First mowing of 2016

A mild winter and warmer than average March caused the grass to green up early in our yard, but since April has been considerably cooler than average, we were able to delay the first mowing until April 13.  As usual, the back yard was probably too long, while the front yard was barely long enough to reach the mower blades.  Historically, the date this year was on the early side of the median, although almost 3 weeks later than the earliest.  Which got me thinking, "I wonder what determines how fast the grass grows in the spring?"  I don't have good data for rainfall in our yard, but since I do have approximate temperature data, I thought growing degree days might be worth investigating. Growing degree days (GDD) are calculated by the number of degrees Fahrenheit the mean daily temperature exceeds 50°F.  So a day in which the mean temperature is 56°F would produce 6 GDD.

In this chart, the green bars represent the first mowing date for each year, while the red line marks the cumulative GDD up until the day I mowed.  Initially I thought the number of GDD might be constant across the data set, implying that the grass is ready to be cut once it has received the right amount of warm weather.  Since an earlier date would probably require more GDD than a later date, it seemed the date and GDD were probably inversely related, hence the degree day data are inverted on the chart.  Although the correlation isn't horrible, the inconsistency across the data set convinced me something was missing. Maybe rainfall data? Maybe something else?

Does the date data correlate better to Heating Degree Days (HDD)? HDD are similar to GDD except they represent the magnitude of the mean temperature below 65°F.  A warm spring would have fewer HDD and presumably an earlier first mowing, while a cold spring would accumulate more HDD before the first mowing, presumably resulting in a later first mow. As you can see, the correlation is pretty close. The HDD data is limited to the days between February 1 and the mowing day, although when the month of January was also included, the correlation looked very similar. The variation from month to month is still unexplained, for instance, why wasn't mowing required earlier in 2009, 2011, and 2016 as the HDD data would seem to suggest?

Those are questions for another day (or another blogger), and even though I think we can correlate HDD to the first mowing date, I'm at a loss to see a method to predict the first mowing date. I guess that gives me something else to ponder until next spring.