Link to Data Chart (ET vs. Air temp)

https://sprinklernerd.com/evapotranspiration-vs-air-temperature

-----------------------------------------------

BAM! We're on fire today, guys. This is the end of ET. End of it. End of ET. Do I have your attention? Think about it. This is the end. The end. This is the end. The end. End, my friends. All right. Hopefully I do have your attention. Because... I just want to make you think. You guys know me by now. I love thinking of things upside down, sideways, backwards.

Just think of something in a way you've never thought of it before, and when I say this is the end of E. T., what does that mean to you? I'm just gonna sit here for a minute and let you think about that. When I say the end of E. T., what does that mean to you?

Hmm. Does it mean we're never going to use ET again? Possibly. However, ET is a uh, known good formula. It's a very detailed calculation that probably took whoever came up with it, Edmund Monti, lots and lots of research to figure out. So, I don't, I personally don't think it's the end of EET in terms of we're never going to use it again.

I guess what I wanted to, what I want you to think about is that is it the right tool to determine how much water we should apply to the landscape right now? For that purpose, I think this is the end of et not today 'cause this is still how we're watering. But if we fast forwarded 20 years, will we be using ET to determine how much water to apply to the landscape?

Right now? At this moment I want you to go out to your landscape, to your project and I want you to water turn on the sprinklers. How long? Should you water the landscape right now? How long should you water it? Will you, will you look at yesterday's wind speed, yesterday's temperature, yesterday's solar radiation, yesterday's humidity?

Will you look at yesterday's Cite environmental data to determine how long to water today. Right now, I don't think you will , so when I say it's the end of et, we are in the transition zone. We are in the transition zone of taking a mechanical. system, a turn on now for 15 minutes every Monday, Wednesday, Friday at 6 a.

m. We're in the middle of that transition to a automated system that can apply the right amount of water. Right now, in real time. And the only way to do that is to put a sensor in the soil. So this is a long winded way of saying my prediction is that this is the end of ET as a real time automation tool.

ET will become the Will become, and actually is right now, the only tool and the best tool to predict water usage, to run calculations, to forecast, to run, uh, forecasting models. It's the only tool for that, but it's not the right tool to automate the irrigation system right now, here, today. So I'm going to start talking more about soil moisture sensors, how they work.

You know, the ups, the downs, the, the, the good, bad, and the ugly. But I kind of wanted to just frame this episode around the end of ET number one to get your attention. But number two, so that you can understand where is it, where is et the right tool for the job? And in my opinion, the right tool for the job is for forecasting, calculating estimating, but it's not designed for real-time watering.

It's not designed to. Be the tool to water until the ground is at field capacity and stop watering. If I told you to go outside right now and only put down enough water so that the soil reaches field capacity and turn it and it turns off, would you really use yesterday's weather data to run a calculation for that?

All right, I'm in the weeds a little bit, but that's the that's my topic for today. And I ran. An experiment that really got me excited. It has to do with ChatGPT, and I asked ChatGPT a really cool question that had some amazing results, and what's kind of fun about ChatGPT is that it doesn't have a It doesn't have a, you know, a horse in the race.

It doesn't care who's right or who's wrong. It's very, uh, sort of factual. You know, so it doesn't have a lens. It's not seeing our industry through, you know, the eyes of one manufacturer or that manufacturer or who's right or who's wrong. You know, it's a very... I'm trying to think of the right word. Factual.

So I asked ChatGPT, instead of using an actual... air temperature sensor like a thermometer to control a thermostat, would it be possible to calculate the air temperature instead of using a thermometer? Okay? And this is a fantastic analogy because it directly correlates to to et. 'cause if ET is the value that we use to water more or water less, that's like saying the temperature goes up or down.

The only difference is the temperature is something that you can, you can measure, right? You can measure the temperature with a thermometer. , just like you can measure the moisture in the ground. With a sole moisture sensor. Yet, we are not doing it. So when I asked ChatGPT, instead of using an actual air temperature thermometer to control a thermostat, would it be possible to calculate the air temperature instead?

Do you want to know what, what the computer said? What, uh, what Mr. ChatGPT said? He said, yes, it is possible to estimate air temperature using alternative methods instead of directly measuring it with a thermometer. However, this is great. However, the accuracy and feasibility of these methods vary, and they might not always be suitable for controlling a thermostat.

Here are a few methods that can be considered. And there's six of them. And when, when I read these, I want you to think about ET, evapotranspiration. I want you to think about the fact that we are calculating, calculating water use. But we are not measuring it. We are not measuring it. We're using third party factors, wind speed, temperature, humidity, solar radiation, air temperature, to calculate the water use.

Okay. So there are. A few methods we can consider for calculating the air temperature to control the thermostat to turn the heat on or off instead of measuring it. Here's the first method. Based on other environmental data, boom! We could control a thermostat based on other environmental data. Does that sound familiar?

Can we control an irrigation controller based on other, other environmental data instead of just how much moisture is in the ground? You could use something called a radiation budget. Using instruments to measure incoming and outgoing radiation can allow one to estimate the temperature. This method is often used in remote sensing applications and meteorology.

You could use heat balance. If you can measure or estimate all the other heat fluxes in an environment like radiation, conduction, convection, latent heat due to evaporation, condensation, you can estimate the temperature Based on the heat balance equation. The frickin heat balance equation. This is like, this is so awesome.

It is exactly like E. T. The evapotranspiration equation. The similarities are like, it's like brother sister. We don't do this with our thermostats, but we do it with our landscape. Alright, the heat balance equation. I'm gonna make that bold again. I'm kind of just reading my notes here. That's a good one.

That's a writer downer. Heat balance equation is like the evapotranspiration equation. Okay, method number two, using electronic components. A resistor. A resistor's resistance changes slightly with temperature. While this isn't typically as accurate as a dedicated temperature sensor, it can give a rough estimate.

Okay? You could also use a diode or transistor. The voltage drops across a diode, or the base emitter junction of a transistor changes with temperature. By measuring this voltage drop, you can estimate the temperature. And that, that correlates a little bit more similar to using a soil moisture sensor because we're using kind of third party measurements to estimate the soil moisture, okay?

Now um, I think, I think I'm going to skip over these other three, you know, you can ask ChatGPT this, this same thing or ask Google BARD, whatever, but the last one is called data modeling. If you have a strong understanding of all the factors influencing temperature in a given setting, like in a building, you can create a model that predicts temperature based on inputs like outside temperature, time of day, insulation.

Machine learning models can also be trained on historical data to predict temperatures. Data modeling, okay? That's what we're doing with our landscape systems. We are not measuring real time moisture in the soil. We are using what might be considered in the heat industry, the heat balance equation and data modeling, but we are not measuring actual moisture for the most part, right?

95%. So, yeah, just think about that. Think about that and think about where we can, where this is going to go. Okay. All, all of this money, this effort placed on, on data modeling, evapotranspiration to adjust runtimes in a controller. This is the end, not here, not today, not 2023, but the prediction is here because if we look at another industry, look at the past and we look at other technology that's out there to control something like our heating and cooling, we are not using data modeling for that.

We are using a real sensor. So in this industry, in our industry, in the future, and now, the best way to water is going to be with a soil moisture sensor. And the only way for this to be the, uh, the, for this to, um, hmm, what's the word? I'm losing my words here. The only way for soil moisture sensors to be everywhere, for every system to use them, they have to be easy to use, easy to install, easy to understand, And they have to be like, frickin cheap.

Okay? Because we can't take a 600 soil moisture sensor and put it on every home in the world. You can't take, you know, 20 600 sensors and put them on one commercial site. You, frankly, you can't, it makes no sense even a sensor that's 250 is too expensive. So in the future, maybe not only does every zone have a sensor, maybe every zone has 10 sensors.

We have to get to a place where the technology is, uh, is for the everyday person and the price is for the everyday person. So that's my prediction. This is the end of E. T. As used to automate irrigation systems. I think we're going to, we're going to pivot this more effort will probably be placed into using ET as a predictive model so that you can go, uh, and run your own water use calculations to understand how much water is being used on this site and how much water should be used on this site based on historical ET.

So you can decide if this site is over watering or under watering and run your models, run your, run your prediction models. Run your ROI models to see if there's an opportunity to enhance the performance of the system. But we're going to get away from using ET to make real time watering decisions, okay?

So I think probably the next episode we'll start talking a little bit more about this and maybe this will become kind of like the, uh, The figurehead episode for my, um, personal opinions that this is a new, you know, I just kind of decided to, to share it and using the thermostat analogy has always been a good one.

Uh, but I never took the time to really dig in or ask, uh, you know, use the tool like chat GPT to tell me. Imagine if, imagine if we didn't have a, we didn't have a, an actual air temperature sensor in the building. How else could we control the thermostat? I can, it came back with, use, you can estimate the temperature based on the heat balance equation.

And if you have a strong understanding of all the factors influencing temperature in a given setting, like in a building, you can create a model that predicts temperature based on inputs, like outside temperature. Wait a minute, you could use outside temperature to predict the inside temperature? Well sure, you just have to know how big is the room, how many people are in there, what's the type of insulation, what directions do the windows face.

We could give it all these inputs, run a data model, and probably get within 10 percent of actual air temperature in the room. Or we could just put a damn air temperature sensor in the room. So, all right, whoo, I got to calm down, cool down a little bit. Just getting excited. This is a lot, a lot of thoughts all coming together with a, um, I don't know, almost like a, a new vision.

So just want to challenge you guys to unthink, relearn, challenge yourself to question what it is you think, you know. Every day can be a new day. Ask yourself, is this the right way? Am I doing this the right way? Do we, are we all thinking the right way? And I think, uh, I'll probably put Maybe I'll just link it in the show notes.

I, uh, I have my own set of, uh, sensor tools and suites and software that, that I use just for kind of, uh, Um, learning about this kind of stuff and I run, um, I run the ET equation for my house using a local weather station. So I have, uh, I record the daily ET every day and I built a model that shows the average air temperature daily.

Overlaid on the daily et and what's really fascinating is that the air, the daily average air temperature when overlaid with et um, the, the similarities are striking. And so in other words, when ET goes up, temperature goes up, when ET goes down, temperature goes down. And it's almost as if. You could use air temperature to adjust your controller.

So let's say we didn't run these ET models and you only had air temperature. You could adjust your run times based on air temperature and it would likely be better than just running it On a time schedule. So I'll take this graph, I'll try to find a way I can, I can put it online and then allow you to click the link so that you can view it.

Uh, just cause it's, it's fascinating. You know, it's almost, uh, makes me think if have we, have we confused the hell out of everybody in this industry by trying to explain evapotranspiration 101 ways, do you need one on site? Do you need one in your local town? There's like, there's all these questions when really, what if there was just a.

Air temperature sensor and we adjusted the runtime, you know, up or down just based on, uh, based on the air temperature because the correlation is striking. And I think if somebody were to ask me, you can't use a weather station, Andy, but you can pick one data point to model from. I would choose air temperature just just like chat.

GP set chat. GPT said that we could use air temperature outside the building. to run and predict the temperature inside the building. Temperature is a key factor. And I also noticed that my house, when the temperature is cooler, you know, specifically because I'm in the Northern climate, we have cool season grass.

So my grass is just healthier. It is, and what I mean healthier, it's less stressed when the air temperature is cool. So when the air temperature is cool, it's almost like. It looks like I've just watered and fertilized, but I did nothing. The air temperature just came down. The grass got happier, healthier, less stressed.

I really don't need to water because the grass is happy just based on air temperature. So it's kind of like, all right, once the daily, once the daily air temperature, let's say for cool season grass, and I don't know what these should be. So I'm kind of just spit balling here, but what if the daily air temperature was below 75 degrees, you know.

Do you need to water or do you need to water like half as much? What matters more, the daily ET or the average air temperature and the type of plant material you're trying to grow? Just, just some interesting questions when you start looking, when you start looking at the data and we have, uh,

Yeah, we have this bias that, you know, E. T. is like a one size fits all. We try to adjust it, uh, based on root depth and plant type and all these factors. I don't know. It's just I'm tired. It's very tiring to know there are other ways to automate the system, yet we continue to push the same agenda. Not that it's an agenda.

It sometimes seems that way. Um, so anyway, there you have it, guys. Thank you so much for listening to my little, uh, So stood in my soapbox for a bit and I'm trying to think if there's anything else that was on my mind related to this. I'm sure that there is, um, you know, it makes me think, uh, of, uh, a quote and I'll actually, maybe I'll play it.

There was a quote from the, um, the episode that I recorded with the CEO of Irrigreen and he made this correlation to, you know, with their developing as similar to, uh, a carburetor. So we don't have carburetors in our car anymore because we have electronic fuel injection. But at one point, there was no electronic fuel injection.

So you kind of had this mechanical You know, carburetor device that control the air intake and it's kind of what's happening here in, in our industry with, with ET. We think that using ET to automate might be like having electronic fuel injection, but, but it's not because it's not measuring anything.

It's, it's, it'd be like having a electronic fuel injector that just calculated everything instead of measuring it. So maybe I'll play that quote, mix it in here. And, uh, Transcribed Yeah. Thanks so much for listening. Please, you know, hit the subscribe button, uh, or to follow. And if, uh, this episode resonated with you and you have a hard time sometimes explaining ET or, or you agree, or, or even if you disagree and you think I'm just full of shit, share this episode with someone, have them listen to it and see what they think.

Love to get this discussion going more so we can just continue to elevate the industry and find the right tools for the job. So hope you guys have a great weekend. Thanks for, thanks for tuning in as always and, uh, catch on the next episode. Bye bye.