The National Climatic Data Center, hailing from the scenic town of
Asheville, North Carolina,USA bills itself 'The World's Largest
Archive of Climate Data', even though it was set up as a suppository
for U.S. climatological data. It has been classified a World Data
Center A for Meteorology by the WMO.

For those interested in deriving Global climatic data for various
locations Worldwide, I'm going to provide my own interpretation of
what NCDC has to offer the consumer in this area.

First, you should know that NCDC generously provides continuously
updated daily summaries of multiple atmsopheric measurements for
several thousand places around the Globe dating back about a decade.
You can access these on the NCDC website at
http://www.ncdc.noaa.gov/oa/climate/...ata.html#DAILY.
The Global Summary of the Day Dataset is available 'anonymously', free
of charge, in accordance with WMO Resolution 40. Global Summary of the
Day is included in NCDC's "Climate Visualization", or CLIMVIS system,
an interactive graphing tool which displays the daily data in easy to
read format. You can also access the 'original' data in plain text
file format.

In theory, this is a wonderful tool to utilize for various purposes in
relation to climate study.

However, I think the award winning NCDC does a poor job of informing
users of its potential limitations.

If you use the data presented, you need to know that the data
undergoes quality control but it is NOT official data. NCDC has no
control over what is submitted by the various originating agencies
(i.e. national meteorological services). NCDC acknowledges that it
derives its information largely from METARs and SYNOPs. Most users are
probably not familiar with the drawbacks of using these reports for
the purposes of serious climatic analysis. NCDC does not attempt to
explain these limitations at all.

Another aspect that needs to be addressed is the system of measurement
used by NCDC tables. The US continues to use its own English system of
measurement for public dissemination despite the fact that it now
conforms to the SI used by everyone else in the World. True, the
website was built for American users but it does a poor job of
providing a 'metric' option. Since the data it received was reported
in metric (and then converted to English by them alone), the original
source material should be present as well. It does provide a global
'metric' file for the entire month in question, a monstrous confusing
flatfile which is merely the English material re-converted to metric
without regards to the original SI units. Some resolution is lost in
this conversion process. 3.7 mm of rain may be converted to .15 inch,
but .15 inches converted back gives us 3.8 mm, for example.
This example may seem trivial and it is but long term sums yield
larger accumulation errors that may deviate from actual totals and
averages.
Fortunately, the resolution of .1 deg Fahrenheit is greater than the
resolution of .1 deg Celsius, so manual conversion of temperature
units is possible. However, in regards to precipitation, .01/inch has
less resolution than .1/mm, so some minor accuracy is lost due to
rounding approximation. Worse though, is the 'mysterious' method NCDC
uses to translate amounts under 1mm/.04 mm: Trace mm amounts are
defaulted to .03 inch just as .7 and .8 mm rightfully are.
NCDC often uses a seemingly random combination of metars and synops to
derive max/min temperatures of the day, which make true monthly return
estimates almost impossible.
METARS do not even usually report daily max/mins and precip. amounts,
only basic hourly readings.
Some stations' daily contributions may have no more than 4 hourly
readings and the greatest or least of these hourly readings is
reported by CLIMVIS as the high or low temperature of the day!

Therefore, for the sake of compiling monthly returns from cumulative
daily returns, CLIMVIS users should definitely tread with caution.
I've noticed many instances of what appeared to be incomplete or
missing daily precipitation amounts in comparison with the SYNOPs on
which they are supposedly based, and in clear conflict with the data
as reported by the respective countries' weather services.

The bottom line is, CLIMVIS might be okay for approximate day by day
weather readings but I do not recommend it as being reliable at all as
an accurate source of daily or monthly climatic data.

Have you sent a copy of this critisism to the NCDC?

Cheers, Alastair.

"Darrell H" wrote in message
m...
The National Climatic Data Center, hailing from the scenic town of
Asheville, North Carolina,USA bills itself 'The World's Largest
Archive of Climate Data', even though it was set up as a suppository
for U.S. climatological data. It has been classified a World Data
Center A for Meteorology by the WMO.

For those interested in deriving Global climatic data for various
locations Worldwide, I'm going to provide my own interpretation of
what NCDC has to offer the consumer in this area.

First, you should know that NCDC generously provides continuously
updated daily summaries of multiple atmsopheric measurements for
several thousand places around the Globe dating back about a decade.
You can access these on the NCDC website at
http://www.ncdc.noaa.gov/oa/climate/...ata.html#DAILY.
The Global Summary of the Day Dataset is available 'anonymously', free
of charge, in accordance with WMO Resolution 40. Global Summary of the
Day is included in NCDC's "Climate Visualization", or CLIMVIS system,
an interactive graphing tool which displays the daily data in easy to
read format. You can also access the 'original' data in plain text
file format.

In theory, this is a wonderful tool to utilize for various purposes in
relation to climate study.

However, I think the award winning NCDC does a poor job of informing
users of its potential limitations.

If you use the data presented, you need to know that the data
undergoes quality control but it is NOT official data. NCDC has no
control over what is submitted by the various originating agencies
(i.e. national meteorological services). NCDC acknowledges that it
derives its information largely from METARs and SYNOPs. Most users are
probably not familiar with the drawbacks of using these reports for
the purposes of serious climatic analysis. NCDC does not attempt to
explain these limitations at all.

Another aspect that needs to be addressed is the system of measurement
used by NCDC tables. The US continues to use its own English system of
measurement for public dissemination despite the fact that it now
conforms to the SI used by everyone else in the World. True, the
website was built for American users but it does a poor job of
providing a 'metric' option. Since the data it received was reported
in metric (and then converted to English by them alone), the original
source material should be present as well. It does provide a global
'metric' file for the entire month in question, a monstrous confusing
flatfile which is merely the English material re-converted to metric
without regards to the original SI units. Some resolution is lost in
this conversion process. 3.7 mm of rain may be converted to .15 inch,
but .15 inches converted back gives us 3.8 mm, for example.
This example may seem trivial and it is but long term sums yield
larger accumulation errors that may deviate from actual totals and
averages.
Fortunately, the resolution of .1 deg Fahrenheit is greater than the
resolution of .1 deg Celsius, so manual conversion of temperature
units is possible. However, in regards to precipitation, .01/inch has
less resolution than .1/mm, so some minor accuracy is lost due to
rounding approximation. Worse though, is the 'mysterious' method NCDC
uses to translate amounts under 1mm/.04 mm: Trace mm amounts are
defaulted to .03 inch just as .7 and .8 mm rightfully are.
NCDC often uses a seemingly random combination of metars and synops to
derive max/min temperatures of the day, which make true monthly return
estimates almost impossible.
METARS do not even usually report daily max/mins and precip. amounts,
only basic hourly readings.
Some stations' daily contributions may have no more than 4 hourly
readings and the greatest or least of these hourly readings is
reported by CLIMVIS as the high or low temperature of the day!

Therefore, for the sake of compiling monthly returns from cumulative
daily returns, CLIMVIS users should definitely tread with caution.
I've noticed many instances of what appeared to be incomplete or
missing daily precipitation amounts in comparison with the SYNOPs on
which they are supposedly based, and in clear conflict with the data
as reported by the respective countries' weather services.

The bottom line is, CLIMVIS might be okay for approximate day by day
weather readings but I do not recommend it as being reliable at all as
an accurate source of daily or monthly climatic data.

Alastair,
How do you say in Britain..Cant be arsed with it.

Do you think I should send a polite inquiry to them?
Perhaps I will.

Their publications dealing with US climatological data are pretty good anyway.

I think what they are doing is trying to summarize all data according
to UTC time and 'the system' defaults to whichever is higher for the
max. temp., metar hourly reading or synop daily max. reading, and
whichever is lower for the min. temp, metar hourly reading or synop
daily min. reading, thats my guess anyway. Daily Prec amounts conflict
because of when data is summarized (in this case 0600+1800 for synops,
vs 2400 for gsod)

Here's an example of this month for London Heathrow:

NCDC GSOD*
Day tmx tmn prec
1 12.0 9.5 10.2
2 14.2 10.3 0.5
3 15.3 11.6 0.0
4 16.6 12.0 1.5
5 13.5 10.0 2.0
6 12.0 7.0 5.1
7 10.2 5.0 0.5
8 9.0 1.0 1.0
9 6.2 -2.4 0.0
10 9.0 1.2 0.0
11 13.0 5.8 0.0
12 11.1 7.9 0.0
13 10.0 7.0 0.0
14 10.0 5.7 1.5

*(F converted to C, in to mm)

Day SYNOP
tmx tmn prec
1 11.8 9.6 7.4
2 14.2 10.3 0.2
3 15.3 11.6 1
4 16.6 12.1 0.4
5 13.5 12.1 2
6 11.9 10.2 5.4
7 10.2 5.7 1
8 8.9 3.3 0
9 6.2 -2.4 0
10 8.6 -1.3 0
11 12.6 5.8 0
12 11.1 7.9 0
13 9.6 7.4 0.4
14 10 6.7 1.8