ABSTRACT -MANY farming activities are dependent on seasonal patterns of rainfall, temperature, streamflow, and other climatological variables. In order to establish levels of risk of these variables to aid in planning farm activities a method was developed for calculating continuous seasonal cycles of probabilities for values of the variables. The probabilities are defined by a mathematical surface based on magnitude of event and month of the year. This surface is a modification of the form-free two-dimensional sliding polynomials. Data consisted of one observation per month for each year of record. These data were standardized using a separate monthly variate transform based on the first three moments for each month. The probability surface was evaluated by least-square smoothing of the twelve months simultaneously. Smoothing was constrained by imposition of boundary conditions for extreme magnitude and for maintenance of full seasonal cyclic continuity. The methodology was tested on long-term monthly data that included total rainfall, total EI, daily maximum streamflow and daily minimum streamflow. The seasonal variation of natural risk for each variable was defined quantitatively and presented in a manner to demonstrate how the information in these seasonal patterns supports many existing agricultural practices. Many of these common practices were developed through years of experimentation; however, with the capability to capture more information from historical records the farmer has the opportunity to improve and refine farming operations with known levels of risk..
Read full abstract