 |
|
|
THE WISCONSIN INTEGRATED
CROPPING SYSTEMS TRIAL: I. CORN YIELDS AND YIELD VARIABILITY (1992-2002).
Josh Posner [1],
Jon Baldock[1] and Janet Hedtcke[1]
INTRODUCTION AND OBJECTIVES
Southern Wisconsin and much of the upper Midwest was
home to mixed grain and livestock production systems from the 1880’s
to early 1960 (Dorner, 1986). Since that time however, with the introduction
of herbicides and chemical fertilizers, farms have tended to become primarily
focused on either livestock or annual grain production systems. At the
same time that this trend in specialized rather than mixed farming has
been taking place, there has been increased scrutiny of farms as sources
of non-point pollution. In 1990, in response to growing public concern
about the environmental impact of this changing agricultural model, a
large scale (25 ha), long-term study entitled the Wisconsin Integrated
Cropping Systems Trial (WICST) was initiated at two locations in southern Wisconsin. The
purpose of the project was to compare alternative grain and forage-based
systems using three performance criteria: 1) productivity; 2) profitability;
and, 3) environmental impact.
The systems were designed to test the agroecological hypothesis that
with increased biological complexity, agricultural systems could be maintained
highly productive but with less reliance on external inputs (Altieri,
1985; Harwood, 1985). As a result, WICST is a nested factorial with
two enterprise types (annual grain production or forage-based livestock
production) and within each, three production strategies: a) low crop
diversity with high inputs; b) medium crop diversity and medium inputs;
and, c) high crop diversity and low inputs (see Figure
1). To adequately test this cropping diversity hypothesis, it was
not possible to either; 1) fix cropping sequence and only vary input
levels (an input management trial); nor, 2) fix input levels and only
vary cropping sequence (a crop rotation trial). In this study, production
management strategies are being compared, so the two factors, sequence
and input level, are fixed simultaneously to represent realistic cropping
systems. This confounding of crop rotation and input level is both
the strength and weakness of the WICST project. It makes it difficult
to identify the exact factor causing variability between treatments,
yet it is the only way to compare production strategies.
The purpose of this, and subsequent articles are to summarize the findings
of the first 12 years of research (1990-2001) on the WICST plots. To
manage the trial, a team consisting of farmers, technical service providers,
extension agents, the two farm managers and researchers from the Michael
Fields Agricultural Institute and University of Wisconsin College of
Agriculture and Life Sciences was formed. A description of how the team
was formed is available in earlier publications (Posner et. al, 1992,
Stevenson et al. 1994).
An obvious initial criterion in comparing production systems is crop
productivity. Three fundamental research questions were asked:
Question #1. Do the low input, biologically
diverse production systems have lower yields than the high input production
systems?
Question #2. Do the low input, biologically
diverse systems have greater annual yield variability than the high input
systems? And
Question #3. If the low input, biologically
diverse systems do have lower yields, do they gradually increase in productivity
over time, approaching that of the high input systems.
At the time this work was started, most of the literature suggested
that organic systems would be less productive than the higher input systems
(Berardi, 1978; Crosson and Ostrov, 1990; Helmers et al., 1986; Klepper
et al, 1977). For example, a survey conducted in Ohio in 1990 indicted
that the certified organic field crop producers (n=19) had yields equivalent
to 72% for corn, 80% for soybeans, 70% for wheat and 68% for hay of their
conventional farming counterparts in the Ohio Farm Household Longitudinal
Survey (n=960) (Batte et al, 1993). However, some studies were available
indicating that organic yields were nearly equivalent to conventional
yields (Lockeretz et al. 1978; Lockeretz et al., 1981; Cacek and Langner,
1986). The former argument, by the mid-90’s, was being used by some
(Avery and Avery, 1996) to argue that the agricultural research focus
must be kept on high input agriculture, as shifting to organic agriculture
would result in massive food shortfalls in the future.
There was less information available about the impact of cropping system
on yield stability. It had been reported, however, that organic systems
did not show increased variability in net returns (Helmers et al., 1986). Nevertheless,
discussions with growers indicated that they were particularly concerned
about the potential for increased weed pressure and reduced nutrient
availability in the low input or organic systems, resulting in economic
losses in some years. And, there was a general consensus that shifting
to organic systems would require a period of transition before yields
would rise. Initial analyses on the Rodale Conversion Trial (1981-1985)
indicated that corn yields were only 75% of conventional yields during
the first four years of the study (Liebhardt et al., 1989). Duffy and
colleagues (1989) did an economic study of the alternative “starting
points” of the organic rotation and concluded that these systems needed
to start with low input crops (small grains) or nitrogen fixing crops
(legumes).
The objective of this report is to compare the yields, variability of
yields, and yield time trends of the cropping systems in the WICST. The
last objective is particularly important in view of the expectation that
cropping systems, especially those with low external inputs, have a period
of transition before reaching more stable output levels (Cady, 1991;
Liebhart et al., 1989; Dabbert and Madden, 1986). Also, it would be
counterproductive to calculate and discuss simple means of crop yields
averaged over time if time trends existed showing increasing or decreasing
yields. As a final note, crop yield is clearly not the only factor to
consider in designing sustainable cropping systems, but it is among the
key items and other reports will examine additional important factors.
MATERIALS AND METHODS
Cropping System Trials and Terminology
By “cropping system” we mean the combination of a
crop rotation and a management philosophy. This is a slightly more general
definition than the one implied, but not stated, in Cady (1991) that
a cropping system is the combination of a crop rotation and a set of
specific management practices. Substitution of a management philosophy
for specific practices allows the flexibility to keep up with rapidly
changing technologies such as changes in varieties, weed control, and
row spacings. We used a panel of farmers and researchers to guide such
changes and ensure that they were consistent with the overall philosophy
of each system (Posner et al., 1992; Posner, Casler and Baldock. 1995). Cropping
system trials such as these may be viewed as fractions of the full factorial
combinations of crop rotations by treatments discussed in Patterson (1965)
and Cady (1991). By choosing only the treatment combinations that are
appropriate for each crop rotation, every cropping system is compared
near its optimal level and problem of impractically large full factorials
that Cady (1991) warned of are avoided. This flexibility and efficient
size of cropping systems trials comes at a price however; that is, the
ability to identify specific causes of differences among systems is mostly
lost. Definitions of the terms: crop rotation, crop sequence, phase,
cycle, and test crop; are consistent with those originated by Cochran
(1939) and Yates (1949 and 1954). Cady (1991) provides a more recent
statement of them.
Experimental Design and Establishment
The WICST study consists of six cropping systems, replicated four times,
at two sites in southern Wisconsin (Fig. 1a). The
cropping systems include three cash grain systems and three forage systems
and within each group the crop rotation ranges from a monocrop to a more
diverse crop mixture and the external inputs from high to low (Table
1 and Fig. 1). Although CS5 is nominally a medium
external input system, it is a low external input system relative to
most conventional systems. Rainfall and growing degree-day units are
reported on Table 2 for the Arlington site and Table
3 for the Lakeland site.
The trials were established in 1989 with each site (25 ha) planted to
corn to improve the uniformity and allow baseline measurements to be
taken. In addition, this homogeneity year was used to identify an adequate
sampling procedure for baseline variables and to block the trial. Based
on the 1989 corn yield map, the trials were laid out with four blocks
and 14 plots (each about 0.3 ha), representing the 14 total phases in
the six systems. Three-meter grass strips separated plots. Due to the
size of the plots, all fieldwork was done with field size equipment requiring
that 17 m alleyways separate the blocks.
Except during the staggered start, which was completed in 1993 (Posner,
Casler, Baldock, 1995), every phase was present every year for all the
crop rotations, thus meeting a core requirement of a crop rotation trial
(Cady, 1991). The staggered start was used to replicate each phase of
the crop rotations in time, thus providing a more powerful analysis of
time trends by cycle than could be accomplished with an even start of
all possible crop sequences for a crop rotation. Posner et al. (1992)
and Posner, Casler, and Baldock (1995) provide additional details on
the design and conduct of the WISCT.
Sites
South central and southeastern Wisconsin primarily lie in Major Land
Resource area 95B (Fig. 1a). (U.S. Dept. of Agric.,
1981). Soils within this land unit are primarily prairie-derived soils
and vary along two gradients, the depth of silt loam loess cap over glacial
till, and internal soil drainage. One site is the somewhat poorly drained
Lakeland Agricultural Complex (LAC), on the Walworth County Farm, and
the other is a well-drained site at the University of Wisconsin Arlington
Research Station (ARS). Both sites had been in a primarily dairy rotation
of corn and alfalfa with manure for the 20 years prior to establishing
the trial. As a result, both sites had high organic matter levels (42-44g/kg),
soil pH levels around 6.8, and high soil test phosphorus (160 mg/kg Bray
I) and soil test potassium (220 mg/kg exchangeable K) values. Approximately
70 miles separate the two sites. A summary of the weather characteristics
over the period of the study is presented in Tables
2 and 3.
Statistical Analyses
RESULTS AND DISCUSSION
Corn Yields
The yields reported in Table 4 were
calculated considering all effects (year, system and blocks) as fixed
effects. As a consequence, these results are descriptive and only apply
to the research sites and the years reported on (1992-2002). The analysis
indicates that corn yields were generally 30 to 40 bu/a higher at Arlington
than Lakeland. The most likely reason for this large difference is that
there were a number of wet and/or cold springs (’92, ‘93, ‘96)
during the trial, and grain yields were low at the more poorly drained
Lakeland site (see Table 3 and Table
4).
To address the question of relative yields of the low and high input
systems (Question #1), a series of linear contrasts
were made. The linear contrast comparing the organic, cash grain system
(CS3) to the higher input cash grain systems (CS1 and CS2) showed a significant
advantage of 19 bu/a at Arlington (P<0.01) and an advantage of 12
bu/a at Lakeland (P<0.01). In the forage systems, the benefit for
high inputs (CS4) over organic inputs (CS5) was 19 bu/a at Arlington
(P<0.01), while at Lakeland it was 25 bu/a (P<0.01).
Corn Yield Variability Over Years
At the outset of the trials it was anticipated that
the low external input systems might show greater variability than the
higher input systems (Question #2). There are
several possible measures of the yield variability. The Mean Square
Error from standard ANOVAs would estimate the plot-to-plot variability,
which would correspond to field-to-field variation. While the disparity
between fields may be of some interest, most farmers would be more concerned
about the variation in total yields from year-to-year. One measure of
the latter is the standard error of the mean calculated in a PROC MIXED
analysis of the model
Yieldij = m + yeari + eij
When the above is used with years as a random factor,
the standard error, or variance, of the overall mean is a measure of
the variability predicted for yields in any year (not just those included
in the study). Table 5 gives the variances of
the mean yields for each system at ARS and LAC. There is some hint of
greater variation with the low input system, CS3, at ARS; however, the
difference is too small to be statistically significant at either site
with Bartlett’s test on equal variances (P=0.8 at ARS and P=1.0 at LAC). Thus,
there the data from the WICST trials do not support the contention that
the low input systems are more variable in corn grain yield than the
higher input systems.
Arlington site. Figure 2 shows the
mean corn yields for CS1 and CS2 for 6 cycles. Initially CS2 yields
were greater than those in CS1, however in the last two cycles, CS1 out
performed CS2. This difference led to significantly different slopes
when linear regression models were fit to the two cropping systems. But
the quadratic model also fit the data and in that model there was no
significant difference between the two systems. The slight decrease
in yields for Cycles 2 and 3 were largely due to the cold, wet weather
in May and June 1992 and 1993 (Tables 2a, 2b). CS1
yields substantially increased with better weather in the later cycles. On
the other hand, CS2 corn yields showed only a small increase in the last
two cycles. This lack of response may have been due to the switch to
a fully no till system in 1994. Differentiating the quadratic model
provides an equation to estimate the slope, that is the change in yields
with time. Using this equation the slope over the last three cycles
(Cycles 4-6) averaged 6.0 bu/a/yr. Additional cycles should determine
if the two systems really do yield similarly in the long run or if CS1
corn yield will continue to improve relative to those in CS2.
Figure 3 illustrates the mean
corn yields for CS1 and CS3 for 4 cycles. Although the data look as
if the quadratic model would provide a good fit as it did for CS1 and
CS2, none of the terms in the quadratic model were statistically different
from zero. Also, the linear slope term for CS3 was not statistically
significant. That leaves the overall CS3 mean, 146 bu/a, as the best
estimate of the corn yields in this system. Regrouping the CS1 data
as if it had 3 phases had little effect on the estimate of increasing
yields. The estimate of a 23.7 bushel/acres increase per cycle, which
translates into a 7.9 bushel/acre increase per year is very close to
that calculated for CS1 above. As a result of the increasing corn yields
for CS1 and the flat yields for CS3 these data project an increasing
advantage for CS1 over CS3. Possible reasons for the flat CS3 productivity
include insufficient N, low availability of soil P and K in cold springs,
and weed competition.
Support for these causes of the constant CS3 yields
comes from the linearly increasing corn yields in the other three-phase
rotation CS5 (Fig. 4). Alfalfa preceding the corn
provided substantially larger N credit in CS5 than did the red clover
cover crop in CS3. Also the applications of manure in CS5 provided N,
P, and K not added in CS3. Finally, the repeated cuttings of the alfalfa
reduced the annual weed pressure in CS5. These advantages led to a mean
yield of 166 bu/a compared to the 144 for CS3 and an average increase
of 11.1 bu/a per cycle (3.7 bu/a/year) versus none for CS3. Nevertheless,
the increase was not as large as in CS1 causing an interaction in which
CS1 yielded less than CS5 initially, but more than CS5 by the end of
the report period (Fig. 4).
A similar interaction occurred between CS1 and CS4
corn yields (Fig. 5). The regression equations predict
a 34-bu/a advantage for CS4 in Cycle 1. However, the advantage is predicted
to switch to CS1 by Cycle 5 despite the fact that both are high external
input systems and CS4 has the benefit of being in a crop rotation with
alfalfa. It will be interesting to see if this crossover materializes
as predicted.
Lakeland site. Due to increased interest
in organic rotation possibilities in SE Wisconsin changes were made in
the two rotations in 2000. As a result, the data can only be analyzed
between 1992 and 1999. In addition, due to the very wet spring in 1996
(Table 3a, 3b), corn was
planted and replanted several times and did not produce an adequate stand
until the June 25 planting. When the corn grain yields for 1996 were
omitted, CS1 and CS2 corn yields at Lakeland followed a pattern of increasing
yields similar to that found at Arlington (Fig. 6). Moreover,
there was no significant difference between the slopes for separate linear
regressions. Thus, a common slope was estimated at 12.9 bu/a per cycle
or 6.5 bu/a/year. This value is very close to the 7.4 bu/a/year increase
found at Arlington for these cropping systems. The difference in intercepts
of 8.3 bu/a was nearly significant.
The very low corn grain yields in 1996 caused an
apparent decrease in productivity for CS1, CS3, and CS5 when the 3-phase
systems were compared. Also, the REML estimates from PROC MIXED were
unstable for these systems, perhaps because there were only two cycles. However,
when the 1996 data was omitted and the least squares estimates from PROC
GLM were used, there were no significant trends in productivity although
CS3 and CS5 showed a small increase for Cycle 2 compared to Cycle 1 (Fig.
7). The reasons for the lack of significant increases in productivity
for CS3 are mostly likely the same as cited for Arlington, but it is
unclear why CS1 and CS5 did not show the same increase in productivity
at both locations.
CONCLUSIONS
We started this yield analysis with three questions: 1) do the low input,
biologically diverse production systems have lower yields than the high
input production systems; 2) do the low input, biologically diverse systems
have greater annual yield variability than the high input systems; and
3) if the low input, biologically diverse systems do have lower yields,
do they gradually increase in productivity over time, approaching that
of the high input systems. Our findings, initially only looking only
at the corn phase in five of the rotations were not as we anticipated.
- Using a fixed effects model, we found that during the reporting period,
the low input systems did result in lower average corn yields—between
10 and 20 bu/a less in the high input grain systems
- The analysis of corn yields over the 12-year period indicates that
the lower input systems do not have significantly higher yield variability
than the high input systems. Although there is a hint at Arlington
that CS3 corn yields were the most variable, it was not significantly
higher than in the other systems. Variability in the more poorly drained
Lakeland site was approximately twice that of the better-drained Arlington
site.
- Corn yields in both the continuous corn and no-till corn-soybean
systems at Arlington showed an upward annual yield trend of 6.0 to
7.5 bu/a/year over Cycles 4-6 and at Lakeland an upward yield trend
of 6.5 bu/a/yr over cycles 1-4. The similarity of the yields and the
yield trends was unexpected as we anticipated that with Best Management
Practices, the corn yields in rotation with soybeans (CS2) would exceed
those of continuous corn (CS1).
- Corn yields in the organic system did not show an upward trend at
either site, thus the overall means adequately characterize the CS3
yields over time (Arlington 146 bu/a; Lakeland 116 bu/a). Again, this
was unexpected as we anticipated that our learning how to manage corn
organically and the gradual change is the soil flora and fauna would
favor a gradual increase in organic corn yields.
- Corn yields from the systems with manure and plow down alfalfa were
generally the highest at both locations. This was expected. What
was surprising at Arlington is that the rate of yield increase for
continuous corn (CS1) was actually higher than with the organic forage
rotation (CS5) and the high input forage rotation (CS4). It will be
interesting to see if the crossover in yields actually materializes
in the future and continuous corn regularly out yields corn in the
forage based rotations.
Fig. 1. Schematic diagram of the 6 cropping systems
of WICST.
Fig 1a. Outline of major Land Resource Area 95B
and the two sites of WICST.
Fig 2. Mean corn yield for CS1 and CS2 by cycle at
Arlington.
Fig. 3. Mean corn yields for CS1 and CS3 by cycle
at Arlington.
Fig. 4. Mean corn yields for CS1 and CS5 by cycle
at Arlington.
Fig. 5. Mean corn yields for CS1 and CS4 by cycle
at Arlington.
Fig. 6. Mean corn yields for CS1 and CS2 by cycle
at Lakeland.
Fig. 7. Mean corn yields for CS1 and CS3 by cycle
at Lakeland.
|
|
Table 1. Details of cropping
systems.
Cash grain systems:
CS1: A monocropped corn (Zea mays) (C) sequence with high
external inputs. A 105-day RM corn hybrid was used. Weeds
were managed with interrow cultivation and herbicides. Insecticides
were annually applied to control corn rootworm (Diabrotica
vigifera) populations. In addition to starter fertilizer,
nitrogen applications were based on Early Spring Nitrate
Test results. After corn harvest, the residues were chopped
and incorporated by chisel plow in the fall.
CS2: A two-phase soybean (Glycine max) (S) corn rotation
denoted S-C with medium external inputs. 2.3 maturity
class soybean varieties and 105-day RM corn hybrids were
used. This rotation was managed with no-till practices
and weeds were controlled with herbicides. In addition
to starter fertilizer, nitrogen application to the corn
phase was based on Pre-Sidedress Nitrate test and the soybeans
were drilled.
CS3: A three-phase rotation of soybean, winter wheat (Triticum
aestivum) (W) with red clover (Trifolium pratense)
cover crop, and corn denoted S-Wrc-C with low external
inputs. A 95-day RM corn hybrid, 1.9 maturity class
soybean variety, and a winter hardy wheat variety were
used. This rotation was managed without chemicals, and
the weeds during corn and soybean were controlled with
rotary hoeing (2 to 4x) and mechanical cultivation (2x). The
fields were chisel plowed after the corn phase. The
corn and soybeans were planted on 76 cm rows and the
winter wheat was drilled after the soybean harvest during
the first two weeks of October. The red clover was seeded
the following winter while the ground was frozen. The
clover biomass was under cut in the fall, and incorporated
prior to corn planting.
Forage-based management summary:
CS4: A four-phase rotation of three years of alfalfa (Medicago
sativa) (A) followed by corn denoted A-A-A-C. The
alfalfa was solo seeded. Weeds were controlled with
herbicides and interrow cultivations during corn. Pyrethroid
insecticides were occasionally applied to control insect
pests. Manure, at the rate of 20 tons acre-1 (15%
d.m., 5 kg N ton-1), was applied in mid November
following the last phase of alfalfa and following corn. A
chisel plow with sweeps incorporated the manure and undercut
the alfalfa roots. This rotation was dropped from LAC
in 1999.
CS5: A three-phase rotation of alfalfa companion seeded with oats
(Avena sativa) and peas (Pisum sativum) (A/O),
alfalfa, and corn denoted A/O-A-C with medium external
inputs. There were no chemical inputs in this system and
weeds were managed by interrow cultivations during corn. Manure,
at the rate of 15 tons acre-1 (15% d.m., 5 kg N
ton-1), was applied and incorporated by chisel
plow after the corn harvest and after the last phase of
established forage. This rotation was dropped from LAC
in 1999.
CS6: A pasture mix of red clover, timothy (Phleum pratense)
and brome (Bromus inermis) denoted P with low external
inputs. The paddocks at both sites were rotationally grazed
with four to five heifers from April to October. The red
clover was periodically re-seeded and the pasture was completely
reestablished at LAC in 1997 due to poor stand quality.
|
|
Table 2. Weather data at Arlington, 1990-2002
vs. 30-yr average.
| 2A.Rainfall data summary (inches) at Arlington (1990-2002
vs. 1961-1990 average) |
| |
1990 |
1991 |
1992 |
1993 |
1994 |
1995 |
1996 |
1997 |
1998 |
1999 |
2000 |
2001 |
2002 |
13-yr mean |
30-yr mean |
| April |
2.49 |
4.52 |
3.96 |
7.06 |
2.28 |
3.37 |
2.64 |
0.65 |
3.71 |
5.95 |
3.38 |
3.14 |
3.30 |
3.57 |
2.84 |
| May |
4.25 |
1.91 |
1.22 |
4.52 |
1.99 |
5.95 |
3.20 |
3.30 |
4.06 |
4.22 |
10.46 |
4.70 |
2.99 |
4.06 |
3.13 |
| June |
6.32 |
2.63 |
1.19 |
6.10 |
7.93 |
2.15 |
7.76 |
4.86 |
6.81 |
4.17 |
7.17 |
6.98 |
4.36 |
5.26 |
3.80 |
| July |
1.57 |
3.75 |
5.80 |
9.40 |
6.08 |
2.81 |
2.42 |
6.00 |
2.14 |
3.43 |
3.43 |
2.92 |
2.95 |
4.05 |
3.42 |
| Aug |
5.36 |
1.78 |
1.91 |
3.20 |
4.03 |
5.02 |
2.83 |
3.20 |
6.71 |
2.53 |
3.35 |
5.39 |
2.93 |
3.71 |
3.88 |
| Sept |
1.22 |
4.70 |
7.46 |
4.20 |
4.65 |
1.78 |
0.86 |
1.61 |
2.98 |
1.41 |
3.14 |
5.22 |
1.94 |
3.17 |
4.28 |
| Oct |
2.29 |
6.75 |
1.26 |
1.17 |
0.50 |
4.19 |
3.29 |
1.35 |
3.41 |
1.38 |
0.80 |
1.65 |
3.92 |
2.46 |
2.38 |
| 2B. Growing Degree Days (base 50) at Arlington (1990-2002
vs. 1961-1990 average) |
| |
1990 |
1991 |
1992 |
1993 |
1994 |
1995 |
1996 |
1997 |
1998 |
1999 |
2000 |
2001 |
2002 |
13-yr mean |
30-yr mean |
| April |
202 |
207 |
85 |
95 |
182 |
90 |
123 |
108 |
137 |
128 |
126 |
184 |
164 |
141 |
148 |
| May |
274 |
461 |
366 |
326 |
346 |
302 |
247 |
192 |
385 |
362 |
386 |
346 |
254 |
327 |
341 |
| June |
552 |
594 |
458 |
449 |
542 |
636 |
535 |
549 |
510 |
805 |
457 |
334 |
272 |
515 |
516 |
| July |
616 |
629 |
492 |
588 |
540 |
696 |
574 |
597 |
655 |
736 |
595 |
674 |
737 |
625 |
645 |
| Aug |
592 |
594 |
475 |
596 |
526 |
775 |
607 |
455 |
651 |
514 |
638 |
639 |
615 |
590 |
583 |
| Sept |
459 |
387 |
343 |
266 |
442 |
371 |
408 |
383 |
478 |
357 |
414 |
324 |
464 |
392 |
383 |
| Oct |
189 |
194 |
167 |
168 |
212 |
217 |
202 |
218 |
175 |
171 |
263 |
325 |
459 |
227 |
184 |
Table 3. Weather data at Lakeland, 1990-2002
vs. 30-yr average.
| 3A.Rainfall data summary (inches) at Lakeland (1990-2002
vs. 1961-1990 average) |
| |
1990 |
1991 |
1992 |
1993 |
1994 |
1995 |
1996 |
1997 |
1998 |
1999 |
2000 |
2001 |
2002 |
13-yr mean |
30-yr mean |
| April |
3.40 |
4.15 |
2.17 |
5.55 |
2.37 |
4.75 |
2.58 |
2.64 |
3.60 |
5.64 |
4.41 |
3.46 |
3.59 |
3.72 |
3.66 |
| May |
5.53 |
2.32 |
0.89 |
1.90 |
0.66 |
2.44 |
5.69 |
3.15 |
4.61 |
5.15 |
10.59 |
3.75 |
2.57 |
3.79 |
3.25 |
| June |
5.27 |
1.56 |
1.34 |
8.50 |
3.71 |
1.57 |
6.61 |
5.36 |
6.76 |
8.79 |
4.14 |
6.19 |
4.56 |
4.95 |
3.88 |
| July |
2.51 |
2.45 |
7.07 |
4.35 |
1.97 |
3.13 |
3.24 |
3.20 |
1.65 |
1.02 |
2.69 |
2.52 |
0.49 |
2.79 |
4.30 |
| Aug |
3.93 |
2.04 |
2.56 |
2.80 |
4.48 |
8.34 |
1.93 |
2.03 |
2.17 |
2.42 |
3.50 |
4.75 |
5.36 |
3.56 |
3.92 |
| Sept |
0.96 |
4.94 |
4.36 |
3.44 |
1.90 |
2.08 |
1.57 |
4.25 |
2.66 |
4.23 |
4.70 |
6.93 |
4.21 |
3.56 |
4.09 |
| Oct |
3.38 |
6.21 |
0.89 |
0.50 |
0.76 |
4.22 |
2.56 |
1.82 |
4.20 |
0.98 |
1.19 |
4.51 |
3.23 |
2.65 |
2.74 |
| 3B. Growing Degree Days (base 50) at Lakeland (1990-2002
vs. 1961-1990 average) |
| |
1990 |
1991 |
1992 |
1993 |
1994 |
1995 |
1996 |
1997 |
1998 |
1999 |
2000 |
2001 |
2002 |
13-yr mean |
30-yr mean |
| April |
191 |
165 |
110 |
86 |
165 |
59 |
117 |
135 |
164 |
154 |
157 |
236 |
188 |
148 |
158 |
| May |
182 |
458 |
331 |
328 |
338 |
261 |
219 |
230 |
462 |
410 |
412 |
384 |
287 |
331 |
365 |
| June |
535 |
617 |
462 |
460 |
562 |
585 |
454 |
575 |
569 |
598 |
548 |
534 |
619 |
547 |
560 |
| July |
601 |
658 |
555 |
653 |
605 |
695 |
673 |
693 |
732 |
847 |
662 |
737 |
804 |
686 |
696 |
| Aug |
614 |
628 |
452 |
630 |
490 |
751 |
681 |
594 |
749 |
648 |
680 |
702 |
724 |
642 |
646 |
| Sept |
437 |
399 |
348 |
250 |
430 |
343 |
416 |
459 |
569 |
448 |
451 |
399 |
548 |
423 |
436 |
| Oct |
159 |
174 |
153 |
156 |
202 |
173 |
198 |
268 |
467 |
241 |
298 |
193 |
149 |
218 |
210 |
Table 4. Mean corn grain yields by cropping
system.
| System |
Arlington, bu/a† |
Lakeland, bu/a‡ |
| CS1 |
162 |
125 |
| CS2 |
168 |
134 |
| CS3 |
146 |
117 |
| CS4 |
185 |
146 |
| CS5 |
166 |
121 |
| Mean |
165 |
128 |
| LSD (10%) |
3.6 |
7.6 |
† Mean yields 1993 through 2002. ‡ Mean yields
1993 through 1998.
Table 5. Variability of mean corn yields by
year within a site.
| System |
Variance of mean
|
| |
Arlington
|
Lakeland
|
| CS1 |
80.7
|
421.6
|
| CS2 |
77.8
|
492.6
|
| CS3 |
139.3
|
411.9
|
| CS4 |
69.4
|
439.1
|
| CS5 |
74.9
|
463.7
|
LITERATURE CITED
Altieri, M.A. 1985. Diversification of agricultural
landscapes: A vital element for pest control in sustainable agriculture.
In T.C. Edens, C. Fridgen, and S. Battenfield (Eds) Sustainable agricultural
and integrated farming systems. Michigan State Univ. Press. East Lansing.
166-184
Avery D. T. and A. Avery. 1996. Farming to
sustain the environment. Hudson Briefing Paper Hudson Institute Indianapolis,
IN. May 1996 Number 190
Batte, M.T., D.L. Forster, and F. J. Hizhusen. 1993. Organic
agriculture in Ohio: An economic perspective. J. Prod. Agric. 5:4:536-542
Berardi, G. M. 1978. Organic and conventional
wheat production: Examination of energy and economics. Agro-Ecosystem
4: 367-76
Cacek, T. and L.L. Langer. 1986. The economic
implications of organic farming. J. of Alternative Agriculture. 1:1:25-29
Cady, Foster B. 1991. Experimental design
and data management of rotation experiments. Agron. J. 83:50-56.
Cochran, W.G. 1939. Long-term agricultural experiments. Supplement
to the Journal of the Royal Statistical Society (London) VI: 104-140.
Crosson, P., and J. Ostrov. 1990. Sorting
out the environmental benefits of alternative agriculture. J. of Soil
and Water Conser. 45:34-41.
Dabbert, Stephan, and Patrick Madden. 1986. The
transition to organic agriculture: A multi-year simulation model of a
Pennsylvania farm. Amer. J. Alternative Agric. 1:99-107.
Doner, P. 1986. Changing structure of Wisconsin
family farming. Wisconsin Academy Review. December 33:1:26-28
Duffy, M., G. Ginder and S. Nicholson. 1989.
An economic analysis of the Rodale Conversion Project: Overview, Staff
Paper No. 212 Dept. of Economics, Iowa State Univ. Ames, Iowa.
Harwood, R. 1985. The integration efficiencies
of cropping systems. In T.C. Edens, C. Fridgen, and S. Battenfield (Eds)
Sustainable agricultural and integrated farming systems. Michigan State
Univ. Press. East Lansing. 64-75
Helmers, G.A., M.R. Langemeier, and J. Atwood. 1986. An
economic analysis of alternative cropping systems for east-central Nebraska. Am.
J. of Alt Ag. 1:4:153-158
Klepper, R., W. Lockeretz, B. Commoner, M.
Gertler, S. Fast, D. O’Leary, and R. Blobaum. 1997. Economic performance
and energy intensiveness on organic and conventional farms in the Corn
Belt: A preliminary comparison. Am. J. of Agric. Econ. 59: 1-12
Liebhardt, W.C., R.W. Andrews, M.N. Culik,
R.R. Harwood, R.R. Janke, J.K. Radke, and S.L. Rieger-Schwartz. 1989. Crop
production during conversion from conventional to low-input methods. Agron.
J. 81:150-159.
Little, R.C., G.A. Milliken, W.W. Stroup, and
R.D. Wolfinger. 1996. SAS system for mixed models. SAS Institute Inc. Cary,
NC.
Lockeretz, W., G. Shearer, R. Keppler, and
S. Sweeney. 1978. Field crop production on organic farms in the Midwest. J.
of Soil and Water Conservation 33: 130-34
Lockeretz, William, Georgia Shearer, and Daniel
H Kohl. 1981. Organic farming in the Corn Belt. Science. 211:540-546.
Patterson, H.D. 1965. The factorial combination of treatments in rotation
experiments. J. Agric. Sci. (Cambridge) 65:171-182.
Posner, J.L., L. Cunningham, J. Doll, J. Hall,
D. Mueller, T. Mulder, R. Saxby, and A. Wood. 1992. The Wisconsin integrated
cropping systems trial: Bridging the cap between station research, the
producer and the consumer. Toward a new paradigm for farming systems
research/extension. Michigan State University. Sept. The Association
for farming system research/extension Working paper set. P389-400
Posner, J.L., M.D. Casler, and J.O. Baldock. 1995. Wisconsin
integrated cropping system trial: combining agro-ecology with production
agronomy. J. Alt Agric. 10:98-107.
Singh, Murari, and Michael J. Jones. 2002.
Modeling yield sustainability for different rotations in long-term barley
trials. J. Agric. Biological, and Environ. Statistics. 7:525-535.
Stevenson, G.W., J.L. Posner, J. Hall, L. Cunningham,
and J. Harrison. 1994. Addressing the challenges of sustainable agriculture
research and extension at land-grant universities: Radially organized
teams at Wisconsin. Am. J. of Alt. Agric. 9:1&2:76-83
U.S. Dept. of Agric. 1981. Land resource regions
and major land resource areas of the United States. Handbook 296. Washington
DC.
Wolfinger, Russell D. 1996. Heterogeneous variance-covariance structures
for repeated measures. J. Agricultural, Bioligical, and Environmental
Statistics. 1:205-230.
Yates, F. 1949. The design of rotation experiments. Comm. Bur. Soil
Sci. Technical Communication No. 46. pp 142-155.
Yates, F. 1954. The analysis of experiments containing different crop
rotations. Biometrics 10:324-346.
[1] J. Posner and J. Hedtcke, Dep. of Agronomy, Univ. of Wisconsin,
Madison, WI 53706, J. Baldock, Agstat, Verona, WI 53593. *Corresponding
author (jlposner@facstaff.wisc.edu).
This paper would not have been possible without the efforts of the
other WICST team members, especially John Hall, Jim Stute, Tom Mulder,
and Scott Alt.
|
|
|
