# Growing Season Dates and Length

The growing season is defined as that part of the year when soil temperatures at 50 cm (20 inches) below the soil surface are higher than biologic zero (5 degrees C, 41 degrees F). As this quantitative determination requires in-ground instrumentation which is not usually available, growing season can be estimated by approximating the number of frost free days. The growing season can be approximated as the period of time between the average date of the last killing frost in the spring to the average date of the first killing frost in the fall. This represents a temperature threshold of 28 degrees F or lower at a frequency of 5 years in 10.

The growing season length is determined from daily minimum temperature values. Threshold surface temperatures of 32, 28, and 24 degrees Fahrenheit are generally used to determine the effects of air temperature on plants using the following commonly accepted classification (National Climatic Data Center, 1984b):

- 32 to 29 degrees F is a
**light** **freeze**: Tender plants killed, with little destructive effect on other vegetation.
- 28 to 25 degrees F is a
**moderate freeze**: Widely destructive effect on most vegetation with heavy damage to fruit blossoms, tender and semi-hardy plants.
- 24 degrees F and less is a
**severe freeze**: Heavy damage to most plants. At these temperatures, the ground freezes solid, with the depth of the frozen ground dependent on the duration and severity of the freeze, soil moisture, and soil type.

It should be noted that temperatures near the ground may be significantly lower than temperatures measured at five feet, the normal height that air temperatures are observed. It is not unusual for surface temperature and air temperature to vary by four degrees or more. For this reason, the WETS program allows users to select the three threshold temperatures.

## Growing Season Definitions

All freeze dates are based upon the season August 1 through July 31 for each threshold temperature. Last spring dates of occurrence for a given year are obtained from the period August 1 of the previous year through July 31 of the given year (e.g., spring season for 1971 runs from August 1, 1970, through July 31, 1971, except for the selected starting year, which begins on January 1).

First fall dates of occurrence are obtained from the period August 1 of a given year through July 31 of the following year (e.g., fall season of 1971 runs from August 1, 1971, through July 31, 1972, except for the selected ending year, which ends on December 31).

Therefore, for purposes of calculating the "growing season" with the WETS program, the climatological year begins on August 1 of the previous year and ends on July 31 of the following year.

This season definition follows that of the National Climatic Data Center (1984b). It coincides more closely with previous definitions of the annual march of temperature, in which the warmest time of year occurs near August 1, and the cold season extends beyond December and into the following winter months. This allows for the first "fall" freeze to occur after December 31, which sometimes happens in warmer climates.

The estimation of freeze probabilities was based upon the work of Thom and Shaw (1958) and Thom (1959), which was later modified by Vestal (1970, 1971).

## Growing Season Dates and Length Probabilities

The average growing season length is shown in the WETS Table as the 50% probability value. Associated with this length are the average dates of the beginning and end of the growing season. The 70% value of growing season length represents the upper bound of the NORMAL category; 70% of years will have a growing season less than or equal to this length, and 30% will have a growing season greater than this length. Associated with the 70% probability value of growing season length are the average dates of the beginning and end of a growing season of this length.

Since average growing season length is determined first (in total days), starting and stopping dates must be calculated. The growing season length calculation does not include the ending date in the fall. Since minimum temperatures usually occur in the morning, the effective last day of the growing season would have been the previous day. Therefore the date of the threshold exceedance would not be included in the growing season calculations.

Starting and ending dates are derived by first determining the "average midpoint date" for each growing season for each year in the selected period. The average probability start and end dates are determined by dividing the average growing season length by two, rounding as appropriate, and then adding and subtracting the resulting number to the "average midpoint date." These values are then converted to the calendar dates shown in the WETS Table. Due to the effects of rounding, leap years, and the use of a 366 day Julian calendar, growing season start and end dates shown in the WETS Table may differ by one day from the growing season lengths.

The 70% starting and ending dates are then determined by taking the difference (in days) between the 70% and the 50% probability growing season lengths, adding half the difference to the 50% probability ending date and subtracting half the difference from the 50% probability beginning date.

Since the minimum temperatures used to determine growing season lengths can be modeled using a normal distribution, the assumption of symmetry in both the 50% and 70% growing season length distributions is valid. Therefore, adding and subtracting the difference in days between the 70% and 50% growing season lengths will provide reasonable results. The 70% probability average beginning and ending dates are to be interpreted as the "normal" growing season for wetland determinations.

The growing season dates for specified temperatures and probabilities are shown in the bottom half of the WETS Table in Columns 11, 12, and 13 (see WETS Table Example and Definitions).

## Accommodating Missing Minimum Temperatures when Calculating Growing Season Dates and Length

Previous research (Ashcroft et al., 1992) has shown that representative last and first frost dates can be calculated from time series that contain missing data. Based on this research and National Water and Climate Center sensitivity tests , the WETS program excludes a year from the calculation if a season (spring or fall) has 9 sequential or 18 random missing minimum temperatures. The number of years excluded for each temperature threshold is shown at the top of each WETS table. The WETS program requires a minimum of 20 valid data years to produce a representative WETS table.

## Threshold Temperature Non-Occurance

Certain areas of the country, Florida or Arizona for example, do not experience one or more of the threshold temperatures in some years. The WETS program adjusts for this situation by using a mixed distribution, binomial and normal, to calculate representative probabilities (Vestal, 1970, 1971). The number of years with non-occurrence are shown at the top of the WETS Table.

A growing season length will not be calculated if the probability of non-occurrence is greater than the preselected probability. If, for example, a temperature of 24 degrees or less was not recorded in 16 out of 30 valid years (probability of non- occurrence equal to 53 percent), a 50 percent probability value could not be calculated. This logic applies to all probabilities calculated by the WETS table.