The freeze/frost probability levels represent the risk with regard to meeting or falling below a certain temperature threshold by a specific date, or within a specified number of days. For example, suppose a 90 percent probability level for the spring season is computed to be March 1 at the 32 degree threshold. This means that nine years out of ten a temperature as cold or colder than 32 degrees is expected to occur later than March 1 during the spring season. For the fall season, the probability level represents the chance of having a temperature as cold or colder earlier than the computed date. The freeze-free probability level indicates the chance of having a longer freeze-free period than the computed number of days. The methods used to compute the freeze probabilities come from two data distributions: a discrete freeze or no-freeze distribution, where a no-freeze annual season is one in which only one or no freeze occurs; and a continuous one of freeze dates for the years of freeze occurrence, where at least two distinct freeze dates occur in a year.
Knowledge of predicted freeze/frost data is extremely important because freezing temperatures can seriously affect commercial and industrial operations, especially those with principal activities in an outside environment. But, the most significant hazards of freeze conditions are in agriculture where late spring freezes, early fall freezes, and short growing seasons can seriously impact plant production and crop volume. The classification of freeze temperatures is usually based on their effect on plants, with the following commonly accepted classification:
Freeze/frost probabilities are presented for several temperature thresholds from 36 °F through 16 °F. Local topography has a most pronounced effect on near ground minimum temperatures, with narrow valleys and enclosed basins experiencing cold air drainage most frequently. Local terrain data can often be used with minimum temperature data and probabilities to give a better assessment of area-wide freeze/frost occurrence.
Freeze/frost probabilities allow the temperature data user to assess the risk of undertaking his operation during a certain time period as opposed to the use of averages or means. This capability is especially useful in agriculture where the risk of crop production associated with planting dates, the length of the growing season, and potential destruction of immature and/or tender vegetation is a determining factor in the farm management strategy. Similar concerns occur in other commercial/industrial operations with regards to scheduling, in order to optimize productivity during favorable temperature conditions. Some examples of temperature-dependent businesses are highway construction, waterworks projects, inland waterway shipping, and large-scale construction. By using the probability data, the user knows the chance of misfortune that accompanies operations at freeze-related temperatures. For example, the Freeze/Frost Occurrence Data tables can be used to determine the chance, by a certain date, of the first frost (36 °F), or the first light freeze (32 °F), or the first moderate freeze (28 °F) in the fall. Analogous data are given for the last occurrence in the spring. The chance of having a period of specified duration (in days) for which the temperature exceeds a specified freeze-related temperature is given in the Freeze-Free Period part of the table. Frost-free periods are those for which the daily minimum temperature exceeds the threshold value of 36 °F.
Reference: