Thursday, July 13, 2017

Dealing with Uncertainty in Manure – Insurance Nitrogen (Just a little bit extra?)

There is uncertainty in every decision we make; all we can do is try to make the best decision we can based on what we know. When it comes to using manure as a fertilizer there are five main sources of uncertainty, these are:

  • Nitrogen need of the crop (due to every growing season being different)
  • Nutrient content of the manure
  • Availability of the manure nutrients to the crop 
  • Nitrogen volatilization during application
  • Application variability and rate uncertainty
How do we deal with this uncertainty? What we really want to do is try to minimize each of these sources, in some cases this may be practical, in others it might be impossible. For example, manure sampling gives us a good indication of the nutrient content of the manure, but as we all know manure isn’t always consistent, so it isn’t a perfect indicator. One way we often hear people handling this is by adding a little extra nitrogen, some insurance nitrogen, but is that really the best approach?
I was wondering about this a little while back, and started to do some calculations to find out how this uncertainty impacted our manure decision making and this lead me down a trail called the ‘value of information.’ Essentially, it asks if you can now make a better decision because of something you learned from it, how much extra value does that add. So I did that for manure sampling (http://themanurescoop.blogspot.com/2014/10/economic-value-of-manure-sampling-and.html) and was pretty happy, but then someone said, even if I sample I may be more confident about what’s in the manure, but there is still some error in that so maybe I should still be putting on a little insurance nitrogen. So I took some time to think on this.
Some background, this is based off higher nitrogen and corn prices than we currently sit at, something like 2013 corn prices but luckily I scaled then so actual profit isn’t shown and as long as the nitrogen to corn price ratio didn’t change the results should be relatively similar.

What I did was use the Maximum Return to Nitrogen calculator (http://cnrc.agron.iastate.edu/) to estimate what my profit would be per acre after field activities and subtracting the value of the nitrogen applied to the crop above the MRTN rate (because that was nitrogen you could have used elsewhere). What I found is that uncertainly in how much nitrogen always lowered our profit as compared to if we had perfect knowledge. Probably because there was always a chance of putting on too much nitrogen or too little nitrogen, where if we really knew the perfect amount we could hit the nail on the head. However, the interesting thing I saw was that if we did have uncertainty, the ideal nitrogen rate was just a little lower than when we had perfect knowledge. There were a few other things I saw though, the chance the manure might have higher nitrogen content worked to our advantage at low application rates, because there was potential that in the area where corn yield was really sensitive to nitrogen application rate that we actually put a bit more on than we were trying to. However, once we got towards that maximum profit zone we didn’t peak quite as high, because the uncertainty meant we were never quite hitting that nail on the head as we always might be a little short or wasting a little.


Figure 1. How does uncertainty change how we should think about nitrogen application? Comparison between perfect and uncertain nitrogen concentration in manure in a corn-soybean rotation (as a note 23% uncertain is what I say using the industry average for your manure is)

Figure 2. How does uncertainty change how we should think about nitrogen application? Comparison between perfect and uncertain nitrogen concentration in manure in a continuous corn rotation.


So what’s the takeaway from all this? The more we know the more accurate we can be in hitting the nail on the head and maximizing our profit – but you already knew that (I mean we say hindsight is 20/20 for a reason), so no surprise there.


The more interesting part was looking at how putting on that insurance N (just a little extra) impacted our maximum profit. Essentially what I saw was that a little extra just doesn’t help us handle that uncertainty. Extra N actually hurts us as we are at a spot on the yield response curve that is relatively flat and the risk of wasting extra nitrogen hurts our profit more than the risk supplying that extra N. 

Tuesday, June 20, 2017

How much manure from a dairy cow?


Happy dairy month!


As you all probably know June is dairy month and that means it’s time for a dairy manure story around here. As I got to thinking about what to discuss - I was working on a manure budget for the state of Iowa (how much manure to do we have, how much manure can we use, and how has this changed with time) and I started pondering about all the changes to agriculture and what it means for manure. Things like moving away from pasture and to confinement operations, adding storages, injecting manure when it is land applied all help us collect a larger fraction of the manure produced and better use it in our crop production systems. However, there is a deeper question, and that is, how has the amount of manure we got from a cow changed?
There are lots of things that could go into this, the production system we use, whether we scrape or flush manure, how much if anytime the cows spend on pasture or a loafing lot, but we are going to take a look at is how much manure the cow would excrete. Using the ASABE Manure Production standard (ASAE D384.2). To get some estimates of milk production the average annual milk production, in terms of pounds per had was obtained from the Survey of Iowa Agriculture for years 1924 through 2016. Over this time period milk production per cow has increased by almost 500% (figure 1); at the same time manure production per cow increased, but only by an estimated 40% (figure 2). This means that the we went getting about 0.1 pounds of milk for every pound of manure to more than 0.5 a pound of milk for every pound of manure, that’s a significant change.


Figure 1. average milk production per cow per year in Iowa.


Figure 2. Average manure production per cow per year in Iowa.

However, we can take a deeper a look at what some of these numbers mean for manure. For example, most of the weight of manure is water – but what about solids, nitrogen and phosphorus? Again we became more efficient here, with cows decreasing the amount of solids, nitrogen and phosphorus they excrete per pound of milk produced by 70-80%.  The actual excretion of solids, nitrogen, and phosphorus did increase, but only by 50% for solids, 16% for nitrogen, and 45% for phosphorus. If we do just a little more math and make some estimate about how much nitrogen is lost during storage, land application, and the availability of the nitrogen, we can make some estimates about how the manure’s nutrient content has changed (figure 3).


Figure 3. Estimated manure nutrient content.


While not entirely accurate as in the 1920’s through the 1960’s slurry manure systems were minimally used as animals were kept on pasture or raised in barns with bedding material and manure was handled as a solid. However, this work does provide some useful insight to understanding how this industry is changing, with animals that are much more efficient than they used to be, making more milk for every unit of manure we now manage. So next time you have a glass of refreshing milk, take a moment to ponder what the next 100 years will bring and what that could mean for how we manage manure.


Friday, May 19, 2017

Compost, Stockpiling, and Fresh Manure – What is happening?

Handling systems affect manure nutrient levels and forms by influencing gaseous emissions, exposure to runoff and leaching, and as a result can influence a manures ability to supply crop nutrients when land applied. Traditionally, on feedlots in Iowa, pens are cleaned periodically and then the manure is stockpiled, either near the lot or in a field and then eventually land applied. In some cases, lots might be scraped frequently and the manure land applied almost immediately. So what difference does each of these make to the manure properties?

 Freshly excreted manures are often very wet, which is especially true during wet periods of the year, like after rainstorms or snowmelt. These wet conditions generally make long-distance transport difficult or almost impossible. This also requires field be available for land application throughout the year, and as such, has generally fallen out of favor as it means the ground can’t be used to support crop production throughout the summer. Though in some cases, through the use of varied rotation, usually including some hay ground, land application can be accomplished. As a result, stockpiling manure becomes more prevalent. In this practice, manure is cleaned form the pen surface and then heaped into stacks, or stockpiles, to await reloading, hauling, and spreading. Stockpiling is generally a passive process but helps in matching manure application timing to better align with typical crop production practices. A third alternative is composting. Composting is an aerobic treatment where the manure is managed so the pile continues to have oxygen in it, providing conditions for microbes and bacteria to break down the material. Composting uses mixing to make a more uniform pile and causes the material to heat, often killing pathogens within the manure and inactivating weed seeds.
Figure 1. Stockpiled beef cattle manure.

When looking at the results, there are two things to look at. The first is the nutrient concentrations and the second is the mass balance.  When it comes to the concentrations, this tells us how far we can afford to move the manure. Higher concentrations mean we can afford to haul it a bit further as it is more nutrient dense. The results indicate that both stockpiling and composting increase the nutrient density relative to fresh manure. Much of this change is due to water loss, so we just aren’t hauling around as much water.

However, if you look at the results a bit closer, you’ll see we actually end up with less nitrogen and phosphorus to land apply from both stockpiling and composting. This occurs because some of the nutrients are lost due to volatilization of the nitrogen during the storage process.  Some of the loss of is due to dust and rainfall runoff during stockpiling and composting for phosphorus. Taken together, these results would show that if we need to transport manure long distances, composting might be a good option, but if we are using the manure on farm and want all those nutrients, stockpiling might be a better choice for your operation. Of course additional factors like consistency of the manure, killing pathogens, or inactivating weed seeds might be additional factors to consider. (these results are a summary of Larney et al., 2006 – Fresh, stockpiled, and composted beef cattle feedlot manure: nutrient levels and mass balance estimated in Alberta and Manitoba).
Table 2. Nutrient concentration of fresh, stockpiled, and composted beef cattle manure.

Dry Matter
Water
Total Carbon
Total Nitrogen
Inorganic Nitrogen
Total P

lb/ton
lb/ton
lb/ton
lb/ton
lb/ton
lb/ton
Fresh
698
1302
216
11.2
2.6
3.2
Stockpiled
856
1144
212
13.2
3.8
4.6
Composted
1280
720
208
18
1
6.6

Table 3. Mass comparison of fresh, stockpiled, and composted beef cattle manure.


Initial Mass
Final Mass
Dry Matter
Water
Total Carbon
Total Nitrogen
Total P

lb
lb
lb
lb
lb
lb
lb
Fresh
1000
1000
349
651
108
5.6
1.6
Stockpiled
1000
636
272
364
67
4.2
1.5
Composted
1000
328
210
118
34
3.0
1.1




Wednesday, May 17, 2017

Manure Application Uniformity



Maybe you have seen some tweets and pictures about manure application uniformity testing over the past years, or hopefully seen some information about upcoming field days where uniformity is going to be a key topic. You might be wondering why we are making a big deal about this, but when it comes to using manure as a fertilizer effectively, understanding how uncertainty impacts your decisions is critical for making the best management decision.

So why should we worry about manure application uniformity? Nitrogen for crop growth can come from multiple sources, the soil organic matter can mineralize and in doing so release mineral nitrogen for the plant. The remaining nitrogen needed to support crop growth comes from applied fertilizer. Years of research have gone into characterizing how crops would respond.  If you take a look at the Nitrogen Rate Calculator it will give you an idea of how a corn crop (in this case corn following soybeans) responds to the addition of nitrogen. What this graph (figure 1) shows is how the addition of nitrogen causes corn to respond. What the figure demonstrates is that we want to apply somewhere around 150 lb N/acre, if we are less than this the yield goes down, if we apply more than this we see minimal yield improvement.


Figure 1. Corn response to added nitrogen for Iowa conditions for a corn crop following soybean.

So what does this all have to do with manure application uniformity? When we are trying to hit 150 lb N/acre does that mean we just need to average that for the field? Probably not, it’s about getting a condition where it’s uniform over the whole field (yes, there might be some soil variations and every field has a bit different response to nitrogen, and weather conditions matter so some year’s crop response to nitrogen is much more drastic than others, but for the sake of argument, let’s work with this curve to see what it means).

For fun as we think about the math behind this problem, let’s assume we have corn planted on 30-inch spacing, our manure toolbar also has 30-inch spacing, and that corn roots only get their nitrogen from the manure application band that was placed next to that corn row. Then let’s figure we applied 150 lb N/acre from liquid swine manure that tested 50 lb available N/1000 gallons, so we were applying 3000 gallons an acre.
Now think about two pieces of equipment, one has a knife-to-knife coefficient of variation of 35% at this application rate, the other has a coefficient of variation of 10%. In both cases let’s figure an 8-knife setup. To give you an idea what this looks like in terms of nitrogen application rates achieved by the different knives and the impact different levels of uniformity have on crop yield let’s run through an example. Both of the tools in this example hit the right application rate on average, but how they do it, in terms of evenness across the toolbar is very different. What I want you to start thinking about is what would this mean for your crop yield from row-to-row and nitrogen leaching.

Table 1. Nitrogen and manure application rates for two pieces of application equipment that that achieve different levels of manure application uniformity.
Knife #
N Application
(lb N/acre)
N Application
(lb N/acre)
Application Rate
(gallons/acre)
Application Rate
(gallons/acre)
% of
Desired Rate
% of
Desired Rate
1
100
150
2000
3000
67
100
2
135
160
2700
3200
90
107
3
165
170
3300
3400
110
113
4
210
160
4200
3200
140
107
5
220
150
4400
3000
147
100
6
180
140
3600
2800
120
93
7
120
120
2400
2400
80
80
8
70
150
1400
3000
47
100
Average
150
150
3000
3000
100
100
St. Dev.
53
15
1060
302
35
10
COV
35
10
35
10
35
10


So let’s do a nitrogen example, using figure 1 (the nitrogen response curve) you can make an estimate of the corn yield that would be achieved from each of the knives (and if you assume a maximum yield of around 200 bushels an acre) can figure out what the field level yields would be.  So if you work through this math, you can find a few interesting results (table 2). Even though we were putting on the same amount of nitrogen on in both cases, because of variation from knife-to-knife we get different average yields per acre. In the case of corn following soybean, yields increased by 2 extra bushels per acre yield from the improved distribution and in the case of continuous corn about 4 extra bushels per acre. But there are other things to notice, the coefficient of variation in corn yield is always much lower than that in the nitrogen application rate. The soil supplies some of the nitrogen and this dampens out the response making everything a bit more uniform, but one thing to keep in mind is that early in the growing season the response might be more visually drastic than what final yields end up showing. Overall I think this asks an interesting question -how good is uniform enough, how does application uniformity uncertainty compare with other uncertainties in crop production, and how does this information help us make better manure decisions?

Table 2. Impact of nitrogen application uniformity on crop yield in corn-soybean (CS) and continuous corn (CC) rotations for machines with coefficients of variation in their application of 35% (left) and 10% (right)
Knife #
Corn Yield (CS)
(bu/acre)
Corn Yield (CS)
(bu/acre)

Corn Yield (CC)
(bu/acre)
Corn Yield (CC)
(bu/acre)
1
187
194
162
179
2
193
195
175
182
3
196
196
183
184
4
198
195
190
182
5
198
194
191
179
6
197
193
186
177
7
191
191
170
170
8
178
194

146
179
Average
192
194
175
179
St. Dev.
7
2
16
4
COV
4
1

9
2


To learn more about this topic and how you can get the most benefit from your manure make sure you register for one of our four field days, get a free lunch, get your manure questions answered, and learn how to set yourself up to maximize the benefit of manure on your farm this fall.

Friday, April 28, 2017

Manure a valuable commodity


When I first started really thinking about the how’s and why’s of how manure is managed one of the best papers to my line of thinking was “Making economic sense of why swine effluent is sprayed in North Carolina and Hauled in Iowa” by Roka, Hoan, and Zering. This paper was written a few years ago, but really what it’s trying to do is understand why people make the decisions they do and trying to look at the economic options from the perspective of each location. The paper does a lot of fancy economics (it was written by economists after all), but in the end it basically says Iowa soils are really good, they grow high value crops, and as a result have a greater need for nutrients and value their manure.

I don’t want to delve too deeply into what they wrote, but instead wanted to focus on something related to it. Manure values aren’t static with time, production practices change, fertilizer prices change, and maybe even our need for these nutrients change. As they do the value we can get from our manure values with it. Over the last couple decades we have seen some pretty big changes in the swine industry, we’ve seen most farms move to we dry feeders, we’ve seen some ration changes that include more DDGS, and we’ve seen increased use of phytase in these rations. Start adding these things together and you get some pretty big changes in manure properties, some increasing nutrient content and some potentially reducing phosphorus content. Along with this we’ve seen some pretty drastic swings in fertilizer prices. This got me thinking – how as the value of swine manure in Iowa changed.

The first figure I have for you is average concentrations of N, P2O5, and K2O in swine finishing manure as a function of time for 2000 through 2015. Starting in 2010 we begin to see a pretty pronounced increase in the manure’s nitrogen content and while P and K have increased, not nearly as much. This has lots of benefits – it means lower application rates to get the same nutrient application, less water to move around to get those nutrients to our field, and even N:P ratios that are more closely aligned with the crop need.

Figure 1. Average N, P2O5, and K2O manure concentrations from swine finishing farms with deep pits in Iowa.

If you start looking at what this means for nutrient value in the manure you’ll find that until around 2005 manure nutrient values were only around $10 per 1000 gallons, but starting in 2007 values skyrocket as nutrient concentrations increased along with higher fertilizer prices. Although it may not be quick as high as back in 2011 when it reached over $50 worth of nutrient value per 1000 gallons, but it still is in excess of $30. To me this means the message stated about the economics of manure management in Iowa that Roka, Hoan, and Zering first presented is even more true today – take advantage of this natural fertility source that that is a more cost effective source of fertility than ever before.


Figure 2. Average value (based on nutrient content) of deep pit swine finishing manure in Iowa.


Friday, March 31, 2017

What is the design basis for feedlot runoff control systems?



When it comes to controlling feedlot runoff we all recognize it is important to keep the manure enriched from reaching streams, rivers, and lakes, but all systems have limits. Build it too big and it gets cost restrictive to construct, build it to small and water quality is at risk. So how do we go about deciding how big is big enough?
Currently government and citizen concern in Iowa has begun to focus more on runoff events from small (non-CAFO) lots that could result in water quality violations and/or fish kills within the receiving water body. A report in 2008 by the Iowa Department of Natural Resources (Iowa DNR) stated that animal waste related issues accounted for over 25% of the reported fish kills. An investigation, conducted by Ubbo Agena (2009) of the Iowa DNR, showed that a large number of the feedlot runoff related fish kills occurred in late summer and early spring following dry periods when stream levels were low or during spring snowmelt. To positively impact water quality, it is important that open beef feedlot runoff control systems are designed to prevent effluent discharges during critical times. As shown in figure 1, this corresponds to times when most of the feedlot runoff is occurring, so that makes sense, but an additional question might be is there any sign that certain storm sizes are causing these issues?


Figure 1. Percentage of feedlot runoff occurring during each of the seasons.

One method to evaluate the events most prone to creating water quality violations would be to identify storm events resulting in feedlot runoff, but not of a size sufficient to result in runoff from the surrounding landscape. Unfortunately, there really doesn’t seem to be much relationship; however, if we look at what size storm we need to design our runoff system to handle to greatly reduce risk of release, we get a pretty strong relationship (figure 2).
Figure 2 shows the number of exceedances of the fraction of runoff from feedlot ratio on the y-axis and the “design storm” size on the x-axis. In each case the graphs show the number of occurrences where the fraction of runoff from feedlot was greater than 0.7, i.e., events where there would be relatively low amounts of dilution water available.  Like the other analyses a one-to-one ratio of feedlot to additional drainage area was assumed. This generalizes results to show how the relative risk of a water quality violation is affected, but does not allow for an evaluation of the absolute risk of a water quality violation for by the different land covers. From these graphs it is clear that the most conservative estimate is to assume that the surrounding area is vegetated with alfalfa or a cool season grass. This was expected as it increased evapotraspiration as compared to the row crops corn and soybeans, keeping the soil profile in the surrounding area drier. This results in fewer runoff events and smaller runoff volumes from the surrounding landscape. Soybeans generally had slightly more runoff from the cropland than corn and thus fewer occurrences when the fraction of runoff from feedlot ratio was exceeded. Also of note is that the vegetation choice has a larger effect on soils that are more prone to runoff (Classes C and D) than for the other classes (Classes A and B). One item of note, this analysis does not consider differences in water quality from the cropland areas and how that it affect runoff quality of the dilution water. If runoff quality was considered it would be relatively safe to assume that the differences approximately balance out, as runoff water from alfalfa/cool season grass fields would be cleaner (lower nutrient and sediment concentrations) than from crop fields.


Figure 2. Effect of vegetation type on the number of times per year the that feedlot runoff fraction exceeded 0.7 for each of the four different hydraulic soil types. Soil hydraulic class C


When I look at this figure the big thing I see is it is important to have a control system that can handle somewhere around a 3-inch storm at this point the level of control from each extra increment of design volume in your control system only offers less improvement in control. Our design recommendation tends to be almost double this and you might be asking, why that is the case. Well if you think about Iowa weather conditions often times we tend to get strings of wet weather and then be dry. So to really control that 2.5 to 3-inch storm, we normally need a little extra capacity built into the system to give us some flexibility about when we apply.