January 29, 2013

Agronomists Notes

Hello Reader,

After six days in Alberta, our Aussie-Kenyan guests now consider -1C to actually be quite warm. What a shock to jump from the combine seat in Kenya to experience snow-covered Alberta for the first time! We’re now huddled in Edmonton for the FarmTech conference and I look forward to chatting with many of you there.

This week we’ll continue the research of grid sampling inside zones versus zone sampling. We’ll also take a look some subscriber comments from last week’s grid versus zone soil sampling article. Then I’ll show you how to reduce erosion in CTF by choosing the right direction to run your tramlines before we finish with technical grain market news.

Have a great week.

Grid sampling versus zone sampling continued

After publishing last week’s article on the case for grid sampling zones, I was sent an article written by an American agronomist who also identified a large variance in nutrient levels in each zone. The study ran additional correlations and posed questions as I have: Is zone sampling the correct approach? If soil nutrient levels are consistent across each zone then yes, zone sampling is the most cost effective way to determine nutrient status. If not, how much do nutrients vary inside each zone and what are the chances of making a wrong fertilizer recommendation?

To show an example, this study measured the variability of pH, P and K using grid versus zone sampling. The zones were separated based on soil type and grid sampled on 1.1 acre grids and compared with composite zone sampling. A statistical analysis was run to calculate the average variation of pH, P and K away from the zone average. Then, a level of error was calculated to determine how much the zone sample over/underestimated the pH, P or K. As you can see in the chart there was a 44% chance of under/overestimating P, a 24% chance of under/overestimating K and 6% chance of under/overestimating pH.

The results of this study, and please take it with a grain of salt, it is only one year and one field, but it basically says you have a 50% chance of getting your P recommendation wrong with zone sampling. You have a 24% chance of making an incorrect K recommendation and very little chance of making a wrong measurement of pH. What's interesting is that I'm finding similar results in my research with large variences in P and K levels inside zones and not as much with pH.

Last week I showed you an NDVI based zone map that compared P levels inside each zone using 1 ac grids. 4 of the 7 zones had a high degree of variability with P and K ranging from deficient to excessive. The American study showed similar results and confirms what I am finding in my analysis. The level of nutrient variability inside of each zone depends on the nutrient you’re looking at. In the case of P, it can vary to a high degree and leave you with almost a 50% chance of making a recommendation error.

In the end, if we’re going to go to all the effort of producing field boundaries, running through NDVI images, dragging EM38 sleds across the field, overlaying yield data, elevation, soil testing, estimating yield in each zone (throw a dart), estimating N mineralization in each zone, calculating fertility recommendations to finally come up with a prescription map, we had better make sure its worth it. So is zone sampling the best method to estimate nutrient levels inside each zone? As I'm finding, it depends on the nutrient you're measuring and depends on the field history. In this particular study, there was a 44% chance of making the wrong P recommendation which is totally unacceptable.

I do believe we need to take another look at how we incorporate a statistically significant number of grid samples inside of each zone vs zone sampling to give us greater accuracy. I understand that it increases the cost and complexity of making recommendations but if the margin of error is too great using zone sampling, then why go to all the effort? Perhaps I'm not a good enough is good enough kind of guy and overcomplicating the process but I do think we have serious room to improve the way we make VR fertility recommendations. SL

Study

This chart shows the margin of error when using zone sampling compared to grid sampling when measuring pH, P and K levels inside each of 7 zones.

Readers comment

Re: “The case for grid sampling variable rate zones”

Last week’s article on grid sampling zones brought on some great conversation and will lead to a lot more. I’ve included a couple great comments but first here are my thoughts. I do believe VR is the future and holds tremendous potential. I just think we’re scratching the surface right now with average to modest ROI’s. The zone map upon all decisions are based should be one that is scientifically sound and repeatable, not one that is built from start to finish with human bias. Create a scientifically sound zone map then massaging the zones to improve logistics makes more sense than someone with a mouse clicking around zones that follow a preconceived bell shaped curve. Then, calculate the number of grid samples required to give you the most statistically accurate picture of background nutrients instead of random blended samples within each zone.

“My experience with VR is limited; we did N on a couple of different occasions, roughly 2500 ac and never did see value. The challenge I found was trying to come up with the rates; in both years we were too high and too low for the conditions. If we had narrowed up the rates you would have wondered what the point of VR was. N response curves would suggest that when working with those higher rates, it could be tough to detect responses. In most of our fields where we have 5-6% OM, we seemed to hit the top of the yield curve for canola at 20lbs below the highest rec'd rates (i.e. 130lbs N, at 150 yields declined). Where we were too low, we created our own variability due to the effects it had on maturation. Optical sensors make so much more sense to me. The problem is trying to work out the logistics.” Adam Gurr, Manitoba

“The grid vs. zone sampling debate has always been a hot topic in precision ag approaches. Nutrient variations can change dramatically within short distances. We have found in certain areas there are similar variations as you have displayed in this example. We have also found in other zoned fields we have very consistent nutrient trends when grid sampling within zones as well. Then there is the issue of weather yield variations in the filed are influence more from nutrient variations or soil type / textures / topography or a combination of both. Ideally we would set a target yield for every grid point based off of all of these factors. When first getting info VR we buried ourselves with data and the deeper we dug the more we realized there are so many more factors and variations to consider. I realized we will never be able to create ideal or perfect VR programs but as long as we continue to be more efficient and continue to increase RIO's even with very basic VR we are moving in the right direction. There are also very different variables to consider from different field areas and even between fields on the same farm. Cool stuff.” Terry Aberhart, Saskatchewan

Tramlines & erosion risk

Control erosion through row direction

I’ve had a number of discussions with subscribers and callers from across the Prairie’s asking about transitioning to CTF. Some are concerned about erosion on the hills. In fact, I’ve been told by PhD’s that CTF will lead to serious erosion in the tramlines. After three years of controlled traffic farming on land with short, steep hills and heavy downpours, I’ve yet to see any type of erosion on our cracking clay, 2:1 soil type. Are the experts wrong or is there a proper direction to run tramlines that reduces the risk of erosion?

It is intuitive to think that tramlines should run across a hill or adjacent to the slope to slow the speed of water runoff. Having tramlines that run in the same direction as the slope should increase the speed of water and increase erosion. Right? Studies have found the exact opposite where tramlines that run in the direction of the slope actually reduce erosion.

The severity of erosion increases when water accumulates in one area, builds up energy, then finds the path of least resistence and cuts a channel into the soil creating erosion. Another way, water collects, builds up volume in between the furrows, and then speeds up as the large body of water rushes out of a small crack in the furrow wall. The end result is a high volume of water moving at a faster speed out a smaller exit which begins to move soil quickly.

The risk of erosion on tramlines running down slope can increase in certain cases. Disc seeders and deep sprayer wheeltracks are the two biggest culprits for increasing erosion risk. In sprayer wheeltracks the ruts allow water to collect and build up energy. Disk seeders on the other hand leave very shallow furrows which allows water to travel faster across the soil surface. Both situations increase the risk of soil movement as the energy and speed of water rises.

To reduce the risk of erosion, it is recommended to manage tramlines to keep them as shallow as possible. This can be done by using tramline renovators or wheeltrack managers. It is also a good idea to rotate your sprayer tracks to minimise the chance of them getting too deep. Narrow spray tires can also cause rutting problems, and should be avoided. These few steps can really help to minize soil erosion.

To see a great example of residue flow and deposition inside tram lines running down slope after a big rain go here. It's by Tim Neale of Precision Agriculture Australia . To learn more about row direction down slopes visit here.

Tim Neale and his crew produce CTF farm layouts with the use of field boundary and elevation maps to maximize efficiency and reduce erosion.

It’s amazing to see how perception is not always reality and running tramlines down the slope definitely goes against conventional thinking. That said, whether random or CTF, erosion can always occur inside tram lines that run across or down slope in extreme situations. Thankfully that is not the norm and not what we're finding after three years of CTF. As outlined there are steps you can take to minimize erosion and running tramlines down the slope is one of them, not across. SL

Picture: Residue deposited on tram lines which slows the speed of water and reduces erosion. Source: Tim Neale