Using Pedotransfer Functions to Predict Soil Properties

In this latest chalk talk video, METER soil scientist and application expert, Leo Rivera, discusses the use of pedotransfer functions (PTFs) for predicting soil properties such as hydraulic conductivity and field capacity.

He explains that while direct measurements are ideal, PTFs can provide rapid, cost-effective alternatives. PTFs use soil texture, particle size distribution, and bulk density as inputs, with accuracy depending on the quality of the database and the input data. There are limitations, such as PTFs not accounting for soil structure and organic matter, which can significantly impact hydraulic conductivity, thus, Rivera recommends using PTFs judiciously. PTFs can be especially helpful for large-scale assessments, and he suggests seeking tools that incorporate more parameters for improved accuracy.

Learn more:

Watch our Soil Moisture Master class→

See METER environmental sensors→

Video transcript

0:00
Hi. My name is Leo Rivera, and this is a METER Chalk Talk.

0:11
Today I want to talk about a topic that I get asked a lot about, and it has to do with how to predict some of these soil properties that we typically measure. So we often have used tools to measure things like soil hydraulic conductivity, retention curves, predicting field capacity and permanent wilting point. And ideally, that’s the best way we can do it, is by making those measurements. But that’s not always an option. So there are times where we need to be able to more rapidly assess soil properties, where we don’t have the budget to assess some of these more expensive soil properties to measure. And there are tools available to make these predictions rather than making the measurements. And I want to talk about those today to make sure that you understand the capabilities of these tools, that they’re out there, but also understand the limitations. And so the primary tool I’m going to talk about is a pedotransfer function, or a PTF.

So typically, with a pedotransfer function, we’re using something like soil texture. So for example, I have here an image of the soil triangle, and I can say I have a clay loam soil. So I’ve got my clay loam right here. And my goal is to predict field capacity. My goal is to predict field capacity for minus 33 kPa, and permanent wilting point, which is minus 1500 kPa, typically to make those measurements, it’s going to take several days to several months, depending on how you choose to make that measurement. But that’s not always an option. So we can use a petal transfer function to take that property, like soil texture, and predict those values. And pedotransfer functions also can be used to predict things like hydraulic conductivity. So you can see an example of a hydraulic conductivity graph here as well. So a pedotransfer function can be a really powerful tool to predict some of these properties that are typically take more time or more expensive to measure, and maybe you don’t have the time to make those measurements. So it’s really important to understand how a pedotransfer function works before utilizing this tool.

So pedotransfer functions utilize databases, whether it’s soil survey or other generated databases, where you have a lot of soil data, and you have all of these data, like texture, density, where they’ve already measured, hydraulic conductivity and some of these other soil properties. It then takes your input and utilizes that database to best predict what those values are that you’re trying to assess. So if I’m trying to predict, for example, field capacity, you can input parameters like soil texture, particle size distribution and bulk density. And you can do this in various orders. You can use soil texture on its own. You can use particle size distribution on its own, or you can combine particle size distribution and soil bulk density together to make these predictions. And it’s going to go into that database and try and make its best prediction based on the data available on that database.

Your pedotransfer function is only going to be as strong as the data that’s in the database, but it’s also only going to be as powerful as how good of an input you give it to predict these these values. So if we’re using soil texture on its own, as you can imagine, if I was predicting a clay as a soil texture, we’ll just use that as our example. If we look at our clay soil on the on the soil texture triangle, that is a huge range of combinations of sand, silt and clay fraction. So as you can imagine, that’s a pretty broad area that you’re trying to predict from, and there’s a higher potential for error in that prediction. Now, if we were to refine that and use the particle size distribution, so if we knew our exact sand, silt and clay fraction that we were trying to predict, we could then refine our predictions. We’re going to get rid of that circle, and we’re going to refine our prediction, saying, our soil has exactly this amount of sand, silt and clay. And that’s going to refine how the pedotransfer function is pulling those data from in the database to predict those values.

But as we know, soil texture on its own and particle size on its own only tells part of the story. So we can further refine that by adding our bulk density into that prediction, which is going to help improve the prediction of either field capacity or hydraulic conductivity. And in some areas, that should be fine. And so as long as you’re happy with that level of error, that’s fine. But, especially when we’re looking at things like hydraulic conductivity, we know there are other factors that play a big role in hydraulic conductivity, such as soil structure. So ideally, we would be including structure in our prediction, and organic matter in our prediction, because we all we know that these play a significant role in how soil transmits water, but most pedotransfer function, tools like Rosetta and Soil View don’t really take these into account.

So when you’re looking at these values, especially trying to assess measurements like hydraulic conductivity, you need to understand these limitations when using these tools. Now, there are other databases and pedotransfer function tools out there that are doing a better job of taking some of these into account. And if you’re going to use those tools, you want to try to make sure, if you’re if you’re really concerned with the accuracy of your values that you’re using pedotransfer function models that take more of these parameters into account. So the more inputs you can have into your prediction, the more accurate you’re likely going to come out with your predictions of these factors.

I just wanted to cover some of the basics on pedotransfer functions, and if you’re going to use them. They’re really powerful tools if we need to use them, especially when we’re trying to characterize large areas. It’s not always feasible to make measurements across these large areas, and they can help give us a little more data to work off of, rather than just the measurements on their own to try and characterize what’s happening across the large watershed, for example. But we need to understand how our inputs can affect the accuracy of those predictions. If you want to learn more about this or other topics, please visit us on our website www.metergroup.com or on our YouTube channel under meter talk talks and thank you for watching.