Precipitation Probabilities in Kansas

We here present precipitation probability across Kansas over 29 long-term surface climate stations. We started from daily observations from 1931 to 2014 inclusively for all selected stations, which daily data were quality-controlled and visually examined. We then transferred daily data into the weekly data. We compared the relative frequency computation and distribution-based computation for obtaining our weekly probabilities for amount of precipitation larger than 0.2”, 0.6”, 1.0” and 2.0”. To reduce some random fluctuations, a three-week moving average method was then used for smoothing this weekly probability in each station. The smoothing process implemented was necessary because temporal variations on a weekly basis time series suffered by any lower quality observations even when the daily observations used were quality-controlled.

Using our statistical analysis for a sufficient long-term period, it is possible to estimate the chance of receiving certain amount of precipitation for a given time period. It should be noted that the method we used is an empirical estimate without using any atmospheric process-driven modeling. Therefore, the results presented here are different from weather forecasting information. Each year the probability of success remains the same regardless of the previous year. For example, the information we presented would help determine whether installation of an irrigation system would be cost effective but it wouldn’t help determine when to start irrigating (an extension publication ‘Chances of Precipitation in Kansas’ by Dean Bark, KSRE Bulletin 461. May 1963).

Xiaomao Lin, Ignacio Ciampitti, Qing Ye, and Mary Knapp in 2016