Python API to get historical data from the NOAA weather station nearest a zip code or latitude and longitude coordinates.

curl, gunzip, Java Runtime Environment / compiler, pyzipcode (if you pass zip codes instead of latitude,longitude)

ftp://ftp.ncdc.noaa.gov/pub/data/noaa/

• 'DIR': wind direction in compass degrees, 990 = variable, reported as '***' when air is calm (spd will then be 000)
• 'SPD': wind speed in miles per hour
• 'GUS': gust in miles per hour
• 'CLG': cloud ceiling--lowest opaque layer with 5/8 or greater coverage, in hundreds of feet, 722 = unlimited
• 'SKC': sky cover -- clr-clear, sct-scattered-1/8 to 4/8, bkn-broken-5/8 to 7/8, ovc-overcast, obs-obscured, pob-partial obscuration
• 'L': medium cloud type
• 'H': high cloud type
• 'VSB': visibility in statute miles to nearest tenth
• 'MW1': manually observed present weather (see ftp://ftp.ncdc.noaa.gov/pub/data/noaa/ish-abbreviated.txt for detail)
• 'MW2':
• 'MW3':
• 'MW4':
• 'AW1': auto-observed present weather (see ftp://ftp.ncdc.noaa.gov/pub/data/noaa/ish-abbreviated.txt for detail)
• 'AW2':
• 'AW3':
• 'AW4':
• 'W': past weather indicator
• 'TEMP': temperature in Farenheit
• 'DEWP': dewpoint in Farenheit
• 'SLP': sea level pressure in millibars to nearest tenth
• 'ALT': altimeter setting in inches to nearest hundredth
• 'STP': station pressure in millibars to nearest tenth
• 'MAX': maximum temperature in Fahrenheit (time period varies)
• 'MIN': minimum temperature in Fahrenheit (time period varies)
• 'PCP01': 1-hour liquid precip report in inches and hundredths -- that is, the precip for the preceding 1 hour period
• 'PCP06': 6-hour liquid precip report in inches and hundredths -- that is, the precip for the preceding 6 hour period
• 'PCP24': 24-hour liquid precip report in inches and hundredths -- that is, the precip for the preceding 24 hour period
• 'PCPXX': liquid precip report in inches and hundredths, for a period other than 1, 6, or 24 hours (usually for 12 hour period for stations outside the u.s., and for 3 hour period for the u.s.) t = trace for any precip field
• 'SD': snow depth in inches

NOAA's raw files have some fixed fields and a richer set of fields with complicated, variable formatting. NOAA provides a reformatting routine which has been modified (static/ishJava.java) to work in a UNIX pipeline within noaahist.py.

Before using this tool, you must compile static/ishJava.java. The Java binary file you create, ishJava.class, must be in the 'static/' directory for noaahist.py to find it, so 'cd' into static/ first.

$ cd static
$ javac ishJava.java

For simple calls, pass command line arguments:

• -d, --date: a single date or a start date and end date in YYYYMMDD format
• -z, --zips: one or more zip codes
• --lats: latitudes
• --lons: longitudes (length and order must agree with --lats args)
• -f, --flds: keys for the data you would like (see NOAA fields below)
• -p: automatically detect number of available processors, N, and run requests in parallel on N-1 processors
• --nprocs: explicitly set how many processors to use (ignored if -p is passed)
• -i, --infile: to run many requests at once, pass in a formatted text file with one request specified per line
• -o, --outfile: redirect comma-separated output lines (defaults to stdout)
• -m, --metadata: feeback on which stations data was pulled from; if outfile is specified, written to outfilename_metadata.txt, else printed to STDOUT

Example:

$ ./noaahist.py -d 19710321 19710323 -z 89109 --lats 34.05 34.893 --lons -118.25 -117.019 -f SPD TEMP -p --outfile fllv.csv -m

This call will create fllv.csv with the requested NOAA weather data, as well as fllv_metadata.txt with information about the stations data was pulled from.

For more complicated calls or for requests with different date ranges or many different locations, pass a pipe-delimited, formatted text file that specifies one request per line.

location name | date OR start_date,end_date | zip or latitude,longitude | comma-separated weather fields

Example:

$ echo 'LasVegas|19710321,19710323|89109|SKC,TEMP' > reqs.txt
$ echo 'WoodyCreek|20050220|39.270833,-106.886111|SPD,SD' >> reqs.txt
$ ./noaahist.py --infile reqs.txt -p -m

This example does use a zipcode, so it will fail if the pyzipcode module is not available. Substitute a latitude and longitude for "89109" to work around that dependency. The location names, 'LasVegas' and 'WoodyCreek', are just for convenient grouping of the returned csv-formatted data. They do not affect what data is pulled from NOAA. There is currently no way to set a location name for calls that don't use the --infile option.

In many cases, NOAA's weather data coverage is sparse, and noaahist.py may not find data for a weather category near a location. To dig further interactively, explore_stations.py contains functions that list the active stations closest to a location in a given year.

>>> from explore_stations import *
>>> stns_with_fld("TEMP", 38.9, -77.0, 2013)

To pull data from a given station and date range, call data_from_station.py from the command line. All the command line args are required.

• -n, --queryname: for convenient grouping of returned data
• -i, --stn_id: the USAF ID of the station from which to pull data
• -f, --flds: one or more field names
• -s, --startdate: date in YYYYMMDD format
• -e, --enddate: date in YYYYMMDD format

$ ./data_from_station.py -n DC_weather -i 724050-13743 -f TEMP SPD -s 20131107 -e 20131110