Translation Specification for Maud data. April 2000, sjw History and preconversion clean up. The original keyed data were received at NSIDC from NCDC. Some QC work was done on the file at NSIDC in preparation for the EWG CD-ROM. In general the corrections were as follows. * about 30 records were labelled as taken in October 1923, but through comparison with Sverdrups location tables these records were actually from September 1923. NSIDC changed the date on these records * In 8 cases a leading '1' before the shelter and mast head air temperatures was assumed to a typo for a '-'. The documentation states, 1 means "negative temperature with brackets". It has not yet been determined what the meaning is of temperature with brackets. * Positions on 30/05/1925 were found to be 75 deg. and 33 minutes, the positions were changed to 75 deg. and 53 minutes. * Some reordering of the records was done to put things in better time sequence. This doesn't effect COADS. This data file was is called maud1.dat and was received from NISDC 7 January 2000. Processing for COADS begins with this file. Preconversion clean up by DSS/NCAR. * All the records were sorted by date and time to obtain a strict chronological order * Several records were found to be misaligned. The following two records were found to be on continuous 240 byte string. They were separated, and MAUD was overwritten into the bytes 1-4 of the 2nd record. MAUD7458N16116E1923103102554-196-199SE 03 10 6 7458N16116E1923103106545-187-189SE 03 10 * Visual inspection of the data shows that once daily positions were assigned to all reports for a given day. Test show this to be true in all but one case. An apparent typo in the longitude was corrected. MAUD7538N16640E1923080502602-015 SW 02 1045 <- fix longitude (16648) MAUD 1923080504 1045 MAUD7538N16648E1923080506613-013 C 00 1045 MAUD 1923080508 1045STR MAUD7538N16648E1923080510604 013 C 00001045 MAUD 1923080512 1045 MAUD7538N16648E1923080514605 008 NNE02 1045 * All reports were checked for valid date (year, month, day, and hour), and location. Reported hours ranged from 02 to 24, note NO reports with 00 hours were found. To avoid report rejection within the COADS data processing, (hour .gt. 23.99 is considered invalid) reports with hour = 24 were assigned hour = 00 and the day was advanced forward by one. * As shown above, in the second half of the Maud time series some reports were keyed without positions. These are typically very sparce data reports (e.g. fields for N and WW ) and occurred at regular two hour intervals between reports with positions. So as to avoid rejecting these reports in COADS processing (all reports must have position) the appropriate constant daily position was assigned. E.g. the above block of data is changed and is shown below. MAUD7538N16648E1923080502602-015 SW 02 1045 MAUD7538N16648E1923080504 1045 MAUD7538N16648E1923080506613-013 C 00 1045 MAUD7538N16648E1923080508 1045STR MAUD7538N16648E1923080510604 013 C 00001045 MAUD7538N16648E1923080512 1045 MAUD7538N16648E1923080514605 008 NNE02 1045 The final record of the Maud time series did not have a position and occurred 2 hours later than the next to last record. This record was assigned the position of the previous record, eventhough it was taken one day later. * Checks were made on ship drift rate to detect position anamolies and possible position errors. Drift checks (position changes) between sequential days revealed a few abnormal shifts in the MAUDS locations. One position was found to have an obvious key error this was corrected. Several other positions were found to have shifts of 50-100kms. The ship movement on days prior to the large shifts were small or zero kilometers. We speculate that these small movements followed by a day with large movement correspond to days when weather conditions did not allow for navigational position determination. We have not attempted to adjust or change these assigned positions. * One cloud report (N) had a value of 19. From surrounding reports this was obviously suppose to be 10. The value was changed. ------------------------------ 2 YR 4 digit year, bytes 16-19 3 MO 2 digit month, bytes 20-21 4 DY 2 digit day, bytes 22-23 5 HR 2 digit hour (no minutes), bytes 24-25 - hours are in "local" time, these were converted to GMT using lmrlib routines rxltut, ixdtnd, and rxnddt 6 TI set = 0, nearest whole hour 7 LON 3 digit degrees, bytes 10-12 2 digit minutes, bytes 13-14 E or W character indicator Convert minutes to decimal degrees, add to 3 digit degrees, test for E or W and adjust to E (0-360) degrees. 8 LAT 2 digit degrees, bytes 5-6 2 digit minutes, bytes 7-8 N indicator (should be no S indicator, but test anyway) Convert minutes to decimal degrees, add to 2 digit degrees, and test to make sure N is specified. 9 LI set = 4, degrees and minutes 10 DCK set = 734, Russian AARI North Pole Stations (EWG data NISDC/NCAR) 11 SID set = 68, Russian AARI North Pole Stations (EWG data NISDC/NCAR) [ DCK and SID need to be updated in online 'lmr') 12 PT set = 9, ice station (manned, including ships overwintering in ice) 18 DI set = 3, 16 of 32-point compass 19 D Convert 3-character string representation of 16 points on the 32 point compass to whole degrees according to Table F2-1 Release 1 COADS Manual. Further, C was mapped to 'calm = 361', and blank is mapped to missing using subroutine wdir16(cdir,dir). The character strings and degree assignments from wdir16 are given below. data comp/'N ','NNE','NE ','ENE','E ','ESE','SE ','SSE', * 'S ','SSW','SW ','WSW','W ','WNW','NW ','NNW', * 'C ',' '/ data deg /'360',' 23',' 45',' 68',' 90','113','135','158', * '180','203','225','248','270','293','315','338', * '361',' '/ 20 WI set = 5, Beaufort force (based on documentation) 21 W Beaufort Wind Force (0-12), bytes 40-41 Convert BF to meters/sec using lmrlib function fxbfms(bf) 24 WW 2 digit code, bytes 46-47. Symbolic notations on the forms were converted to numeric values in this dataset by NCDC. The 2 digit codes assigned included 62, 72, 95, 87, 45, 05, 38, 80, 85, 65, 75, 68, and 69. This is according to the hardcopy documentation and digital images of these documentation (MAUD-1.gif, MAUD-2.gif, MAUD-3.gif, maud.htm). 27 SLP 3 digits (pressure in mm - 700mm), bytes 26-28 Convert to hPA by adding 700 and using lmrlib function fxmmmb(pmm) 28 T1 set = 0, degrees Celsius and tenths 29 AT 4 digits, bytes 29-32 byte 29, - = negative, blank = positive, 1 = negative temperature with brackets 2 = positive temperature with brackets bytes 30-32, degrees C to tenths [Processing at NSIDC resolved byte 29 indicators 1 and 2 to positive or negative. There were approx. 8 occurrences of these indicators. As noted above, the meaning of temperatures in brackets has not been determined.] 34 N 2 digit cloudiness (00-10), bytes 44-45 Convert to oktas cloudiness with lmrlib function ixt1ok(n10). 56 II set = 10, composite information from early ship data 57- 64 ID(8) 4 character string, bytes 1-4 Scott - we need a way to connect scanned or .html documents to these transpecs. E.G. I have the following files which are digital copies, in either .gif or .html of the original NCDC documentation for the MAUD data. I garnered these from NSIDC. MAUD-1.gif MAUD-2.gif MAUD-3.gif maud.htm Supplemental data: [fill in numerical values once processing is finalized] data byte1/1,5,7,9,10,13,15,16,20,22,24,26,29,30,33,34, * 37,40,42,44,46,48,52,55,59,62,66,69/ data byte2/4,6,8,9,12,14,15,19,21,23,25,28,29,32,33,36, * 39,41,43,45,47,51,54,58,61,65,68,120/ n1=1 n2=4 csup(n1:n2)=line(byte1(15):byte2(16)) !Mast head air temperature len=3 n1=n2+1 n2=n2+len csup(n1:n2)=line(byte1(17):byte2(17)) !wind direction - 3 characters len=2 n1=n2+1 n2=n2+len csup(n1:n2)=line(byte1(18):byte2(18)) !beaufort wind force len=2 n1=n2+1 n2=n2+len csup(n1:n2)=line(byte1(19):byte2(19)) !precipitation NOTE: See hard copy (or digital copies) for precipitation. Something does not seem correct. The units should probably be cm and not mm as noted. If mm a maximum of 9.9 mm per 24 hour period would fit into this 2-byte field - this seems like an unrealistic maximum for the Arctic. len=4 n1=n2+1 n2=n2+len csup(n1:n2)=line(byte1(22):byte2(22)) !cloud type - first reported len=3 n1=n2+1 n2=n2+len csup(n1:n2)=line(byte1(23):byte2(23)) !cloud movement for first reported len=4 n1=n2+1 n2=n2+len csup(n1:n2)=line(byte1(24):byte2(24)) !cloud type - second reported len=3 n1=n2+1 n2=n2+len csup(n1:n2)=line(byte1(25):byte2(25)) !cloud movement for 2nd reported len=4 n1=n2+1 n2=n2+len csup(n1:n2)=line(byte1(26):byte2(26)) !cloud type - third reported len=3 n1=n2+1 n2=n2+len csup(n1:n2)=line(byte1(27):byte2(27)) !cloud movement for 3rd reported len=52 n1=n2+1 n2=n2+len csup(n1:n2)=line(byte1(28):byte2(28)) !remarks