about 1940-on (with 
earlier data also).

	At present NCAR has daily data from CAC for 1979-present (DS 512).  This will 
be re-done from the basic data in order to add elements, to give data for 4 lost months, to 
add at least 2 earlier years, and to fix problems during 1977-79.  Later on , we hope to start 
projects that will extract daily data for the earlier years.

	The USAF at Asheville also has a dataset of summary of day data (TD34) that has 
daily data for 1931-74 with about 206, 270 station year - months of data.  This was key 
entered from forms received from countries.

______________________________________

Appendix A

CAC Daily Data 
25 November 1991


They had a meeting a week ago with Paul Dyke.

	%	CAC soon will send us the daily data that we lack for July 1989-on.

	%	CAC actually sent us data for 1977-on.  They added more verbal information at 
the start of 1977, but still sent us old format for awhile.

	%	Verne is working 10 hours per week, which will increase to 15 hours.

		--	He will decide on the missing four months
		--	We will tell him about tapes for 3-hour and 6-hour data

	%	NCDC has synop tapes from January 1973 (weekly tapes).

	%	NCDC has $25K to process data for 1977-on soon.

	%	Dyke has a contract with Rockefeller Institute to prepare data for 1974-on.

	%	DS 463 is our Dick Davis set of 1967-on surface data.  3-month/tape, about 
17,600 blocks per tape (65 bytes/report) (50 reports/block).  So there are about 
58 MB for a 3-month period.

	%	Hope to determine all 1974-present data again and get extra parameters, as well 
as to clean up problems in early years (about 1977-1979).

	%	We have the Dick Davis data from 1967-on.

		--	Dave asked about this
		--	We could send test tapes
		--	They will check on this


Appendix B

CAC Daily Data
January 2, 1992


Talked with Dave Miskus, CAC, on January 2, 1992

	%	Texas has Jun. and Sep 1986 (missing at NCAR), but they are missing some 
information that we have.

	%	Today CAC will finish getting the 1991 dailies and monthly data into a good 
format

	%	Will send data for July 1989 through 1991 in a few days
		--	on Exabyte
		--	on round tapes


cc:	Willie May
	Rol Madden

______________________________________

Appendix C

CAC Daily Data
28 January 1992


Visit by Paul Dyke and Bill Komar, Jr., Texas A&M, on 28 January 1992.

	1.	Surface 3-hour synoptic

		They have several programs coming up for which the basic 3-hour data 
will be needed.  Many studies will be needed to see the various weather 
elements for the diurnal cycle.

		Mike Hutchins will visit from Australian University.  He does diurnal 
work.  He wants to visit NCAR for 2 or 3 weeks, probably during the summer 
of 1992.

			Verne (at CAC) needs 3-hour data for the one missing month (July 1984?).  
Then he will calculate the daily data.

			The 3-hourly surface data from NMC has the starting dates:

			%	Jul. 1976 in NCAR archives
			%	Jan. 1973 in NCDC archives		

	2.	Texts

		%	CD-ROM:   Gave him the text.  Their CD-ROM will be out in 60 days.
		%	Send storage paper
		%	Send compute text

	3.	Routines

		%	He received Gbytes today
		%	He would like a good sort routine--this takes a long time

	4.	Reanalysis

			Paul is very interested.  Send him:

			--	copy of paper for COADS workshop
			--	copy of December 1991 bulletin paper

	5.	Media

			They like to use Exabyte tapes.

	6.	Daily data for entire period

			Data for the entire period (Sep. 1977 through Dec. 1991) will fit on one 
Exabyte tape (the 5,000 MB type).  In this case, the character data has been 
compressed using the PKZIP routine, which gets good reviews for PCs.


______________________________________

Appendix D

Status of Data at CAC
4 March 1992


	CAC has defined a new and cleaner format for the daily data.  They will convert data 
from Jan. 1987-now to that format.  This has already been done for 1990 and 1991.  But 
they are having systems problems on their VAX and that is slowing the progress.  They 
will then send us the data for 1987-on.

	There are still two missing months, I think these are July and Sept. 1984.  Asheville 
has sent the basic data to them.  But it may be in some merged form.  They are not sure 
what they have as yet.

	Project to decode all of the data again.  They will probably put in a Global Change 
proposal (that includes Madden and Shea at NCAR) to do the new decode and extend data 
back in time.

	(From talk with Dave Miskus, CAC)





cc:	Willie May
	Rol Madden



__________________________________________________________________________
22 FEB 93

 Steve Piper at Scripps is working on a one degree gridded tmin, tmax
 set derived from the old format version of this dataset.
__________________________________________________________________________
[Top of page]