The Wisconsin System for Evaluating Corn Silage

Proceedings of the 2005 Idaho Alfalfa and Forage Conference
THE WISCONSIN SYSTEM FOR EVALUATING CORN SILAGE
Richard J. Norell, Ph.D.1
SUMMARY
The University of Wisconsin has developed a spreadsheet (Milk2000) for comparing corn silages
varying in yield per acre and feed quality. Milk2000 calculates two important values for comparing
hybrids: milk/ton and milk/acre. Milk per ton is an estimate of forage quality and is primarily driven
by starch content, starch digestibility, and NDF digestibility. Milk per acre combines the estimate of
feed quality (milk/ton) and crop yield into a single term. Milk per acre is calculated by multiplying
milk per ton times dry matter yield per acre. Selecting corn hybrids with above average milk/acre
values results in selecting hybrids that produce high yields of high quality silage. Milk2000 is an
excellent tool for summarizing variety trial data on corn silage.
INTRODUCTION
Corn breeders and seed companies have done a marvelous job in developing corn hybrids for a wide
variety of conditions. Corn growers have the opportunity to select hybrids differing in agronomic and
nutritional characteristics. Key agronomic traits include: maturity rating, dry matter yield, grain
yield, standability, disease resistance, pest resistance, herbicide resistance, drydown rate, and
staygreen. Nutritional characteristics used to evaluate silages have undergone change in the last few
years. In the past, the key nutritional characteristics were: crude protein, acid detergent fiber (ADF),
neutral detergent fiber (NDF), and estimated energy content (usually based on ADF equations). More
recently, it has been recognized that silages differ in grain content, grain digestibility, amount of
stover, and stover digestibility. These differences can be included in selection criteria by evaluating
silage dry matter content, starch levels, and NDF digestibility (NDFd).
The large number of selection traits can be rather daunting and confusing to many corn growers. A
spreadsheet developed by the University of Wisconsin (Milk2000) has taken some of the drudgery
out of comparing hybrids. Results of this spreadsheet are used by several Midwestern universities
and private seed companies for reporting results of their hybrid trials. The primary objectives of this
paper are: 1) review the Milk2000 spreadsheet, 2) discuss how to use the results in hybrid selection,
and 3) discuss how to estimate the economic value of silages varying in quality and yield.
USING Milk2000
Where can I get Milk2000?
Milk2000 is a Microsoft Excel spreadsheet. It can be downloaded from the University of Wisconsin
at the following web address: (http://www.uwex.edu/ces/forage/pubs/milk2000.xls). You must have
Microsoft Excel to run the spreadsheet.
_______________
1
Richard J. Norell, Ph.D., Extension Dairy Specialist, University of Idaho, Idaho Falls, ID. Published In:
Proceedings Idaho Alfalfa and Forage Conference 7-8 February 2005, Twin Falls, ID, University of Idaho
Cooperative Extension.
44
Proceedings of the 2005 Idaho Alfalfa and Forage Conference
What are the required inputs?
Milk2000 has a separate page for alfalfa hay and corn silage. Required inputs are different for alfalfa
and corn silage. Make sure you go to the corn silage page. The required inputs are: corn processed
(yes or no), dry matter %, crude protein %, NDF %, in-vitro NDF digestibility-48 hour, starch %,
bound protein (NDF-CP %), ash %, ether extract %, and crop yield (tons of dry matter per acre). If
laboratory results are not available, the authors suggest using: 1.3% for NDFCP, 4.3% for ash, and
3.2% for ether extract.
How does Milk2000 estimate milk/ton?
Milk2000 uses several equations in the process of estimating milk/ton of dry matter. Full details are
provided at the following web address: (http://www.uwex.edu/ces/forage/pubs/milk2000.htm). Key
points are discussed below. Milk2000 calculates TDN based on a summative equation. Nutrients in
the summative equation include: crude protein, NDF, fat, starch, and non-starch carbohydrates.
These nutrients are multiplied by varying digestibility estimates (Table 1). Crude protein, fat, and
non-starch carbohydrates have a fixed digestibility estimate. Total tract starch digestibility is
predicted from the whole-plant DM content using a regression equation developed from previously
published data. NDF digestibility is determined with a 48-hour in-vitro digestion of the sample and is
expressed as a percentage of total NDF. TDN is converted to Net Energy at 3x maintenance (NEL) by
equation.
Milk2000 estimates silage dry matter intake based on NDF concentration and NDF digestibility;
assuming a 1350-pound cow consuming a 30% NDF diet. For corn silage hybrid trials, base dry
matter intake is adjusted by 0.374 pounds for each percentage unit change in NDF digestibility above
or below the average NDF digestibility for the trial. Energy intake from silage is calculated by
multiplying silage dry matter intake by the estimated NEL concentration of the silage. Milk/ton of dry
matter is then estimated by first subtracting the forage portion of cow maintenance energy
requirements from estimated energy intake and then dividing by 0.31 (the NEL requirement per pound
of milk).
How does Milk2000 estimate milk/acre?
Milk/acre is calculated by multiplying the milk/ton estimate times the observed silage yield
(expressed in tons of dry matter/acre). Milk/acre provides an index of corn silage performance based
on both yield and quality.
What are the effects of increasing dry matter percentage on silage energy content and milk/ton?
Milk2000 adjusts starch digestibility based on whole plant dry matter content (Table 2). Net Energy
for Lactation and milk/ton estimates are also shown in Table 2 for silages varying in dry matter
content from 30% to 45%. Crude protein, starch, NDF, and NDFd were held constant across dry
matter concentrations and were set at the average value for the 2004 Wisconsin hybrid trials. Net
energy for lactation declined 0.07 Mcal/lb as unprocessed silage increases from 30% dry matter to
45% dry matter. The drop in energy concentration was only 0.03 Mcal/lb for processed silage across
the same dry matter range. Net Energy decreases due to the negative relationship between whole
plant dry matter content and starch digestibility (Table 2). Milk/ton is very sensitive to changes in
45
Proceedings of the 2005 Idaho Alfalfa and Forage Conference
estimated energy content. Milk/ton declined 145 pounds for processed corn silage over the range of
silage dry matters in Table 2. The milk/ton estimate for unprocessed silage declined 481 pounds from
30% to 45% dry matter.
What are the effects of varying starch and NDFd on energy content and milk/ton?
Net Energy (NEL) and milk/ton estimates are shown in Table 3 for silages varying in starch content
(range 25 to 35%) and NDF digestibility (range 57 to 67%). Crude protein and NDF concentration
dry matter concentrations were set at their respective average value for the 2004 Wisconsin hybrid
trials and were held constant across the combinations of starch and NDF digestibility. A tenpercentage unit increase in starch content had a smaller effect on estimated NEL concentration (+0.03
Mcal/lb) than a ten-percentage unit increase in NDF digestibility (+0.05 Mcal/lb). Milk/ton increased
202 pounds with a ten-percentage unit increase in starch and 408 pounds with a ten-percentage unit
increase in NDF digestibility. Why the twofold difference? Milk2000 increases NEL content and
estimated dry matter intake with improved NDF digestibility. Conversely, Milk2000 only increases
NEL with improved starch concentration.
What are the limitations of Milk2000?
The calculations for milk/ton and milk/acre are very responsive to small changes in nutrient
composition and yield. When comparing two silages, only use nutrient values and yields that are
statistically significant between the hybrids. In variety trials with replications, use the reported LSD
(Least Significant Difference) estimates to determine significance between hybrids. University of
Wisconsin routinely reports LSD10% values in their variety trials. When the difference between two
hybrids is more than the LSD, there is a 9 out of 10 chance that the hybrids are statistically different.
For the last two years, typical LSD10% values within University of Wisconsin trials are ~100 to 150
pounds for milk/ton and ~2,000 to 2,500 pounds for milk/acre.
Milk/ton is a useful indicator of forage quality and can be used to rank forages. However, it should
not be viewed as a direct estimate of the cow’s ability to produce milk while fed that particular
forage. Milk production varies due to stage of lactation, genetics, environment, and ration
constraints. It is impossible to include all these factors into a simplified spreadsheet.
Milk2000 assumes that corn silage is the only forage in the animal’s diet. The effects of silage
quality are more evident in high silage diets than in blended diets of alfalfa hay and corn silage.
Producers are less likely to see a productive response when high quality silages are included at a low
rate in the diet (example 15% of dry matter intake versus 45% of dry matter intake).
Net energy estimates are higher in Milk2000 for high starch, high NDFd silages at 30 to 35% dry
matter than previous estimates with ADF equations or the 2001 NRC equation. Milk/ton of silage dry
matter will be overestimated if the calculated NEL concentration is too high. Silages with superior
feed quality (high starch, high NDFd) will be ranked correctly with Milk2000 in hybrid trials but their
superiority over conventional silages will tend to be overestimated.
Milk2000 adjusts both NEL and dry matter intake in response to increasing NDF digestibility.
Current research trials support enhanced dry matter intakes with improved fiber digestibility but not
an improvement in NEL concentration.
46
Proceedings of the 2005 Idaho Alfalfa and Forage Conference
How do you use milk/ton and milk/acre in selecting hybrids?
Remember, milk/ton provides a relative indication of silage quality and milk/acre combines quality
and yield into a single term. Milk/ton, milk/acre, and nutrient composition vary widely between
varieties (Table 4). Summarized milk/ton and milk/acre data provides information for understanding
general relationships between groups of hybrids and individual hybrids.
Relative comparisons can be made between specialty hybrids by comparing milk/ton and milk/acre to
the average for all hybrids. For example, let’s look at the hybrid summary from Wisconsin in Table 5
and let’s compare relative performance of brown mid-rib (bmr) silage to the average for all hybrids.
The average bmr silage was 300 lbs higher in milk/ton and 2900 lbs lower in milk/acre than the
average hybrid. Milk/acre decreased due to lower dry matter yield in bmr silages. Based on this
simple comparison, we understand that bmr silages have above average feeding value but lower
tonnage yields.
Let’s look at some of the other general relationships in Table 5. Milk/ton and milk/acre were below
the all hybrid average for waxy and high oil corn, suggesting lower feed quality and lower tonnage
yields for these two specialty varieties. Leafy hybrids were similar to the all hybrid average in
milk/ton and milk/acre, suggesting average feed quality and yield. Nutri-dense varieties were higher
in milk/ton but similar in milk/acre to the all hybrid average, suggesting higher feed value but
slightly lower yield.
Milk/ton and milk/acre can be helpful in the initial screening between hybrids. Let’s use Table 6 as
an example and assume that we are dealing with a seed company with only seven hybrids. Dry matter
yields, NDFd, starch %, milk/ton, and milk/acre are shown for each hybrid. Which hybrid would you
choose? Let’s rank them from high to low on milk per ton (D, E, F, G, B, A, C) and milk per acre (G,
F, E, B, A, C, D). Three varieties quickly drop off the preference list due to low milk/ton and
milk/acre ratings (A, B and C). Hybrid D ranks highest in milk/ton and lowest in milk/acre. I would
like to feed that quality of silage but probably couldn’t pencil out positive returns with such a low
yield. Hybrid D should be dropped from the preference list.
We have narrowed the list down to three (G, F, and E). Hybrid G has the highest dry matter yield and
starch content but lowest milk/ton in the top three. Hybrid G was probably a little more mature at
harvest, accounting for higher yield but lower milk per ton (reduced starch digestibility). These top
three are therefore very similar and I would evaluate other agronomic considerations to finalize my
choice.
From a dairy nutrition perspective, I would use alternative criteria depending on diet and target
animal. If the silage is grown strictly for growing heifers and dry cows, then feed quality (milk/ton
rating) of the silage is not very critical. The main goal would be to maximize yield of silage at
reasonable cost. If the silage will be fed to lactating cows at a high rate (say 40% of the diet dry
matter), then feed quality (milk/ton rating) becomes very critical. I would look for hybrids with high
NDFd and average starch levels. If the silage will be blended say 50:50 (dry matter basis) with alfalfa
hay, then a good dual-purpose blend is a logical choice with at least average or better starch and
average NDFd.
Can Milk2000 be used to set economic values for corn silage varieties?
Extension workers at the University of Wisconsin have made relative income comparisons between
silage varieties with the Milk2000 spreadsheet. Milk income per acre was calculated by multiplying
47
Proceedings of the 2005 Idaho Alfalfa and Forage Conference
the milk/acre estimate times a $0.105 milk price ($10.50/hundredweight). Example results from this
calculation are shown in Table 5. The average estimated milk income for all hybrids was $2,562 per
acre. Relative comparisons can be made between hybrids or compared with the average. For
example, the estimated milk income for average brown mid rib (bmr) hybrids was $2,258 per acre,
$304 below average (2,562-2,258 = 304). Brown mid-rib varieties were well above average in
estimated milk income per ton of silage dry matter, but due to lower dry matter yields, estimated
income per acre was below average.
There are limitations to using Milk2000 for this purpose. Care should be taken to only make
comparisons between hybrids that differ statistically in their milk/ton and milk/acre estimates. For
the last two years, typical LSD10% values within University of Wisconsin trials are ~100 to 150
pounds for milk/ton and ~1,500 to 2,500 pounds for milk/acre. Estimated milk income per acre
should be assumed to be equal if hybrid differences in milk/acre are below LSD10% values.
Are there other approaches for estimating relative values for corn silage?
Many different approaches have been used to derive relative values for corn silage. Three methods
are compared below.
Minnesota has developed a simple equation for valuing silages with a base price and adjustments for
starch content and NDF digestibility. The Minnesota method assumes an average corn silage with
29% starch, 41% NDF, and 58% NDF digestibility. They assume that every 1% change in starch
(DM basis) is equivalent to 0.5 bushels of corn per ton of DM and every 1% unit of NDF digestibility
supports 0.6 lb of milk production. The formula for calculating relative value is shown below along
with an example calculation.
$/ton DM = Base price/ton DM + Starch adjustment + NDF adjustment
Starch adjustment = ((% starch (DM basis) – 29%) x .5 bushels) x corn $/bushel
Example 34% starch and corn price of $2.10/bushel
Adjustment = ((34 – 29) x 0.5) x $2.10 = $5.25/ton of silage DM
NDF digestibility adjustment
Example 54% NDF digestibility (48 hour in vitro) and milk at $13.50/cwt
Adjustment = 54% x .6 x .1350 = $4.37/ ton silage DM
Base price - established based on planting, agronomic and harvest cost
Example - $20/ton at 33% DM = $60.61/ton DM
$/ton of corn silage DM = base + Starch Adj + NDF dig Adj.
= $60.61 + $5.25 + $4.37
= $70.23/ton DM or $23.18/ ton at 33% DM
The second method for developing relative prices between silages is to compare ration costs between
a base ration and alternative forages. The University of Wisconsin has developed a spreadsheet
(Pricer) for making these calculations. Pricer was developed in the late 80’s and modified in the 90’s.
Further modification was needed to use current energy estimates rather than digestible dry matter. I
have modified the program to use the Milk2000 energy estimates. Inputs include: milk production,
milk price, alfalfa price, corn price, by-pass protein price, alfalfa hay price, and silage nutrient
48
Proceedings of the 2005 Idaho Alfalfa and Forage Conference
composition. The Modified Pricer spreadsheet calculates relative prices for silage on a dry ton and
wet ton basis.
The third method sets the value of a feed to equal to the sum of its’ feed nutrients. The University of
Wisconsin has a spreadsheet (Feedval4) that calculates the value/pound for TDN, degradable protein,
by-pass protein, fat, calcium and phosphorus using common dairy ingredients. Feedval4 multiplies
the composition of each feed times the value per pound and calculates total value on a dry matter and
as fed basis. Ohio State University has a new computer program (Sesame) that calculates nutrient
value and estimates relative values using 2001 Dairy NRC estimates.
Relative values, estimated by the above three methods, are shown in Table 7 for corn silage varying
in starch content and NDFd. The Minnesota method increases silage value by $15 per ton of DM
when starch content increases from 25 to 35% but only increases silage value by $0.81 per ton when
NDFd increases from 57% to 67%. The Minnesota method ignores potential differences in starch
digestibility. Using the Modified Pricer spreadsheet, a ten-percentage unit increase in starch increases
the value/ton of DM by $2.63 while a ten percentage unit increase in NDF digestibility increases the
value/ton by $5.26. Modified Pricer uses MILK2000 energy estimates and can account for
differences in starch digestibility. Changes in relative values due to increasing starch ($2.20) or
NDFd ($3.20) were smaller with the Feedval4 spreadsheet. Feedval4 can use energy estimates from
Milk2000 and therefore can reflect differences in starch digestibility.
Which method is best?
Each method has strengths and weaknesses. The Minnesota method appears to overvalue increasing
starch content and undervalue increasing NDFd compared to the other two methods. Modified Pricer
and Feedval4 can use TDN estimates from Milk2000 and therefore can account for differences in
starch and NDF digestibility. Both may overvalue silages with high NDF digestibility relative to
actual cow performance. Recent research suggests that cows consume more dry matter and produce
more milk due to higher energy intake from high NDFd silages rather than increased energy content
of the silage. In my opinion, Modified Pricer or Feedval4 probably provide more realistic relative
values between silages varying in starch or NDFd.
49
Proceedings of the 2005 Idaho Alfalfa and Forage Conference
Table 1. Digestibility coefficients used in Milk2000 to estimate TDN concentration of silages.
Nutrient
Digestibility
Factor
Starch
Variable
Estimated by equation
NDF
Variable
Estimated by 48 hr in-vitro
digestibility
Fat
Fixed
93%
Protein
Fixed
97%
Non starch carbohydrates
Fixed
98%
Table 2. Effect of silage dry matter content on starch digestibility, estimated Net Energy
and Milk/ton in the Milk2000 spreadsheet.1
Starch Digestibility
NEL (Mcal/lb)
Milk/ton
DM %
Unproc2
Proc2
Unproc
Proc
Unproc
Proc
30
94.7
95.0
0.75
0.75
3645
3645
35
86.3
91.0
0.74
0.75
3572
3645
40
78.0
86.6
0.71
0.74
3363
3578
45
70.0
82.2
0.68
0.72
3164
3469
1
Assumed silage quality: 35% DM, 7% CP, 31% starch, 47% NDF, 62% NDFd, 1.3% NDFCP,
4.3% Ash, and 3.2% Fat
2
Unproc = unprocessed silage, Proc = processed silage.
Table 3. Effect of varying starch content and NDFd on estimated Net Energy and Milk/ton.
Milk/ton
Starch %
NDFd %
NEL (Mcal/lb)
25
62
0.72
3448
30
62
0.74
3547
35
62
0.75
3646
31
57
0.71
3364
31
62
0.74
3572
31
67
0.76
3772
Assumed silage quality: 35% DM, 7% CP, 47% NDF, 1.3% NDFCP, 4.3% Ash, and 3.2% Fat
Table 4. Average values of corn hybrid means (reps per location=3) in University of
Wisconsin performance trials conducted between 1995 and 2003 (n=3255).
Standard
Measurement
Mean
Deviation
Minimum
Maximum
Forage yield (T/A)
7.9
1.7
2.9
13.5
Forage moisture (%)
62.3
6.1
37.4
84.8
NDFd (%)
61.6
3.6
46.4
80.6
Starch (%)
31.4
5.3
5.1
47.2
Crude Protein (%)
7.3
0.8
5.0
10.9
NDF (%)
46.4
4.4
34.6
61.2
Source: Lauer, 2004
50
Proceedings of the 2005 Idaho Alfalfa and Forage Conference
Table 5. Impact of various corn silage hybrids on estimated milk per ton and per acre
using Milk2000.
Milk
Milk
Milk
Milk
1
lbs/acre
$/acre1
Hybrid
n
lbs/ton DM
$/ton DM
Bmr
12
3410
358
21500
2258
Bt
130
3140
330
25000
2625
High oil
12
3040
319
22500
2363
Nutri-Dense
10
3240
340
24300
2552
Leafy
70
3110
327
24600
2583
Waxy
56
3090
325
22600
2373
All hybrids
2407
3110
327
24400
2562
Source: Lauer, 1995-2000 UW Silage Trials (Cited by Shaver)
1
Calculated using a $10.50 milk price.
Table 6. Example variety trial data with yield, milk/ton and milk/acre selection variables.
Hybrid
DM Yield
NDFd %
Starch %
Milk/ton
Milk/acre
A
9.5
59
23
3250
30800
B
9.3
61
37
3360
31400
C
9.4
56
34
3190
30300
D
6.6
68
31
3720
24500
E
9.7
65
34
3560
34500
F
10.1
62
35
3480
35400
G
10.7
62
37
3380
36100
Trial Avg
9.4
62
33
3420
31857
Table 7. Comparison between the Minnesota Method, Modified Pricer, and Feedval4 spreadsheet for estimating relative values for corn silages varying in starch and NDFd %.
MN Method
Modified Pricer
Feedval4
Starch
NDFd
$/ton
$/wet
$/ton
$/wet
$/ton
$/wet
%
%
DM
ton
DM
ton
DM
ton
25
62
59.02
20.66
59.18
20.71
72.60
25.50
30
62
66.52
23.28
60.50
21.17
73.70
25.84
35
62
74.02
25.91
61.81
21.64
74.80
26.18
31
57
67.62
23.67
57.87
20.25
72.20
25.28
31
62
68.02
23.81
60.50
21.17
73.80
25.85
31
67
68.43
23.95
63.13
22.10
75.40
26.41
Assumed silage quality: 35% DM, 7% CP, 47% NDF, 1.3% NDFCP, 4.3% Ash, and 3.2% Fat
51