Index of /ftp/ndp026d
Name Last modified Size Description
![[DIR]](https://webarchive.library.unt.edu/eot2008/20080920191056im_/http://cdiac.ornl.gov/icons/back.gif)
Parent Directory -
![[DIR]](https://webarchive.library.unt.edu/eot2008/20080920191056im_/http://cdiac.ornl.gov/icons/folder.gif)
cat01/ 14-Jul-2004 14:23 - Station Info
cat02/ 16-Mar-2003 08:58 - Mean Annual
cat03_07/ 16-Mar-2003 08:59 - Mean Seasonal
cat08_09/ 16-Mar-2003 09:00 - Mean Monthly
cat10_13/ 16-Mar-2003 09:01 - Synoptic Hours
cat14/ 16-Mar-2003 09:17 - Annual,Diurnal Cycles
cat15_30/ 16-Mar-2003 09:07 - Seasonal Means
cat31_42/ 16-Mar-2003 09:11 - Monthly Means (Day)
![[TXT]](https://webarchive.library.unt.edu/eot2008/20080920191056im_/http://cdiac.ornl.gov/icons/text.gif)
ndp-026d.txt 26-Mar-2003 16:26 106K Full Documentation
![[ ]](https://webarchive.library.unt.edu/eot2008/20080920191056im_/http://cdiac.ornl.gov/icons/layout.gif)
ndp026d.pdf 13-May-2004 16:11 17M Full Documentation
readme.txt 16-Mar-2003 13:10 48K Brief Documentation
START_README.NDP-026D*****2002-Feb******
This file contains 812 lines of up to 80 characters.
This file is a brief documentation of the data archive:
*
Carole J. Hahn and Stephen G. Warren, 2002: CLOUD CLIMATOLOGY FOR LAND STATIONS
WORLDWIDE, 1971-96. NDP-026D, Carbon Dioxide Information Analysis Center,
Oak Ridge National Laboratory, Oak Ridge, TN.
(Email to authors: hahn@atmo.arizona.edu; sgw@atmos.washington.edu.)
-
CONTENTS:
I. INTRODUCTION
A) General Description of Data Set
B) Data Source and Stations Included
C) Caution in using this Data Set
-
II. DATA SET CONTENTS AND DATA FORMATS
A) Cloud Type Definitions
Reference to Tables 1 and 2
B) Contents and Organization
Reference to Table 3
C) Data Formats
Reference to Tables 4 and 5
D) Brief Description of Data Files
(1) Averaging methods
(2) Specific Comments on the Data Files
-
III. IMPORTANT NOTES ON USE OF THIS DATA SET
A) Stations with Bogus Amount-When-Present
B) Minimum Observations, the Missing-Value Code, and the Acode
-
IV. REFERENCES
-
V. TABLES
Table 1: Cloud Information Contained in Synoptic Weather Reports
Table 2: Cloud Type and Weather Type Definitions Used
Table 3: Data Organization
Table 4: Format of Header Records
Table 5: Formats of Data Records
Table 6: Terms and Abbreviations Used
*______________________________________________________________________________*
I. INTRODUCTION
A)_ General Description of Data Set
This archive provides a climatology of clouds, based on surface synoptic weather
reports, for 5388 land stations around the globe. It includes multi-year annual,
seasonal and monthly averages for day and night separately; seasonal and monthly
averages by year; averages for eight times per day; and analyses of the first
harmonic for the annual and diurnal cycles. Averages are given for total cloud
cover, 9 cloud types (5 low, 3 middle, 1 high), and clear-sky frequency. Cloud
amounts and frequencies are given for all types, non-overlapped amounts are
provided for middle and high clouds, and mean base heights are computed for low
cloud types.
This archive differs from previous cloud climatologies in this series (NDP-026
and NDP-026A) in several ways. It uses the illuminance criterion of Hahn et al.
(1995) so as to minimize the night-detection bias, making it possible to prepare
a climatology of cloud types for both day and night and to meaningfully evaluate
diurnal cycles of the cloud types. Some cloud types that were grouped together
in the earlier climatology are reported separately here: we now distinguish
between As and Ac and between Sc, St and fog. Finally, this is a climatology
for individual land stations, not grid boxes, so that trends and diurnal cycles
can be evaluated without biases which may arise when using data from more than
one station within a box.
This document provides the minimum description of the archive needed to read
and use the data. For brevity, discussions in this text contain numerous
abbreviations which are defined in the text or in the various tables provided.
Table 6 gives an alphabetical listing of definitions of many of the terms used.
B)_ Data Source and Stations Included
The data source for this analysis was the "Extended Edited Cloud Reports
Archive" (EECRA, Hahn & Warren, 1999), also available from CDIAC as NDP-026C.
The data records of over 11,000 stations were examined. Initially, stations
were selected for use in this climatology if they met two criteria: 1) they had
at least 20 observations (obs) in at least 15 years (over the 26-yr period) for
either January or July, and 2) the number of "night" obs (see Table 6) was at
least 15% of the total obs. (Night obs average about 30% of the total because of
screening by the illuminance criterion.) These criteria were designed to select
stations that would be most likely to have sufficient periods of record for
trend analyses and sufficient night obs for diurnal analyses. The number of
stations meeting these criteria was 5158. However, this group of stations left
vast land areas south of 30 N underrepresented. We therefore included also 230
stations south of 30 N (mostly in Africa and South America) which did not meet
the original two criteria. File 1 of the archive (see below) lists the stations.
C)_ Caution in using this Data Set
It is important to note the cautions described below so as to avoid erroneous
use of the data. For example, not checking the number of obs when required
could lead to using unrepresentative values, and not checking for the "missing
value code" (a negative number; see below) could lead to erroneous analyses.
The full documentation is available from CDIAC (see above).
-
II. DATA SET CONTENTS AND DATA FORMATS
A)_ Cloud Type Definitions
Table 1 lists the cloud information contained in a synoptic weather report.
These quantities, along with the station identification and the time of the
report, are the basic data used to create this climatology.
Table 2 lists the cloud types analyzed for this climatology and provides their
definitions in terms of the synoptic code as defined by the World Meteorological
Organization and as modified in the EECRA. Precipitation codes are also given
because they are used in our definitions of nimbostratus and cumulonimbus cloud
types. Preliminary analyses of the various cirriform cloud types suggested that
the assignment of high clouds to these various types is not globally uniform, so
we group all high clouds together in this dataset.
B)_ Contents and Organization
As shown in Table 3, the archive is divided into 42 file categories (referred to
simply as "Files"; the actual number of physical data files will be larger as
described below). The category divisions are based on content as suggested by
the category names. Thus, File 1 contains station identification information,
File 2 contains mean-annual cloud averages for the various cloud types, Files
3-7 contain mean-seasonal averages (of amount, frequency, etc, respectively),
and so on. Files 2-14 contain the multi-year averages, while Files 15-42 contain
averages for individual years, either seasonally (Files 15-30) or monthly (Files
31-42), which can be used for analyses of trends.
Each file contains a series of one or more "station cloud data groups" (SCDGs).
A data group consists of 5388 data records (one for each of 5388 stations,
ordered by station ID) and a header record which identifies the group:
| Header record identifying data group
| Data record for first station
| Data record for second station
| etc. for 5388 stations
The header record specifies the season, cloud type, cloud parameter and data
format for the data records in the group.
The order in which data groups appear within a multi-group file is:
first season, first type, averages for all stations;
first season, second type, averages for all stations; etc.
This order is indicated by the order of listing in Table 3.
The order in which the data groups appear for the various cloud types is as
shown for File 2 in Table 3. This same sequence (also given in Table 4) is used
for all the files. However, some quantities are not applicable for all types.
For example, there is no "amt" for clear-sky frequency so Cr does not appear in
File 3. So for many files only a subset of the 13 types is given; these subsets
are listed in a footnote to Table 3.
Each data group is assigned a unique sequence number (the SCDG number) which is
indicated in Table 3 and contained in the header record (see Table 4). There are
862 SCDGs in this archive. While it may be convenient to know the sequence
number, it is not essential when searching a file for a desired data group
because the group header record contains information which uniquely defines the
group contents. This is described in the next section.
C)_ Data Formats
The header record is described in Table 4. It defines the content of a data
group by specifying the parameters: type, pcode, year and SN. (Here all data are
for land stations, so LO=1 and SIZE=0 in all headers.) Each header also has a
unique SCDG number which indicates the file category number and the sequence of
the group within the file. The header record also indicates (by number) the
format needed to read the data records. The formats used are defined in Table 5.
All data were written as integers, but if they are read with the specified
formats, the proper floating point values will be obtained. For example, the
integer "1234" would be read as "12.34" if read under fmt 121 (F6.2) or "1234."
if read under fmt 122 (F6.0). Amounts and frequencies are given as percent;
base heights are given in meters.
D)_ Brief Description of Data Files
(1) Averaging methods.
Averages of cloud amt, fq, awp, nol and hgt (defined in Table 6) are given, in
specified files with specified formats, for individual synoptic hours, for
daytime (Dy), for nighttime (Nt), and for the average of day and night (DN). No
minimum number of obs (min) was applied; an "average" is reported even if only
one observation was available (except for amt of mid and hi clouds as discussed
below). The number of obs contributing to each average is included in the data
record (see formats in Table 5). If no obs are available for a particular
average (Nobs=0), the missing value code (Mcode) is inserted for that average.
One must check for the missing value code and/or the Nobs before using the data.
The daily average (avgDN) is computed by one of three methods depending on the
variable averaged. For fq and amt, avgDN is computed as the average of the Dy
and Nt values unless NobD or NobN is less than a specified minimum (which may be
100 or less as indicated below for each file category). This method of forming
avgDN weights day and night equally; it is the preferred method, if sufficient
obs are available for both day and night, because there are generally fewer obs
at night due to use of the illuminance criterion. If NobD or NobN is less than
the min then, an average is computed using all available obs, regardless of time
of day. A flag, the "averaging code" (Acode), is included in the data record to
identify the DN-averaging method employed. Acode values and their meanings are
given in Table 6.
For base height, avgDN is always computed as the simple average of all the obs
of a low cloud type when it is present, regardless of time of day. The Acode
in this case does not represent the averaging method but does indicate the
relationship between NobD, NobN and a specified "minimum" as above.
Amounts of middle and high clouds. Because the synoptic code allows reporting
of only two amounts even if clouds are present at all three levels (Table 1), it
is possible for the amount of a middle or high cloud to be indeterminate even if
the cloud is visible. Therefore we compute an amount-when-present (awp) from the
obs for which the amount can be determined ("number of computable obs", NC) and
obtain the cloud amount as:
amt= fq x awp. (Eqn. 1)
Thus there may be obs from which to compute fq but no obs from which to compute
awp or insufficient obs from which to compute amt (see "mina" in Table 6 and
full documentation for further explanation). In these cases Nobs will be
greater than zero, fq will be given, but amt will be assigned the Mcode. (If
fq=0, amt=0.) Again, it is necessary to check the Mcode before using the data.
AwpDN is computed as amtDN/fqDN (if fq=0, awp=Mcode). This preserves the
relationship in Eqn. 1 but, in general, awpDN computed in this way does NOT
equal (awpD+awpN)/2.
For low clouds, amt (or awp) is given as Nh in the synoptic weather report
(Table 1). Thus the amount of a low cloud type can be determined in the same way
as total cloud amount - it is not necessary to first compute awp. Nevertheless,
we compute awp for low clouds as amt/fq and include it in the archive.
(2) Specific Comments on the Data Files.
Refer to Tables 3 and 5 throughout this discussion.
File 1 provides information about the land stations used. Data records (Format
111) for the 5388 selected stations are listed in ascending order of the station
identifiers (StaID) which here range from 01001 to 98851. Latitude and longitude
are given in degrees (-90 to 90N, 0-360 E) to 2 decimal places, and elevation is
given in meters. The variables ny1, fy1 and ly1 give the number of years of
Januarys with at least 20 obs, the first of such years (e.g. "71" for 1971) and
the last such year. The variables ny7, fy7 and ly7 give the same information for
July. The "station data code" (SDC) indicates the following for a station:
SDC meaning
2 Jan and Jul have at least 15 yrs with at least 20 obs and NobD/NobN < 6
1 Jan or Jul has at least 15 yrs with at least 20 obs and NobD/NobN < 6
0 Neither month has at least 15 yrs with at least 20 obs, Nob D/N unspecified
-1 Jan or Jul has at least 15 yrs with at least 20 obs and NobD/NobN > 6
-2 Jan and Jul have at least 15 yrs with at least 20 obs and NobD/NobN > 6.
A station reporting reliably both day and night for many years in all seasons
will have SDC=2; there are 5053 such stations. Finally, the variable b5c gives
the grid box number on the "5c" grid which we use (and define) elsewhere (NDP-
026, NDP-026A) but do not use here. The variable b5c can be ignored.
File 2 gives the annual average Dy, Nt, and DN amounts of all the cloud types,
including the Low and Mid levels (sum of amounts of all cloud types within each
level as defined in Table 2), and the clear-sky frequency. Annual averages were
computed by averaging the seasonal values from Files 3 and 4. A seasonal value
contributed to the annual average if there were at least 100 obs for the season.
NSN (Format 121) is the number of seasons contributing. There were 5383 stations
that had 4 seasons contributing to Tc, for example. The Acode assigned here for
avgDN was based on the Acodes of the seasonal averages contributing. Acode was
assigned as 2 if all seasons contributing to the annual average had Acode=2. If
any contributing season had Acode=3, then Acode=3 also for the annual average.
Acode=1 does not apply here. Acode=0 if no seasons had 100 obs. 5338 stations
had Acode=2 for Tc while 5278 stations had Acode=2 for Hi amt.
Files 3-5 contain the mean-seasonal (multi-year) averages for amt, fq and awp,
respectively. Amounts for the Low and Mid levels (LoL and MiL) are included in
File 3. File 5 does not include awp for Tc, for which none is computed, or for
Fo (sky obscured by fog) which is, by definition, always 100%. Fo is included in
both Files 3 & 4 even though fq_Fo = amt_Fo. These averages were obtained by
summing, seasonally, all obs within the span of years for each station; they
were not obtained by averaging individual season averages. Note that, because
the sky may be overcast with lower clouds, LOBS >= MOBS >= HOBS. Nobs in Format
121 for awp is NC, the number of occurrences of a cloud type for which an amount
was computable (for low clouds NC=NTy). The Nobs given in a data record for amt
is the same as that for fq, though amt may be missing if NC is inadequate (NC
<mina). The minimum used in computing avgDN was 100 for fq and amt, and 50 for
awp. For Tc in DJF, 5352 stations had Acode=2, while for awp Hi, 4877 stations
had Acode=2.
Files 6 & 7 contain the mean-seasonal averages for two additional variables: the
non-overlapped amounts (File 6) of the four middle and high cloud types, and the
base heights (File 7) for the four low cloud types St, Sc, Cu, Cb (hgt for Fo
is, by definition, zero). Nobs for NOL is larger than MOBS or HOBS (see full
documentation) and Nobs for hgt may be less than NTy since h (Table 2) is not
always reported. Format 122 differs from Format 121 only in that Amt and Fq are
given to two decimal places while Hgt is given to whole meters. The min used for
avgDN for non-overlapped amounts was 100, and the consequent number of stations
with Acode=2 is 5352 for Hi in DJF, for example. Base height was computed from
all available obs, as explained above. Using a min of 50 to compute Acode, 4463
stations have Acode=2 for Sc in DJF. Fewer stations had Acode=2 for other types
which occur less frequently.
Files 8 & 9 contain the mean-monthly averages for amt and fq. The min used in
the computation of avgDN for these files was 75. The number of stations with
Acode=2 for January, for example, is 5283 for Tc and 5059 for Hi. These files
were used in computing the annual cycles in File 14.
Files 10-13 give cloud variable averages (seasonally) for the eight synoptic
hours. Formats 138 and 139 contain fewer variables than Format 121 used above
but eight data lines (one for each hour) are given for each station. No min was
applied for computing these averages (except mina for amt) so the user must
check Nobs (and Mcode for amt) to evaluate the representativeness of the average
given. These files were used in computing the diurnal cycles in File 14.
File 14 gives the phase, amplitude (AMP), and variance accounted for (VAF) of
the first harmonic of the annual (DN avgs only) and diurnal cycles of cloud amt
and fq. The "phase" is the time of maximum of the fitted cosine curve. Formats
140 and 148 differ only in that the label "140" is used to signify that the
values of phase and NT are representative of months whereas the label "148" is
used to signify that phase and NT are representative of hours of the day.
Annual cycles were computed from the monthly (DN) averages in Files 8-9 and are
given for a station only if all 12 months had Acode=2 (min 75 both day and
night). Mcode was inserted for the variables if NT (the number of months) was
less than 12. In this way, annual cycles were obtained for 5240 stations for Tc,
4913 stations for Fq_Hi and 4829 stations for Amt_Hi.
Diurnal cycles were computed from the 3-hourly averages in Files 10 & 11. They
are given for a station if each of the 8 hours had a minimum of 75 obs or if
each of the four 6-hourly times (0,6,12,18 GMT) had the minimum 75 obs. A
station with 8 hours by this test was then tested for the ratio N6/N3, where N6
is the total number of obs at the 6-hourly times and N3 is the total number of
obs at the intermediate 3-hourly times. If this ratio exceeded 4.0 then the
diurnal cycle was computed from only the four 6-hourly averages in order to
reduce a possible bias which may result if reports are made at the intermediate
3-hourly times only in special weather conditions. Diurnal cycles were computed
for 4755 stations for Tc in DJF, for example. Of these, 993 cycles were computed
from only 4 hours; of these, 16 had 75 obs for all 8 hours but failed the N6/N3
ratio test.
Files 15-30 provide Dy, Nt and DN averages of the cloud variables for individual
years (1971-96) of each season. Format 126 (and 127) is similar to Format 121
(and 122) used in Files 2-9 except that 26 data lines (one for each year listed
in ascending order) are given for each station. The min used in the computation
of avgDN for these files varied with the cloud type analyzed because LOBS >=MOBS
>=HOBS, and Nobs becomes limiting as the period of averaging becomes small. If
a station made only one report each night, the maximum NobN would be about 90
for a 3-month season. On average, about 60% of these reports will be excluded
by the illuminance criterion, leaving about 40 obs. Furthermore, reports are
occasionally missing from the source data set because of failures of data
transmission or archival. Considering these factors, we specified the following
mins for avgDN:
35 for Tc, Cr, low clouds, NOL and Fq_Ns;
30 for As, Ac and Amt_Ns;
25 for Hi;
20 for Hgt (used for Acode only; hgt is computed from all available obs).
Using these mins, for DJF for example, the number of stations with Acode=2 in 15
or more years for Tc is 4119 while the number for Amt_Hi is 3536. These files
can be used in analyzing trends in cloud cover. (Note: Mins were not applied for
Dy and Nt averages so NobD and NobN must be consulted to choose representative
avgs. The Mcode must also be checked for middle and high cloud amounts.)
Files 31-42 give monthly averages for individual years (1971-96) for selected
cloud variables. Because more than half of nighttime reports are excluded by
the illuminance criterion, nighttime averages for a single month cannot be fully
representative of that month. Therefore we give only daytime averages for cloud
variables for individual months. It is then convenient to include the three
cloud variables (AMT, FQ, AWP) in a single data record (Format 162). The data
record for a single year includes NOBS, AMT, FQ, AWP and NC. NC is the number
(note exception below) of occurrences of a cloud type for which AWP was
computable (and Amt = Fq * AWP). Again, to allow for user flexibility, no min
is applied in presenting these averages so the user is responsible for checking
Nobs (the maximum possible NobD per month is 124) to determine reliability of
Fq, and for checking NC to determine the reliability of AWP and Amt. Two special
situations, involving AWP for middle and high cloud types, exist for which NC is
assigned a code value (<0) instead of the number of computable occurrences.
Values of -1 or -2 are used to indicate that one of two types of bogus values
for AWP has been used to compute Amt. This is explained more fully in Section
III A. NC is not applicable for Tc and Cr and is there assigned the value -9.
Files 1-42. The size of the entire archive is about 1.5 gigabytes. The File
Categories 2-42 (Table 3) each contain a fairly large amount of data so they are
physically divided into smaller files for ease of handling. The resulting files
are given names that indicate data content. The names contain the File Category
number, the category name abbreviation, and several name extensions to indicate
the season and/or cloud type(s) whose station cloud data group(s) (SCDGs) is
(are) contained in the file. The total of 862 SCDGs are contained in 433 files.
Examples of file names and contents (refer to Table 3, including footnotes):
"02_MACA.tc" contains mean annual total cloud amount (SCDG 02001).
"03_MSCA.41.mil" contains amount for middle level cloud for DJF (SCDG 03011).
"04_MSCF.43.mh" contains frequency of occurrence for the four middle and high
cloud types for JJA (SCDGs 04027-04030).
"06_MSUU.14" contains non-overlapped amount for the four middle and high
cloud types for MAM (SCDGs 06005-06008).
"08_MMCA.04.low" contains amount for the five low cloud types
for April (SCDGs 08032-08036).
"11_MSFT.44.cr" contains clear sky frequency by synoptic hour
for SON (SCDG 11031).
"14_HARM.aa.tc" contains the parameters of the annual cycle
for total cloud amount (SCDG 14001)
"14_HARM.df.44.mh" contains the parameters of the diurnal cycle
for frequency of mid and hi clouds for SON (SCDGs 14097-100).
"16_SMCA.14.sc" contains the amount of stratocumulus cloud
for MAMs for each of the years 1971-96 (SCDG 16004).
"42_MNYD.12.hi" contains the daytime Amt, Fq and AWP of high cloud
for Decembers for each of the years 1971-96 (SCDG 42011).
A complete list of the files and full documentation are available from CDIAC.
-
III. IMPORTANT NOTES ON USE OF THIS DATA SET
A)_ Stations with Bogus Amount-When-Present
China. Because of problems in China's reporting procedures in the 1970's (see
full documentation), we assigned to AWP, for middle cloud types for the years
1971-79, a value obtained by averaging AWP for 1980-89 (averaged and applied for
each of the 12 months separately). This is indicated in Files 31-42 by the value
"-2" assigned to NC (Format 162) for all China stations (StaId begins with "5")
for data records for the years 1971-79. Thus any interannual variations of
middle cloud amounts in China for these years will be due solely to interannual
variations of frequency.
Indonesia and South America. During preliminary analyses, we discovered that
there were two equatorial regions in which the ratio NC/NTy for upper cloud
types (the number of times the cloud amount was computable, divided by the
number of times the cloud was present) was quite small (<0.25 compared to 0.7
globally). Our analysis suggested that the average AWPs obtained from this small
sample were unrepresentative. We chose therefore to apply appropriate (see full
documentation) mean values to AWP for the stations in the affected region.
The values used for AWP are 98% for Ns, 80% for As, 51% for Ac, and 46% for Hi.
The stations affected lie in an irregular region between latitudes 10N to 10S
and between longitudes 95E to 175E (includes Indonesia and other islands) and in
a smaller region of South America from 0 to 10N and 55 to 60W.
StaIDs for the 155 stations affected are:
48077 48096 48108 48110 48112 48300 48303 48325 48327 48328 48330 48331 48351
48352 48353 48354 48356 48357 48375 48376 48377 48378 48379 48381 48383 48400
48403 48405 48407 48425 48426 48430 48431 48432 48450 48455 48456 48459 48460
48462 48475 48477 48480 48500 48501 48517 48532 48550 48551 48552 48564 48565
48567 48568 48569 48580 48583 48601 48615 48620 48647 48657 48665 48698 48917
59995 81200 81202 81225 81250 81251 81253 91203 91212 91217 91317 91324 91334
91348 91353 91356 91366 91367 91371 91376 91408 91413 91425 91434 94044 94085
96001 96009 96011 96015 96035 96073 96075 96091 96109 96145 96147 96163 96171
96179 96221 96237 96249 96253 96295* 96413 96421 96441 96449 96471 96491 96581
96633 96685 96743* 96747* 96797* 96805* 96839 96925 96933 97008 97016 97048
97072 97086 97096 97146 97180 97230 97260 97300 97340 97388* 97390* 97395* 97406
97430 97502 97530 97560 97630 97690 97698 97724 97748 97760 97900* 97980* 98618
*: Added to a preliminary list of 145 stations after completion of Files 2-14.
These "bogus" values of AWP appear in the Dy, Nt and DN avgs in File 5 for these
stations. In Files 31-42 a "-1" appears in the NC variable. Interannual
variations of middle or high cloud amounts for these stations will be due solely
to interannual variations of frequency.
B)_ Minimum Observations, the Missing Value Code, and the Acode
No minimum number of observations was applied to the recording of averages for
the individual synoptic hours, the day average or the night average. This allows
the user to aggregate the data in any manner. However, this also REQUIRES THE
USER TO CHECK the sum of Nobs against a user-specified min, and to check an Amt
for Mcode before using the data. (The amount of mid or hi clouds may be
"missing" even when Nobs>0 as mentioned above and explained in more detail in
the full documentation.) The Acode is a convenient tool for evaluating DN
averages if one accepts the mins applied in creating this archive.
-
IV. REFERENCES
Hahn,C.J., S.G.Warren and J.London, 1995: The effect of moonlight on
observation of cloud cover at night, and application to cloud climatology.
J. Climate, 8, 1429-1446.
Hahn,C.J., and S.G.Warren, 1999: "Extended Edited Synoptic Cloud Reports from
Ships and Land Stations Over the Globe, 1952-1996". NDP-026C, Carbon Dioxide
Information Analysis Center, Oak Ridge National Laboratory, Oak Ridge, TN.
-
V. TABLES
*********-*********-*********-*********-*********-*********-*********-*********-
TABLE 1. CLOUD INFORMATION CONTAINED IN SYNOPTIC WEATHER REPORTS^
________________________________________________________________
Symbol Meaning Codes#
________________________________________________________________
N total cloud cover 0-8 oktas
9= sky obscured
Nh lower* cloud amount 0-8 oktas
h lower* cloud base height 0-9
CL low cloud type 0-9
CM middle cloud type 0-9
CH high cloud type 0-9
ww present weather 00-99
Ix present weather indicator 1-6
________________________________________________________________
^ Reports are made 8 times per day: 00, 03, 06, 09, 12, 15, 18, 21 GMT.
# Any category for which information is lacking is coded as "/".
* The "lower" cloud is the middle level if there are no low clouds.
*********-*********-*********-*********-*********-*********-*********-*********-
TABLE 2. CLOUD TYPE AND WEATHER TYPE DEFINITIONS USED
________________________________________________________________________________
Shorthand Extended
Level notation Meaning Synoptic codes codes#
________________________________________________________________________________
Tc total cloud cover N = 0-9 0-8
Cr completely clear sky N = 0
Ppt precipitation ww= 50-75,77,79,80-99
D drizzle 50-59
R rain 60-69
S snow 70-75,77,79
Ts thunderstorm or shower 80-99
Low CL=
Fo sky obscured by fog / with N=9 and 11
ww=10-12,40-49
St stratus 6,7
Sc stratocumulus 4,5,8
Cu cumulus 1,2
Cb cumulonimbus 3,9,
or N=9 with ww=Ts 10
Mid CM=
Ns nimbostratus 2,7, or N=9 with ww=DRS 12,11,10
/ with ww=DRS and CL=0,7 10
/ with ww= RS and CL=4-8 10
As altostratus 1; 2 if not DRS
Ac altocumulus 3,4,5,6,8,9; 7 if not DRS
High CH=
Ci or Hi cirriform clouds 1-9
________________________________________________________________________________
# Used in the source data set, the EECRA (NDP-026C). Extended codes are shown
where they differ from synoptic codes. In the extended code the value "-1",
rather than "/", is used to signify missing information.
*********-*********-*********-*********-*********-*********-*********-*********-
TABLE 3. DATA ORGANIZATION FOR LAND STATION CLOUD ARCHIVE,1971-1996
________________________________________________________________________________
File Num of SCDG Data
Cat.* SCDGs numbers# Contents** (file category name abbreviation) Format
_____ ______ _________ _______________________________________________ ________
0 README 80
STATION ID (STID)
1 1 01001 LAND_STATION_ID, Lat,Lon, Elev, Yrs 111
MEAN-ANNUAL AVERAGES (MACA)
2 13 02001-13 Mean-Annual CLOUD AMOUNTS & CLEAR-SKY FREQ
ANN:
1 Tc 121
5388 Stations
2 Cr
5388 Stations
3 Fo "
4 St "
5 Sc "
6 Cu "
7 Cb "
8 Ns "
9 As "
10 Ac "
11 Hi "
12 Mid Level "
13 Low Level "
MEAN-SEASONAL AVERAGES
3 48 03001-48 Mean-Seasonal Cloud AMOUNTS (MSCA) 121
1-12 DJF: 12 TYPES, 5388 Stations
13-24 MAM: 12 TYPES, 5388 Stations
25-36 JJA: 12 TYPES, 5388 Stations
37-48 SON: 12 TYPES, 5388 Stations
4 40 04001-40 Mean-Seasonal Cloud FREQUENCY (MSCF) 121
1-10 DJF: 10 TYPES, 5388 Stations
11-20 MAM: 10 TYPES, 5388 Stations
21-30 JJA: 10 TYPES, 5388 Stations
31-40 SON: 10 TYPES, 5388 Stations
5 32 05001-32 Mean-Seasonal AMOUNT-WHEN-PRESENT (MSAW) 121
1-8 DJF: 8 TYPES, 5388 Stations
9-16 MAM: 8 TYPES, 5388 Stations
17-24 JJA: 8 TYPES, 5388 Stations
25-32 SON: 8 TYPES, 5388 Stations
6 16 06001-16 Mean-Seasonal NON-OVERLAPPED AMTS MID & HI (MSUU) 121
4 SEASONS, 4 TYPES, 5388 Stations
7 16 07001-16 Mean-Seasonal BASE HEIGHTS LOW CLOUDS (MSHL) 122
4 SEASONS, 4 TYPES, 5388 Stations
MEAN-MONTHLY AVERAGES
8 120 08001-120 Mean-Monthly Cloud AMOUNTS (MMCA) 121
12 MONTHS (JAN-DEC), 10 TYPES, 5388 Stations
9 120 09001-120 Mean-Monthly Cloud FREQUENCY (MMCF) 121
12 MONTHS, 10 TYPES, 5388 Stations
MEAN-SEASONAL by HOUR (MSAT, MSFT, MSUT, MSHT)
10 40 10001-40 Mean-Seasonal CLOUD AMOUNTS by SYNOPTIC HOUR
4 SEASONS, 10 TYPES, 5388 Stations, 8 Times 138
11 40 11001-40 Mean-Seasonal CLOUD FREQ by SYNOPTIC HOUR
4 SEASONS, 10 TYPES, 5388 Stations, 8 Times 138
12 16 12001-16 Mean-Seasonal NOL MID,HI by SYNOPTIC HOUR
4 SEASONS, 4 TYPES, 5388 Stations, 8 Times 138
13 16 13001-16 Mean-Seasonal BASE HGT LOW by SYNOPTIC HOUR
4 SEASONS, 4 TYPES, 5388 Stations, 8 Times 139
HARMONIC ANALYSES (HARM)
14 100 14001-100 Annual and Diurnal Cycles, First Harmonic
1-10 ANNUAL CYCLE AMOUNT (DN)
10 TYPES, 5388 Stations 140
11-20 ANNUAL CYCLE FREQUENCY (DN)
10 TYPES, 5388 Stations 140
21-60 DIURNAL CYCLE AMOUNT
4 SEASONS, 10 TYPES, 5388 Stations 148
61-100 DIURNAL CYCLE FREQUENCY
4 SEASONS, 10 TYPES, 5388 Stations 148
SEASONAL-MEAN AVERAGES (SMCA, SMCF, SMUU, SMHL)
each SEASON: 10 TYPES, 5388 Stations, 26 Years
15 10 15001-10 Seasonal-Mean CLOUD AMOUNT, DJF 126
16 10 16001-10 Seasonal-Mean CLOUD AMOUNT, MAM 126
17 10 17001-10 Seasonal-Mean CLOUD AMOUNT, JJA 126
18 10 18001-10 Seasonal-Mean CLOUD AMOUNT, SON 126
each SEASON: 10 TYPES, 5388 Stations, 26 Years
19 10 19001-10 Seasonal-Mean CLOUD FREQUENCY, DJF 126
20 10 20001-10 Seasonal-Mean CLOUD FREQUENCY, MAM 126
21 10 21001-10 Seasonal-Mean CLOUD FREQUENCY, JJA 126
22 10 22001-10 Seasonal-Mean CLOUD FREQUENCY, SON 126
each SEASON: 4 TYPES, 5388 Stations, 26 Years
23 4 23001-4 Seasonal-Mean NOL MID, HI, DJF 126
24 4 24001-4 Seasonal-Mean NOL MID, HI, MAM 126
25 4 25001-4 Seasonal-Mean NOL MID, HI, JJA 126
26 4 26001-4 Seasonal-Mean NOL MID, HI, SON 126
each SEASON: 4 TYPES, 5388 Stations, 26 Years
27 4 27001-4 Seasonal-Mean BASE HEIGHT, DJF 127
28 4 28001-4 Seasonal-Mean BASE HEIGHT, MAM 127
29 4 29001-4 Seasonal-Mean BASE HEIGHT, JJA 127
30 4 30001-4 Seasonal-Mean BASE HEIGHT, SON 127
MONTHLY-MEAN AVERAGES, DAY (MNYD)
each MONTH: 11 TYPES, 5388 Stations, 26 Years
31 11 31001-11 Monthly-Mean Daytime Cloud AMT,FQ,AWP, JAN 162
32 11 42001-11 Monthly-Mean Daytime CLOUD AMT,FQ,AWP, FEB 162
33 11 42001-11 Monthly-Mean Daytime CLOUD AMT,FQ,AWP, MAR 162
34 11 42001-11 Monthly-Mean Daytime CLOUD AMT,FQ,AWP, APR 162
35 11 42001-11 Monthly-Mean Daytime CLOUD AMT,FQ,AWP, MAY 162
36 11 42001-11 Monthly-Mean Daytime CLOUD AMT,FQ,AWP, JUN 162
37 11 42001-11 Monthly-Mean Daytime CLOUD AMT,FQ,AWP, JUL 162
38 11 42001-11 Monthly-Mean Daytime CLOUD AMT,FQ,AWP, AUG 162
39 11 42001-11 Monthly-Mean Daytime CLOUD AMT,FQ,AWP, SEP 162
40 11 42001-11 Monthly-Mean Daytime CLOUD AMT,FQ,AWP, OCT 162
41 11 42001-11 Monthly-Mean Daytime CLOUD AMT,FQ,AWP, NOV 162
42 11 42001-11 Monthly-Mean Daytime CLOUD AMT,FQ,AWP, DEC 162
________________________________________________________________________________
* Abbreviations and non-standard terms are defined in Table 6.
# Data group (SCDG) numbers encoded in header format 120 are:
(file_category_number x 1000) + (sequence-number within file category).
Order of data groups in a multi-group file is: hold
LEFT (upper) group qualifier constant and increment RIGHT (lower) qualifier.
** Not all types are given in every file. The rule applied is:
"13 types" means Tc, Cr, Fo, St, Sc, Cu, Cb, Ns, As, Ac, Hi, MiL, LoL.
"12 types" means Tc, Fo, St, Sc, Cu, Cb, Ns, As, Ac, Hi, MiL, LoL.
"11 types" means Tc, Cr, Fo, St, Sc, Cu, Cb, Ns, As, Ac, Hi.
"10 types" means Tc, Fo, St, Sc, Cu, Cb, Ns, As, Ac, Hi (for Amt).
or means Cr, Fo, St, Sc, Cu, Cb, Ns, As, Ac, Hi (for Fq).
" 8 types" means St, Sc, Cu, Cb, Ns, As, Ac, Hi (for Awp).
" 4 types" means Ns, As, Ac, Hi (for NOL),
or means St, Sc, Cu, Cb (for Hgt).
*********-*********-*********-*********-*********-*********-*********-*********-
TABLE 4. GROUP HEADER RECORD FORMAT (Format 120) AND CODES*
USED FOR LAND STATION CLOUD CLIMATOLOGY ARCHIVE#
___________________________________________________________________________
Format I5 I5 I3 I2 I3 I2 I5 I3 I4
Parameter SCDG NSTA SIZE LO TYPE PCODE YEAR SN FMT
Values 10001 5388 0 1=Land 1=Tc 0=AWF (1951) 0=ANN 111
| (820) (5) (2=Ocean) 2=Cr 1=Amt | 121
42011 (1820) (10) (3=Global) (3=Ppt) 2=Fq 1971 1=Jan 122
11=Fo 3=AWP | | 126
12=St 4=NOL 1996 12=Dec 127
13=Sc 5=Hgt (5296) 138
14=Cu 7196 41=DJF 139
15=Cb 42=MAM 140
21=Ns 43=JJA 148
22=As 44=SON 162
23=Ac
30=Hi
20=MiL
10=LoL
___________________________________________________________________________
* Terms are defined in text or in Tables 2 & 6.
# Values in parentheses are not used in this Land Station Archive but are
included to show generalized header record to be used also in future
land and ocean gridded cloud climatologies.
*********-*********-*********-*********-*********-*********-*********-*********-
TABLE 5. DATA FORMATS* FOR READING
LAND STATION CLOUD CLIMATOLOGY ARCHIVE
________________________________________________________________________________
Format Variables (Num of characters Files# in
number & Format in record) which used
______ ________________________________________________________________________
80 Text. nnnA80 (80) F0
------
110 I5 I5 I3 I2 i3 i2 I5 i3 I4 (32) F1,Header
SCDG NSTA SIZE LO -9 -9 YR -9 FMT
111 I5 F6.2 F6.2 I5 I3 I3 I3 I3 I3 I3 I3 I5 (48) F1,Data
StaID LAT LON ELEV ny1 fy1 ly1 ny7 fy7 ly7 SDC b5c
------
120 I5 I5 I3 I2 I3 I2 I5 I3 I4 (32) Headers
SCDG NSTA SIZE LO Type Pcode YR SN FMT
121 I5 I7 F6.2 I7 F6.2 I7 F6.2 I2 (46)
StaID NobD AvgDy NobN AvgNt NobDN AvgDN Acode F3-6,8-9
StaID NSNd AvgDy NSNn AvgNt NSNdn AvgDN Acode F2
122 I5 I7 F6.0 I7 F6.0 I7 F6.0 I2 (46) F7
StaID NobD AvgDy NobN AvgNt NobDN AvgDN Acode
126 26(I5 I7 F6.2 I7 F6.2 I7 F6.2 I2) (26x 46) F15-26
StaID NobD AvgDy NobN AvgNt NobDN AvgDN Acode
127 26(I5 I7 F6.0 I7 F6.0 I7 F6.0 I2) F27-30
138 8( I5 I3 I6 F6.2) (8x 20) F10-12
StaID Hr Nobs Avg
139 8( I5 I3 I6 F6.0) F13
140 I5 F5.2 F5.2 F4.1 I3 F4.1 (26) F14
StaID PHASE AMP VAF NT AVG
148 " " " " " "
162 26( I5 I3 I4 F6.2 F6.2 F6.2 I4 ) (26x 34) F31-42
StaID Yr NobD AmtD FqD AwpD NCD
________________________________________________________________________________
* Abbreviations are defined in text or in Table 6. The value "-9" in Format 110
means that the variable (used in Format 120) is not applicable.
# File categories (F1-F42) are listed in Table 3.
*********-*********-*********-*********-*********-*********-*********-*********-
TABLE 6. TERMS AND ABBREVIATIONS USED
________________________________________________________________________________
Term Meaning and description
---- ------------------------------------------------------------
Acode "Average code" for avgDN; codes for Nobs Dy & Nt:
0=no data {missing value entered},
1= avg all obs {nD+nN < min},
2= (avgD+avgN)/2 {nD>min and nN>min},
3= avg all obs {nD+nN>min}.
AFW Amount, frequency, amount-when-present.
AMP Absolute amplitude of harmonic (not normalized).
amt Average amount of cloud cover, given in percent.
ANN Annual.
avg Average (of amt, Fq, AWP, NOL or height).
avgDy, Nt Mean value of day or night obs.
avgDN Average over day and night.
AWP Amount-when-present.
Cat. Category.
D(y) Abbreviation or suffix meaning "daytime".
day(time) Local time 06-18. Abbreviations used are Dy and D.
DJF December, January, February.
elev Station elevation in meters.
FMT Data format number (see Table 5).
F(q) Frequency of occurrence.
GMT Greenwich Mean Time.
hgt Low cloud base height (given in meters).
hr(s) Hour(s). (00, 03, 06, 09, 12, 15, 18, 21 GMT)
HOBS Number of obs with cloud information for high level.
JJA June, July, August.
lat Latitude (-90 to 90 degrees North).
lon Longitude (0 to 360 degrees East).
LOBS Number of obs with cloud information for low level.
LoL Sum of all clouds in the low level.
Low Low level cloud types (Fo, St, Sc, Cu, Cb).
MAM March, April, May.
Mcode Missing value code (q.v.).
mean seasonal Average over several years for a season.
Mid Middle level cloud types (Ns, As, Ac).
MiL Sum of all clouds in the middle level.
min Minimun number of obs used for averaging or reporting.
mina = Min NC required for computing amts for Hi or Mid clouds.
= min * Fq * 0.6.
missing value code The integer -90000 (-900 for hgt and harmonic parameters).
Put in data record where no legitimate value is computed.
mns Months. (Jan,Feb,Mar,Apr,May,Jun,Jul,Aug,Sep,Oct,Nov,Dec)
MOBS Number of obs with cloud information for middle level.
n Number (of observations); used with other abbreviations.
(Distinguished from symbol for total cloud cover.)
NC Number of obs with cloud type present and amount computable.
night(time) Local time 18-06. Abbreviations used are Nt and N.
Nobs Number of observations; generic for LOBS,MOBS,HOBS, NTy, NC.
NOL Non-overlapped amount; the amount of a middle or high cloud
visible from below.
NSN Number of seasons contributing to the annual average.
NSTA Number of land stations for which data are given (5388).
N(t) Abbreviation or suffix for "nighttime". (Distinct from NT.)
NT Number of hrs used (4 or 8) for diurnal harmonic analysis
or number of months used (12) for annual harmonic analysis.
NTy Number of times a cloud type was reported present.
Num Number.
obs Cloud reports or observations.
Pcode Parameter code: 0=AWF, 1=Amt, 2=Fq, 3=AWP, 4=NOL, 5=Hgt.
PHASE Phase of first harmonic (time of maximum).
Diurnal: 0-24 hours mean solar time of station lat,lon;
when indeterminate (AMP=0), value was set to "-899".
Annual: month (0.5 to 12.4 [1.0 = middle of January, etc.];
0 if AMP=0).
SDC Station data code. Codes selected characteristics of data
for a particular station. See text for File 1 discussion.
seasonal mean Average for an individual year for a particular season.
SGRP Station group number. Aid in identifying data.
SN Season or month indicator in header record or file names.
sn(s) Season(s) (DJF, MAM, JJA, SON).
SON September, October, November.
StaID 5-digit, WMO station identification number.
upper cloud Clouds in middle or high levels.
VAF Percent variance accounted for by the first harmonic.
YEAR Year(s) that apply to data group. Coded as 19yr or as yfyl
for multi-year averages where yf=yr of the first year and
yl=yr of the last year of the period analyzed.
(For example, "7196" means 1971 to 1996.)
yr(s) Year(s).
________________________________________________________________________________
*
END__README.NDP-026D_*********-*********-*********-*********-*********-*********