Pedosphere 24(4): 476–486, 2014
ISSN 1002-0160/CN 32-1315/P
c 2014 Soil Science Society of China
Published by Elsevier B.V. and Science Press
Effects of Tillage and Crop Residue Management on Maize Yields
and Net Returns in the Central Mexican Highlands
Under Drought Conditions∗1
R. ROMERO-PEREZGROVAS1,2 , N. VERHULST1 , D. DE LA ROSA3 , V. HERNÁNDEZ1 , M. MAERTENS2 ,
J. DECKERS2 and B. GOVAERTS1,∗2
1 International
Maize and Wheat Improvement Center (CIMMYT), P.O. Box 6-641, Mexico D.F. 06600 (Mexico)
of Earth and Environmental Sciences, University of Leuven, Celestijnenlaan 200 E, Leuven 3001 (Belgium)
3 Mathematics Research Center (CIMAT), Jalisco S/N, Valenciana, Guanajuato, Mexico CP 36240 (Mexico)
2 Department
(Received June 27, 2013; revised October 5, 2013)
ABSTRACT
In the subtropical highlands of Central Mexico, where the main crop is maize (Zea mays), the conventional practice (CP) involves
tillage, monoculture and residue removal, leading to soil degradation and unsustainable use of natural resources and agricultural
inputs. Conservation agriculture (CA) has been proposed as a viable alternative in the region, based on reduction in tillage, retention
of adequate levels of crop residues and soil surface cover and use of crop rotation. This study began in 2009 when the highlands
of Central Mexico suffered from a prolonged drought during vegetative maize growth in July–August, providing an opportunity for
the on-farm comparison of CA with CP under severe drought conditions which 21 climate change models projected to become more
frequent. Under dry conditions, CA resulted in higher yields and net returns per hectare as early as the first and second years after
adoption by farmers. As an average of 27 plots under farmers’ management in 2009, the maize yields were 26% higher under CA (6.3
t ha−1 ) than under CP (5.0 t ha−1 ). 2010 was close to a normal year in terms of rainfall so yields were higher than in 2009 for both
practices; in addition, the yield difference between the practices was reduced to 19% (6.8 t ha −1 for CA vs. 5.7 t ha−1 for CP). When
all the 2009 and 2010 observations were analyzed in a modified stability analysis, CA had an overall positive effect of 3 838 Mexican
Pesos ha−1 (320 $US ha−1 ) on net return and 1.3 t ha−1 on yield. After only one to two years of adoption by farmers on their fields,
CA had higher yields and net returns under dry conditions that were even drier than those predicted by the analyzed 21 climate
change models under a climate change scenario, emission scenario A2.
Key Words:
climate change, conservation agriculture, conventional practice, emission scenario, modified stability analysis
Citation: Romero-Perezgrovas, R., Verhulst, N., De La Rosa, D., Hernández, V., Maertens, M., Deckers, J. and Govaerts, B. 2014.
Effects of tillage and crop residue management on maize yields and net returns in the Central Mexican highlands under drought
conditions. Pedosphere. 24(4): 476–486.
INTRODUCTION
Subtropical highlands of the world have become increasingly densely populated and intensively cropped
over recent centuries. Agricultural sustainability problems resulting from soil erosion and fertility decline
have arisen throughout these agro-ecological zones due
to conventional practices based on mechanical soil disturbance, monoculture and removal of crop residues
(Lal, 1993). Soil degradation and lack of soil cover
lead to extensive erosion and rain runoff events, making the farmers’ production system very vulnerable
to dry conditions which happen periodically. Maintai∗1 Supported
ning crop productivity is further complicated by existing and predicted climate change effects, such as rising temperatures, decreasing rainfall and increasing
frequency and intensity of dry periods (Carvalho and
Jones, 2013). In order to satisfy increasing agricultural
output demands, the subtropical highlands around the
globe need to produce more crops, with less water,
higher temperatures and increasing input costs and
with minimal available land that could be converted
to agriculture. Cropping systems need to be adapted
to face these challenges.
In 1991, the International Maize and Wheat Improvement Center (CIMMYT) established a long-term
by a scholarship from the Mexican National Science Commission (CONACYT), the CGIAR Research Program on Climate
Change Agriculture and Food Security (CCAFS), and the Mexican Secretariat of Agriculture, Livestock, Rural Development, Fisheries
and Food (SAGARPA).
∗2 Corresponding author. E-mail: [email protected].
TILLAGE AND RESIDUE MANAGEMENT EFFECTS ON MAIZE
experiment in the highlands of Central Mexico to develop cropping systems that can preserve natural resources, have higher and stable yields and at the same
time are more robust against drought and more profitable for the farmer (Govaerts et al., 2005). Among
the evaluated practices in the highlands were the conventional practice (CP), based on tillage, monoculture
of maize and removal of crop residues for use as animal fodder, grazing or otherwise burned, and the conservation agriculture (CA), based on i) reduction in
tillage to zero tillage; ii) retention of adequate levels
of crop residues for soil surface cover and iii) use of
crop rotation. CA increased physical, chemical and biological soil quality compared to CP (Govaerts et al.,
2006a, b, 2007, 2008, 2009). Aggregation and infiltration were higher under CA than CP (Govaerts et al.,
2009), resulting in higher soil water content under CA,
which allowed for the buffering of short dry periods
(Verhulst et al., 2011). Since periodic droughts were an
important constraint in rainfed conditions in the highlands, CA had higher and more stable yields than CP
(Govaerts et al., 2005; Verhulst et al., 2011). However,
there was a lack of information regarding the results of
these practices under farmers’ conditions and management, and it was questioned if CA cropping systems
are an economically sound solution for farmers outside
the experimental stations. Moreover, it is important
to assess if farmers practicing CA can have positive
economic (net returns) results in the short term that
can help to quickly overcome the investments of the
technological change.
In 2007–2008, an extension program led by CIMMYT started working with farmers in the highlands of
Central Mexico to adapt and compare CA-based practices with the conventional farmers’ practices, both under the management of the same farmer. Plots were
established with farmers in their fields, conceived as
a medium- to long-term comparison and as training
spots for neighboring farmers for a large-scale adoption
strategy. The intense drought in 2009 allowed for the
first time in this region a comparison of CA with CP in
farmers’ fields and under climate conditions predicted
to become more common in the whole of Mexico and
North America (Carvahlo and Jones, 2013). In 2010,
the rainfall was close to the historical long-term average, which also allowed comparison between the CP
and CA systems in a scenario similar to an average
year.
The objective of this study was to evaluate whether
CA results in the short term in higher yields and net
returns for farmers in the highlands of Central Mexico under different climatic and production conditions,
477
in a year with extreme drought and a close-to-average
rainfall year. A secondary objective was to compare the
2009 drought with the projections of climate change
models to assess if the drought was as severe as the
projected rainfall patterns for the region under a climate change scenario A2.
MATERIALS AND METHODS
Study area
This study was conducted in the highlands of Central Mexico, in an area that comprises the States of Hidalgo and Mexico, which are located in 18◦ 21 –20◦ 23
N and 98◦ 35 –100◦ 17 W. The climate is sub-humid
with some variations, and temperatures are semi-hot
to temperate (Améndola et al., 2006). A 4–6 month
wet season (May to September), with rainfalls on average between 350 and 800 mm, was followed by a
dry winter season (Sayre et al., 2001). The dominant soil groups are Andosols, Phaeozems, Vertisols,
Regosols and Cambisols (Sotelo-Ruı́z et al., 2011). The
area has been intensely cropped for many centuries,
in particular in high valleys (1 500–3 000 m above sea
level). However, soil erosion and decline in soil chemical and physical fertility have led to serious problems of production sustainability. Rainfed agriculture
is the most common practice, but irrigation is also applied in a minority of areas. One of the main water
sources for irrigation is sewage from the urban centers
(Jiménez, 2005). Crops are planted before the start
of the main rainfall season. The rains are often intense and short and there can be significant drought
periods with crop water stress during the growing
season (Sayre et al., 2001). Maize (Zea Mays) is the
main crop in terms of cultivated area and production
volume. Beans (Phaseolus vulgaris), wheat (Triticum
aestivum), barley (Hordeum vulgare) and oats (Avena
sativa) are also grown (Sayre et al., 2001; Fischer et al.,
2002). Only one cycle for grain cultivation is possible
under rainfed conditions, i.e., during the rainy season,
while two grain harvests a year are obtained under irrigation.
In the subtropical highlands of Central Mexico, farmers applying conventional practices (CP) under rainfed conditions leave the soil bare for more than six
months each year because almost all crop residues are
directly removed for fodder, grazed, and/or burned.
Fields are subject to tillage, mainly with small tractordrawn disc plows/harrows and field cultivators after
harvest, during the dry season for weed control and
in the weeks before sowing. Sub-soiling to a depth of
60–70 cm is increasingly used by farmers in the region.
478
R. ROMERO-PEREZGROVAS et al.
Experimental plots
Economic and agronomic data analysis
The data were gathered from 27 on-farm CA-CP
comparative plots. These plots were established in
2007–2008 in 21 different municipalities of the states
of Mexico and Hidalgo (Fig. 1). They were established
in fields of farmers who were willing to participate and
had contact with a local technician trained in the application of CA. The average plot size for the 27 plots
was 1.5 ha with a minimum of 0.5 ha and a maximum
of 8.2 ha. All plots were divided into two parts as treatments. In one part, the farmers used their conventional
practice and in the other they applied the CA principles as recommended by the technicians. The farmer
had to commit to leave a minimum of 30% residue for
soil cover after zero tillage was implemented. Fertilization, seed quality, plant density and the rest of the
management remained equal for both parts of the plot;
the number of weed control operations was equal and
in most of the cases with the same method (chemical). However, in three plots, CP had mechanical weed
controls, whereas CA had chemical controls (Table I).
The farmers that participated were innovative and
considered good farmers by the local technicians, which
explains why the CP practices had already relatively
high maize yields when compared to the average maize
yield of less than 3 t ha−1 in the region.
The economic analysis was based on partial budgets (Boughton et al., 1990). Not all costs involved in
the production process were measurable (machinery
depreciation, infrastructure investment, tool replacement, etc.). Hence, the assessment focused on costs
varying between CA and CP. Net return was used to
measure and compare profitability between the two
systems. All variable costs of the two systems were
registered by a local technician using a standardized
field book. Variable costs for inputs and labor were
valued at market prices, even when inputs were subsidized or when family labor was used, in order to avoid
distortions when farmers used family labor or free inputs such as fertilizer, seeds or diesel in one of the two
systems.
The grain yield was measured using a randomized
sampling method for both CA and CP. For each plot,
ears were hand harvested in five random samples of 3 m
in length and one row wide, measuring the width of the
row and calculating the sampling area. The ears harvested were dried, shelled and weighed. A subsample
of 100–150 g was oven-dried at 75 ◦ C, and to determine the grain yield, the moisture was at 12% (v/w)
H2 O. The price for the harvest and fodder units was
the actual price received by the farmer. It was the same
Fig. 1 Location of the 21 Central Mexican municipalities in the highlands where 27 conservation agriculture-conventional practice
plots were established.
TILLAGE AND RESIDUE MANAGEMENT EFFECTS ON MAIZE
479
TABLE I
General management of the 27 conservation agriculture (CA)-conventional practice (CP) comparative plots in 21 municipalities of the
Central Mexican highlands
Plot
Fertilization
N
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
59
41
27
0
152
100
0
82
138
122
80
115
200
132
110
184
0
92
98
0
0
0
51
0
0
0
0
P2 O5
kg ha−1
34
0
18
0
57
69
57
92
0
0
0
0
23
55
46
0
0
0
0
0
0
0
0
0
0
0
0
K2 O
0
0
0
0
6
0
0
0
0
0
0
0
47
28
0
0
0
0
0
0
0
0
0
0
0
0
0
No. of
weed
control
operations
Irrigationa)
2d)
1d)
1
1
2
2
2
1
2
1
2d)
1
1
1
2
2
2
2
2
2
2
1
2
2
1
2
2
No
No
No
Partial
Partial
Partial
No
Partial
No
No
No
No
No
Partial
Partial
Partial
Fulle)
Partial
Full
Fulle)
Fulle)
Fulle)
Fulle)
Fulle)
Fulle)
Fulle)
Fulle)
No. of tillage interventions
in CP plots
Percentage of residues left in the
fields
≤ 50 cmb)
> 50 cmc)
CA (cover)
CP (incorporation)
3
1
4
2
3
3
3
3
4
2
2
2
4
3
4
2
2
2
3
2
2
2
2
3
3
1
1
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
1
1
0
0
0
1
1
1
1
1
1
100
100
80
30
30
50
100
30
100
100
100
100
100
30
40
60
30
30
30
30
30
40
30
30
30
50
40
50
0
40
50
30
0
50
0
40
0
0
0
0
0
0
20
0
0
0
0
0
0
0
0
0
0
0
a) Full
= 3 or more irrigations; Partial =1 or 2 irrigations; No = rainfed.
tillage operations such as harrow plow or disc plow.
c) Deeper tillage operations such as sub-soiling.
d) In these plots the weed control operations were differentiated: with chemical methods for CA and with mechanical methods for CP.
e) With urban sewage water.
b) Shallow
price in each plot for both systems, and consequently
no standardized price was used. When part of the fodder or yield was consumed or used by the household,
the unit was accounted for as sold at the same price as
the other units. Crop residues used as soil cover in CA
or incorporated to the soil in CP were not accounted
as sold.
The gross return, variable cost and net return were
calculated as follows:
Ri = xi × pgrain
+ zi × presidue
i
i
(1)
where Ri is the gross return obtained from crop production; i is the farmer number (i = 1, . . . , N ); xi is
is the grain sale price; zi is the
the grain yield; pgrain
i
is the residue unit
number of residue units and presidue
i
price.
cvar
= (cprep
+ csow
+ cfert
+ cweed
+ cins
i
i
i
i
i +
i
+ charv
)
cwat
i
i
(2)
is the variable cost; cperp
is the variable cost
where cvar
i
i
of the mechanical or chemical soil preparation; csow
i
is the variable cost of sowing (seed, machinery rent,
is the variable cost of fertilization;
diesel, etc.); cfert
i
cweed
is
the
variable
cost of weed controls; cins
is the
i
i
variable cost of insecticide products and applications;
is the variable cost of irrigation (electricity, lacwat
i
is the variable cost of
bor and water quota) and charv
i
harvesting and bailing.
NRi =
1
(Ri − cvar
i )
ai
(3)
where NRi is the net return and ai is the area in
hectare.
480
Climate change modelling
Downscaled outputs from 21 Special Report on
Emissions Scenarios (SRES) models were used for analysis, and data were provided by the International
Center for Tropical Agriculture (CIAT) (Ramirez and
Jarvis, 2008). The timeframe chosen spans the period
from 2040 to 2069, and is utilized as an average of
this same period, referred to below as the 2050s. Each
of the scenarios represents a possible version of the
future, differentiated by social, political, environmental and technological developments in global societies
which are supposed to have a potential impact on the
climate. The emission scenario A2 is commonly known
as “business as usual”, as it assumes that emissions of
carbon and other greenhouse gases will not be reduced;
it is considered to be a medium- to high-emission scenario. This scenario is defined by high global population estimates, with no efforts to reduce global warming being employed, high energy use and a more regional approach to solving social and environmental
issues (Nakicenovic et al., 2000). The data were downscaled to a 2.5-min resolution (ca. 5 km) using an empirical statistical approach. For this, linear or other
relationships are established between historically observed climate data at local scales, such as meteorological station measurements and climate model outputs
(Ramirez and Jarvis, 2008). ArcGIS software (Ormsby
et al., 2009) was used to calculate averages and standard deviations of model outputs for rainfall average
for all 21 models. The differences between future predictions and current long-term average values (1950–
2000) were calculated using the WorldClim 1.4 dataset,
also at a 2.5-min resolution as a reference (Hijmans et
al., 2005). The zonal statistics function of the ArcGis
9.3.1 Spatial Analyst was used to derive the predicted
average changes of July–August rainfall for each of the
21 municipalities where CA plots were established.
For 2009 and 2010 data, the accumulated rainfall
for July–August registered (on a daily base) in the
meteorological stations of the Mexican National Commission of Water (CNA) was used for each of the 21
municipalities. Three sets of results, historical rainfall,
2009 and 2010 registered rainfall and predicted rainfall,
were compared at the municipal scale.
Statistical analysis
The economic and yield data were processed using
R 2.13.2 software for the application of modified stability analysis (MSA) following the method described
by Hildebrand (1984). For the application of MSA,
an environmental index (EI) was built using the mu-
R. ROMERO-PEREZGROVAS et al.
nicipal mean maize yield for ten years (1998–2007),
reported by the National Service of Agricultural and
Fishing Information (SIAP) for each of the 21 municipalities where comparative plots were established (Table II). The plots were numbered consecutively from
the lowest to highest average yield and assigned a place
in this ranking; the position in the ranking constitutes the specific EI. When irrigated and rainfed plots
were established in the same municipality, differentiated means for both rainfed and irrigated plots from
the same municipality were used. In all those cases,
the rainfed plot had a low EI and the irrigated one a
high EI. When two rainfed or two irrigated plots were
established in the same municipality, the plots were
assigned the same EI. In total, these resulted in 25
different environments.
The yield and net return data were arranged according to EI and regressions were performed using
ordinary least squares (OLS) separately on yield and
net return, adding each of the CA and CP plots from
2009 and 2010 as individual observations to increase
the degrees of freedom and fit of the model. The total
number of observations for the regressions was 108. In
the same regressions, the factors year (2009) and treatment (CA) were used as covariates. MSA was used because specific environmental and management factors
of each plot, such as climate, soil conditions, farmers’
knowledge and quality of the technical advice, could
influence the response to CA or CP.
It is important to note that the values of coefficient
of determination (R2 ) and adjusted R2 in the MSA
could not be interpreted in the same way as for another regression analysis because plot yield values were
highly correlated to the EI values even if the EI values
were built with exogenous data as suggested by Hildebrand (1984). Observations for higher EI values normally have larger yields and net returns for both years
and treatments. After these calculations, two models
with interaction terms were tested. In the first model
the interaction was built using the year 2009 and CA
treatment in order to assess if CA had a buffer effect
for yields and net returns in dry years across all the
environments. The second model included the interaction between EI and CA, assuming that CA positive
effects diminished in irrigated areas with high input
systems.
RESULTS AND DISCUSSION
Yields and net returns in 2009 and 2010
Conservation agriculture demonstrated positive effects on yield compared to CP, in spite of being only
TILLAGE AND RESIDUE MANAGEMENT EFFECTS ON MAIZE
under first or second year of adoption (Table II).
TABLE II
2009–2010 maize yields,10-year average yields and environmental index (EI) values of the 27 conservation agriculture (CA)conventional practice (CP) comparative plots in the Central Mexican highlands
Plot
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
EIa)
Yield
Year 2009
Year 2010
CA
CP
CA
3.6
2.7
1.4
1.7
0.8
2.3
1.5
1.2
3.0
2.5
6.9
0.7
0.5
3.0
2.4
7.0
0.6
11.4
10.9
5.3
6.0
2.3
10.2
10.6
15.8
10.9
10.3
t ha−1
7.1
4.6
6.4
4.3
3.2
2.2
3.5
2.6
2.2
1.5
3.7
2.9
4.6
2.3
3.3
2.8
5.6
4.1
6.6
3.9
8.1
6.7
2.2
1.8
1.4
1.2
5.2
4.3
5.1
3.4
9.1
8.3
1.4
1.6
12.1
11.7
11.5
11.1
6.3
5.9
9.4
8.2
3.6
2.9
10.4
10.6
10.9
11.1
15.8
15.2
10.8
10.7
11.1
10.6
7.0
6.5
2.7
3.3
0.9
3.8
3.9
2.8
5.4
6.4
7.9
0.9
0.8
4.9
4.8
8.8
0.5
11.9
10.8
5.9
8.8
3.4
9.6
10.0
16.2
10.4
10.9
a) Assigned
10-year average
CP
4.1
4.0
1.9
2.5
1.4
2.9
2.0
1.7
3.6
3.8
4.4
1.3
1.0
3.1
3.4
4.8
0.8
6.9
7.1
4.2
4.7
2.7
6.2
6.6
7.5
6.6
6.9
16
15
6
8
4
10
7
5
13
14
18
3
2
11
12
20
1
23
24
17
19
9
21
22
25
22
23
481
as a consequence of drought stress and the recently
(2007–2008) adopted CA-based practices resulted in a
43% higher maize yield (5.0 t ha−1 ) than CP (3.5 t
ha−1 ). The treatments CA and CP intercepted at a
confidence coefficient of 75% (Fig. 2b). However, under
high-input irrigated conditions as well as in the environments where production-limiting factors severely
affected crop growth for both treatments, there were
no significant positive effects of CA on yield. For the
year 2010, both treatments had higher yields than in
the dry year 2009: the average yield was 6.8 t ha−1 for
CA and 5.7 t ha−1 for CP, with the difference between
the two practices decreasing to 19%. The two treatments intercepted at a confidence coefficient of 65%
(Fig. 2c).
using the 10-year average yield.
Conservation agriculture produced higher yields
than CP in farmers’ fields under the dry conditions
of 2009. Averaged over 25 environments, maize yield
was 26% higher under CA: 6.3 t ha−1 for CA vs. 5.0 t
ha−1 for CP (Fig. 2a). The treatments CA and CP intercepted at a confidence coefficient of 70%. This was
because in some plots, production-limiting factors such
as hail, pest incidence or other external factors greatly
reduced the yield for both treatments. Moreover, the
environments with high EI values were the ones where
farmers had access to irrigation and used that to control drought stress. Consequently, there were farmers
who obtained high or low yields depending on the environment more than the management practice. The
overall yield difference of 26% had its main driver in
the plots with EI values ranging from 5 to 18. In these
plots, water was the main production-limiting factor
Fig. 2 Distribution of confidence intervals for maize yield under
conventional practice (CP) and conservation agriculture (CA)
for all 25 studied environments in 2009 (a), the environments
with environmental index values ranging from 5 to 18 in 2009
(b), and all 25 studied environments in 2010 (c).
The higher yields under CA had a positive effect on
the net returns obtained by farmers. However, selling
or using straw from the CP treatment as fodder gene-
482
R. ROMERO-PEREZGROVAS et al.
rated some extra income. For the CA treatment farmers had to leave a certain amount (on average 56%)
of their residues as soil cover (Table I), so they sold or
used as fodder lower quantities of residues than for the
CP treatment. However, higher yields in combination
with lower production costs of CA resulted in larger
net returns for CA using partial budgets. The use of
CA had a positive effect of 3 838 Mexican Pesos (MXN)
ha−1 (320 US$ ha−1 ) compared to that of CP, while
the year 2009 used as a covariate showed a negative effect of −1 986 MXN ha−1 (165 US$ ha−1 ) for both CA
and CP compared to the year 2010. The former indicates that CP resulted in a lower net return per hectare
compared to CA especially when dry conditions were
present. The difference in net return between CA and
CP was reduced in the close-to-average rain year 2010.
The CP practice was less resilient compared with the
CA practice when rainfall was reduced such as occurred
in July–August 2009.
When the interaction terms year 2009 × CA and EI
× CA were included in the regression analysis, there
was no conclusive evidence to demonstrate that CA
had larger yields and net returns throughout all the
environments (Table III). The effect of CA was unclear
when all the environments and the two years were analyzed together; both interactions were not statistically
significant, adding evidence that CA had larger effects
in the dry year 2009 and in the plots where full irrigation was not available (Table III).
An on-station experiment similarly in the highlands of Central Mexico showed clear differences in
soil quality between CA and CP after 10 years (Govaerts et al., 2006b). During the extended dry period
in 2009 (62–83 d after planting), the soil water content
stayed above or near the permanent wilting point under CA in the on-station experiment, whereas under
CP, values were below the wilting point for 3 weeks
with severe implications for maize growth (Verhulst et
al., 2011). This resulted in large yield differences of
up to 4.71 t ha−1 between CA and CP with a wheatmaize rotation, whereas treatments with zero tillage
but without residue retention had even lower yields
than CP (Verhulst et al., 2011).
However, it is important to note that the onstation experiment was established in 1991 and the
2009 drought occurred 18 years after the start of the
experiment (Verhulst et al., 2011). Our data was obtained under farmers’ conditions and management, and
TABLE III
Modified stability analysisa) with two interaction terms for maize yields and net returns in the Central Mexican highlands under
conservation agriculture (CA) and conventional practice (CP) in 2009 and 2010 using factors year (2009) and treatment (CA) as
covariates
Source
Item
Mean
Standard error
t value
P value
Significance level
Yield
EIb)
CA
Year 2009
0.40
1.281
−0.518
0.0255
0.3538
0.354
15.757
3.621
−1.466
0.0000
0.0005
0.1458
***
***
EI
CA
Year 2009
CA × year 2009
0.4018
1.042 88
−0.75263
0.467 39
0.0255
0.5065
0.5019
0.7093
15.757
2.059
−1.499
0.659
0.0000
0.0421
0.1458
0.5117
***
**
EI
CA
Year 2009
EI × CA
0.423 96
1.8808
−0.51894
−0.4428
0.0361
0.7769
0.3544
0.0514
11.744
2.421
−1.464
−0.867
0.0000
0.0173
0.1463
0.3877
***
**
EI
CA
Year 2009
1 194.3
3 838
−1 986.5
100.9
1 400.4
1 401.4
11.832
2.741
−1.418
0.0000
0.0072
0.1594
***
***
EI
CA
Year 2009
CA × year 2009
1 194.3
3 559.8
−2 259.6
546.1
100.4
2 008.9
1 990.7
2 814.6
11.776
1.772
−1.135
0.194
0.0000
0.0794
0.2590
0.8465
***
*
EI
CA
Year 2009
EI × CA
1 281.2
6 194.5
−1 986.5
−173.9
142.9
3 075.2
1 403.1
202
8.967
2.014
−1.416
−0.861
0.0000
0.0466
0.1599
0.3913
***
**
Net return
*, **, ***Significant at P < 0.05, 0.01 and 0.001, respectively.
a) Adding each of the CA and CP treatments from 2009 and 2010 as individual observations to increase the degrees of freedom and fit
of the model. The total number of observations for the regression was 108.
b) Environmental index.
TILLAGE AND RESIDUE MANAGEMENT EFFECTS ON MAIZE
only one to two years after CA had been implemented.
It was unlikely for physical, chemical and biological
soil quality to have changed in such a short period
(one to two years) and thus the yield difference was
probably due to physical effects of the residues on the
soil surface, i.e., breaking rain drop impact, slowing
down runoff and reducing evaporation. The yield gap
between CA and CP was not as large as that after
18 years of different management practices in the onstation experiment (Verhulst et al., 2011), but the tendency was the same, positioning CA as a system more
resilient to drought. In years without drought, the differences between CA and CP also diminished considerably under experimental conditions (Govaerts et al.,
2009; Verhulst et al., 2011). The results under farmers’
conditions reported here showed the same tendency
(higher yields for CA than CP) but in smaller magnitudes. Additionally to the fact that it takes time before
management practices change soil quality, the implementation of CA by farmers and technicians involves
a learning process. In this study, they achieved positive effects of CA on yield and net return compared to
CP; however, it can be expected that CA results will
further improve as the learning process advances.
The literature suggests that CA-based cropping
systems can contribute to sustainable agriculture but
are not a panacea (Hobbs, 2007; Groisman et al., 2005;
Giller et al., 2009; Kassam et al., 2009). Some benefits of CA systems, such as erosion control, have been
widely acknowledged (Hobbs, 2007), but others such
as yield increaments, carbon sequestration or positive
economic effects are still under assessment and currently under discussion (Giller et al., 2009). CA-based
cropping systems for large-scale commercial farms have
been systematically adopted around the world, with
Argentina, Brazil, Australia and USA having been presented as successful cases of adoption, due primarily
to cost reductions and secondly to environmental benefits (Kassam et al., 2009). However, when focusing
on yield increments, the main potential areas are dry
tropical or subtropical regions where the main limiting
factor for agricultural production is lack of water. Under such conditions, rainfalls normally face a degree
of uncertainty and, as a result of climate change, dry
spells are becoming increasingly common, while the
timing of the rainfalls is predicted to continue changing (Groisman et al., 2005). It is here that CA holds
a great potential for assuring crop production, acting
as a robust technology which can increase resilience to
climatic shocks such as drought and help farmers to
sustain their yields and economic returns per hectare,
even in years with dry periods during the growing sea-
483
son. One of the main problems for implementing CA in
the region is the competition for crop residues. However, in this study net returns for CA were higher than
those for CP, in spite of the fact that with CA the farmers had to leave a certain percentage of their residues
as soil cover. The value of residues on the soil compared
to other uses increases under drought conditions. This
could become even more important in the future when
drought occurrences increase in frequency.
While it is clear that no single technology can solve
all the problems faced by agriculture in the tropical and
subtropical highlands of the world, CA can serve as a
solid base on which other technologies can be implemented. Some authors argue that the positive effects
of CA on yield are not realized until the medium and
long term (Knowler and Bradshaw, 2007), while the
results outlined in this study suggested a short-term
(one to two years) positive impact under rainfed conditions with dry periods. In a drier future in the Central Mexican highlands and beyond, CA could be an
alternative to adapt to this change, producing greater
yields, with more net returns for farmers even if they
have to leave part of their residues as soil cover. Under
irrigated conditions with high inputs (defined in MSA
as environments with high EI values), the results did
not show a significant positive effect of CA on net return or yield compared to CP in this short-term period
(one to two years).
Climatic conditions in 2009 and 2010 and predictions
for the 2050s
The historical data (1950–2000) for the 21 municipalities where the CA-CP plots were established reflects the fact that the July–August period was the
core of the rainy season in the region. Almost half of
the annual rainfall occurred in these two months (Table
IV). At that time, the maize crop in the region is usually at the vegetative stages. Between 1950 and 2000,
the eight driest municipalities had an average rainfall
between 140 and 205 mm in July–August. Eight others had an average between 206 and 271 mm, three
between 272 and 337 mm and two between 338 and
390 mm (Fig. 3a). Compared to the historical data, a
strong drought during the July–August period in 2009
resulted in 61%–80% reduction of rainfall in five municipalities, 41%–60% in eight, 21%–40% in five and
20%–0% in three (Fig. 3c). For the 2050s, the analyzed
models predicted that average rainfall in July–August
will also decrease substantially in most municipalities,
although the decrease was predicted to be smaller than
in 2009. Predictions are that nine municipalities will reduce 21%–40% of the average rainfall during the July–
484
R. ROMERO-PEREZGROVAS et al.
TABLE IV
Historical, 2009 and 2010 rainfall and rainfall of the 2050s projected using 21 climate change models for July–August and the annual
total in the conservation agriculture-conventional practice comparative plots of 21 municipalities in the Central Mexican highlands
Municipality
Rainfall
July–August
Annual total
1950–2000
2009
2010
2050s
231
315
274
233
348
350
201
143
145
146
238
147
198
220
225
158
250
149
209
212
222
115
237
154
164
180
185
104
89
37
60
89
52
110
207
167
151
155
147
87
59
47
221
303
266
229
341
343
195
106
148
153
240
151
211
240
243
153
285
169
130
190
148
180
230
214
159
247
364
177
120
139
121
243
170
215
242
229
165
183
102
167
168
151
1950–2000
2009
2010
2050s
576
763
739
614
823
852
574
356
465
463
786
504
575
583
562
452
772
484
381
452
390
493
723
709
605
813
815
562
362
233
440
956
580
412
616
797
524
827
524
185
524
188
525
780
742
641
846
865
612
390
479
501
879
521
547
582
613
480
776
419
430
448
392
501
663
665
589
749
886
505
334
447
430
794
534
593
595
573
457
718
464
305
434
382
mm
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
August period, another ten will reduce 0%–20% and
another two are expected to gain some additional rainfall (1%–20%) in July–August (Fig. 3b). In 2010, rainfall was close to the historical average. Eight municipalities had a slight rainfall reduction of between 0% and
5%, four had some additional rainfall of 1%–5%, three
6%–10%, three 11%–15%, and two 16%–20% (Fig. 3d).
Some studies found a positive tendency for more
intense rainfall events and consecutive dry days in the
whole area of Mexico and neighboring Central American countries (Aguilar et al., 2005; Groisman et al.,
2005). In addition, Groisman et al. (2005) found a substantial decrease in precipitation over the highlands of
Central Mexico over the last 30 years. In the same period, the frequency of rain events above 75 mm increased by 110%. Consequently, there is not merely a
decrease in average annual rainfall for the area, but
rainfall patterns have also changed, with more intense
events as well as longer dry periods now being observed. This poses a challenge for farmers practicing
rainfed agriculture because the risk of crop failure resulting from a drought period or a single event producing excess rain has increased. The data of this study
support these findings. The year 2009 was not a very
dry year in terms of annual rainfall average for the
highlands and some heavy rain events occurred outside
of the historical core of the rainy season with July–
August being extremely dry during this year (Table
IV).
While the drought of 2009 was, for the majority of
the municipalities, even more severe than predictions
for the region under the A2 climate change scenario,
rainfall was predicted to become more erratic during
the year, and drought extremes, like the one of 2009,
to become more frequent. The predicted changes in
climate could severely affect agricultural production in
Mexico if its cropping systems fail to adapt to these
new conditions. According to Parry et al. (2004), grain
yields in Mexico could be reduced by 30% by 2080 if
currently used production systems are maintained.
CONCLUSIONS
Drought in the Central Mexican highlands in July–
August 2009 resulted in an even drier situation than
that predicted for the 2050s, while the 2010 rainfall
was similar to the historical average. In 2009, the overall average maize yield was 26% higher for CA than for
CP; however, the difference increased to 43% when excluding from the analysis the environments with high
EI values, which were under irrigated, high-input conditions, or those with low EI values, where production
TILLAGE AND RESIDUE MANAGEMENT EFFECTS ON MAIZE
485
Fig. 3 Historical rainfall average (1950–2000) for the July–August period by municipality in the Central Mexican highlands where
conservation agriculture-conventional practice plots have been established since 2007 (a), comparison of percentage of the July–August
historical rainfall and the projections of 21 climate change models for the 2050s (b), comparison of percentage of the July–August
rainfall in 2009 and the historical average (c) and comparison of percentage of the July–August rainfall in 2010 and the historical
average (d).
was severely affected. In spite of the short time after
adoption of CA (one to two years), the CA system
resulted in more resilient maize production than the
CP system for a variety of conditions and locations. In
2010, the yield differences between the two systems
decreased to 16%. A modified stability analysis of the
data of both systems and years showed that the CA
system had a positive effect on yield and net return.
These positive effects of CA were achieved under conditions where full irrigation was not available. In the
irrigated areas, drought effects were buffered and the
effect of CA diminished or disappeared. Given that in
these high-input areas the market is heavily distorted
by subsidies, further research is needed to quantify the
impact of CA in the absence of these subsidies. Under
farmers’ management, CA proved to be more resilient
and generate higher yields and higher net returns, in
comparison with CP under a severe drought scenario
under rainfed conditions. In a near future CA would
help farmers to face some of the consequences of climate change such as drought and changing rainfall patterns.
ACKNOWLEDGEMENT
We thank Mr. José Luis Salgado, Mr. Francisco
Olguı́n, Ms. Andrea Chocobar and Mr. Dagoberto Flores from the International Maize and Wheat Improvement Center (CIMMYT), Mexico for technical assistance.
REFERENCES
Aguilar, E., Peterson, T. C., Ramı́rez Obando, P., Frutos, R.,
Retana, J. A., Solera, M., Soley, J., González Garcı́a, I.,
Araujo, R. M., Rosa Santos, A., Valle, V. E., Brunet, M.,
Aguilar, L., Álvarez, L., Bautista, M., Castañón, C., Herrera, L., Ruano, E., Sinay, J. J., Sánchez, E., Hernández
Oviedo, G. I., Obed, F., Salgado, J. E., Vázquez, J. L., Baca,
M., Gutiérrez, M., Centella, C., Espinosa, J., Martı́nez, D.,
Olmedo, B., Ojeda Espinoza, C. E., Núñez, R., Haylock, M.,
Benavides, H. and Mayorga, R. 2005. Changes in precipitation and temperature extremes in Central America and
northern South America, 1961–2003. J. Geophys. Res.- Atmos. 110: 1984–2012.
Améndola, R., Castillo, E. and Martı́nez, P. A. 2006. Mexico,
country pasture/forage resource profiles, FAO. Available on
line at http: //www.fao.org/ag/AGP/agpc/doc/Counprof/
Mexico/Mexico.htm (verified on May 22, 2012).
486
Boughton, D., Crawford, E., Krause, M. and de Frahan, B.
H. 1990. Economic Analysis of on-farm Trials: A Review of
Approaches and Implications for Research Program Design.
Michigan State University, Michigan.
Carvalho, L. M. V. and Jones, C. 2013. CMIP5 simulations of
low-level tropospheric temperature and moisture over the
tropical Americas. J. Climate. 26: 6257–6286.
Fischer, R. A., Santiveri, F. and Vidal, I. R. 2002. Crop rotation, tillage and crop residue management for wheat and
maize in the sub-humid tropical highlands: II. Maize and
system performance. Field Crop. Res. 79: 123–137.
Giller, K. E., Witter, E., Corbeels, M. and Tintonell, P. 2009.
Conservation agriculture and small farming in Africa: the
heretics’ view. Field Crop. Res. 114: 23–34.
Govaerts, B., Mezzalama, M., Sayre, K. D., Crossa, J., Lichter,
K., Troch, V., Vanherck, K., De Corte, P. and Deckers, J.
2008. Long-term consequences of tillage, residue management, and crop rotation on selected soil micro-flora groups
in the subtropical highlands. Appl. Soil Ecol. 38: 197–210.
Govaerts, B., Mezzalama, M., Sayre, K. D., Crossa, J., Nicol, J.
M. and Deckers, J. 2006a. Long-term consequences of tillage,
residue management, and crop rotation on maize/wheat
root rot and nematode populations in subtropical highlands.
Appl. Soil Ecol. 32: 305–315.
Govaerts, B., Mezzalama, M., Unno, Y., Sayre, K. D., LunaGuido, M., Vanherck, K., Dendooven, L. and Deckers, J.
2007. Influence of tillage, residue management, and crop
rotation on soil microbial biomass, and catabolic diversity.
Appl. Soil Ecol. 37: 18–30.
Govaerts, B., Sayre, K. D. and Deckers, J. 2006b. A minimum
data set for soil quality assessment of wheat and maize cropping in the highlands of Mexico. Soil Till. Res. 87: 163–
174.
Govaerts, B., Sayre, K. D. and Deckers, J. 2005. Stable high
yields with zero tillage and permanent bed planting? Field
Crop. Res. 94: 33–42.
Govaerts, B., Sayre, K. D., Goudeseune, B., De Corte, P.,
Lichter, K., Dendooven, L. and Deckers, J. 2009. Conservation agriculture as a sustainable option for the central
Mexican highlands. Soil Till. Res. 103: 222–230.
Groisman, P. Y., Knight, R. W., Easterling, D. R., Karl, T.
R., Hegerl, G. C. and Razuvaev, V. N. 2005. Trends in intense precipitation in the climate record. J. Climate. 18:
1326–1350.
Hijmans, R. J., Cameron, S. E., Parra, J. L., Jones, P. G. and
Jarvis, A. 2005. Very high resolution interpolated climate
surfaces for global land areas. Int. J. Climatol. 25: 1965–
1978.
Hildebrand, P. E. 1984. Modified stability analysis of farmer
managed, on-farm trials. Agron. J. 76: 271–274.
R. ROMERO-PEREZGROVAS et al.
Hobbs, P. R. 2007. Conservation agriculture: what is it and
why is it important for future sustainable food production?
J. Agr. Sci. 145: 127–137.
Jiménez, B. 2005. Treatment technology and standards for agricultural wastewater reuse: a case study in Mexico. Irr. Drain. 54: S23–S33.
Kassam, A., Friedrich, T., Shaxson, F. and Pretty. J. 2009. The
spread of conservation agriculture: justification, sustainability and uptake. Int. J. Agr. Sust. 7: 292–320.
Knowler, D. and Bradshaw, B. 2007. Farmers’ adoption of conservation agriculture: a review and synthesis of recent research. Food Policy. 32: 25–48.
Lal, R. 1993. Tillage effects on soil degradation, soil resilience,
soil quality, and sustainability. Soil Till. Res. 27: 1–8.
Nakicenovic, N., Alcamo, J., Davis, G., de Vries, B., Fenhann,
J., S. Gaffin, K. Gregory, A. Grübler, Jung, T. Y., Kram,
T., La Rovere, E. L., Michaelis, L., Mori, S., Morita, T.,
Papper, W., Pitcher, H., Price, L., Riahi, K., Roehrl, A.,
Rogner, H.-H., Sankovski, A., Schlesinger, M., Shukla, P.,
Smith, S., Swart, R., van Rooijen, S., Victor, N. and Dadi,
Z. 2000. IPCC Special Report on Emissions Scenarios. Cambridge University Press, Cambridge.
Ormsby, T., Napoleon, E., Burke, R. and Groessl, C. 2009. Getting to Know ArcGIS Desktop. Updated for ArcGIS 9.3.
Basics of ArcView, ArcEditor, ArcInfo. ESRI Press, Redmond.
Parry, M. L., Rosenzweig, C., Iglesias, A., Livermore, M. and
Fischer, G. 2004. Effects of climate change on global food
production under SRES emissions and socio-economic scenarios. Global Environ. Chang. 14: 53–67.
Ramirez, J. and Jarvis, A. 2008. High resolution statistically
downscaled future climate surfaces. Available online at http:
//gisweb.ciat.cgiar.org/GCMPage (verified on November
10, 2010).
Sayre, K. D., Mezzalama, M. and Martinez, M. 2001. Tillage,
crop rotation and crop residue management effects on maize
and wheat production for rain fed conditions in the altiplano of central Mexico. In Garcı́a-Torres, L., Benites, J.
and Martinez-Vilela, A. (eds.) Conservation Agriculture, A
Worldwide Challenge. Proceedings of the I World Congress
on Conservation Agriculture. pp. 575–585.
Sotelo-Ruı́z, E., A., González, Cruz, G., Moreno, F. and Cárdenas. C. 2011. The soils of Mexico states and their updates to the World Reference Base for Soil Resources 2006.
Revista Mexicana de Ciencias Forestales. 2: 71–84.
Verhulst, N., Nelissen, V., Jespers, N., Haven, H., Sayre, K. D.,
Raes, D., Deckers, J. and Govaerts, B. 2011. Soil water content, maize yield and its stability as affected by tillage and
crop residue management in rainfed semi-arid highlands.
Plant Soil. 344: 73–85.