"Don't try to jump on his bandwagon"

Jon Scott wrote to me about my recent post on a poor way to fertilise.

"While this superintendent has solved his problem of nitrogen input by monitoring salinity level that has worked for him, this is probably a very unique situation. It may be relevant to other golf courses where similar salt levels exist, but there are too many variables to draw general conclusions. Thus, I would focus on salt levels as related to this situation and not extrapolate. What he has said may be relevant to similar situations, but it all depends on the salt levels."

I agree, and I meant to make that clear in the original post. Let me try now to explain in clear terms.


If there is a salinity problem at a site, then one will always want to minimize the salinity in the soil. If one is always trying to minimize salinity in the soil, then it is impossible to use any measure of salinity as a criterion for fertilizer application.

In a case where there is not a salinity problem at a site, it might sound reasonable to try to use salinity as an index of nutrient content in the soil. However, there are three big problems with this, and these I did describe in the original post. First, most turf managers don't want fluctuating nutrient supply; second, salinity says nothing about which nutrients are there; and third, the salinity measurements from soil moisture meters, whether EC or a salinity index, are so affected by the water content of the soil that using the salinity of non-saline soils to make decisions about fertilizer is like chasing a target that moves randomly.

I like using soil moisture meters to measure the water content in the soil. I think it is useful to assess the salinity of the soil with the meter too, if that function is available. But I don't think it is a good idea to make fertilizer decisions based on soil salinity.

I replied to Jon that "I think it is ridiculous but tried to be as polite as possible."

He wrote back:

"You, trying to be polite? Don’t lose your edge ... I think you need to clarify how unique this situation is so that others don’t try to jump on his bandwagon. His premise is flawed when applied outside of his operation."

"Is there a particular reason why you think it's a poor way to fertilise?"

A correspondent wrote:

"I'm hoping to get your thoughts on something I came across today.

I was discussing greens fertilising whilst at a friend's course this morning. He went on to get his new toy, the [...]. He's started to use the salinity level reading as an indicator to fertilise. So he's found a number that he's happy with that the turf looks hungry, applies a granular fertiliser and then waits for the number to drop back down to his threshold number again and repeats.

I'm not sure about this method as I've never come across it before plus I've never really looked into the salinity levels of my soils. I would just prefer to use gp and feel for when the plant needs something and adjust accordingly. But maybe he's onto something.

I would love to get your feedback on this if you're not too busy."

I replied that "I think that is a poor way to decide when to fertilize. Or what to fertilize with."

Then came a few more questions:

"Is there a particular reason why you think it's a poor way to fertilise?

If he's getting the results he desires, does that still make it poor? A reason he gave me about fertilising with a granular is [...] that by fertilising this way, he will encourage his perennial poa rather than poa annua."

First, the idea of deliberately managing soil nutrients to fluctuate up and down seems like the opposite of what most turf managers would like to accomplish.

I think most would ideally try to keep nutrient supply and growth as consistent as possible, rather than trying to cause them to fluctuate.


Second, it's changes in N that make grass grow, and then P and K and Ca and Mg and all the rest get taken up by the grass according to how much the grass is growing. So it makes sense to know the quantities of nutrients supplied, and also the quantities of the nutrients in the soil. But measuring the salinity of the soil doesn't tell which nutrients are there. It just gives the total quantity of salt.

Third, I have some concerns about the salinity number itself. The soil moisture meters that measure electrical conductivity at the same time are measuring the electrical conductivity of the water in the soil, and that measurement is strongly influenced by the amount of water in the soil. When the soil is drier, the meters give a low electrical conductivity reading, and when there is more water in the soil, even though there is no salt added, the electrical conductivity goes up.


This chart shows some measurements I made over a four day period on test plots on a golf course nursery green (pictured above). On a Sunday, I measured the soil VWC and the EC. Then I added irrigation water with 137 ppm salt, and I measured soil VWC and EC again. On Tuesday, there was a typhoon with 121 mm rain. On Wednesday, I measured the VWC and the EC again.

The EC as measured by the soil moisture meter is influenced by the water content of the soil.


One might say that is useful, because it gives some idea of the EC as the plant sees it. But if one makes that argument, then it is difficult to simultaneously make the argument that the EC is a useful criterion for determining when to supply fertilizer, because it is clear that the EC measurement is affected by the soil water content independently of the quantity of nutrients in the soil.

It is possible to adjust the EC measurement by incorporating the soil water content and the EC into a unitless measurement. The salinity index can be obtained by taking the EC, dividing it by the VWC, and multiplying by 100. This value takes into account both the EC and the amount of water in the soil at the time the EC was measured. There is not such a direct relationship between the VWC and the salinity index. But for those same data as shown in the previous chart, the salinity index also shows higher values with more VWC, even though no salt was added. In fact, after the typhoon's 121 mm of rain, one might expect leaching of nutrients, and a lower salinity index. But the opposite happened here, as shown in this chart.


And fourth, the follow-up question about if he's getting the desired results, is that still a poor way to fertilize? If he is getting the desired results, then fine, keep doing it. At some point it comes down to personal preference, because one can get good results in a lot of different ways. My preference, and what I think is a better way to determine when to supply fertilizer, involves monitoring the grass conditions, supplying N to produce the desired growth rate, and ensuring the grass is supplied with enough of each nutrient to meet the grass requirements. I expect such an approach is easier and will result in lower nutrient applications.

And about perennial Poa vs annual Poa, I'd be looking to supply a consistent amount of nutrients to the grass, rather than a fluctuating amount, because I expect the more ruderal biotypes of Poa annua would be more competitive with fluctuating nutrient supplies and with periodic granular fertilizer applications.

MacKenzie's fundamental principle of greenkeeping

I taught two seminars yesterday at the Philippine Golf Course Management conference. The first was about irrigation water requirement. The slides are here, and I made this Shiny app with data from 2013 through 2016 for Manila, Cebu, and Baguio.


In the second presentation I spoke about MLSN after 5 years. I explained what soil test interpretation is, why the MLSN guidelines were developed, and explained how they work.

This surprised me

I was making some calculations today about irrigation water requirement. I looked at Manila, where the normal annual rainfall is 1877 mm, and Cebu, where the normal annual rainfall is 1260 mm. These data from Manila are the 30 year average from 1961 to 1990, and at Cebu a 20 year average from 1971 to 1990.

For my calculations, I was looking at the past 10 years, from 2007 to 2016. I wanted to show the variation in irrigation water requirement at both locations, based on a calculation of the daily soil water balance.



I was able to get the daily precipitation from the GHCN (Global Historical Climatology Network) daily summaries by using the rnoaa package in R. For Manila, the annual precipiation (summed from the daily amounts) for the past 10 years ranged from 1381 mm to 2932 mm, with a mean annual amount of 1908 mm. That's pretty close to the normal of 1877 mm. For Cebu, there wasn't as much rain. The lowest year of the past 10 had 682 mm, the most was 1713 mm, and the mean was 1276 mm. Also pretty close to the normal of 1260 mm.


So what surprised me? Cebu gets less precipitation than Manila. The year with the most precipitation (in the last 10 years) at Cebu still had less rain than an average year at Manila. With those kind of differences, I expected the irrigation water requirement to be more at Cebu than at Manila. It rains less at Cebu, so more irrigation should be required, right?

The calculations don't work out that way. The reason is the way the rain is distributed through the year. Manila has more pronounced dry seasons and wet seasons. Cebu has dry and wet seasons too, but the dry seasons have more rain than at Manila.

For more, see the full presentation.

A Shiny app with adjustable rootzone characteristics and irrigation rules

I made this Shiny app that calculates the daily soil water balance.

The idea of the app is to change the soil conditions, specifically the rootzone depth and the field capacity, to see how changes in those parameters influence the irrigation requirement.

And the irrigation "rules" can be changed too. When will irrigation water be added? How much water will be added? What is the crop coefficient? What is the distribution uniformity of the irrigation system?

Then the soil conditions and the specific irrigation "rules" are matched to a year of weather data from a location, to see how any changes influence the amount of water required to satisfy the rules.


Estimating irrigation water requirement for different soil conditions

In previous posts, I wrote about the daily soil water balance and irrigation frequency using Bangkok (DMK) weather data, and showed how changes in irrigation "rules" can change a predicted irrigation water requirement, that time using Phuket (HKT) weather data.

Thailand conveniently has golf courses adjacent to many of its major airports, so I can imagine turf being maintained at that location, and then make the calculations using data from the airport.

Let's go to Chiang Mai. The Star Dome Golf Club sits right next to Chiang Mai International Airport (CNX).

Now I want to consider fairways, and specifically the soil type of the fairway. There are a number of advantages to growing fairways in soil, rather than in sand, and using surface and subsurface drainage, and perhaps a bit of sand topdressing, to create the desired playing surface. One of the advantages to growing in soil is a lower irrigation water requirement.

I'll use 2015 weather data from CNX. Let's imagine a fairway with a 20 cm rootzone depth, a field capacity of 40%, and irrigation at 20% to return the soil back to field capacity. With 2015 weather data, that gives an expected irrigation requirement of 718 mm.

That's a deep and infrequent irrigation regime, and with a 20 cm rootzone depth, about 40 mm will be required at each irrigation event. That's a lot of water. It would probably be more reasonable to do more frequent irrigation.

And that can save water too. For example, with that same rootzone depth and field capacity, but now irrigating at 24% to increase VWC to 30% (supplying about 12 mm at each irrigation event), the expected irrigation requirement goes down to 674 mm.

In SE Asia, it is common to sandcap fairways. For example, see this course now under construction in Thailand:

What would the irrigation requirement be for a sand rootzone at CNX in 2015? I'll keep the same rootzone depth and the same crop coefficient and distribution uniformity, just changing how much water is held in the rootzone because of the sand. I'll estimate that a fairway sand will have a field capacity of 20% (I think that is a generous estimate) and that irrigation will be supplied at a VWC of 10% to return the soil to field capacity. That gives an estimated irrigation requirement of 909 mm.

For simplicity, let's say that for the soil rootzone, the irrigation requirement is 700 mm, and for the sand rootzone it is 900 mm. Let's say this water requirement is for 10 ha of irrigated fairways. For the soil fairway condition, that gives an irrigation requirement of 70,000 m3. With a sandcapped fairway, an extra 20,000 m3 are required. Plus the energy to pump the extra water.

What happens to the irrigation water requirement after changing the irrigation "rules"?

I've shown how calculation of the daily soil water balance, matched to precipitation data, can be used to estimate the irrigation water requirement for a given set of irrigation "rules." That is, if I calculate how much water is in the soil (details about the calculation method here), carefully adding in the amount added by rainfall, and subtracting the amounts lost to drainage or evapotranspiration, I can determine when and how much irrigation is required.

And the irrigation rules are things like the quantity of water I will apply at each irrigation, the soil's field capacity, at what quantity of soil water will I reapply irrigation, the distribution uniformity of the irrigation system, and so on.

The first set of calculations I showed were for a location in Bangkok. Now let's go south, to the island of Phuket, and look at the irrigation water requirement using weather data from recent years. I got the data from the Phuket International Airport (HKT), which is just north of Blue Canyon Country Club. I'll imagine that these calculations are for a hypothetical stand of turfgrass at that location.

This is a view over the Canyon course looking north, with the control tower for HKT visible in the top left corner.


Now I will calculate the irrigation requirement using the weather data from HKT in 2015. First, I'll start with a scenario of:

  • rootzone depth at 10 cm
  • field capacity of 25% VWC
  • irrigation threshold of 12% VWC -- when the soil is predicted to drop below 12%, an irrigation event is triggered
  • each irrigation is set to return the soil to field capacity
  • the crop coefficient used to adjust the reference evapotranspiration to crop evapotranspiration is 0.7
  • the distribution uniformity of the irrigation system is 0.75

Calculating the water balance for every day of the year with those conditions, the annual irrigation water requirement is 644 mm.

That would be a classic deep and infrequent irrigation regime. For that same location and same weather data, what happens if I change to a light and frequent approach? Now I'll irrigate at 15%, rather than at 12%, but I will add only enough water at each irrigation to reach 20% VWC in the top 10 cm, rather than 25%. In this case, the annual irrigation requirement drops to 620 mm.

What might happen if I start using a (or use an improved) soil surfactant? I could reasonably expect that the spatial variability in soil water content would be reduced, and that the soil would be easier to rewet after drying. I can go back to the original deep and infrequent rules, but now with the surfactant use I will let the irrigation threshold drop down to 10%, instead of the more conservative 12%. With the surfactant, I think that is a reasonable and safe adjustment. Now what happens? The irrigation water requirement drops from 644 mm to 605 mm.

What happens if I can get the roots to grow a little deeper? If I then increase the rootzone depth from 10 cm to 12 cm, the irrigation water requirement goes from 620 mm down to 569 mm.

Here's a way to make a substantial drop in the irrigation water requirement -- improve the distribution uniformity of the irrigation system. If I improve the DU from 0.75 to 0.8, while keeping the other rules as in the previous scenario, the irrigation water requirement goes from 569 mm to 533 mm.

And if I then go back to frequent irrigation rules, in this case irrigating at 15% and adding water to increase the top 12 cm to 20%, the irrigation water requirement is 529 mm.

Simulating irrigation frequency at the world famous "snake" course

The "snake" course, and simulation using the daily soil water balance

Many of you will have seen the Kantarat Golf Course when flying into Bangkok. Maybe you've played it. It's a cool course, set between the two runways at Don Mueang International Airport in Bangkok.

It is commonly called the snake course, and I can confirm there are a lot of snakes out there.

image from c1.staticflickr.com

And then there are the planes, and the crossing of active taxiways.



The most common problem with irrigation water quality is high salinity, and the solution to that problem is adjusting the quantity of water supplied. At the end of yesterday's seminar, I switched from talking about water quality, and discussed the application of the daily soil water balance in managing irrigation water quantity.

I used the snake course as a hypothetical location, because I had a set of daily data from the weather station at Don Mueang (DMK).

More about irrigation frequency

I've written previously about whether it is better to do deep and infrequent irrigation, or whether it might actually be better to irrigate frequently in small amounts. I applied the daily soil water balance to work through this for a location at DMK.

Let's say we are growing grass at DMK and have a 10 cm rootzone depth and then the weather happens as it did every day at that location in 2015.

I'll have some plan of how I'm going to irrigate, too. Let's say there is a field capacity of 25%, and I expect the grass may wilt when the volumetric water content (VWC) is less than 10%. I will try to irrigate to keep the soil from dropping below 12%, and every time I irrigate, I will fill the soil back to field capacity. When I do that, with the details as shown here, for example using a crop coefficient (Kc) of 0.7 and a lower quartile distribution uniformity (DULQ) of 0.75, I can then simulate the soil water content day by day through the year. I do that by stepping through each day of the year, with the evapotranspiration and precipitation as it happened, adding irrigation as required by the rules I've set. Doing that for 2015, the irrigation requirement is 1011 mm and the median VWC through the year is 19.7%.

image from c1.staticflickr.com

I can also simulate the soil water content and irrigation required for a different set of rules, but for the same soil and weather. I did that, for those same 2015 weather data, now irrigating at 14% rather than at 12%, but instead of supplying enough water to raise the soil back to field capacity, I only add enough to increase the soil water to 18%. When I do this, the irrigation requirement drops to 970 mm, and the median VWC goes to 15.5%.

image from c1.staticflickr.com

I checked this for 2016 data, and the results were similar: a total 949 mm of irrigation required and median VWC for the year of 20.2% using deep and infrequent rules, 889 mm irrigation and a median VWC of 15.6% with light and frequent irrigation.

image from farm1.staticflickr.com

image from farm4.staticflickr.com

My presentation on irrigation water quality

Yesterday I taught a seminar about irrigation water quality.

Here are some links related to that presentation.