Can we add GDD (Growing Degree Day) to the future software?

10 posts / 0 new
Last post
Can we add GDD (Growing Degree Day) to the future software?

I (Carl) don't really belong here with code writers and engineers - but I think something useful to beekeepers would be the current GDD for their hive location. Hivetool will be recording and graphing ambient temperature, so it is possible to calculate the current total for a year? This would give beekeepers a number to note when particular forage comes into bloom and weight increases in the hive.

In time each area could develop a predictive tool for the onset of flow.

What say you code writers?

The day's low temperature is subtracted from the high, then 10 deg. C is subtracted from the remainder. Negative numbers = 0.
I suggest the year begins at Winter Solstice, Dec. 21.

Paul's picture
GDD Analysis

Attached is graph John did for the Athens, Ga USA hive in 2012. We have pretty good data for that hive for 4 years. We should be able to start predicting flows for certain regions where we do have enough data.

If we could get dates of when different species started blooming from local beekeepers and horticulturalists, that would would give more detail.

(Tmax - Tmin)/2 - Tbase

which calculates the average, is pretty crude. The area under the curve would be better.

Also, Tbase varies with species and that would be another parameter to explore in trying to get the best fit.

Maybe we need to add a GDD column to the the daily summary table?

Using December 21st as the start date messes up slicing the data by year i.e. having the data for 2014 would not allow you to calculate GDD starting on Dec 21 - But we could start on Jan 1.

Note that you need two pieces of information:
1. What is the current GDD value at my hive? (this could come from the local hive computer), and
2. At how many GDDs do the nectar flows in my area start? (this would come from analysis of the historical data in the data warehouse.)

Esa's picture
this would be interesting

this would be interesting idea to play with, and would be valuable info for producers of different flower honey.

Could that base value be made configurable ? seems that here in Finland we are using as base +5C (40F) for calculations, found some values calculated e.g for different apple varieties and wild berries like blue- and lingonberry.

Nate's picture
Great idea, where to implement??

There's a lot of different writeups showing different values - one of the better ones I found is:

is it necessary to use different base temps, it probably adds to the complexity. It kind of depends on where the calcs are porfromed - is it on the fly calcs configured by the user, or would these be stored in the back end DB?

If the GDD is added... there's the capability then to start adding a prediction curve to the trend... based on last years weight vs GDD profile, extend the trend out two weeks to predict the future weight

I guess the GDD shouldn't be

I guess the GDD shouldn't be live. We could do the calculations in the back in batch per hive once a day or so. And then update a static graph. That reduces the load on the server a bit :)

Or are their arguments for live calculations?

I still have GDD on my task list on the Data warehouse a s well.

Thoughts on GDD

I don't think it needs to be "live" or graphed. Maybe just in the table that comes up with each hive? Total automatically updated daily. Maybe a place to insert beekeeper notes "Black Locust bloomed today" - a way to compare bloom dates from year to year in the beekeepers microclimate. I was speaking with a friend who remarked there is a 'chill' calculation as well - for things like fruit set in orchards, but I don't think that would be as relevant to beekeepers.

Paul's picture
Chill calculations

I hadn't thought about it but I am aware that some species require an amount of cold to flower. We need to look into that. It could help explain why some years have good nectar flows and some years poor flows. (Prior year dependencies.)

According to Wikipedia it's the plants that develop next year's buds in the summer. Does that include maple, poplar, gums, or berries?

"Chilling unit in agriculture is a metric of a plant's exposure to chilling temperatures. Chilling temperatures extend from freezing point to, depending on the model, 7 °C (45 °F) or even 16 °C (60 °F).[3] Stone fruit trees and certain other plants of temperate climate develop next year's buds in the summer. In the autumn the buds go dormant, and the switch to proper, healthy dormancy is triggered by a certain minimum exposure to chilling temperatures. Lack of such exposure results in delayed and substandard foliation, flowering and fruiting. One chilling unit, in the simplest models, is equal to one hour's exposure to the chilling temperature; these units are summed up for a whole season. Advanced models assign different weights to different temperature bands."

It could be that Chill Units determine the start date for accumulating GDD.

Paul's picture
Calculating GDD

I suggest two steps:
1. Research GDD on the data center. Determine which parameters (Tbase, start date) best predict the nectar flow each year.
2. After we play (get the database schema correct) and correlate, then incorporate it as a daily forecasting tool.

I'm still thinking about it, but, at least for research, the starting date, Tbase (and Tmax and Tmin) should be set-able by the user. Tmax may vary by species and could be tuned for specific research. I think the best starting date and Tbase will vary by region, Tmax and Tmin by species, but that is just a guess.

That implies being able to calculate it and display it on the fly. Possibly calculate it several different ways and display them for comparison. Correlate nectar flows with different GDD calculations. See which parameters (Tbase, start date) best predict the nectar flow each year.

There are different ways to calculating GDD. This short paper discusses three methods and the short comings of the common method.

1. Average (Simple if just have high/low thermometer, but flawed.)
2. BE (Complicated but better than Average. Best is method 3.)
3. "Electronic Weather Data Collection devices don't need to go through these arithmetic calculations.
Instead, these devices record temperatures every few minutes. These can then be programmed to
simply take the temp reading, subtract the base threshold and accumulate the readings for the day.
Hence, these results will be the most accurate; however, keep in mind that virtually all of the research
equating pest and plant development to degree day accumulations has been done based on the BE
method of accumulation."

Here is longer more technical article discussing inconsistencies of applicaation in research

From both of these articles it appears that much of the research either doesn't specify the method or uses different methods and will be useless to us. Which is OK, since we have better data and a better way to measure the affect of GDD and can build our own tables.

Emil's picture
Weather forecast

I'm not sure how to use this in GDD,but the Norwegian government gives out free forecast in XML format for almost every place around the world. I know many sailing people use this site for forecast when they are at sea. Take a look at
but there is some limitations like minimum 10 minutes between sampling. Maybe this could be included in the data warehouse to find out whats happening in the future, and not just what has been.
Try to translate the pages by google or ask me for info if needed
Here is a sample for Charlotte NC
Here is the list for 16000 places, or you can use


I'm a little lat to the party. The HiveControl version calculates and graphs the air temp based GDD. While Soil is often a better indicator, Air GDD isn't that far off, and after one season of recording the GDD and noting what is blooming will assist with next season predictions.


Log in to post comments