TITLE OF THE DATA SET United States Historical Climatology Network (U.S. HCN) Monthly Temperature and Precipitation Data DATA CONTRIBUTORS David R. Easterling, Thomas R. Karl, Elaine H. Mason, Pamela Y. Hughes, National Oceanic and Atmospheric Administration, National Climatic Data Center Asheville, North Carolina 28801, U.S.A. and David P. Bowman Dyntel Central Zone Program, National Climatic Data Center Asheville, North Carolina 28801, U.S.A. SOURCE AND SCOPE OF THE DATA One of the objectives in establishing the U.S. HCN was to detect temporal changes in regional rather than local climate. Therefore, only stations not influenced to any substantial degree by artificial changes in their local environments were included in the network. Some of the stations in the U.S. HCN are first-order weather stations, but the majority were selected from the ~5,000 stations in the U.S. cooperative weather station network. To be included in the U.S. HCN a station had to be currently active (in 1987), have at least 80 years of mean monthly temperature and total monthly precipitation data, and have experienced relatively few station moves and equipment changes (see Appendix A for a complete listing of the stations in the U.S. HCN). An additional criterion used in selecting the 1221 U.S. HCN stations that sometimes compromised the preceding requirement, was the desire to have a uniform distribution of stations across the continental United States. The U.S. HCN database contains station histories, monthly temperature data {mean, mean minimum, average [i.e., (monthly minimum + monthly maximum) / 2], and mean maximum}, and total monthly precipitation data that were compiled from digital and non-digital data sets archived at the National Climatic Data Center (NCDC). These data sets originated from a variety of sources, including climatological publications, universities, federal agencies, individuals, and data archives. All stations were quality controlled by NCDC through the use of outlier and areal edits. Each station record was then corrected for time-of-observation (TOB) differences, instrument changes, instrument moves, and station moves (Karl and Williams 1987). The effects of urbanization were then removed and four urban temperature files produced (Karl et al. 1988). A unique feature of these data are that within most temperature and precipitation data files, both original (or estimated) and adjusted data are given along with confidence factors for each adjusted data value. Another unique feature of the database is that in comparison with the long periods of record, only a small portion of the data are represented as missing. In order to make the U.S. HCN record as serially complete as possible, values for stations missing data have been estimated using data from neighboring stations. Since 1900, over 50% of the 1221 stations have data records that are serially complete (i.e., have original, or estimated, and adjusted data). DATA FORMAT This NDP provides 27 files: 6 FORTRAN input/output routines; 6 SAS input/output routines; 1 station inventory file; 1 station history file; 4 U.S. HCN files containing monthly mean, mean minimum, average (i.e., [monthly minimum + monthly maximum] / 2), and mean maximum temperatures; 1 file containing total monthly precipitation data; 3 quality assessment files for the U.S. HCN temperature and precipitation data files; 4 temperature files (mean, mean minimum, average, and mean maximum) that have been adjusted for urban heat island effects; and this documentation file. The format and contents of each data file are described below. STATION INVENTORY FILE The station inventory file (INVENT94.ASC) provides a list of the stations in the U.S. HCN at the time the data set was compiled. Appendix A of the documentation provides a complete listing of the station inventory file. The file contains 1221 lines of data and provides essential information about each station. Each record contains the state number; station number; latitude and longitude coordinates; station elevation; station name; two letter state code; beginning year of record in the station history file; ending year of record in the station history file; beginning year of record for the minimum, mean, average, maximum temperatures, and precipitation in the U.S. HCN files; and the beginning year of record for minimum, mean, average, and maximum temperatures in the urban files. This file has no missing values. The file may be read by the following FORTRAN format: 10 READ (5,100,END=999) STACOD, LAT, LONG, ELEV, 1 STANAME, STATE, BYRSHF, EYRSHF, HCNMN, HCNAVG, 1 HCNAV2, HCNMX, HCNPCP, URMN, URAVG, URAV2, URMX GOTO 10 C 100 FORMAT(1I6,2F8.2,1I6,1X,1A30,1A2,11(1X,1I4)) Stated in tabular form, the format of the station history files is as follows Table 3. Variable formats for INVENT94.ASC _______________________________________________________________ Variable Variable Variable Column name type width start end _______________________________________________________________ STACOD Integer 6 1 6 LAT Real 8 7 14 LONG Real 8 15 22 ELEV Integer 6 23 28 STANAME Character 30 30 59 STATE Character 2 60 61 BYRSHF Integer 4 63 66 EYRSHF Integer 4 68 71 HCNMN Integer 4 73 76 HCNAVG Integer 4 78 81 HCNAV2 Integer 4 83 86 HCNMX Integer 4 88 91 HCNPCP Integer 4 93 96 URMN Integer 4 98 101 URAVG Integer 4 103 106 URAV2 Integer 4 108 111 URMX Integer 4 113 116 _______________________________________________________________ where STACOD is the unique station code formed by combining the two-digit state number [state numbers range from 1 to 48 (e.g., 01 = Alabama, 02 = Arizona, . . . 48 = Wyoming)] and the four-digit station number (values range from 0008 to 9933); LAT is the station latitude in decimal degrees; LONG is the station longitude in decimal degrees; ELEV is the ground elevation at the station, expressed in whole feet above or below mean sea level (values range from -194 to 9065); STATION is the full name of the cooperative station as of 1994; STATE is a two-letter state abbreviation (e.g., AL = Alabama, AZ = Arizona, etc.); BYRSHF is the beginning year of record in the station history file; EYRSHF is the ending year of record in the station history file (9999 indicates the station is still in operation); HCNMN is the beginning year of record for a station in the minimum temperature data file; HCNAVG is the beginning year of record for a station in the mean temperature data file; HCNAV2 is the beginning year of record for a station in which both the maximum and minimum temperature data are available; HCNMX is the beginning year of record for a station in the maximum temperature data file; HCNPCP is the beginning year of record for a station in the precipitation temperature data file; URMN is the beginning year of record for a station in the minimum temperature, urban heat island effect data file; URAVG is the beginning year of record for a station in the mean temperature, urban heat island effect data file; URAV2 is the beginning year of record for a station in which both the maximum and minimum urban heat island effect temperature data are available; URMX is the beginning year of record for a station in the maximum temperature, urban heat island effect data file. STATION HISTORY FILE The station history file (SHF94.ASC) provides valuable information concerning each station in the U.S. HCN. This file documents station moves and instrument changes, lists station observers and observation times, and identifies suspect fields. When this file is used in conjunction with the station inventory file and quality assessment files the user may accurately assess the quality and suitability of each station's data for long-term climate analyses. The file may be read by the following FORTRAN format: 10 READ (5,100) STACOD, STATE, DIVISION, STANAME, 1 COUNTY, XREF 20 READ (5,110,END=999) STACOD2 BACKSPACE 5 IF (STACOD .NE. STACOD2) GOTO 10 READ (5,115) STACOD2, MOBEG, DAYBEG, YRBEG, 1 MOEND, DAYEND, YREND, (SUSP(I),I=1,15), LATDEG, LATMIN, 1 LONGDEG, LONGMIN, DISTANCE, DPLUNIT, DIRECT, 1 ELEV, DISTPO, DPOUNIT, DIRECTPO, NAME, QUALIF, 1 (INSTRU(I),I=1,36), PCPOT, TMPOT, PCPHT, TMPHT, 1 (PUB(I),I=1,16), OBSNAME, NUMOBS C 100 FORMAT(1I6,1X,1A2,1X,1I2,1X,1A30,1X,1A16,1X,1A25) 110 FORMAT(1I6) 115 FORMAT(1I6,2(2(1X,1I2),1X,1I4),1X,15A1,1X,1I3,1X,1I2,1X, 1 1I4,1X,1I2,1X,1I3,1A1,1A3,1X,1I5,1X,1I4,1A1, 1 1A3,1X,1A28,1X,1A10,1X,36A1,2(1X,2A2),1X,16A1,1X, 1 1A46,1X,1I2) Stated in tabular form, the format of the station history file is as follows. Table 4. Variable formats for SHF94.ASC ______________________________________________________________ Variable Variable Variable Column name type width start end ______________________________________________________________ Station header record STACOD Integer 6 1 6 STATE Character 2 8 9 DIVISION Integer 2 11 12 STANAME Character 30 14 43 COUNTY Character 16 45 60 XREF Character 25 62 86 Station data record STACOD2 Integer 6 1 6 MOBEG Integer 2 8 9 DAYBEG Integer 2 11 12 YRBEG Integer 4 14 17 MOEND Integer 2 19 20 DAYEND Integer 2 22 23 YREND Integer 4 25 28 SUSPLAT Character 1 30 SUSPLONG Character 1 31 SUSPLOC Character 1 32 SUSPELEV Character 1 33 SUSPPO Character 1 34 SUSPNAME Character 1 35 SUSPQUAL Character 1 36 SUSPINST Character 1 37 SUSPTIME Character 1 38 SUSPHTS Character 1 39 SUSPPUBS Character 1 40 SUSPBEG Character 1 41 SUSPEND Character 1 42 SUSPOBS Character 1 43 SUSPOTHR Character 1 44 LATDEG Integer 3 46 48 LATMIN Integer 2 50 51 LONGDEG Integer 4 53 56 LONGMIN Integer 2 58 59 DISTANCE Integer 3 61 63 DPLUNIT Character 1 64 DIRECT Character 3 65 67 ELEV Integer 5 69 73 DISTPO Integer 4 75 78 DPOUNIT Character 1 79 DIRECTPO Character 3 80 82 NAME Character 28 84 111 QUALIF Character 10 113 122 AI Character 1 124 CRS Character 1 125 DT Character 1 126 EVA Character 1 127 FP Character 1 128 HYTHG Character 1 129 MN Character 1 130 MX Character 1 131 NRIG Character 1 132 NSRG Character 1 133 NSS Character 1 134 RRIG Character 1 135 RRNG Character 1 136 SDE Character 1 137 SG Character 1 138 SRG Character 1 139 SS Character 1 140 TG Character 1 141 DGT Character 1 142 TB Character 1 143 EVO Character 1 144 MMTS Character 1 145 TELSY Character 1 146 HYGRO Character 1 147 HY6 Character 1 148 HY8 Character 1 149 SFP Character 1 150 SRRNG Character 1 151 SSG Character 1 152 SSRG Character 1 153 STB Character 1 154 AMOS Character 1 155 AUTOB Character 1 156 PSY Character 1 157 PCPOT Character 2 161 162 TMPOT Character 2 163 164 PCPHT Character 2 166 167 TMPHT Character 2 168 169 BULLETW Character 1 171 COMBBUL Character 1 172 CLIMDATA Character 1 173 RIVSTAGE Character 1 174 HYDROBUL Character 1 175 PRECDATA Character 1 176 SNOWBULL Character 1 177 NOTPUB Character 1 178 CWB Character 1 179 MONTHREV Character 1 180 STATEPUB Character 1 181 LCD Character 1 182 BQ Character 1 183 SGPD Character 1 184 WWR Character 1 185 MYB Character 1 186 OBSNAME Character 46 188 233 NUMOBS Character 2 235 236 ______________________________________________________________ where STACOD is the unique station code formed by combining the two-digit state number [state numbers range from 1 to 48 (e.g., 01 = Alabama, 02 = Arizona, . . . 48 = Wyoming)] and the four-digit station number (values range from 0008 to 9933); STATE is the two-letter state abbreviation (e.g., AL = Alabama); DIVISION is the state U.S. climatological subdivision number [e.g., Vermont (STACOD = 43) has three climatological subdivisions (01 = Northeastern, 02 = Western, and 03 = Southeastern)]; STANAME is the most current station name; COUNTY is the county in which the station is presently located; XREF is a station cross-reference, representing the cooperative station number of the station or the county name that the current station moved to or from; STACOD2 is the unique station code formed by combining the two-digit state number [state numbers range from 1 to 48 (e.g., 01 = Alabama, 02 = Arizona, . . . 48 = Wyoming)] and the four-digit station number (values range from 0008 to 9933); MOBEG is the month the data record started (values range from 01 to 12, and missing values are represented by 99); DAYBEG is the day the data record started (values range from 01 to 31, and missing values are represented by 99); YRBEG is the year the data record started (values range from 1738 to 1994); MOEND is the month the data record ended (values range from 01 to 12, and missing values are represented by 99); DAYEND is the day the data record ended (values range from 01 to 31, and missing values are represented by 99); YREND is the year the data record ended (values range from 1759 to 1994, and 9999 indicates that the station is still in operation at this location). The next 15 variables represent suspect fields in the station history file. The values for these variables will be either 0 or 1. Values of 1 represent field(s) flagged as suspect by the pre-key editor. In the FORTRAN program this data is stored in an single-character, 15-element array named SUSP. 1. SUSPLAT Latitude 2. SUSPLONG Longitude 3. SUSPLOC Distance from previous location 4. SUSPELEV Elevation 5. SUSPPO Distance from Post Office location 6. SUSPNAME Station name 7. SUSPQUAL Qualifier added to the station name 8. SUSPINST Instruments 9. SUSPTIME Observation time 10. SUSPHTS Instrument heights 11. SUSPPUBS Publications 12. SUSPBEG Beginning date 13. SUSPEND Ending date 14. SUSPOBS Observer 15. SUSPOTHR Other observers LATDEG is the degree portion of the station latitude (North). LATMIN is the minute portion of the station latitude (North). LONGDEG is the degree portion of the station longitude (West) (Note: negative signs are missing). LONGMIN is the minute portion of the station longitude (West). DISTANCE is the distance in tenths of miles or city blocks from the previous station location (e.g., 015 = 1.5 miles); if DPLUNIT = "B" then this value represents 15 city blocks; unknown directions are represented by 999. DPLUNIT is the units indicator for DISTANCE (i.e., blank = miles, B = city blocks). DIRECT is the direction on a 16-point compass of a station move from the previous location. Location of the temperature instrument defines the official station location. Values may be blank, character, or numeric. Unknown direction is represented by 999. For example, blank = first record of new station or distance and direction unknown; 015 NW = station moved 1.5 (variable DISTANCE) miles (variable DPLUNIT) NW from previous location; 000 000 = no change in station or instrument location; 000 ESE = moved <0.1 mile east-southeast (ESE) from previous location; 000 999 = moved <0.1 mile, direction unknown; 900 ESE = temperature instrument moved, precipitation instrument did not move or both instruments were moved to the same location; 800 000 = precipitation instrument moved but the temperature instrument did not; and 999 NW = distance unknown, direction NW. ELEV is the ground elevation at the station, expressed in whole feet above or below mean sea level (values range from -194 to 9,843 feet). DISTPO is the distance in tenths of miles or city blocks from the nearest Post Office (e.g. 015 = 1.5 miles); if DPOUNIT = "B" then this value represents 15 city blocks, unknown distances are represented by 9999. DPOUNIT is the units indicator for DISTPO (i.e., blank = miles, B = city blocks). DIRECTPO is the direction on a 16-point compass from the nearest Post Office. Values may be either blank, character, or numeric. Unknown directions are represented by 999.9. For example: blank = distance and direction unknown; 0015 NW = 1.5 (variable DISTPO) miles (variable DPOUNIT) NW of post office; 0000 NW = <0.1 mile NW from post office; 0000 999 = <0.1 mile from post office, direction unknown; and 0000 000 = at the post office. NAME is the full station name. QUALIF is a qualifier or description that is added to the proper name of the station (e.g., Asheville 2N). The next 36 variables represent the instrumentation of the station and the station's classification. If an instrument was used at a particular station or if a particular classification is appropriate for that station, the variable will have a value of 1; if it was not used, the variable will have a value of 0. In the FORTRAN program this data is stored in an single-character, 36-element array named INSTRU. 1. AI Additional instrument (wind, pressure, etc.) 2. CRS Cotton region shelter 3. DT Dry bulb thermometer 4. EVA Class "A" evaporation station 5. FP Fisher-Porter gage 6. HYTHG Hygrothermograph 7. MN Minimum thermometer 8. MX Maximum thermometer 9. NRIG Nonrecording river gage 10. NSRG Nonstandard rain gage 11. NSS Nonstandard shelter 12. RRIG Recording river gage 13. RRNG Recording rain gage 14. SDE Snow density gage 15. SG Storage gage 16. SRG Standard rain gage 17. SS Standard shelter (official) 18. TG Thermograph 19. DGT Digital thermometer 20. TB Tipping bucket gage 21. EVO Other than class "A" evaporation station 22. MMTS Maximum/minimum temperature system 23. TELSY Telemetry system 24. HYGRO Hygrothermometer (type unknown) 25. HY6 Hygrothermometer, H06X series 26. HY8 Hygrothermometer, H08X series 27. SFP Shielded Fischer-Porter gage 28. SRRNG Shielded recording rain gage 29. SSG Shielded storage gage 30. SSRG Shielded standard rain gage 31. STB Shielded tipping bucket gage 32. AMOS Automated meteorological observing system 33. AUTOB Automated observing station 34. PSY Psychrometer (Alaska only) 35. not assigned at present 36. not assigned at present PCPOT is the observation time (rounded to the whole hour) for precipitation. TMPOT is the observation time (rounded to the whole hour) for temperature. Values for PCPOT and TMPOT may be either numeric, character, or alphanumeric. Possible values and their meaning include the following: 01 to 24 = nearest whole hour; 99 = missing; SR = sunrise; SS = sunset; XXHR XX = number of hours observed per day; 9XX9 XX = hours observed, unknown for both temperature and precipitation; TRID = tri-daily observations, Average Temperature = [7 AM + 2 PM + (2 x 9 PM)] / 4; RS = crop season=sunrise (SR) and other months = sunset (SS). PCPHT is the height of the precipitation instrument aboveground, rounded to the nearest foot. TMPHT is the height of the temperature instrument aboveground, rounded to the nearest foot.Values for PCPHT and TMPHT may be given in either numerals or characters. Numerical values range from 01 to 99. Potential values include the following: 01-97 = actual height; 98 = >= 98 feet; 99 = missing; and RF = roof, height unknown. The next 16 variables represent publications. If the data from a particular station appeared in a publication the variable will have a value of 1; if not, the variable will have a value of 0. In the FORTRAN programs, these data are stored in an single-character, 16-element array named PUB. The variables and their corresponding publications are as follows: 1. BULLETW Bulletin W 2. COMBBUL Combined Bulletin 3. CLIMDATA Climatological Data 4. RIVSTAGE Daily River Stages 5. HYDROBUL Hydrologic Bulletin 6. PRECDATA published as hourly precipitation data 7. SNOWBULL Snow Bulletin 8. NOTPUB not published 9. CWB Report to the chief of the U.S. Weather Bureau 10. MONTHREV Monthly Weather Review 11. STATEPUB published in state publications 12. LCD Local Climatological Data 13. BQ Bulletin Q, 1870-1903. 14. SGPD Storage Gage Precipitation Data, Western United States 15. WWR Weekly Weather Review 16. MYB U.S. Meteorological Yearbook OBSNAME is the observer's name (may include more than one name per record). NUMOBS is the number of additional known observers during the time of record for a station (values range from 0 to 28). TEMPERATURE AND PRECIPITATION DATA FILES Monthly data have been extracted from digital and nondigital data sets archived at the NCDC. The data are sorted by element: one file each for monthly maximum temperature, mean temperature, average temperature [i.e., (monthly minimum + monthly maximum)/2], minimum temperature, and total precipitation. The total monthly precipitation and temperature data files are revisions of the files distributed with the last revision of this NDP. These updates extend the period of record from 1987 through 1994. Each record of these four files contain 1 year of monthly data for a given station. These monthly data sets come from various sources as defined by the source field in the station history file. At most, each file will contain four lines of data per year for each station. These lines include (1) original measured data; (2) time of observation corrections, (3) data adjusted for station moves, instrument changes, and other factors influencing the homogeneity of the station record; and (4) confidence estimates for these adjusted values. These files have not been adjusted for urbanization effects. If a station moved between 1987 and 1994, then the adjusted and estimated data for the station will differ from the revision 1 and 2 data files for the entire period of record. The precipitation data have been completely revised due to the correction of logical errors in the precipitation adjustment programs. The average temperature file is new at this revision of the NDP and may contain two data records per year, per station. The "+" record is the TOB mean temperature data calculated from the TOB maximum temperature and TOB minimum temperature data. The "A" record is the FILNET (see p. 36) mean temperature data calculated from the "A" rows in the maximum and minimum temperature data files. The files may be read by the following FORTRAN format. All data are expressed in hundredths of degrees Fahrenheit, and missing values are represented by -9999. 10 READ (5,100,END=999) STACOD, YEAR, ELEM, 1 TYPE, (HCNDAT(I),FLAG(I),FLAG2(I),FLAG3(I), 1 FLAG4(I),I=1,13) GOTO 10 C 100 FORMAT(I6,1X,I4,1X,I1,A1,13(I5,4A1)) During a few years original data were not available and only estimated data and confidence factors are included. Stated in tabular form, the format of the data files are as follows. Table 5. Variable formats for files HCN94MAX.ASC, HCN94MEA.ASC, HCN94AV2.ASC, HCN94MIN.ASC, and HCN94PCP.ASC ______________________________________________________________ Variable Variable Variable Column name type width start end ______________________________________________________________ STACOD Integer 6 1 6 YEAR Integer 4 8 11 ELEM Integer 1 13 TYPE Character 1 14 JAN Integer 5 15 19 JANFLAG1 Character 1 20 JANFLAG2 Character 1 21 JANFLAG3 Character 1 22 JANFLAG4 Character 1 23 FEB Integer 5 24 28 FEBFLAG1 Character 1 29 FEBFLAG2 Character 1 30 FEBFLAG3 Character 1 31 FEBFLAG4 Character 1 32 MAR Integer 5 33 37 MARFLAG1 Character 1 38 MARFLAG2 Character 1 39 MARFLAG3 Character 1 40 MARFLAG4 Character 1 41 APR Integer 5 42 46 APRFLAG1 Character 1 47 APRFLAG2 Character 1 48 APRFLAG3 Character 1 49 APRFLAG4 Character 1 50 MAY Integer 5 51 55 MAYFLAG1 Character 1 56 MAYFLAG2 Character 1 57 MAYFLAG3 Character 1 58 MAYFLAG4 Character 1 59 JUN Integer 5 60 64 JUNFLAG1 Character 1 65 JUNFLAG2 Character 1 66 JUNFLAG3 Character 1 67 JUNFLAG4 Character 1 68 JUL Integer 5 69 73 JULFLAG1 Character 1 74 JULFLAG2 Character 1 75 JULFLAG3 Character 1 76 JULFLAG4 Character 1 77 AUG Integer 5 78 82 AUGFLAG1 Character 1 83 AUGFLAG2 Character 1 84 AUGFLAG3 Character 1 85 AUGFLAG4 Character 1 86 SEP Integer 5 87 91 SEPFLAG1 Character 1 92 SEPFLAG2 Character 1 93 SEPFLAG3 Character 1 94 SEPFLAG4 Character 1 95 OCT Integer 5 96 100 OCTFLAG1 Character 1 101 OCTFLAG2 Character 1 102 OCTFLAG3 Character 1 103 OCTFLAG4 Character 1 104 NOV Integer 5 105 109 NOVFLAG1 Character 1 110 NOVFLAG2 Character 1 111 NOVFLAG3 Character 1 112 NOVFLAG4 Character 1 113 DEC Integer 5 114 118 DECFLAG1 Character 1 119 DECFLAG2 Character 1 120 DECFLAG3 Character 1 121 DECFLAG4 Character 1 122 ANN Integer 5 123 127 ANNFLAG1 Character 1 128 ANNFLAG2 Character 1 129 ANNFLAG3 Character 1 130 ANNFLAG4 Character 1 131 ______________________________________________________________ where STACOD is the unique six digit station code consisting of a two-digit state number [values range from 01 to 48 (i.e., 01 = Alabama, 02 = Arizona, . . . 48 = Wyoming)] and the cooperative station network number (values range from 0008 to 9933); YEAR is the year of record; ELEM is the data element to identify the data file (1 = monthly maximum temperatures, 2 = monthly minimum temperatures, 3 = monthly mean temperatures or monthly average temperatures, and 4 = total monthly precipitation); TYPE is the data type code (blank = original data; + = time of observation adjustments; A = station moves, MMTS corrections, missing data adjustments; and C = confidence factor). Original data are the manuscript or "sources" data that have been archived at NCDC. These data have been manually corrected as needed when checks, such as the annual total vs monthly sum and the average temperature file [i.e., (monthly maximum + monthly minimum)/2] versus mean temperature file, have indicated suspect data. These data are as "clean" as possible with respect to the source (see FLAG2). Adjusted data, along with the confidence factors, are a product of four major computer programs. The station history data and a network of the best correlated nearby stations are used in all these routines. First, the original data are input into a time-of-observation (TOB) debiasing routine so that the data will be consistent with a midnight-to-midnight observation schedule (Karl et al. 1986). Secondly, the TOB corrected data are input into the MMTS program which debiases the data taken from the new Maximum/Minimum Temperature System (MMTS) (Quayle et al. 1991). Thirdly, the data is then entered into the station history adjustment program (SHAP) which debiases the data with respect to instrument changes (other than the MMTS conversion) that may affect the temperature and precipitation data (Karl et al. 1987). The final routine, FILNET, uses the SHAP data and fills in missing original data (i.e., estimates data values based on highly correlated neighboring stations) and then calculates the adjusted data for periods where station changes occurred too often for the SHAP program to make the corrections needed to debias the data. In the FORTRAN code, the monthly and annual data variables are stored in a 13-element integer array named HCNDAT. JAN-DEC are the monthly mean temperature or total monthly precipitation estimates. Original and adjusted or estimated temperature data are expressed in units of hundredths of degrees Fahrenheit but are only measured to the nearest tenth of a degree. Precipitation is in hundredths of inches. For adjusted data, one-half the maximum plus one-half the minimum do not necessarily exactly equal the mean. The mean temperature is adjusted independently of the adjustments of the maximum and minimum values (Karl et al. 1986). A -9999 value indicates that data were not available or adjustments were not made for a given data variable. Confidence factors for temperature are in units of hundredths of degrees Fahrenheit. For precipitation, a dimensionless multiplication factor in hundredths is provided for calculation of confidence intervals. A -9999 indicates that no confidence factors were calculated. Temperature confidence intervals are calculated by subtracting the temperature confidence factor from the adjusted value to obtain the lower end of the confidence interval (16%, -1 standard deviation). The upper end of the confidence interval (84%, +1 standard deviation) is obtained by adding the confidence factor to the adjusted data. For the precipitation confidence intervals, the lower end of the confidence interval (16%) is obtained by dividing the adjusted value by its confidence factor, and the upper end of the confidence interval (84%) is obtained by multiplying the adjusted value by its confidence factor. ANN is the annual mean of the monthly temperature or total annual precipitation estimates. Original and adjusted temperature data are expressed in units of hundredths of degrees Fahrenheit or hundredths of inches (all 12 months must have data for this variable to be calculated). In the FORTRAN code the following variables are stored in four, 4-element, single-character arrays named FLAG, FLAG2, FLAG3, and FLAG4 (in the SAS code the data variables are named FLAG1, FLAG2, FLAG3, and FLAG4). FLAG CODES FOR THE BLANK ROW, THE ORIGINAL DATA (JAN-DEC)FLAG1 is the code for the number of daily values not available in computing the monthly value (flag values are Blank = 0, A = 1, B = 2, C = 3, . . . I = 9; "." = data is estimated). If the source code flag (FLAG2 -see below) equals 0 or 1, then FLAG1 = I indicates that 1 to 9 days of data are missing. (JAN-DEC)FLAG2 is the data source code. The codes and their meaning are as follows: 0 = NCDC Tape Deck 3200, Summary-of-the-Day Element Digital File; 1 = NCDC Tape Deck 3220, Summary-of-the-Month Element Digital File; 2 = Means Book Smithsonian Institute, C. A. Schott (1876, 1881, up to 1931); 3 = Manuscript -Original National Climatic Data Records; 4 = CD Climatological Data, published monthly by NCDC; 5 = Climate Record Book for a description, see History of Climatological Record Books, U.S. Department of Commerce, Weather Bureau, U.S.G.P.O., 1960; 6 = Bulletin W Summary of the Climatological Data for the United States by Section by F. H. Bigelow, U.S. Weather Bureau, 1912; and Summary of the Climatological Data for the United States, (Bulletin W - 2nd ed.); 7 = LCD Local Climatological Data, published monthly by NCDC; 8 = State climatologists various sources; B = Professor Raymond Bradley see Climatic Fluctuations of the Western United States During the Period of Instrumental Records by Bradley, Barry, and Kiladis, Contribution No. 42, Department of Geography and Geology, University of Massachusetts, 1982; D = Dr. Henry Diaz A compilation of data from Bulletin W, LCD and NCDC Tape Deck 3220 (1983); G = Professor John Griffith -primarily from Climatological Data; and Blank = computed value from two of three temperature elements. (JAN-DEC)FLAG3 is a precipitation indicator for trace precipitation (T = Trace). This flag is not used in the temperature files and is blank. (JAN-DEC)FLAG4 is a code for outliers. The codes and their meaning are as follows: S = data are outliers, but NCDC did not have enough confidence to invalidate the observation [the data in question are between 3.5 and 5.0 standard deviations from their mean offset with respect to its nearest neighbors (Karl et al. 1986)]; and X = data are outliers and are represented as missing (>5.0 standard deviations from their mean offset with respect to the station's nearest neighbors); Blank = data are not outliers. ANNFLAG1 is an indicator that daily values were not available when the annual value was computed. The codes and their meaning are as follows: Blank = indicates that data were not missing; and I = indicates that 1 to 9 days of data are missing in at least 1 month of the year. ANNFLAG2 is the same source code used for the original data. ANNFLAG3 is used as an indicator for trace precipitation (T = Trace). ANNFLAG4 is not used. FLAG CODES FOR ROW "+," THE TIME-OF-OBSERVATION ADJUSTED DATA (JAN-DEC)FLAG1 is the same as the code used for the original data in FLAG1. In the average temperature file (HCN94AV2.ASC), a code of "." means that the maximum or minimum data had a code of "."; otherwise, the code is the code of the maximum/minimum data with the greatest number of missing days. (JAN-DEC)FLAG2 is the same source code used for the original data. In the average temperature file, this flag is not used and is blank. (JAN-DEC)FLAG3 is a precipitation indicator for trace precipitation (T = Trace). For temperature data, this flag is an indicator of the quality of the available observation times for a given station. The temperature codes and their meanings are as follows: F = information concerning the observation times for the station during the year was suspect or "flaky" (the temperature data provided when this flag code appears should be used with caution); and G = information concerning the observation times for the station during the year was "good," and the information was judged to be accurate; Blank = observation times were not available for the station during that year. In the average temperature file, this flag represents the worst code for the maximum and minimum data files: blank indicates that the maximum and/or minimum flags were blank; "F" indicates that the maximum and/or minimum flags were "F"; and "G" indicates that the maximum and minimum flags were "G". (JAN-DEC)FLAG4 is the same code as used for FLAG4 in the original data. In the average temperature file, this flag represents the worst code for the maximum and minimum data files: "S" indicates that the maximum and/or minimum flags were "S"; and "X" indicates that the maximum and/or minimum flags were "X"; and blank indicates that the minimum and maximum data were not outliers; ANNFLAG1, ANNFLAG2, and ANNFLAG4 are not used. ANNFLAG3 is used as an indicator for trace precipitation (T = Trace). FLAG CODES FOR ROW "A," THE ADJUSTED DATA (JAN-DEC)FLAG1 is the same code as used for FLAG1 in the original data. (JAN-DEC)FLAG2 is the same code as used for FLAG2 in the original data. (JAN-DEC)FLAG3 is an indicator for the station move bias correction (added by the SHAP program). This flag is set to blank in the precipitation data file. The indicators and their meaning are Blank = no temperature data adjusted -treated as a station move; and O = temperature data adjusted. (JAN-DEC)FLAG4 indicates outliers and missing data estimates. The codes and their meanings are as follows: Blank = original data do not have outliers; S = original data have outliers, but NCDC did not have enough confidence to invalidate the observation; M = no original data are available, but an estimate is provided that is consistent with the data adjusted by using the nearest neighboring stations; or, the data were in excess of 5.0 standard deviations from their mean offset with respect to the station's nearest neighbors; and E = original data are available, but data were estimated by using nearest neighbors because (1) the data were between 3.5 and 5.0 standard deviations from their mean offset and 1 or more days in the month were missing; or (2) adjustments of original data were inappropriate (< 5 years between potential inhomogeneities), so nearest neighbors were used to estimate the data consistent with its 1994, or last known, location. In the average temperature file, this flag represents the worst code for the maximum and minimum data files: "S" indicates that the maximum and/or minimum flags were "S"; "M" indicates that the maximum and/or minimum flags were "M"; and "E" indicates that the maximum and/or minimum flags were "E"; and blank indicates that the minimum and maximum data do not have outliers; ANNFLAG1, ANNFLAG2, ANNFLAG3, and ANNFLAG4 are not used. FLAG CODES FOR ROW "C," THE CONFIDENCE FACTORS (JAN-DEC)FLAG1 is not used. (JAN-DEC)FLAG2 is a code representing the number of the move, counting back from 1994 (0, 1, 2, . . . 9, A = 10, B = 11, . . . P = 25). (JAN-DEC)FLAG3 represents the significance level at which the initial adjustment was made. Possible flag codes and their meanings are as follows: Value Sigma Confidence interval (%) 1 1.0 16 to 84 2 2.0 5 to 95 3 2.57 1 to 99 4 3.3 0.1 to 99.9 (never used at this revision) 5 3.75 0.01 to 99.99 These numeric flag values are used only in the temperature data files. Other possible values are as follows: X = algorithm was unable to adjust the data; U = algorithm was unable to adjust the entire series because of the low station density of the network, but an estimate for the missing data is given by using neighboring stations; A = precipitation data was adjusted by using annual values; or S = precipitation data was adjusted. (JAN-DEC)FLAG4 is not used. ANNFLAG1, ANNFLAG2, ANNFLAG3, and ANNFLAG4 are not used. The preceding information is all that is needed to use the data on a first-look basis. The flags and the data may be interpreted as follows: (1) if the adjusted data and confidence factor are present (not equal to -9999), then the data have successfully passed through all the adjustments and fill-ins and may be considered NCDC's "best" estimate, or (2) if the original data and TOB data are present but the adjusted data and confidence factor are missing, the temperature data have passed through only the TOB program (i.e., station history adjustment and the fill-in procedures could not estimate the adjusted data). DATA QUALITY ASSESSMENT FILES Three data quality assessment files provided were generated by NCDC to allow the user to identify the stations of the highest quality. The information in these files, along with the information in the station history and data inventory files, enable the user to identify the stations with the longest, most complete records. Complete listings of the quality assessment files, by state, are provided in Appendix B of the documentation. An example is given in Appendix B to demonstrate how to interpret and use these files to select stations suitable for climatological analyses. These files provide information for the HCN temperature and precipitation data files only, not the temperature files that have been adjusted for heat island effects. The statistics in these files were generated for the data through 1994. Quality assessments of the minimum and maximum temperature data have been combined into one file. The minimum and maximum temperature quality assessment file may be read by using the following FORTRAN format: 10 READ (5,100,END=999) STATE, STANAME, STATENUM, 1 STANUM, STADIV, LAT, LONG, ELEV, BPORDATE, R1, 1 PERORIG, R2, USCDATE, R3, ASCDATE, R4, PERASCD, 1 R5, PDBPOR, PDR1, APDMX, PDR2, APDMN, PDR3, 1 PDASCD, PDR4, APDASCMX, PDR5, APDASCMN, PDR6, 1 CONFMX, R6, CONFMN, R7, CONS40MX, R8, CONS40MN, 1 R9, CONSFMX, R10, CONSFMN, R11 GOTO 10 C 100 FORMAT(1X,A2,1X,A28,1X,I2,1X,I4,1X,A2,1X, 1 A6,1X,A7,2X,I6,1X,A7,1X,I1,1X, 1 F5.1,2(1X,I1,1X,A7),1X,I1,1X, 1 F6.1,1X,I1,2X,I2,2X,I1,4X,I2,3X, 1 I1,3X,I3,3X,I1,1X,I3,1X,2(I1,3X, 1 I3,3X),I1,1X,F5.2,1X, 1 I1,1X,F5.2,1X,2(I1,3X,I2,3X), 1 I1,2(4X,I2,4X,I1)) The mean temperature and precipitation quality assessment files may be read with the following: 10 READ (5,100,END=999) STATE, STANAME, STATENUM, 1 STANUM, STADIV, LAT, LONG, ELEV, BPORDATE, R1, 1 PERORIG, R2, USCDATE, R3, ASCDATE, R4, PERASCD, 1 R5, PDBPOR, PDR1, APDBPOR, PDR2, PDASCD, 1 PDR3, APDSCD, PDR4, CONF, R6, CONS40, 1 R7, CONSF, R8 GOTO 10 C 100 FORMAT(1X,A2,1X,A28,1X,I2,1X,I4,1X,A2,1X, 1 A6,1X,A7,2X,I6,1X,A7,1X,I1,1X, 1 F5.1,2(1X,I1,1X,A7),1X,I1,1X, 1 F6.1,1X,I1,2X,I2,2X,I1,4X,I2,3X,I1,3X, 1 I3,3X,I1,3X,I3,3X,I1,1X,F5.2,1X,I1, 1 2X,I2,2X,I1,2X,I2,2X,I1) Table 6 shows the contents of the maximum and mean temperature file. Table 7 shows the contents of the mean temperature and total precipitation files. Note that the two files have identical formats through variable PDR1. Table 6. Variable formats for files QA94MXMN.ASC ______________________________________________________________ Variable Variable Variable Column name type width start end ______________________________________________________________ STATE Character 2 2 3 STANAME Character 28 5 32 STATENUM Integer 2 34 35 STANUM Integer 4 37 40 STADIV Character 2 42 43 LAT Character 6 45 50 LONG Character 7 52 58 ELEV Integer 6 61 66 BPORDATE Character 7 68 74 R1 Integer 1 76 PERORIG Real 5 78 82 R2 Integer 1 84 USCDATE Character 7 86 92 R3 Integer 1 94 ASCDATE Character 7 96 102 R4 Integer 1 104 PERASCD Real 6 106 111 R5 Integer 1 113 PDBPOR Integer 2 116 117 PDR1 Integer 1 120 APDMX Integer 2 125 126 PDR2 Integer 1 130 APDMN Integer 3 134 136 PDR3 Integer 1 140 PDASCD Integer 3 142 144 PDR4 Integer 1 146 APDASCMX Integer 3 150 152 PDR5 Integer 1 156 APDASCMN Integer 3 160 162 PDR6 Integer 1 166 CONFMX Real 5 168 172 R6 Integer 1 174 CONFMN Real 5 176 180 R7 Integer 1 182 CONS40MX Integer 2 186 187 R8 Integer 1 191 CONS40MN Integer 2 195 196 R9 Integer 1 200 CONSFMX Integer 2 205 206 R10 Integer 1 211 CONSFMN Integer 2 216 217 R11 Integer 1 222 ______________________________________________________________ where STATE is the two-letter state abbreviation (e.g., AL = Alabama); STANAME is the full name of the cooperative station; STATENUM is the state code (e.g., 01 = Alabama, 02 = Arizona, . . . 48 = Wyoming); STANUM is the cooperative station number (the unique station code, STACOD, used in the other data files is formed by combining the two-digit state number and the four-digit station number); STADIV is the division number [the cooperative station number and division number are separated by a hyphen (e.g., 1084-07)]; LAT is the latitude, expressed in degrees and minutes (e.g., 31-04); LONG is the longitude (West), expressed in degrees and minutes (e.g., 91-17); ELEV is the station elevation, expressed in feet above or below mean sea level; BPORDATE is the first month and year that data became available for a station (e.g., 12/1891); R1 is a decile ranking to judge the length of the period of record (BPORDATE) for a particular station in relation to the rest of the network [values range from 0 to 9, with the lowest deciles given to the longest records (i.e., 0 implies the lower 10 percentile, 1 implies 10-20 percentile, and 9 implies 90-100 percentile)]; PERORIG is the percentage of original data available from the beginning period of record (BPORDATE) through 1994 for a given station (a useful indicator of the amount of missing and estimated data for a given station); R2 is a decile ranking to compare the percentage of original data available for one station with the percentage of original data available for the other stations in the network [values range from 0 to 9, with the lowest deciles assigned to stations having the greatest percentage of original data available (i.e., 0 implies the 1-10 percentile, 1 implies the 10-20 percentile, and 9 implies the 90-100 percentile)]; USCDATE is the date the original data become serially complete, or the date from which there are no missing data in the record for a particular station (99/9999 indicates that the original data never became serially complete and no data exist for 12/1994); R3 is a decile ranking to compare the length of the serially complete original data in relation to the rest of the network [values range from 0 to 9 with the lowest deciles assigned to the stations with the fewest estimated adjusted data (i.e., 0 implies the 1-10 percentile, 1 implies 10-20 percentile, and 9 implies 90-100 percentile)]; ASCDATE is the date at which the adjusted data become serially complete (99/9999 indicates that the adjusted data never became serially complete and no data exist for 12/1994); R4 is a decile ranking to compare the length of the serially complete adjusted data record with other stations in the network [values range from 0 to 9 with the lowest deciles assigned to the stations with the fewest estimated data values (i.e., 0 implies the 1-10 percentile, 1 implies 10-20 percentile, and 9 implies 90-100 percentile)]; PERASCD is the percentage of adjusted data that were estimated since the adjusted serially complete date (ASCDATE); if ASCDATE = 99/9999, then PERASCD = -999.90; R5 is a decile ranking to judge the percentage of the estimated data since the adjusted serial complete date (ASCDATE) [values range from 0 to 9, with the lowest deciles assigned to the stations with the lowest percentages of estimated data (i.e., 0 implies the 1-10 percentile, 1 implies 10-20 percentile, and 9 implies 90-100 percentile)]; PDBPOR is the number of potential discontinuities counted from BPORDATE (from the station history file); PDR1 is a decile ranking to judge the number of potential discontinuities in the record [values range from 0 to 9, with the lowest deciles assigned to the stations with the fewest discontinuities (i.e., 0 implies the 1-10 percentile, 1 implies 10-20 percentile, and 9 implies 90-100 percentile)]; APDMX is the number of potential discontinuity adjustments since BPORDATE for the maximum temperature record; PDR2 is a decile ranking to judge the number of potential discontinuities adjusted for in the maximum temperature record since BPORDATE; APDMN is the number of potential discontinuity adjustments for since BPORDATE for the minimum temperature record; PDR3 is a decile ranking to judge the number of potential discontinuities adjusted for in the minimum temperature record since BPORDATE; PDASCD is the number of potential discontinuities counted from ASCDATE (from station history file), if ASCDATE = 99/9999, then PDASCD = -99.00; PDR4 is a decile ranking to judge the number of potential discontinuities since ASCDATE; APDASCMX is the number of potential discontinuity adjustments made since ASCDATE in the maximum temperature record, if ASCDATE = 99/9999, then APDASCMX = -99.00; PDR5 is a decile ranking to judge the number of potential discontinuity adjustments made since ASCDATE in the maximum temperature record; APDASCMN is the number of potential discontinuity adjustments made since ASCDATE in the minimum temperature record, if ASCDATE = 99/9999, then APDASCMN = -99.00; PDR6 is a decile ranking to judge the number of potential discontinuity adjustments made since ASCDATE in the minimum temperature record; CONFMX is the average confidence interval from ASCDATE for the maximum temperature record, if ASCDATE = 99/9999, then CONFMX = -9.99; R6 is a decile ranking to judge the average confidence interval from ASCDATE for the maximum temperature record; CONFMN is the average confidence interval from ASCDATE for the minimum temperature record, if ASCDATE = 99/9999, then CONFMN = -9.99; R7 is a decile ranking to judge the average confidence interval from ASCDATE for the minimum temperature record; CONS40MX is the consistency of the adjusted maximum temperature data for each station compared with the 20 most highly correlated of the station's 40 nearest neighbors over the past 40 years; R8 is a decile ranking to judge the consistency of the adjusted maximum temperature data for each station with respect to the 20 most highly correlated of the station's 40 nearest neighbors; CONS40MN is the consistency of the adjusted minimum temperature data for each station compared with the 20 most highly correlated of the station's 40 nearest neighbors over the past 40 years; R9 is a decile ranking to judge the consistency of the adjusted minimum temperature data for each station with respect to the 20 most highly correlated of the station's 40 nearest neighbors; CONSFMX is the consistency of the adjusted maximum temperature data for each station with respect to the 20 most highly correlated of the station's 40 nearest neighbors over the full period of record (since BPORDATE); R10 is a decile ranking to judge the consistency of the adjusted maximum temperature data since BPORDATE for each station with respect to the 20 most highly correlated of the station's 40 nearest neighbors; CONSFMN is the consistency of the adjusted minimum temperature data for each station with respect to the 20 most highly correlated of the station's 40 nearest neighbors over the full period of record (since BPORDATE); R11 is a decile ranking to judge the consistency of the adjusted minimum temperature data since BPORDATE for each station with respect to the 20 most highly correlated of the station's 40 nearest neighbors. CONS40MX, CONS40MN, CONSFMX and CONSFMN were assessed by calculating the trend of the difference (or log ratio) series of a station with each of its 20 nearest neighbors. If the station is consistent with its neighbors, no overwhelming trend of sign between the 20 trend values should be found. The number of station pairs with positive trends, minus negative trends, is given. A value above 12 has about a 1% chance of occurrence (if the station is consistent with its neighbors over time). A large value, 14 or more, may imply there is an undocumented move or local modifications of the local environment (i.e., heat island effects, growth of trees, etc.). Table 7. Variable formats for files QA94MEAN.ASC and QA94PCP.ASC ______________________________________________________________ Variable Variable Variable Column name type width start end ______________________________________________________________ STATE Character 2 2 3 STANAME Character 28 5 32 STATENUM Integer 2 34 35 STANUM Integer 4 37 40 STADIV Character 2 42 43 LAT Character 6 45 50 LONG Character 7 52 58 ELEV Integer 6 61 66 BPORDATE Character 7 68 74 R1 Integer 1 76 PERORIG Real 5 78 82 R2 Integer 1 84 USCDATE Character 7 86 92 R3 Integer 1 94 ASCDATE Character 7 96 102 R4 Integer 1 104 PERASCD Real 6 106 111 R5 Integer 1 113 PDBPOR Integer 2 116 117 PDR1 Integer 1 120 APDBPOR Integer 2 125 126 PDR2 Integer 1 130 PDASCD Integer 3 134 136 PDR3 Integer 1 140 APDSCD Integer 3 144 146 PDR4 Integer 1 150 CONF Real 5 152 156 R6 Integer 1 158 CONS40 Integer 2 161 162 R7 Integer 1 165 CONSF Integer 2 168 169 R8 Integer 1 172 ______________________________________________________________ The descriptions for variables STATE through PDR1 are as defined following Table 6. The remaining variables are defined as follows: APDBPOR is the number of potential discontinuity adjustments made since BPORDATE; PDR2 is a decile ranking to judge the number of potential discontinuity adjustments made since BPORDATE; PDASCD is the number of potential discontinuities counted from ASCDATE (from station history file), if ASCDATE = 99/9999, then PDASCD = -99.00; PDR3 is a decile ranking to judge the number of potential discontinuities since ASCDATE; APDSCD is the number of potential discontinuity adjustments made since ASCDATE (from station history file), if ASCDATE = 99/9999, then APDSCD = -99.00; PDR4 is a decile ranking to judge the number of potential discontinuity adjustments ma de since ASCDATE; CONF is the average confidence interval since ASCDATE, if ASCDATE = 99/9999, then CONF = -9.99; R6 is a decile ranking to judge the average confidence interval since ASCDATE; CONS40 consistency of the adjusted data for each station with respect to the 20 most highly correlated of the station's 40 nearest neighbors over the past 40 years; R7 is a decile ranking to judge the consistency of the adjusted data for each station with respect to the 20 most highly correlated of the station's 40 nearest neighbors over the past 40 years; CONSF consistency of the adjusted data for each station with respect to the 20 most highly correlated of the station's 40 nearest neighbors over the full period of record; R8 is a decile ranking to judge the consistency of the adjusted data for each station with respect to the 20 most highly correlated of the station's 40 nearest neighbors over the full period of record. CONS40 and CONSF are assessed by calculating the trend of the difference (or log ratio) series of the station of interest with each of its 20 nearest neighbors. If the station is consistent with its neighbors, no overwhelming number of trend in sign should be found. The number of station pairs with positive trends, minus negative trends, is given. A value above 12 has about a 1% chance of occurrence (if the stations are truly consistent with their neighbors over time). A large value, 14 or more, may imply there is an undocumented moves or modifications of the local environment have occurred. TEMPERATURE FILES ADJUSTED FOR URBANIZATION EFFECTS This document provides four additional temperature files {minimum, mean, average [i.e., (monthly maximum + monthly minimum)/2], and maximum} adjusted for the same biases mentioned in the other temperature files (i.e., station moves, instrument changes, time-of-observation biases, etc.) and also for urbanization effects. The urbanization biases have been removed by using the techniques of Karl et al. (1988). Each station has the same period of record as in the corresponding temperature files (that containing original data, adjusted data, estimated data, and confidence estimates). In these files "urban" only adjusted data are given. No original data or confidence factors are provided. All data are expressed in degrees Fahrenheit, and missing values are represented by -99.99. All four files are formatted in the same manner and may be read by using the following FORTRAN format: 10 READ (5,100,END=999) STACOD, YEAR, (MONTH(I), 1 I=1,12),WIN, SPR, SUM, FAL, ANN GOTO 10 C 100 FORMAT(I6,1X,I4,17(1X,F6.2)) Stated in tabular form, the files are formatted as shown in Table 8. Table 8. Variable formats for files UR94MAX.ASC, UR94MEAN.ASC, UR94AV2.ASC, and UR94MIN.ASC ______________________________________________________________ Variable Variable Variable Column name type width start end ______________________________________________________________ STACOD Integer 6 1 6 YEAR Integer 4 8 11 JAN Real 6 13 18 FEB Real 6 20 25 MAR Real 6 27 32 APR Real 6 34 39 MAY Real 6 41 46 JUN Real 6 48 53 JUL Real 6 55 60 AUG Real 6 62 67 SEP Real 6 69 74 OCT Real 6 76 81 NOV Real 6 83 88 DEC Real 6 90 95 WIN Real 6 97 102 SPR Real 6 104 109 SUM Real 6 111 116 FAL Real 6 118 123 ANN Real 6 125 130 ______________________________________________________________ where STACOD is the unique station code formed by combining the two-digit state number [state numbers range from 1 to 48 (e.g., 01 = Alabama, 02 = Arizona, . . . 48 = Wyoming)] and the four-digit station number (values range from 0008 to 9933); YEAR is the year of record; JAN - DEC are the monthly adjusted temperature estimates expressed in degrees Fahrenheit (missing values are represented by -99.99; estimates have been adjusted for station moves, instrument changes, time-of-observation differences, MMTS, and urbanization effects); and SPR, SUM, FAL, WIN, ANN are the seasonal temperature estimates calculated using the JAN-DEC data, values are expressed in degrees Fahrenheit (missing values are represented by -99.99.). The seasons are defined as follows: Winter (WIN) = December of the preceding year and January and February; Spring (SPR) = March, April, and May; Summer (SUM) = June, July, and August; Fall (FAL) = September, October, and November; and Annual (ANN) = January 1 to December 31. REFERENCES Karl, T. R., and C. N. Williams, Jr. 1987. An approach to adjusting climatological time series for discontinuous inhomogeneities. Journal of Climate and Applied Meteorology 26:1744 1763. Karl, T. R., C. N. Williams, Jr., P. J. Young, and W. M. Wendland. 1986. A model to estimate the time of observation bias associated with monthly mean maximum, minimum and mean temperatures for the United States. Journal of Climate and Applied Meteorology 25:145 160. Karl, T. R., H. F. Diaz, and G. Kukla. 1988. Urbanization: Its detection and effect in the United States climate record. Journal of Climate 1:1099 1123. Quayle, R. G., D. R. Easterling, P. Y. Hughes, and T. R. Karl. 1991. Effects of recent thermometer changes in the cooperative network. Bulletin of the American Meteorological Society 72:1718-1724.