There are three data options that can be used to calculate probability density functions and cumulative distribution functions for precipitation at a station and to calculate
likelihood of recovering a precipitation deficit or reaching a threshold.
1) Observed data in station record: Use sums of precipitation observations from each period in the stations record equal to the specified future "recovery" period to build
2) Random daily samples: Build 1,000 “synthetic periods” equal to the length of the specified future "recovery" period by taking random daily samples from the station’s data.
Example: if the future period specified were Jan 1-March 31 at a station with a record 1920-present, samples might be sums of precipitation values for days as follows:
Sample 1: Jan 1 1994 + Jan 2 1937 + Jan 3 1976 + ... + Mar 30 1942 + Mar 31 2009
Sample 2: Jan 1 1966 + Jan 2 2001 + Jan 3 2010 + ... + Mar 30 1949 + Mar 31 1951on through sample 1,000.
3) Analog periods: For date ranges in the station’s record that are equal to the date range specified for the observed period (From/To), “analog periods” are defined as
those periods where precipitation total is within +/- 1, 2, or 3 deciles of the current observed period precipitation value as specified by the user. The distribution generated is built from the periods of those particular years/periods that correspond to the future "recovery" period specified.
Define analog periods as those with precipitation totals
decile(s) of the current accumulation period total.
Spaghetti plot: Generates a graph showing precipitation accumulation for observed period attached to all precipitation traces for the selected
future "recovery" period to create a plume of possible outcomes by the end of the future "recovery" period. Based on observed data in station record only.
Probability density function: Generate a probability density function for precipitation totals in a station’s record corresponding to the specified future "recovery" period
and using the data option chosen. Each bar’s height represents the likelihood of precipitation values within the corresponding 1-inch bin specified on the x-axis.
PDF is normalized such that area of bars is equal to 1.
Cumulative distribution function: Generate a CDF for precipitation totals in a station’s record corresponding to the specified future period and using the
data option chosen.