From http://www1.ncdc.noaa.gov/pub/data/cirs/drought.README 
                      TIME BIAS CORRECTED DIVISIONAL
                    TEMPERATURE-PRECIPITATION-DROUGHT
                                INDEX

                              (TD-9640)



                              MARCH 1994




The major parameters in this file are sequential "Time Biased Corrected" state
climatic division monthly Average Temperatures (Deg.F. to 10ths), Precipitation
(Inches to 100ths), and Palmer Drought Indices (PDSI, PHDI, PMDI, and ZNDX). 
Period of record is 1895 through latest month available, updated monthly.

Monthly averages within a climatic division have been calculated by giving
equal weight to stations reporting both temperature and precipitation within a
division. In the U.S., observers at cooperative stations often take one
observation per day, and the ending time of the climatological day at any
station can vary from station-to-station as well as year-to-year.  Differences
of the 24-hour period over which each observer reports his or her maximum and 
minimum temperature as well as the average temperature [(max + min)/2] affects 
the calculated monthly mean temperature.  Karl, et al. (1986), describe the 
biases that this introduces. These potential biases were rectified by 
adjusting for these varying observation times.  The model described by Karl, 
et al. (1986), was used to adjust the climate division averages such that all 
stations end their climatological day at midnight; i.e., climatological day 
coincides with calendar day.  The time of observation was determined at each 
station within a climate division during January of the years 1931, 1941, 1951,
1965, 1975, and 1984 for the states of California, Colorado, Illinois, Indiana,
New York, North Carolina, and Washington.  The fraction of observers recording 
at various hours of the day was calculated and interpolated for intervening
years (extrapolated for subsequent years).  For these seven states, the ending 
time of observation was grouped into three categories:  AM, PM, and MD.
The AM category included observers who ended their climatological day between
3 AM and 11AM; the PM category between noon and 9 PM; and the MD category
between 10 PM and 2 AM; all local standard time.  The fraction of observers in
these categories was calculated, and it was assumed the 7 AM observation time 
best represented the AM category; the 5 PM observation time, the PM category; 
and midnight for the MD category.  The reason for the simplification was to 
test if a faster method, requiring significantly less bookkeeping and 
keypunching, could not provide nearly as good results as calculating the 
fraction of observers at each of the 24 hours of the day.

The time of observation bias model was run by using the latitude and longitude 
of each of the centroids of the climate divisions.  The output from the model 
was the time of observation bias, with respect to a midnight-to-midnight 
climatological day, for each of the possible ending hours of the climatological
day.  Each climate division's monthly average was then adjusted by weighting 
the bias at any given hour by the fraction of stations within the climate 
division observing at that hour, and subtracting the result for the reported 
monthly mean temperature.

Differences of the biases were small (< 0.3 Deg. F.) for those calculated by
categorizing the ending time of observation into three categories compared to
those obtained from calculating the fraction of stations with observation times
at each of the 24 hours of the day.  This is attributed to the preponderance of
AM observation times falling between 6 AM and 9 AM, and PM observation times
falling between 4 PM and 7 PM.  As a result, by assuming 7 AM observation for 
all AM stations and 5 PM for all PM stations, a good estimate of the median 
bias is obtained for all AM or PM observations.  Furthermore, nearly all the MD
stations observed at midnight.

It should also be noted that the borders of the climate divisions in 1951 were 
not consistent with those defined in 1965.  Due to the substantial additional 
effort it would have required locating each station within three or four 
climate divisions, as defined today, the change in the statewide percentage of 
AM, PM, and MD observation times was applied in equal proportions to all 
climate divisions prior to and including 1951.

Based on small differences between the two methods of estimating the time of
observation bias, the simpler categorical procedure was used for all climate
divisions.  This should effectively eliminate most of the biases (over 2 Deg.F)
in some climate divisions that have become part of the divisional averages.  
These biases affect both trends and actual estimates of divisional averages.  
Reference Karl, et al. (1986):  "A model to estimate the time of observation 
bias associated with monthly mean maximum, minimum, and mean temperatures for 
the Unites States" (Thomas R. Karl, Claude N. Williams, Jr., and Pamela J.
Young, National Climatic Data Center, and Wayne M. Wendland, Illinois State 
Water Survey, Journal of Climate and Applied Meteorology, January 1986, 
American Meteorological Society, Boston, MA).  The time of observation bias 
model which can be used for the maximum, minimum, and mean monthly temperature 
at any location in the contiguous U.S. [program NCC*F11SRC (OBTBIAS Code) and 
NCC*F11REL (OBTBIAS)], is available from the NCDC.

Historical drought data have been added to this file for the period 1895 to
present.  The file is updated monthly.  All drought data are calibrated using 
the period 1895-1990 (cf. Karl, 1986; Journal of Climate and Applied 
Meteorology, Vol. 25, No. 1, January 1986).  Drought data include:  

1.  Palmer Drought Severity Index (PDSI)

 This is the monthly value (index) that is generated indicating the severity
 of a wet or dry spell.  This index is based on the principles of a balance
 between moisture supply and demand.  Man-made changes were not considered in
 this calculation.  The index generally ranges from -6 to +6, with negative
 values denoting dry spells and positive values indicating wet spells.  There
 are a few values in the magnitude of +7 or -7.  PDSI values 0 to -.5 =
 normal; -0.5 to -1.0 = incipient drought; -1.0 to -2.0 = mild drought; -2.0
 to -3.0 = moderate drought; -3.0 to -4.0 = severe drought; and greater than -
 4.0 = extreme drought.  Similar adjectives are attached to positive values of
 wet spells.  This is a meteorological drought index used to assess the
 severity of dry or wet spells of weather.

2.  Palmer Hydrological Drought Index (PHDI)

 This is the monthly value (index) generated monthly that indicates the
 severity of a wet or dry spell.  This index is based on the principles of a
 balance between moisture supply and demand.  Man-made changes such as
 increased irrigation, new reservoirs, and added industrial water use were not
 included in the computation of this index.  The index generally ranges from -
 6 to +6, with negative values denoting dry spells, and positive values
 indicating wet spells.  There are a few values in the magnitude of +7 or -7. 
 PHDI values 0 to -0.5 = normal; -0.5 to -1.0 = incipient drought; -1.0 to -
 2.0 = mild drought; -2.0 to -3.0 = moderate drought; -3.0 to -4.0 = severe
 drought; and greater than -4.0 = extreme drought.  Similar adjectives are
 attached to positive values of wet spells.  This is a hydrological drought
 index used to assess long-term moisture supply.

3.  Palmer "Z" Index (ZNDX)

 This is the generated monthly Z values, and they can be expressed as the
 "Moisture Anomaly Index."  Each monthly Z value is a measure of the departure
 from normal of the moisture climate for that month.  This index can respond
 to a month of above-normal precipitation, even during periods of drought. 
 Table 1 contains expected values of the Z index and other drought parameters. 
 See Historical Climatology Series 3-6 through 3-9 for a detailed description
 of the drought indices.

4.  Modified Palmer Drought Severity Index (PMDI)

 This is a modification of the Palmer Drought Severity Index.  The
 modification was made by the National Weather Service Climate Analysis Center
 for operational meteorological purposes.  The Palmer drought program
 calculates three intermediate parallel index values each month.  Only one
 value is selected as the PDSI drought index for the month.  This selection is
 made internally by the program on the basis of probabilities.  If the
 probability that a drought is over is 100%, then one index is used.  If the
 probability that a wet spell is over is 100%, then another index is used.  If
 the probability is between 0% and 100%, the third index is assigned to the
 PDSI.  The modification (PMDI) incorporates a weighted average of the wet and
 dry index terms, using the probability as the weighting factor.  (Thomas R.
 Heddinghause and Paul Sabol, 1991; "A Review of the Palmer Drought Severity
 Index and Where Do We Go From Here?," Proceedings of the Seventh Conference
 on Applied Climatology, pp. 242-246, American Meteorological Society, Boston,
 MA).  The PMDI and PDSI will have the same value during an established
 drought or wet spell (i.e., when the probability is 100%), but they will have
 different values during transition periods.



               Table 1    Classes for Wet and Dry Periods


Approximate 
Cumulative                                                           
Frequency               Range                                     Range
    %                   PHDI                  Category              Z         

  > 96                > 4.00                Extreme wetness      > 3.50

    90-95               3.00,  3.99         Severe wetness         2.50,  3.49

    73-89               1.50,  2.99         Mild to moderate       1.00,  2.49
                                                    wetness

    28-72              -1.49,  1.49         Near normal           -1.24,  0.99

    11-27              -1.50, -2.99         Mild to moderate      -1.25, -1.99
                                                    drought

     5-10              -3.00, -3.99         Severe drought        -2.00, -2.74

  <  4                <-4.00                Extreme drought      <-2.75





State Code Table. 
                             Range of values of 01-91.

                             01 Alabama                 28 New Jersey
                             02 Arizona                 29 New Mexico
                             03 Arkansas                30 New York
                             04 California              31 North Carolina
                             05 Colorado                32 North Dakota
                             06 Connecticut             33 Ohio
                             07 Delaware                34 Oklahoma
                             08 Florida                 35 Oregon
                             09 Georgia                 36 Pennsylvania
                             10 Idaho                   37 Rhode Island
                             11 Illinois                38 South Carolina
                             12 Indiana                 39 South Dakota
                             13 Iowa                    40 Tennessee
                             14 Kansas                  41 Texas
                             15 Kentucky                42 Utah
                             16 Louisiana               43 Vermont
                             17 Maine                   44 Virginia
                             18 Maryland                45 Washington
                             19 Massachusetts           46 West Virginia
                             20 Michigan                47 Wisconsin
                             21 Minnesota               48 Wyoming
                             22 Mississippi
                             23 Missouri   
                             24 Montana    
                             25 Nebraska   
                             26 Nevada     
                             27 New Hampshire




			    FILE FORMAT


Element          Record
Name             Position    Element Description

STATE-CODE          1-2      STATE-CODE as indicated in State Code Table as
                             described in FILE 1.  Range of values is 01-91.

DIVISION-NUMBER     3-4      DIVISION NUMBER - Assigned by NCDC.  Range of
                             values 01-10.

ELEMENT CODE         5       1 = Precipitation
			     2 = Temperature
			     5 = PDSI
			     6 = PHDI
			     7 = ZNDX
			     8 = PMDI

YEAR                6-9      This is the year of record.  Range is 1895 to
                             current year processed.

JAN-VALUE          10-16     Palmer Drought Index format.
			     Range of values b-20.00 to bb20.00.  Decimal
                             point retains a position in 7-character field. 
                             Missing values in the latest year are indicated
                             by -999.99.

			     Monthly Divisional Temperature format.
                             Range of values b-50.00 to b140.00 degrees
                             Fahrenheit.  Decimals retain a position in the
                             5-character field.  Missing values in the latest
                             year are indicated by b-99.90.

			     Monthly Divisional Precipitation format.
                             Range of values 00.00 to 99.99.  Decimal point
                             retains a position in the 5-character field. 
                             Missing values in the latest year are indicated
                             by bb-9.99.

FEB-VALUE          17-23     

MAR-VALUE          24-30     

APR-VALUE          31-37     

MAY-VALUE          38-44     

JUNE-VALUE         45-51     

JULY-VALUE         52-58     

AUG-VALUE          59-65     

SEPT-VALUE         66-72     

OCT-VALUE          73-79     

NOV-VALUE          80-86     

DEC-VALUE          87-93     

