Last week’s post on false spring had me thinking about when I should actually start planting things in the ground. I was always told “Mother’s Day” is the safest time to plant. Climate data is free, so let’s figure out the earliest planting date in Dayton, Ohio. Feel free to use the NOAA data to repurpose this repo to find the best frost date in your area!
Methodology
- Import Climate Data into a DF, from our friends at NOAA.
- Create a Monthly/Weekly Breakdown of Frost Events (sub-32-degree temperatures).
- Create a Last Frost Data-Table, and a histogram to show the probability of frost occurrence after a certain date.
Results
Monthly/Weekly Breakdowns of Frosty Events
Here is a monthly and weekly breakdown of frost events. While useful to identify certain controls, I adjusted it via the mean days in both the weeks and months with frost events, to control for weird week 53 nonsense and normalize some of the data.
Another way to look at this, is that on average, more than ~70% of the days in Month 1 (January) have temperature minimums of less than 32 degrees, versus ~10% of days in Month 4 (April) have frost events.
Yet, if you think about it, ONE frost is enough to kill my plants, so what I’m really concerned with isn’t the prevalence of frost events; but what the last frost event typically is. So I created a new data-table ordered by the last frost event per-year, and boom chicka boom, here you go.
I also included a probability line, for easy understanding of the distribution – essentially what is the likelihood of a frost occurring after this this. Looks like it dips below 20% at week 17.
Guess what week it is, right now. That’s right. Week 17.
See you in the garden.
Stay Curious!
Data Sources
Climate Data Online (CDO) – The National Climatic Data Center’s (NCDC) Climate Data Online (CDO) provides free access to NCDC’s archive of historical weather and climate data in addition to station history information. | National Climatic Data Center (NCDC). n.d. Retrieved April 16, 2026. https://www.ncei.noaa.gov/cdo-web/.
Leave a Reply