®
Request a quote
Request a quote

Why Your GDD Estimate Isn’t Good Enough—and How to Fix It

undefined

Dr. Colin Campbell discusses what you need to know for more accurate models, so you can be confident in your management decisions.

How to nail your estimates and act at the right time

When you use inaccurate data, the further you are into the growing season, the greater the estimate will differ from reality. For longer season crops, the difference could be quite significant, which is a problem because plant maturity, flowering, and pest/disease GDD targets often have tight windows.

In this 20-minute webinar, Dr. Colin Campbell discusses what you need to know for more accurate models, so you can be confident in your management decisions.

 

      Next steps

      Questions?

      Our scientists have decades of experience helping researchers and growers measure the soil-plant-atmosphere continuum. 

      Transcript

      Brad Newbold  0:09  

      Hello, everyone, and welcome to “Why Your Growing Degree Day (GDD) Estimate Isn't Good Enough and How to Fix It.” Today's presentation will be about 20 minutes and we'll hear from Dr. Colin Campbell, who will discuss what you need to know for more accurate growing degree day models, so you can be confident in your disease and pest management decisions. Colin Campbell has been a research scientist at METER for 19 years following his PhD at Texas A&M University in soil physics, and is currently serving as Vice President of METER Environment. He is also adjunct faculty with the Department of Crop and Soil Sciences at Washington State University, where he co-teaches environmental biophysics, a class he took over from his father Galen, nearly 20 years ago. Colin’s early research focused on field scale measurements of CO2 and water vapor flux, but has shifted toward moisture and heat flow instrumentation for the soil plant atmosphere continuum. So without further ado, I'll hand it over to Colin, to get us started.

      Colin Campbell  1:07  

      Thanks, Brad. It's a pleasure to be with you today and talk about this subject, which I'm particularly interested in, and I'll tell you why. The picture on the left here is of my apple tree in my yard. And I really love this apple tree. It produces some of the sweetest and tastiest Golden Delicious apples that you'll ever find. And, of course, one of my biggest concerns is having those apples affected by any pest damage.

      So, on the right hand side, I'm showing a golden delicious apple, and there's a little hole in that apple. And that's caused, as I understand it, by a codling moth that's particularly interested in investing in apples and pears. And so that's the little hole that larva makes. And there's a pupa coming out of the apple in the middle there that it's eaten. And once it's done that, it goes over and produces its cocoon on the trunk of the tree. Now, every year, I'm really excited to eat the apples off my tree. And so I pay someone to come by and actually spray the tree to make sure that pests like the codling moth stay away. But it's interesting, I can't remember a time where every single apple in my apple tree was not infested with codling moths. Every apple had at least one hole. And I can't tell you how discouraging that is. I was so interested in eating these apples that I started picking a few and just eating off the side that didn't have the hole in it, but I don't really enjoy that so much. So why, when I'm paying a person to come by and spray the tree and kill it–specifically the codling moth so I don't have holes in the apples–why is my entire crop of apples being destroyed by the codling moth?

      Well, it comes down to one very important thing and I'm going to talk about that in just a moment. First, I want to talk about why we would use growing degree days or GDD. So, here's the concept, and it's really simple: plants and insects–insects being cold blooded– develop on a clock that includes not only time, but both time and temperature.

      This combination is called thermal time. A very imaginative name, I suppose, but it really points out that we're talking about time that adds together the temperature component, and it has the units of degree days or some people really get kind of adamant that they call them day degrees, but we're going to use degree days in our discussion here. So, the relationship between specific stages of development for these plants and these insects. This has been studied intensely for many species, many plants and many insects. And we know this pretty well as it turns out.

      So knowing the degree days that is required for a specific organism to move through a specific stage is a powerful yet simple management tool where we can control things by understanding when to, let's say, apply nitrogen to a plant, or when to apply a pesticide to an insect to manage our system optimally. So, before we launch into some of our discussion, we need to talk about what are some of the fundamentals of this thing called thermal time or growing degree days. So as I mentioned, all growth happens on a time and temperature clock, but it has to happen above this line here at the bottom which we call Tbase or the bottom threshold temperature.

      Now, anytime the temperature is above this baseline that will allow the organism we're interested in to grow. And it grows more as the temperature is further above this line. And so there are a couple of ways that we can actually add together these temperatures. One is what we call an averaging method. And it just takes this maximum or in some cases, the upper threshold, depending on what model we're using, and this base here, and the minimum, and it works out this kind of orangey, pinkish area here. And that's basically the one day total of degree days that we can add to this insect or plant's development.

      There's another methodology that we can use, it's actually kind of here in blue, and it's grayed out here. This is called the Baskerville-Emin (BE) method, it just fits a sine function to your data to your maximum and minimum, and then integrates under a curve. Now, these techniques were developed in a time where you typically didn't have as high quality temperature measurements as we do now. Now I actually make measurements, maybe every 15 minutes out in the field, we certainly can do way better than that. But you could also integrate under that one if you wanted to. And, of course, that would give you more accuracy in your degree day. But the point I'm trying to make is that these two methods are out there, the BE method (the Baskerville-Emin) is a little bit better, a little bit more accurate, but their outputs are fairly similar. And I've done these calculations in several different datasets.

      Well, in any presentation, we have to have the obligatory equation. And I showed this slide to a friend who wasn't familiar with this. And he said, “oh my gosh, now I'm checking out of your presentation.” Please don't check out. This equation simply says what I just told you about in that picture on the last slide, but if you're going to do this mathematically, maybe you're interested in just seeing how this works. So, this is thermal time, 𝜏n and it's the daily summation, so i is just one day. So each day, we sum the average temperature, or Tmax plus Tmin divided by 2, that's average temperature, and subtract from that the base temperature, and then we have that value, and we multiply it by Δt which is time. So this is 1, for one day. So, we have the average temperature minus the base temperature times 1. And then we just keep adding those up every day.

      Now, I actually put a little spreadsheet on this slide to show you what it looked like day after day. But it got a little complicated, so I left that off. But you can kind of see how this goes, I'm happy to share some data with you. So if you want to try this on your own, I got a little spreadsheet you can work on, just let me know. So, growing degree days are really simple to calculate. As I mentioned, maximum and minimum temperature wherever we're interested in. Some of these daily warming units, the 𝜏n that the degree days, as I mentioned, the difference between the average and the base temperature. And when you accumulate enough thermal time, that organism is going to advance from one stage to another.

      Now you might ask, why do we know this? Well, we know this because many researchers have gone and studied these things in the field, whatever they're interested in. Over time, they’ve done a significant number of studies. They're all out there in literature, and you can go find them. And while I was preparing for this talk, I had to go do this on my own. So, I was interested in codling moths as you saw earlier, so that I could tell my spray person, “Hey, you come in right now and spray my moths so I get some apples to eat that I like.”

      So, I studied codling moth a little carefully, as I mentioned, in apples and pears. Now, how do you know when to start adding up degree days? Well, that starts with what we call a bio fix. So, you put in your tree a little trap that has a pheromone in it for the moth of interest. And once you trap moths on these sticky traps two successive nights, that is time zero so that starts this i = 1, that's day one. And so after that, we're just adding additional degree days, day after day. This lower threshold for the codling moth is 10°C, upper threshold is 29.4°C. Now, if you don't like Celsius, maybe you're a Fahrenheit person. That's just fine. All of these are done in both of them.

      There are a couple of great resources to look at this. The University of California has a lot of great information on this in their agriculture area; just Google that. Or Washington State University, here where I live in the state of Washington, also has that. So, you can get a whole bunch of these things; of course not just for codling moth. That was my interest: how to eat good apples off my tree. So here's a stage–eggs, larvae, pupae, and adult–and here are the degree days, in units of “°Cd” for it to go through these various stages.

      One thing that I didn't really think about a lot about my own apple tree is what the person who comes and sprays actually is using to kill these codling moths, because different sprays have different impacts. So, one spray might work on the larva stage, and another may work on the pupae,  and something else may just be for the eggs. And so if you're not spraying at the time that it's hitting this particular stage, your efforts are meaningless. And quite frankly, that's exactly what's happening in my tree at my house, that I needed to call up my friend and just say, “Hey, listen, I can measure your growing degree days and I'll let you know when it's time to come to my house.” I haven't done that yet, but I plan to. So, this is not just for insects. This is for a lot of different things.

      So, degree days are a management tool. And probably many of you have had experience in using these in various areas. For example, I really liked this illustration I found on the internet. This is from a published article called irrigated wheat by Rawson and MacPherson, 2000. And this just shows some of the growing degree days needed for different stages of wheat. And the reason I really liked this is because when I was a young budding biologist, as a sophomore in high school, I needed a project for biology. And I said, because my dad was an environmental biophysicist, I always assumed that he should do most of the work in my science classes. Of course, he assumed the opposite. And he gave me a little project where I had a weather station out in this field, full of wheat, and I used growing degree days to predict when the different leaf stages of the wheat were.

      So, we've actually got a little wheat plant growing. So first leaf, second leaf, third leaf, fourth leaf here. I predicted when this fourth leaf would actually emerge based on growing degree days. And you know what, it worked really well. And as a sophomore in high school, I was completely shocked and like, “Things operate on a clock?”, and they do. I've never forgotten that experience, you can see a lot more degree days go into actually growing a wheat plant, as it shows there. And we're not going to spend time going down deep into this, but if you need to make an operation, for example, I studied rice during my PhD. And it turned out that applying nitrogen treatment right here in the rice–rice has a similar set of stages, not exactly the same but but similar to what you see here–applying nitrogen before heading really did make a difference in the biomass accumulation of the rice, but they actually applied a nitrogen treatment out here after heading. And it turned out that it was a complete waste of money. And we talked to the grower about that.

      And that all can be predicted and let you know when you're going to have to do that based on this nice little degree day calendar there. So, there are important management steps that need to be taken at specific phenological stages, no matter if we're talking about plants or insects. And our degree day modeling that we've talked about so far allows us to know when to act. But timing is everything.

      As we talked about in that first example, if we just go out there and spray for codling moth when we have time, you know, it's on the list and drive by my house and just spray when the list comes up, we don't know if we're going to have that maximum effect that we need. In fact, often we don't. So, accurate data feeding into these models provides a stable platform from which to make decisions.

      So, let's talk about some ways that we can gather this weather data, specifically temperature data, that's actually occurring right now, and whether or not that's going to meet the needs to create this stable platform of action. 

      There are a few different sources of weather data for growing degree models. There's virtual data, we often talk about gridded data for models. So, they take several weather stations that might be in a large area, and they basically grid the earth in between and and use statistics to try to figure out what the temperature is at a specific location. There are other ways to do this, like regional weather. Something that's nearby our field of interest, for example, or we can even put things in the field, and we're going to talk about this at the end, but all of those are not created equally. And we're going to talk about the difference between a more accurate system, like an aspirated–meaning flowing air over a sensor or a corrected version of that–or a non aspirated–something that's just sitting passively in the sun and what errors can come from that.

      Estimates of thermal time are only as good as the data they're fed. And so, one of the things that I loved recently was this publication by the Ag Weather Net, and I can send you a link, here's the link, if you want to copy it down, you can also just search for this analysis and you'll find it out on the web. So, these are DarkSky, this is one of those virtual weather sources.

      These are the growing degree days from DarkSky minus the actual observed degree days at the Ag Weather Station at Sakuma station. And this starts on January one. So this is an accumulation over time, and this is the era. So, for a while, not much error going on, but for these different years, you can see that, for example, the last year they did this, in 2019, we just have massive differences between the DarkSky estimate and the observed growing degree days. Other years were a little less, for example, there in 2013, there was no problem. But this graph clearly shows that sometimes we're way off the beaten path. So, how well could we do with predicting this weather in natural systems out there.

      As I mentioned, many are using this virtual weather to feed into these disease and pest models. This study was done, as I mentioned, by AgWeatherNet by Dr. Dave Brown and a few colleagues. And what they show is a significant error in their estimate. So let's take a specific case: we have several station-season combinations. We have a few stations, we're comparing this to, not just that Sakuma station, but others. And then we go several seasons, so we're not just one year but multiple years. And we combined these statistically, and said okay, with dark sky versus the actual station reading, if the actual degree days were 375 growing degree days, how would the dark sky estimate differ from the reality on the ground? So, zero would be no degree day difference, and we can see that many were fairly good at predicting, or at 0 days.

      But as you see, as we get further from 0 out toward 10, or -10, what we do see is that there are still predictions that go all the way out to 10 day errors. So, if you were adding up your degree days, and you needed to do an operation, if we use virtual weather, we might say at least one of these sites in one year, maybe that's over here, we would be 10 days off in our operation. And so their point was that 33% of the station season-combinations showed more than five calendar day error, so we'd be off by five days. And I don't know exactly, because I haven't been taking count, but I assume that's exactly what is happening in my house when we try to spray our apple tree.

      So the take home: virtual weather alone cannot replace an actual infield weather station. So, that's a direct quote from that article by Dr. Brown. So, what about regional stations? There's a nice network of regional stations around in agricultural areas here in Washington: it's AgWeatherNet. And one of the efforts they've been working on lately is actually to add more stations to that to get better data. But let's say we couldn't do that. And we chose to just use our regional station, instead of putting one directly out in our orchard or our field. Could one of these regional stations provide accurate enough data for our field?

      Here on the right, I'm comparing a regional station to an infield station for an irrigated field. So, here's air temperature and this is time. And we're just talking about a month and a half worth of data here. I selected this because it's toward the summertime, we got heavy growth going on, and I wanted to say, “hey, what's the difference here?” And so the darker colors: this red, and this black here, they represent this regional weather station and the orange the lighter color, the orange here and the light blue here, this was an infield station. And so we see when it's cooler. Do that badly but when it starts to warm up, not surprisingly, we see a consistent overestimating bias by the regional weather station compared to the infield station that probably wouldn't surprise surprise us based on our knowledge of evapotranspiration in the field and that generally cooling things.

      So, what does that mean in terms our estimates for degree days? So, I just summed the degree days across the six week period. And I actually did it for wheat, because it's one I know well. And what we found was the regional weather station would have estimated about 545 degree days while they infield weather station is about 60 degrees days less at 486. So, we would have actually overestimated the development of our wheat in this field.

      Now, that's not all we can do. I love and have used a lot the infield system, but often the question comes to me, “Hey, aren't all temperature measurements that we make equal? What's the difference? Why should I buy, for example, an ATMOS 41 and make a measurement there versus just this nice little passive shield?” Well, they're not all created equal.

      And so again, we're looking at temperature, we're looking over the same time period, so we have the ATMOS 41 I showed in the last slide compared to regional now we're comparing the same thing against a passive shield. So the ATMOS 41 actually corrects the air temperature based on its radiation energy balance and the wind speed. The passive shield just hides the temperature sensor below a plastic, a white plastic plate or at louvered plates. And again, we see the systematic overestimate by the passive shield in blue, compared to the orange. Now you might ask me, “Hey, wait a second, how do you know it's an overestimate?”

      So, we've made many comparisons of the ATMOS 41 to an aspirated temperature sensor, and we know and we have data to show that this is true. The passive shield typically overestimates by about two plus degrees Celsius while the ATMOS 41 is generally at about half a degree error or less. So, we can see and I didn't calculate the degree day error, they're closer, as you can see from that, that, that slide here, or the graph here, they are closer, they're just not the same. There's a systematic over estimate, by the lower cost, non-aspirated temperature system.

      So what's the take home from all of this discussion? Well, here it is: plants and insects move through their phenological stages according to thermal time. And I have this beautiful picture drawn on the right hand side of my codling moth example going from eggs on the bottom right here, those eggs produced pupae that dig into the apple they eat the apple seed they come out as larvae they go down and can overwinter on the bark or they actually do a second second life cycle during the summer so they can be during the summer on the bark and then they emerged as adults and then they go back on the fruit lay their eggs on the leaves and we get this life cycle going.

      So, if we can understand that cycle that we have there, we can apply management practices to make sure we don't waste both costly and environmentally challenging operations on these codling moth for example. And we can also get the apples for example that we want. So critical timely action can only be taken on crops and pests using accurate degree days, and degree days that we get from virtual and regional weather information can be off as I mentioned, by multiple days making any action we do potentially wasted. Local temperature data can provide good degree day estimates, but the best accuracy comes from aspirated or corrected air temperature. And that's all I have for you today.

      Brad Newbold  24:14  

      All right. Thanks, Colin. That's going to wrap it up for us today. Thanks again for joining us. We hope you enjoyed this discussion. And for more information on what you've seen today, please visit us at metergroup.com, and stay tuned for future METER webinars. Thanks again, stay safe, and have a great day.


       

      Presenter

      Dr. Colin Campbell has been a research scientist at METER for 20 years following his Ph.D. at Texas A&M University in Soil Physics. He is currently serving as Vice President of METER Environment. He is also adjunct faculty with the Dept. of Crop and Soil Sciences at Washington State University where he co-teaches Environmental Biophysics, a class he took over from his father, Gaylon, nearly 20 years ago. Dr. Campbell’s early research focused on field-scale measurements of CO2 and water vapor flux but has shifted toward moisture and heat flow instrumentation for the soil-plant-atmosphere continuum.

       

      Case studies, webinars, and articles you’ll love

      Receive the latest content on a regular basis.