Frost Statistics

Description:

This routine creates a table of the first fall and last spring occurrence of a specified temperature value each year for a series of years. The user specifies the starting and ending years, using a starting year of 0000 to begin with the beginning of the period of record and/or selecting an ending year of 9999 to specify the end of the period of record. The user also specifies the temperature threshold of interest (in degrees F) and type of distribution to use for the summary statistics (either normal or Pearson III).

The output consists of two parts: the header and the table. The first portion of the header section is the CLIMOD standard header. The next portion of the header provides the threshold temperature, spring/fall cutoff date (always August 1), requested years, and type of distribution selected. Some of the data used in the analysis (especially for recent months) might be preliminary.

The table consists of a row of data for each year. For each row, the columns of data are: date of last spring occurrence of a minimum temperature less than or equal to the threshold, what the temperature was on that day, and the number of missing days encountered before finding a temperature that reached or exceeded the threshold. The search for the last spring occurrence begins with July 31 and works backwards looking for a temperature that meets the threshold criteria. Missing days are accumulated. The search ends when a daily minimum temperature is found that meets the criteria (less than or equal to threshold), 18 days with missing temperatures are encountered (result="missing"), or August 1 of the previous year is reached (result="no").

The same information is presented for the first fall occurrence. The methodology is similar, except the search begins with August 1 and works forward no further than July 31 of the following year. "Season length" is the number of days between the spring and fall dates.

Summary statistics are provided at the bottom of the table. The probability of occurrence is the number of years where the threshold is exceeded divided by the number of years of good data (years not listed as "missing"). The earliest and latest occurrences are provided. Also the 90% (9 out of 10 years), 50% (5 out of 10 years) and 10% (1 in 10 years) dates are computed based on fitting either the normal or Pearson III distribution to the data. These probabilities represent the chances of observing a temperature as cold or colder than the threshold later in the spring or earlier in the fall than the indicated date. There must be at least 20 years of non-missing data available for the probabilities to be computed.

The season lengths in the summary section are based on the season length time series in the table, not the values to their left in the summary section. "Earliest" is the shortest season length and "Latest" is the longest. Probabilities represent the chances of a longer than indicated season length.

A brief description of the contents of each column appear at the bottom of the output.


Price

The price for this product is $0.50/summary.

Sample

Frost Statistics

Station: GREAT BARRINGTON 5 SW State: MA ID: 193213
Latitude: 42.15 degrees Longitude: -73.42 degrees Elevation: 817 feet
Station period of record: 12/01/1973 - 12/11/2001

CLIMOD Product: Frost Statistics Creation Time: 04/15/2002 12:13 EDT
Threshold temperature: 32 degrees F Spring/Fall Cutoff Date: August 1
Requested years: 1973 - 2001 Distribution: Normal

     Last Spring Occurrence      First Fall Occurrence          
Year Date Temp Missing    Date Temp Missing     Season Length 
1974 May 21 32 1    Sep 24 22 0    125 days
1975 May 8 31 0    Sep 15 28 1    129 days
1976 May 24 32 2    Sep 23 29 3    121 days
1977 May 16 31 0    Sep 30 32 1    136 days
1978 May 8 30 0    Sep 10 30 0    124 days
1979 Jun 26 31 0    Sep 20 25 0    85 days
1980 Jun 11 29 0    Sep 9 31 0    89 days
1981 May 21 31 2    Sep 21 32 6    122 days
1982 May 6 32 0    Aug 22 32 0    107 days
1983 May 18 27 0    Sep 15 30 0    119 days
1984 May 19 31 0    Sep 7 32 0    110 days
1985 May 9 25 0    Oct 7 31 0    150 days
1986 May 12 32 0    Sep 17 31 0    127 days
1987 May 14 30 0    Sep 26 31 0    134 days
1988 May 4 30 0    Sep 29 30 0    147 days
1989 Apr 29 30 0    Sep 25 30 0    148 days
1990 Apr 22 32 0    Sep 19 32 0    149 days
1991 Apr 26 30 0    Sep 22 32 0    148 days
1992 May 7 28 0    Sep 24 31 0    139 days
1993 Apr 29 30 0    Sep 20 30 0    143 days
1994 May 28 31 0    Oct 3 30 0    127 days
1995 May 9 31 0    Sep 16 32 0    129 days
1996 May 15 30 0    Oct 4 26 0    141 days
1997 May 8 31 0    Sep 22 30 0    136 days
1998 missing --    missing --    --
1999 missing --    Oct 6 32 0    --
2000 May 1 32 0    Sep 29 29 0    150 days
2001 May 8 31 0    Oct 7 32 0    151 days
                   
Prob. occ. 100%       100%      
Earliest Apr 22       Aug 22       85 days
90% Apr 25       Oct 5       107 days
50% May 13       Sep 22       130 days
10% May 31       Sep 9       153 days
Latest Jun 26       Oct 7       151 days
                   


For additional assistance, you can contact the HPRCC.