Field Crops Research 257 (2020) 107934

Contents lists available at ScienceDirect

Field Crops Research

journal homepage: www.elsevier.com/locate/fcr

The dos and don'ts of no-till continuous cropping: Evidence from wheat yield T
and nitrogen use efficiency
Oswaldo R. Ernsta,*, Armen R. Kemanianb, Sebastián Mazzillia, Guillermo Siri-Prietoa,
Santiago Dogliottid
a
Departamento de Producción Vegetal. Facultad de Agronomía, Estación Experimental Mario Alberto Cassinoni, Universidad de la República, Ruta 3, km 363, Paysandú
60000, Uruguay
b
Department of Plant Science, The Pennsylvania State University, 116 ASI Building, University Park, PA 16802, USA
d
Departamento de Producción Vegetal, Facultad de Agronomía, Universidad de la República, Av. Garzón 780, 11200 Montevideo, Uruguay

ARTICLE INFO ABSTRACT

Keywords: When crop-pasture rotation is converted to a single fallow/soybean or winter crop/soybean annual cropping,
No-till wheat grain yield declines progressively as the annual cropping phase lengthens, regardless of the tillage system.
Sustainable intensification This decline can be attributed to i) depletion of the soil nutrient supply capacity and ii) subtle but cumulative
Yield gap degradation of soil physical properties. The objectives of this study were to disentangle and quantify the lim-
Cropping system
itations on wheat yield imposed by these processes, and to identify the cropping sequence that preserves soil
Wheat
quality and enables high wheat yield. Wheat was grown for two years at three nitrogen (N) fertilization rates (0,
80 and 190 kg ha−1) in soils after a 20-years experiment with six cropping systems. The cropping systems are
crop-pasture rotations with tillage (ROT_CT) or no-till (ROT_NT), continuous cropping with no-till and high
frequency of sorghum and maize (CC_NTC4) or soybean and sunflower (CC_NTC3) or winter fallow (CC_NTWF),
and continuous annual cropping under conventional tillage (CC_CT). Soil quality was assessed based on chemical
(soil organic carbon, total soil N concentration and potentially mineralizable N) and physical properties (field
water infiltration rate and soil aggregate stability). In each system, we estimated the yield gap due to N supply
(YgN ) limitations and the yield gap due to soil properties other than N supply limitations (Y gothers ), so that the total
yield gap (YgT ) is the sum of YgN and Y gothers . Systems that degraded chemical and physical properties had lower
yield, grain N concentration and fertilizer N use efficiency (NUEf, kg of grain kg−1 of N added). Only two
systems, ROT_NT and CC_NTC4, achieved Ymax (7.2 Mg ha-1). For these two systems YgT = YgN . For the other
systems, the percentage of YgT explained by Y gothers varied between 23 % and 50 %. Rotations that increased the
soil N supply (N uptake with no N fertilizer) also increased NUEf. Wheat under ROT_NT reached the maximum
yield obtained under CC_CT with 45 % less N fertilizer (104 vs 190 kg ha−1) and higher NUEf (50 vs 27 kg kg−1).
Comparing ROT_NT and CC_NTC4 to other continuous no-till cropping systems (CC_NTWF and CC_NTC3), the N
fertilizer required was increased from 104 and 107 to 152 and 163 kg ha-1, respectively. In conclusion, rotating
annual crops under no-till is not enough to preserve soil productivity. Sustainable intensification under con-
tinuous no-till would require either re-balancing crop sequences towards crop-pasture rotations or a shift to-
wards a lower frequency of soybean in favor of higher frequency of maize and sorghum in the summer phase of
the rotation.

1. Introduction Paruelo, 2020), largely driven by no-till soybean (Glycinemax (L.)

Merr.) and fallow cropping systems. However, wheat (Triticum aestivum
Land use has changed drastically during the last 20 years in the L.) is still the most important winter crop, occupying 6 M ha (www.fao.
South American Pampas (Modernell et al., 2016), a region that covers org/faostat). Following these regional trends, the annual cropped area
more than 700 thousand km² in central-east Argentina, southern Brazil, in Uruguay increased from 0.4 to 2 M ha in the same period, with
and Uruguay. From the year 2000 onward, the total cropped area in- soybean as the main crop. However, 30%–40% of the soybean area is
creased by 28 % (Redo et al., 2012; Modernell et al., 2016; Baeza and sown following a winter cereal crop, in a double annual winter crop/

⁎
Corresponding author.
E-mail address: oernst@fagro.edu.uy (O.R. Ernst).

https://doi.org/10.1016/j.fcr.2020.107934
Received 15 January 2020; Received in revised form 4 August 2020; Accepted 10 August 2020
Available online 29 August 2020
0378-4290/ © 2020 Elsevier B.V. All rights reserved.
O.R. Ernst, et al. Field Crops Research 257 (2020) 107934

soybean sequence (DIEA, 2019). This increase in cropped area occurred explained strictly by nutrient limitations which could be overcome by
mainly by shifting cropping systems from crop-pasture rotations to increasing fertilization rates. However, after 5 years, further reductions
continuous annual cropping under no-till (Franzluebbers et al., 2014; in wheat yield were associated to a deterioration in soil properties that
Wingeyer et al., 2015). This shift had the following causes and ratio- could not be compensated by increasing the nutrient supply. These
nale. The first one was the introduction in the nineties of no-till tech- studies were carried out on the dominant crop rotation in the region,
nology, which was touted as an improvement over conventional tillage where soybean is the main summer crop followed by a winter fallow or
due to its potential to reduce soil erosion, increase C sequestration and wheat. Two questions remain unanswered after these studies: which
reduce energy consumption (Peiretti and Dumanski, 2014). The second soil properties are responsible for the increase in Yg that cannot be
one was the proposition that, since no-till succeeded, intensifying the compensated by increasing nutrient supply? Will the Yatt decrease and
cropping sequence by reducing or eliminating the pasture phase in the the Yg increase in all no-till annual cropping sequences, or are there
rotation was not only agronomically sound but environmentally sus- crop sequences (for example, replacing soybean with summer cereals)
tainable. Unbeknown in the earlier phase of this transition was that that can set soils in trajectory that preserve or enhance soil properties?
shifting cropping systems to continuous annual cropping might affect In other words: can we re-design the soybean-based annual cropping
soil quality, crop performance and production system sustainability system that currently prevails in the region and direct it towards a
through processes of soil degradation that take time to build up and sustainable agricultural intensification pathway?
become measurable (Munkholm et al., 2013). Long-term cropping system experiments provide unique opportu-
A growing body of evidence emerging in the last 20 years supports nities to evaluate the impact of changes in soil properties induced by
the hypothesis that removing perennial pastures from no-till rotations cropping systems on Yatt, Yg, and input use efficiency. In this study, we
may gradually reduce soil quality and limit productivity (Alvarez and used a 20-y crop rotation and tillage experiment to test hypotheses
Steinbach, 2009; Tourn et al., 2019 in Argentina; Boeni et al., 2014; da concerned with the effect of cropping system on soil properties, wheat
Silva et al., 2014; Salton et al., 2014 in Brazil; García-Préchac et al., yield, Yg and N use efficiency (NUE). The experiment included crop–-
2004; Ernst and Siri-Prieto, 2009; Ernst et al., 2016; Pravia et al., 2019 pasture rotation and continuous cropping under conventional tillage
in Uruguay). The main message from this body of work was that despite and no tillage, and contrasting crop sequences under continuous no-till.
its benefits, no-till alone was no panacea. The current no-till continuous The core proposition is that under continuous no-till cropping systems,
annual cropping systems may contravene the current sustainability the strategies for enhancing soil quality include eliminating fallow
paradigm, which states that sustainable agricultural intensification re- periods by using double cropping or cover crops (Novelli et al., 2013,
quires increasing food production from existing farmland in ways that 2017), increasing residue inputs and yield by increasing N fertilization
lower environmental impact and do not undermine the farmland pro- (Varvel, 2006), and incorporating maize (Zea mays L.) and sorghum
duction capacity (Garnett et al., 2013). (Sorghum biocolor (L.) Moench) as crops with high biomass production
There is a lack of a cohesive framework that explains the pro- and greater and deeper root systems (Huggins et al., 2007; Salvo et al.,
gressive decline in grain yield observed in no-till cropping systems 2010; Mazzilli et al., 2015; Jackson et al., 2017).
when the perennial pasture phase of the rotation is removed (Aparicio The objectives were to: (i) quantify long-term effects of cropping
and Costa, 2007; Ernst et al., 2016). Some evidence points to changes in systems (tillage system and crop sequence) on soil properties and wheat
soil physical properties. For example, the total porosity of the top yield, (ii) identify the soil properties that are most associated with the
horizon in no-till systems is lower than under pasture, with macropores observed reduction in wheat yield and NUE, and (iii) identify cropping
oriented in parallel to the soil surface and limiting soil water infiltration systems that would enable sustainable intensification under no-till.
(Sasal et al., 2006; Alvarez et al., 2014; Sasal et al., 2016). However, it
is difficult to assign the loss of soil productivity to changes in specific 2. Materials and methods
chemical or soil physical properties that could be altered by manage-
ment. There is also a lack of a quantitative understanding about the 2.1. Description of study area and experimental setup
impacts of the loss in soil productivity on nutrient use efficiency.
Increasing the productivity of existing cropping systems requires The long-term trial that serves as the basis for this experiment is
identifying the causes of yield gaps (Yg), i.e. the difference between the located at the experimental station Mario A. Cassinoni of the College of
attainable (Yatt) and actual yield in production systems (van Ittersum Agronomy (Universidad de la República), located 10 km south of
and Rabbinge, 1997; van Ittersum et al., 2013). In its current for- Paysandú (32° 21′ S and 58° 02′ W; elevation 61 m) in the northwest of
mulation, this framework ignores the effects of gradual changes in soil Uruguay, which corresponds broadly with the eastern edge of the South
properties that might occur in the short-term (10 years or less) and can American Pampas. The climate is meso-thermal sub-humid with a mean
alter Yatt, Yg and nutrient use efficiency. Although soils are usually daily temperature of 25° and 13 °C for summer and winter, respectively,
grouped taxonomically in genoforms, soil properties are dynamic al- and annual precipitation of 1200 mm distributed on average uniformly
lowing for the same taxonomic group to have multiple phenoforms as within the year, but with large intra- and inter-annual variation. Water
defined by Bouma and Droogers (1999), which are distinguished by deficits occur frequently between October and March and water sur-
changes in chemical and physical properties. The Yatt is a dynamic pluses between May and August. The soil at this site is a fine, mixed,
property related to the phenoform. When comparing two soils that have active, thermic Typic Argiudoll (USDA), with a slope of 1%. Total soil
different rotations (and therefore phenoforms) and that differ in the depth is 90 cm with an A horizon 20 cm deep, pH (H2O) 5.7, and clay,
yield of a given crop, Russelle et al. (1987) named that portion of the silt and sand contents of 289, 437, 273 g kg−1, respectively. Clay
yield that cannot be compensated for with synthetic chemicals (and content increases progressively to 320 g kg−1 and 460 g kg−1 in the
therefore associated to chemical properties) the “rotation effect”. These 20−40 cm and 40−90 cm layers, respectively. Estimated total plant
concepts were adapted by Ernst et al. (2016) using frontier analysis available water capacity is130 mm. There are no physical restrictions to
with a database containing 1072 records of wheat fields in Uruguay, rooting depth in the top 90 cm of soil.
and by Ernst et al. (2018) implementing 80 on-farm trials in farmers’ A detailed description of the cropping system treatments has been
rainfed wheat fields, to evaluate the impact of the agricultural in- reported by Ernst and Siri-Prieto (2009) and Salvo et al. (2010), and
tensification process, quantified as the number of years under con- only details relevant for the current work are summarized here. The
tinuous no-till cropping systems, on wheat Yatt and Yg. The first study experiment started in autumn 1993 combining two rotation systems,
reported that lengthening the annual cropping phase of crop-pasture continuous cropping (CC) and crop-pasture rotation (ROT), and two
rotations decreased wheat Yatt and increased Yg. The second study re- tillage systems, conventional tillage (CT) and no-till (NT). The resulting
ported that the Yg increase during the first 5 years after a pasture was four cropping systems were arranged in a randomized block design with

2
O.R. Ernst, et al. Field Crops Research 257 (2020) 107934

Table 1
Description of annual crop and pasture sequences in each cropping system from 1993 till 2012 in the long-term experiment located at Paysandú, Uruguay. P is a
three-years grass-legume pasture. WC/SC is a winter crop (wheat or barley) followed by a summer crop (sunflower or sorghum until 1999). The subscripts C3 or C4
indicate broadleaf (soybean and sunflower) or cereal (maize and sorghum) summer crops, respectively. Oat was grown in some treatments as a winter cover crop
(June to September). SF and WF are summer fallow and winter fallow respectively. Col is Colza grown as winter crop. Wh is wheat and Sb is soybean. Wheat crops
evaluated in this paper are indicated with bold letters.
Cropping system treatment name after 2008

Period ROT_NT CC_NTC4 CC_NTWF (*) CC_NTC3 ROT_CT CC_CT

1993/95 WC / SC WC / SC WC / SC WC / SC WC / SC WC / SC
1996/98 P WC / SC P WC / SC P WC / SC
1999 WC/SF WC / SF WC / SF WC / SF WC / SF WC / SF
2000/01 WC / SCC4 WC / SCC4 WC / SCC3 WC / SCC3 WC/SCC3 WC / SCC3
2002/04 P WC / SCC4 P WC / SCC3 P WC / SCC3
2005/07 WC / SCC3 WC / SCC4 WC / SCC4 WC / SCC3 WC/SCC3 WC / SCC3
2008/10 P WC / SCC4 WF / Sb WC / SCC3 P WC / SCC3
Winter 2011 / Summer 2012 Wh / Sb Wh / Sb WF / Sb Wh / Sb Wh / Sb Wh / Sb
Winter 2012 / Summer 2013 Oat / Sb Oat / Sb WF / Sb Oat / Sb Oat / Sb Oat / Sb
Winter 2013 / Summer 2014 Wh / Sb Col / Sb Wh / Sb Col / Sb Wh / Sb Col / Sb Wh / Sb Col / Sb Wh / Sb Col / Sb Wh / Sb Col / Sb
Winter 2014 Col Wh Col Wh Col Wh Col Wh Col Wh Col Wh

* Under ROT_NT from 1993 to 2004.

Table 2
Soil use indicators for the period 1993 to 2012, equivalent to 40 growing seasons (2 per year). The cropping systems’ names are those after 2008. The pasture, annual
crop, and winter fallow fractions are the fraction (0 to 1) of time under each land use in the sequence.
Cropping system ROT_NT CC_NTC4 CC_NTWF CC_NTC3 ROT_CT CC_CT

Annual crop instances 17 33 22 33 17 33

Annual crops per year 0.85 1.65 1.1 1.65 0.85 1.65
Fraction of the time
Pasture 0.5 0 0.25 0 0.5 0
Annual crop 0.425 0.825 0.55 0.825 0.425 0.825
Winter fallow 0.075 0.175 0.2 0.175 0.075 0.175
Ratio of instances to total number of growing seasons
Sorghum or maize .075 .400 .075 .075 .075 .075
Soybean or sunflower .200 .075 .400 .400 .200 .400
Wheat, barley or oats .150 .350 .075 .350 .150 .350
Winter fallow .075 .175 .300 .175 .075 .175
Total .500 1.00 .850 1.00 .500 1.00

three replications. Two additional plots by block were assigned to (Hordeum vulgare L.) in the winter season (June to November) followed
ROT_NT and CC_NT respectively. Then, from 1993–1999, each re- by sunflower or sorghum during the summer season (December to
plication had two plots under CT and four under NT (Table 1). April). From 2000 onward, the experiment stabilized with six treat-
Continuous annual cropping with conventional tillage (CC_CT). A se- ments by reassigning treatments under no-till (Table 1, years
quence of double annual cropping system (summer crops sown fol- 2000–2001). These two additional treatments (or cropping systems)
lowing winter crop) with the soil tilled using a combination of chisel differ in the summer crop used in a double annual crop sequence. One
plow and disk harrow (depending on the year) to a depth of 20−25 cm included only broad-leaf summer crops with C3 photosynthesis (soy-
followed by disking to a depth of 10−15 cm during April, before bean and sunflower), and the other cereal crops with C4 photosynthesis
sowing each winter crop. After the winter crop harvest, the soil is tilled (maize and sorghum). The winter crops were wheat or barley. From
using a disk harrow to a depth of 15−20 cm, followed by field culti- 2000–2011, eleven C3 or C4 summer crops were sown under CC_NT and
vator to a depth of 10−15 cm and then sown with a summer crop in seven under ROT_NT (CC_NTC4 and CC_NTC3 vs ROT_NTC4 and
December, sunflower (Helianthus annus L.), soybean, maize or sorghum. ROT_NTC3, Table 1). In 2008, following regional trends, while winter
Crop-pasture rotation with conventional tillage (ROT_CT). The rotation crop/soybean under ROT_NT was sown with pasture, winter crop/sor-
was 3.5 years of annual crops and 2.5 years grass and legume pasture. ghum and maize under ROT_NT was changed to a new cropping system
The pasture was under-seeded with the winter crop in the same composed of continuous winter fallow/soybean under no-till (CC_NTWF,
planting operation in 1995 and 2002. The sod-legume pasture consisted Table 1, years 2008–2010). As result, while CC_NT with a high fre-
of birdsfoot trefoil (Lotus corniculatus L.), white clover (Trifolium repens quency of sorghum and maize had a frequency of C4 crops of 0.4 (or 1
L.), and tall fescue (Festuca arundinacea Schreb.). During the annual C4 crop every 1.25 years; Table 2), under the other treatments this
cropping phase both, tillage operations and crop sequence were the frequency was 0.075 (or 1 C4 crop every 6.7 years).
same as CC_CT. Winter fallow period occurred each year in which a winter crop was
Continuous cropping with no-till (CC_NT). The same crop sequence as skipped, from end of April to beginning of November (harvest of the
CC_CT but sown with no-till and using glyphosate at rates of 1.5–2.0 kg soybean to sowing of the next soybean). Soil during winter was main-
a.i. ha−1 depending on weed infestation and weather conditions in lieu tained weed free by applying glyphosate. Following Rossister and
of tillage for weed control. Bouma (2015), we defined ROT_NT as the reference cropping systems,
Crop-pasture rotation no-till (ROT_NT). The same crop sequence and the more frequent cropping system used prior to 2002, including double
pasture as ROT_CT, but both crops and pasture are sown under no-till annual cash crop where sunflower, soybean, sorghum and maize were
and with weed control based on glyphosate as in CC_NT. sown immediately after harvest of winter crops (mainly wheat).
From 1993–1999, the annual crop sequences had wheat or barley Currently, winter crop/soybean and fallow/soybean under CC_NT

3
O.R. Ernst, et al. Field Crops Research 257 (2020) 107934

represent the most common cropping systems in Uruguay. layer was estimated following Waring and Bremner (1964). Field water
In this study we evaluated the performance of wheat crops during infiltration rate (INF) was determined using the method developed by
2013 and 2014 winter growing seasons. To avoid a possible wheat yield the Soil Quality Institute (1999). A 15-cm diameter ring was pressed 8
penalty from sowing wheat into wheat straw (Carignano et al., 2008; cm vertically into the topsoil. The ring was lined up with plastic wrap
Mazzilli et al., 2016), wheat was always preceded by soybean. During and 450 mL of distilled water applied. The plastic wrap was then re-
2012, except for CC_NT plots under winter fallow followed by soybean, moved to begin wetting and to homogenize soil moisture. After in-
all experimental plots had oat (Avena sativa L.) as winter cover crop filtration was completed, the procedure was repeated, recording the
followed by soybean (Table 1). During the winter of 2013, plots were time required for all water to infiltrate into the soil. Three subsamples
split in two and sown with either wheat (evaluated in the current study) were taken at each plot. To evaluate wet aggregate stability, we used
or colza (canola, [Brassica napus L]); both crops were followed by the wet sieving procedure of Yoder (1936) as modified by Kemper and
soybean during the summer. During winter 2014, the wheat evaluated Rosenau (1986). Three samples from the 0–10 cm soil layer were col-
in this study followed the colza-soybean sequence. The prior wheat crop lected from each main plot. Immediately after collection, aggregates
in these plots occurred in the winter of 2011. between 4.5 and 9.5 mm were separated from the composite sample by
At establishment, the experiment occupied 1 ha with individual gently pressing large clods and collecting the resulting aggregates.
plots of 50 × 10 m, thereby allowing the use of field-scale equipment Moist aggregates (30 g) of between 4.5 and 9.5 mm were spread evenly
for all operations. Wheat was maintained free of weeds, pest insects and on the uppermost sieve of a nest of 4.5, 2.8, 2.0, 1.0, 0.6 and 0.3-mm
diseases with timely applications of herbicides, insecticides and fungi- diameter. To avoid sudden rupture of the aggregates, the screens were
cides, and amended with non-limiting amounts of P, K and S fertilizer. lowered to wet the base of the topmost sieve to allow aggregates to
During 2013 and 2014 we evaluated wheat yield and NUE under become completely wet (10 min) by capillarity. The water level in the
three N fertilization rates and two water supply levels within each shaker was adjusted so that aggregates on the uppermost sieve were just
cropping system in a split-plot design. Cropping system was the main submerged on the highest point of the cycle. Samples were subjected to
plot (60 m2 inside each original plot) and the minor plot (2 × 5 m) 40 strokes per minute for 15 min with the amplitude of the action set at
treatments were a factorial of three N (urea) by two water supply levels, 8 cm. The soil remaining on every sieve after 15 min was transferred
where N0, N80 and N190 represent 0, 80 and 190 kg ha−1 of N, and into a beaker and oven-dried at 105 °C for 48 h and then weighed. The
Rainfed or Irrigated represent limited or not limited water supply levels strength of aggregates in water was calculated as mean weight diameter
(Appendix 1). Irrigation was triggered when soil available water < 65 (MWD, mm).
% in the top 40-cm of soil. Water was supplied through drip irrigation; i=7
the irrigation volume averaged 20–25 mm per event, with four events MWD = Xi × Wi
in 2013 and three events in 2014. i=1
We defined the N uptake at physiological maturity under N0 as “soil where, Xi is the average diameter of the openings of two consecutive
N supply” and N uptake and wheat yield under N190 as “yield not sieves, and Wi is the weight ratio of aggregates remained on the ith
limited by N”. The treatment N80 corresponds to the average N ferti- sieve (g/g). The multipliers used in this study were 0.45, 0.8, 1.5, 2.4,
lization rate used by wheat producers under ROT_NT in the region 3.65 and 7 mm for the sieves 1–6, respectively, and 0.15 mm for the
(Ernst et al., 2016). remaining soil.
To quantify the wheat response to soil attributes modified by the
long-term effect of the cropping systems (i.e. by the soil phenoform) 2.3. Grain yield, harvest index and N uptake
independently of the tillage system, all treatments were sown under no-
till from 2012. Seeding dates were July 6 of 2013 and June 27 of 2014. Grain yield (adjusted to 0% moisture) was measured using a plot
The wheat cultivar was Baguette 501, which ranked in the top five grain harvester on 6 m2 per plot. Additionally, 1 m2 per plot was
yielding cultivars in the trials of the National Testing Network of Wheat sampled at physiological maturity to measure grain and straw (St) mass
Cultivars (INASE, 2015) (Table 3). and used to estimate the harvest index and to determine both N grain
(N gr
conc
) and N straw (Nstconc ) concentration (Kjeldhal).
2.2. Soil quality assessment and meteorological data
2.4. Quantifying the effect of cropping system on yield gap
All soil properties were measured on samples collected from each
The maximum yield obtained across systems in the N190 treatment
main plot in June 2013. Soil organic carbon (SOC) (Nelson and
was used as benchmark and defined as maximum yield (Ymax ). Two
Sommers, 1996), and soil total N concentration (STN) were analyzed on
yield limiting factors, N supply and soil properties other than N supply,
dried samples that were sieved through a 2 mm mesh. Each sample was
were assumed to operate additively within each cropping system. This
a composite of 10 sub-samples of the top 0−5 and 5−20 cm soil layers.
is a simplification applied for expediency, as it is well understood that
Treatment effects were analyzed separately for each soil depth. Poten-
yield-limiting factors tend to interact, for example by making the re-
tially mineralizable N (PMN, mg N-NH4 kg−1) in the 0−10 cm soil
sponse to N dependent on soil physical conditions. To make compar-
isons at a wheat yield level achievable by all systems, we estimated
Table 3
Nitrogen fertilizer rates applied (kg ha−1) at sowing, Zadoks 22 (Z22, early both the N fertilizer rate and the NUE based on the added fertilizer at
tillering), Zadoks 30 (Z30, jointing) and Zadoks 50 (Z50, heading) phenological the yield achieved by the treatment with lowest Yatt. The yield reduc-
stages in the fertilized plots. tion due to N supply was calculated as the difference between yield
under N190 and yield under N0 (YgN = YN 190 YN 0 ). Following Ernst
Phenological stage (Zadoks et al., 1974) Nitrogen fertilization rate
et al. (2018), yield reduction due to properties other than N supply, was
N80 N190 calculated as the difference between maximum yield and yield under
—————— kg ha−1 —————— N190 (Y gothers = Ymax YN190 ). This term represents the portion of the
yield gap that cannot be recovered by adding N fertilizer. For each
Sowing 25 40 cropping system, the total Yg is therefore the sum of the two compo-
Z22 25 50 nents: YgT = YgN + Y gothers .
Z30 30 60
To address the extent to which Ymax approached the potential yield
Z50 40
Total 80 190 (Yp), we estimated the latter using the Cropsyst simulation model
(Stöckle et al., 2003, version 4.19.06), calibrated to local conditions

4
O.R. Ernst, et al. Field Crops Research 257 (2020) 107934

(Mazzilli et al., 2013). Daily solar radiation, minimum and maximum Table 4
temperature, relative humidity, wind speed and precipitation, were Top soil properties after 20 years in six cropping systems: crop-pasture rotations
retrieved from an automatic weather station located at the same ex- under no-till (ROT_NT) or under conventional tillage (ROT_CT), continuous
perimental station. Soil hydraulic properties were estimated from annual cropping under no-till (CC_NTC4, CC_NTC3, and CC_NTWF) and under
conventional tillage (CC_CT). The soil properties are: soil organic carbon (SOC)
pedotransfer functions in the model using as inputs soil texture and SOC
and soil total nitrogen (STN) concentration at 0-5 and 5-20 cm, and potentially
for five soil layers of 20-cm for a total depth of the soil profile of 1 m.
mineralizable nitrogen (PMN, mg N-NH4 kg−1 soil), water infiltration rate (INF)
and aggregates mean weight diameter (MWD, mm) at 0-10 cm.
2.5. N uptake, N supply and N use efficiency
Cropping system ROT_NT CC_NTC4 CC_NTF CC_NTC3 ROT_CT CC_CT

N use efficiency has two primary components: efficiency of ab- SOC0−5 (g kg−1) 28 a 28 a 27 ab 30 a 25 b 18 c
sorption or uptake, and efficiency with which total N absorbed is uti- SOC5−20 (g 19 a 19 a 19 a 18 a 18 a 17 a
lized to produce grain (Moll et al., 1982). Differences in N fertilizer kg−1)
requirements between cropping systems to reach the same yield STN0−5 (g kg−1) 2.0 ab 2.1 a 2.0 ab 1.9 b 2.1 a 1.9 b
STN5−20 (g 1.7 a 1.7 a 1.6 ab 1.6 ab 1.5 b 1.5 b
modifies NUE by affecting soil N supply, grain yield and grain protein
kg−1)
(Huggins and Pan, 1993). To explain NUE differences between cropping PMN (mg kg−1) 20 b 16 c 16 c 22 b 28 a 17 c
systems we estimated the terms N uptake, N supply and N uptake deficit INF (cm h−1) 3.1 a 2.5 b 1.8 c 1.7c 2.0 bc 1.3 d
(Nup, Nsu and NupD, respectively; kg ha−1 of N), and NUE as follows: MWD (mm) 2.13 a 1.29 b 1.34 b 2.08 a 1.47 b 0.64 c

Nh = N grconc × Y N extraido no grão x produtividade C4 = winter crop/sorghum and maize as summer crops; C3 = winter crop/
soybean and sunflower as summer crops; WF = winter fallow/soybean.
Nst = Nstconc × St N na palhada x produtivdade palhada

Nup = Nh + Nst uptake random effect. Soil properties were analyzed using a similar model
without the year effect. Means were compared using Fisher’s protected
Nsu = Nf + NsuN 0 supply least significant difference (LSD) test at the 5% probability level.

ref
NupD = Nup Nup Não da para usar pq n tem um experimento de referencia 3. Results
NUEf = Y / Nf
3.1. Soil properties
fexp = Ngr / Nsu
After 20 years, we observed statistically significant effects of crop-
where Nup (kg ha−1) is the sum of the N harvested in grain (Nh) and N in ping system on SOC only in the topsoil layer (0−5 cm) and small but
straw (Nst), (the N in roots was not considered in these calculations), significant effects on STN in both layers (0−5 and 5−20 cm) (Table 4).
N0
Nsu is the N uptake from N0 plots and is an indicator of soil N supply, All cropping systems under no-till had SOC > 27 g kg−1 in the top 5 cm
NupD was defined as the difference in N uptake between each cropping of soil. However, ROT_NT had the highest INF and MWD of all cropping
system and Nup ref
(defined as the N uptake under ROT_NT and N190), and systems. In the last 20 years, this system was only 10 years under an-
Nf is the N applied as fertilizer. The NUEf (kg kg−1) assesses efficiency nual crops (Table 2). Continuous annual cropping under not till
based on grain yield and N from fertilizer. The fraction fexp is the pro- (CC_NT) with winter crop/sorghum and maize or winter fallow/soy-
portion of N supply exported from each cropping system. To capture bean, reduced potentially mineralizable N and water infiltration rate.
differences among cropping systems on total N absorption efficiency While ROT_CT had lower top-soil organic carbon than no-till systems, it
(NAE), we use the slope of the regression between N uptake and N had the highest potentially mineralizable nitrogen and similar water
supply including a cropping system dummy variable to estimate the infiltration rate to no-till systems. The worst values in soil quality in-
slope for our reference system (ROT_NT, the more frequent cropping dicators were observed after 20 years under CC_CT.
system in this region prior to 2002) and the deviations (or delta) from
this slope for the other systems; the intercept was forced to the origin.
3.2. Wheat yield
To compare systems at equivalent yield, we estimated both Nf and
NUEf that would have been achieved using the minimum yield and N
Wheat yield differed between years (5.7 vs 4.7 Mg ha−1 for 2013
uptake obtained across systems in the N190 treatment (YminN190 and
and 2014, respectively), cropping system, N application and their in-
minN190
Nup ), i.e. using as target yield that obtained with N190 in the
teraction (P ≤ 0.05, Fig. 1). Total irrigation amounted to 87 and 60 mm
system that yielded the least (CC_CT at N190). In this way, all com-
for 2013 and 2014, respectively (Appendix 1). However, yield did not
parisons are made with a yield achievable in any system. The corre-
sponding Nf (NYminN190 ) and NUEf (NUEfYminN190 ) are calculated as: differ between Rainfed and Irrigated treatments (5.1 vs 5.2 Mg ha−1).
f
Since this was the case for other response variables as well, we removed
minN190
Nsurplus = Nup Nup water availability from further analyses.

NYminN 190 Nsurplus

f = 190 NAE 3.3. Wheat yield, N supply, yield gap and N deficit
minN 190
NUEfYminN 190 = Y NfYminN 190 Cropping system and N fertilizer rate significantly affected wheat
yield and N uptake (P ≤ 0.05), generating a quantifiable Yg and N
where Nsurplus represents the additional N uptake supplied by soil plus uptake deficit (Fig. 1). The highest YN 0 was obtained under ROT_NT (4.7
fertilizer under each cropping system (Nsurplus is always positive or Mg ha−1). Two continuous cropping systems under no-till (winter crop/
zero). sorghum and maize and winter fallow/soybean) reduced YN 0 to 4.1 Mg
ha−1 (P ≤ 0.05). When soybean was a frequent summer crop or the
2.6. Statistical analysis system was under conventional tillage (ROT_CT), YN 0 was further re-
duced to 3.7 Mg ha−1 (P ≤ 0.05). The lowest YN 0 was obtained under
Wheat yield, N uptake, and N use efficiency components were CC_CT (3.2 Mg ha−1). On average, wheat yield increased in N80
analyzed using a linear mixed model, where year, cropping system, N compared to N0 for all systems (P ≤ 0.05). However, YN80 was reduced
fertilization rate and water supply are fixed effects and blocks is a from 5.3 Mg ha−1 under ROT independently of tillage system to 4.9 Mg

5
O.R. Ernst, et al. Field Crops Research 257 (2020) 107934

Fig. 1. Wheat yield (left panel) and N uptake (right panel) at three N fertilization rates (N0 = white, N80 = shadow, and N190 = dotted pattern) in response to 20
years of crop-pasture rotation under no-till (ROT_NT) or conventional tillage (ROT_CT), continuous annual cropping under no-till with maize and sorghum as summer
crops (CC_NTC4), sunflower or soybean as summer crops (CC_NTC3), or winter fallow (CC_NTWF); and continuous annual cropping under conventional tillage (CC_CT).
The yield gap terms in the figure are: YgN = due to reduced soil N supply, Y gothers = due to deterioration of soil properties other than soil N supply capacity, and
YgT =YgN + Y gothers . Ymax = maximum yield obtained under not-limiting N supply. NupD is to difference in Nup at N190 between each cropping system and ROT_NT.
Different letters indicate significant differences at P < 0.05 in wheat yield and Nup between cropping systems (not between N application rates).

ha−1 under all CC_NT, and to 4.0 Mg ha−1 under CC_CT (P ≤ 0.05). Table 5
ROT_NT showed higher N uptake at all N fertilization levels. Estimated effect of cropping system on total yield gap (YgT ), yield gap due to N
Compared to this cropping system, N uptake at N0 was reduced on supply (YgN ), yield gap due to limited N supply (YgN 80 ), and due to factors other
average by 20 % under the three continuous annual cropping systems than N supply (Y gothers ). Relative total yield gap (RYgT ) was estimated as
under no-till (125 vs 100 kg ha−1, P ≤ 0.05). Under conventional til- 100 ×YgT / Ymax . The fraction of the RYgT explained by N supply (FYgN ) or factors
lage, N uptake at N0 was reduced by 28 % when crops were rotating other than N supply (FY gothers ) were also estimated so that FYgN + FY gothers = 1.
with pasture (ROT_CT) and by 46 % under continuous annual cropping The proportion of YgN explained by N80 (FYgN80 ) was estimated as
(CC_CT) (125 vs 89 and 67 kg ha−1, P ≤ 0.05). (YgN YgN 80 )/ YgN . The systems are described in the caption of Table 4.
The Ymax (7.2 Mg ha−1) was achieved at YN190 only by ROT_NT and
CC_NT with high frequency of winter crop/sorghum and maize, while System YgN Y gothers YgT YgN80 RYgT FYgN FY gothers FYgN 80
CC_NT including winter crop/soybean or winter fallow/soybean had Mg ha−1
YN190 similar to that of ROT_CT and averaged 6.1 Mg ha-1, 15 % or 1.1
ROT_NT 2.4 0.07 2.5 1.6 0.35 0.96 0.04 0.33
Mg ha-1 below Ymax (P ≤ 0.05). The YN190 under CC_CT was 5.2 Mg ha-1, CC_NTC4 3.0 0.00 3.0 2.3 0.41 1.00 0.00 0.23
28 % or 2.0 Mg ha-1 below Ymax. Yield differences between ROT and CC CC_NTWF 2.1 1.0 3.1 1.1 0.43 0.68 0.32 0.48
under winter crop/sorghum and maize, both under no-till, with CC CC_NTC3 2.0 1.4 3.4 0.8 0.47 0.59 0.41 0.60
including winter fallow/soybean or winter crop/soybean correspond to ROT_CT 2.7 0.8 3.5 1.1 0.49 0.77 0.23 0.60
2.0 2.0 4.0 1.2 0.56 0.50 0.50 0.40
N uptake deficit of 40 kg ha-1. Comparing ROT and CC under conven- CC_CT

tional tillage with no-till, the N uptake deficit increased 17 and 64 kg

ha-1, respectively (Fig. 1).
Table 6
The Ymax measured in our experiment (7.2 Mg ha−1) compares well
N absorption efficiency (NAE, kg kg−1) estimated as the slope of wheat N up-
with the potential wheat yield estimated with CropSyst (Stöckle et al., take regressed against N supply for the six cropping systems studied in this
2003) for the two years (7.0 Mg ha−1). The yield gap explained by N experiment. The NAE for the crop-pasture rotation under no-till (ROT_NT ) was
supply (or YgN ) was 2.4 and 3.0 Mg ha-1 for ROT_NT and CC_NT under set as the reference and the delta slope estimated for the other systems. The two
winter crop/sorghum and maize (the two systems with the highest years were pooled for this analysis. The systems are described in the caption of
YN190 ), 2.7 Mg ha-1 for ROT_CT, and 2.0 Mg ha-1 for the other cropping Table 4.
systems (Table 5).
NAE Std. error P value
Under no till, ROT and CC under winter crop/sorghum and maize, kg kg−1
the yield reduction was all attributable to YgN (FYgN of 0.96 and 1, re-
spectively; Table 5). Under other continuous cropping systems, the ROT_NT 0.74 0.02 < 0.001
percentage of Yg explained by factors other than soil N supply increased Delta CC_NTC4 0.10 0.03 0.001
from 32 % under CC_NT including winter fallow/soybean to 50 % when Delta CC_NTWF −0.02 0.03 0.693
Delta CC_NTC3 −0.02 0.03 0.693
CC was under conventional tillage, while it was only 23 % for rotating Delta ROT_CT 0.10 0.03 0.001
crop-pasture under conventional tillage (ROT_CT). Delta CC_CT −0.09 0.03 0.05
N uptake was linearly related to N supply and the slope was affected
by the cropping system (Table 6). The slope between N supply and N
uptake represents the N absorption efficiency (NAE). Using the NAE of (N gr
conc
, Nh, NUEf and fexp) were affected by cropping system and N fer-
0.74 kg kg−1 for ROT_NT as the benchmark, we found that two of the tilizer rate (P ≤ 0.05, Table 7). Once the N limitations were removed
three continuous cropping systems under no till (under winter crop/ (N190) all cropping systems had the same N grain concentration (≈ 24
soybean and winter fallow/soybean) had the same NAE; however, NAE g kg−1). However, when N uptake was reduced by cropping system
increased to 0.84 kg kg−1 including sorghum and maize under CC_NT (CC_NT and CC_CT) or N fertilizer rates, differences among systems
and ROT_CT (P < 0.05) and decreased to 0.65 kg kg−1 under CC_CT were large (P ≤ 0.05), with N grain concentration varying between 16
(P < 0.05). and 20 g kg−1. The fraction of N supply exported varied from 0.66 to
The N grain concentration, amount of N harvested, NUE based on 0.48; it was significantly lower under N190 than under N80. Increasing
grain yield and N from fertilizer and the fraction of N supply exported soil N supply (N uptake under N0) improved NUE from fertilizer, which

6
O.R. Ernst, et al. Field Crops Research 257 (2020) 107934

Table 7 4. Discussion
Cropping system effect on wheat grain nitrogen concentration (N grconc ), har-
vested nitrogen (Nh), nitrogen use efficiency (NUE) based on the ratio of grain We demonstrated that cropping systems that degraded soil proper-
yield to N from fertilizer (NUEf), and fraction of the nitrogen supply exported ties reduced both, attainable wheat yield (≈ YN190) and NUE. The po-
from each cropping systems (fexp) under the three N fertilizer rates (N0, N80 tential wheat yield estimated using CropSyst (7.0 Mg ha−1) matched
and N190). The cropping systems are crop-pasture rotation under no-till our measured maximum wheat yield (YN190 under ROT_NT and CC_NT
(ROT_NT ) or conventional tillage (ROT_CT ); three continuous annual cropping
under winter crop/sorghum and maize), lending certainty to our Yg
systems under no-till (CC_NT), (CC_NTC4 = winter crop/sorghum and maize;
estimates. By definition, potential yield is achieved under non-limiting
CC_NTC3 = winter crop/soybean; CC_NTWF continuous winter fallow/soybean),
and continuous annual cropping under conventional tillage (CC_CT ).
water and nutrient availability and with no biotic stress (van Ittersum
et al., 2013). Potential yield estimates using crop growth models should
Cropping System not be affected by soil quality. In this study we ensured that cropping
ROT_NT CC_NTC4 CC_NTWF CC_NTC3 ROT_CT CC_CT
system specific YN190 were achieved under non-limited water and nu-
trient availability and with no biotic stress. Consequently, the differ-
conc
N gr (g kg−1) ence between potential yield and the cropping system specific YN190
N0 21 a 17 c 20 a 17 c 20 b 16 c must represent the impact in yield caused by a change in soil properties
N80 23 a 21 b 21 b 22 b 21 b 21 b that are independent of the nutrient supply. We started the long-term
N190 24 a 24 a 24 a 25 a 25 a 24 a
experiment in 1993 on a soil that after 20 years of contrasting cropping
Nh (kg ha−1)
N0 100 a 72 c 83 b 64 c 72 c 51 d systems, rendered new phenoforms (Droogers and Bouma, 1997;
N80 126 a 103 b 107 b 109 b 111 b 85 c Rossiter and Bouma, 2018).
N190 171 a 173 a 149 b 144 b 159 b 124 c Our measurements revealed that cropping systems affected physical
fexp (MWD and INF) and chemical properties (SOC, STN and PMN) and had
N80 0.61 ab 0.57 b 0.58 b 0.63 a 0.66 a 0.58 b
N190 0.54 ab 0.59 a 0.51 ab 0.51 ab 0.57 a 0.48 b
a significant impact on YN190 . Two systems, ROT_NT and CC_NT under
NUEf (kg kg−1) winter crop/sorghum and maize, had YN190 similar to potential yield
N80 69 a 61 b 64 ab 62 ab 66 ab 50 c and averaging 7.1 Mg ha−1, but other systems gradually degraded soil
N190 38 a 38 a 33 b 30 b 34 b 27 c properties so that YN190 decreased to 6.1 Mg ha−1 (CC_NT with winter
crop/soybean or winter fallow/soybean, and ROT_CT) and finally to 5.2
Numbers followed by different letter in the same row are significantly different
Mg ha−1 (CC_CT) (Fig 2). When cropping systems affected only STN
(P ≤ 0.05).
and PMN, 100 % of YgT was explained by N supply. However, these soil
All differences in N grconc , Nh, and NUEf among N rates in the same column are
significant (P ≤ 0.05). N supply indicators were poorly related to YgN , implying that cropping
systems captured more than just effects of soil total N concentration and
mineralization potential (R2 of 0.5 and 0.1, respectively). Most likely
varied from 69 to 50 kg kg−1 (kg of grain per kg of N added) for N80,
other soil N supply indicators, more sensitive and directly related to
and from 38 to 27 kg kg−1 for N190 (ROT_NT and CC_CT, respectively).
mineral N release, or multiple soil properties (or related soil functions)
must be incorporated to improve the soil N supply capacity estimation
related to wheat yield in the short time. When soil physical properties
3.4. N fertilizer required to reach the same grain yield and NUEf in each
were also affected, Y gothers explained up to 50 % of YgT (ROT_NT and
cropping system
CC_NT under winter crop/sorghum and maize vs CC_CT (Table 5). As
we sowed all treatment under no-till, we are quantifying residual effects
Compared to CC_CT under N190, the N fertilizer required under
of cropping systems on Yatt, Yg and Yg composition. These effects reflect
ROT_NT to reach the same yield was reduced by 86 kg ha−1 (190 vs 104
the loss in productivity in response to subtle but cumulative soil de-
kg ha-1), increasing NUE of N applied as fertilizer from 38 to 50 kg kg-1
gradation imposed by some of the long-term cropping systems. While
(Table 6 and Table 8 respectively). Furthermore, the CC_NT under
the effects of tillage systems and crop- pasture rotation on soil quality
winter crop/sorghum and maize required significantly less N fertilizer
indicators are well documented, in this research we quantified the ef-
than CC_NT under winter crop/soybean to achieve the same yield (107
fect throughout the aggregated effect on attainable wheat yield (kg
vs. 163 kg ha-1).
ha−1). Or in other words, attainable yield is a biological indicator of
Insert Table 8.
soil quality.
According to Newman et al. (2010) and Müller et al. (2012), N is the
most important yield restriction under actual wheat management in the
Table 8 eastern Pampas. However, Ernst et al. (2018) found in experiments
Nitrogen uptake surplus (Nsurplus; see text for calculations), and both expected N conducted in producers fields that segregating YgT into YgN and Y gothers
fertilizer required in each cropping system to match the wheat yield in the revealed a loss of soil quality in response to the prevailing cropping
CC_CT system (5(NfYminN190 = 5.2 Mg ha−1) and expected N fertilizer use effi- systems. This research confirms the results obtained by Ernst (2018),
ciency (NUEYminN190 ). The cropping systems are crop-pasture rotation under no- implying that if one only measures YgN , then the YgT can be under-
f
estimated under the current no-till cropping system centered on soy-
till (ROT_NT ) or conventional tillage (ROT_CT ); three continuous annual
cropping systems under no-till (CC_NT), (CC_NTC4 = winter crop/sorghum and bean in the summer annual phases. Ignoring limitations to wheat yield
maize; CC_NTC3= winter crop/soybean; CC_NTWF continuous winter fallow/ other than those derived from limiting N supply may lead to an over-
soybean), and continuous annual cropping under conventional tillage (CC_CT ). estimation of attainable yield (for the soil is not expressing the true
potential of the environment), and an inefficient use of N if the target
Cropping system
yield is overestimated.
ROT_NT CC_NTC4 CC_NTWF CC_NTC3 ROT_CT CC_CT Our estimated YgT varied from 2.5 Mg ha−1 for ROT_NT to 4.0 Mg
ha−1 for CC_CT, and was ≈ 3 Mg ha−1 in continuous no-till systems
Nsurplus (kg ha−1) 64 70 28 20 47 0 (Table 5). All cropping systems had lower N grain concentration at N0
(kg 104 107 152 163 134 190
NYminN190
f and N80 than ROT_NT (Table 6). This reflects a rarely discussed con-
ha−1) sequence of soil degradation: a loss of nutritional quality of the har-
50 49 34 32 39 27
NUEYminN190
f vested grain, or a reduction in NUE from fertilizer that forces higher
(kg kg1) fertilization rates to reach a given N grain concentration. A yield-

7
O.R. Ernst, et al. Field Crops Research 257 (2020) 107934

focused interpretation would simply acknowledge that the reduced N water and air movement that, in turn, directly affect crop productivity
supply of a degraded soil would increase Yg, however we clearly show (Kasper et al., 2009; Kumar et al., 2012). In our research, wheat yield
that there is also a prominent impact on YgN and N grain concentration, and N grain concentration became indicators of such changes in soil
because the limiting N supply causes a N grain dilution. In other words, properties. In practical terms, our results provide evidence that in the
increasing YgN (both due to low N supply or low NUE), carries with it a soil used in this experiment, water infiltration rate and soil aggregate
decreasing N grain concentration and an additional monetary loss: in stability should be maintained over certain critical thresholds (Table 4).
bread wheat production, N grain below a desirable threshold get a price The soil N supply capacity depletion can be inferred from the reduction
penalty (Grahmann et al., 2014). in soil total N and N potentially mineralizable N measured in this study
The Ymax of 7.1 Mg ha−1 was obtained in systems where N uptake (Table 4). These thresholds are soil and methodology dependent. We
was higher than 230 kg ha−1, implying that reaching this yield require recognize that the infiltration rate measured in this experiment is not
about 30 kg of N per Mg−1 of grain of wheat with of 24 g of N kg−1. the saturated hydraulic conductivity, but a related variable. The 2 cm
Degrading soil conditions reduced N uptake, (Table 7), but all cropping h−1 measured under the two best cropping systems (ROT_NT, and
systems produced 36 kg kg-1 (Appendix 2), indicating that N uptake CC_NT under winter crop/sorghum and maize) is twice as large as that
deficit was the main cause of yield reductions (Fig. 1). Our results also corresponding to the saturated hydraulic conductivity obtained using
showed that the N uptake deficit resulted from reduced NAE at pedotransfer functions from Saxton and Rawls (2006) for soils with this
equivalent N supply (Table 6), reducing NUE from fertilizer (Table 8). texture. The pedotransfer functions, however, may underestimate
Thus, even when the N is in the soil, the soil degradation reduces the N macropore flow, which can dominate under perennial pastures. A re-
recovery by wheat. While to obtain 5.2 Mg ha−1 of wheat under vealing outcome of our experiment is that no-till, high residue input
ROT_NT required an N supply of 229 kg ha−1 (125 kg of soil Ns +104 and densely rooted maize and sorghum, as opposed to soybean, seem to
kg of Nf), the same wheat yield under CC_CT required an N supply of provide similar benefits on soil physical properties to perennial pas-
257 kg ha−1 to compensate for lower soil N supply and lower N re- tures, and may allow decoupling grain production from pastures and
covery (67 kg of soil Ns +190 kg of Nf) (Table 8). animal production without degrading soil properties.
Increased N fertilizer requirements to reach the same yield imply One of the pillars of ecological intensification proposed by Cassman
increasing risks of N losses per unit of product and per unit area through (1999) is the maintenance or improvement of soil quality, defined as
volatilization, denitrification, leaching or runoff, particularly when N the capacity of soils to sustain biological productivity, while ensuring
supply exceeds the removal through harvest (McLellan et al., 2018). environmental, plant and animal health (Doran and Parkin, 1994;
This is illustrated by comparing ROT_NTN80 to CC_CTN190 (YminN190). Blum, 2005). Our results suggest that rotating annual crops under no-
The fraction of N supply exported from each cropping system (fexp) was till is not enough to fulfill this aspiration, and that progress towards a
reduced from 0.61 to 0.48, implying that N potential losses were in- sustainable intensification of agriculture under a continuous no-till
creased from 80 to 131 kg ha−1, while the grain yield was almost the cropping system would require re-balancing crop sequences towards
same. crop-pasture rotations under no-till (Ernst et al., 2018; Pravia et al.,
Under current N fertilization rates used by most producers (N80), 2019) or a shift from the dominance of soybean in annual no-till system
wheat yield would be limited by N supply in all systems, and severely so (under winter fallow/soybean or winter crop/soybean) towards systems
in winter crop/soybean or winter fallow/soybean annual sequences with higher frequency of maize and sorghum (winter crop/sorghum
under no-till, which are the most common cropping systems used by and maize as summer crop). While the later no-till systems would have
wheat producers in the region in the last 15 years. To close the Yg, and reduced N supply and will require higher N inputs through fertilization,
assuming as a thought experiment that it is possible for every cropping soil physical properties will remain above thresholds that will not
systems to reach Ymax of 7.1 Mg ha−1 maintaining N grain concentra- compromise the expression of the highest yields, and will likely carry
tion (24 g kg−1) (i.e. the performance of wheat under ROT_NT and forward other environmental benefits associated with healthy soils.
N190; Fig. 1, Table 7) and with the NAE obtained for each system
(Table 6), the N fertilizer rate would have to increase to at least 239 kg 5. Conclusions
ha−1 under these two CC_NT, but only to 210 kg ha−1 under ROT_NT.
Crop-pasture rotation is an ancient cultural management practice Our study revealed that there are two pathways towards main-
(Russelle et al., 1987) that maintains or improves soil quality, but the taining soils that can achieve the highest yields under continuous no-
specific technology and crop sequence used in the rotation alters both till: rotating crops with pastures (ROT_NT) or annual continuous
economic and environmental outcomes. Our reference ROT_NT system, cropping with high frequency of maize and sorghum in the summer
maintained soil aggregates stability, water infiltration rate and PMN at phase of the rotation. The rotation with pastures has an additional
the highest levels, requiring less N fertilizer to reach Ymax, albeit with benefit: lower N fertilizer requirements than continuous double annual
fewer grain harvests per rotational cycle than continuous annual winter crop/sorghum and maize under no-till. However, it reduces the
cropping. Compared to the traditional ROT_CT widely used until 1990, number of grain harvests through the rotation cycle. The net economic
the switch to ROT_NT at the end of the 20th century in the eastern result depends on the profit from grain or from animal production
Pampas, represented a step towards sustainable intensification. How- during the pasture phase. Continuous annual cropping systems under
ever, after 2002, the market-driven shift towards CC_NT with a high no-till with high frequency of sorghum and maize, is more dependent
frequency of soybean (CC_NTC3 and CC_NTWF), seems to work in the on N fertilizer inputs; the lack of pastures that include legumes reflects
opposite direction. In fact, soybean became the dominant crop in Ur- in lower PMN. All other systems limited wheat yield through a de-
uguay, Argentina, Bolivia, Paraguay and Southern Brazil (Rio Grande gradation of soil physical properties (expressed through INF and MWD)
do Sul, Santa Catarina, and Parana States), representing 63 % of total and required higher N fertilizer to reach both higher yields but also
cultivated area versus 19 % and 12 % of maize and wheat respectively higher N grain concentration, expressing a severe drop in NUE from
(FAOSTAT, 2016). Our results suggest that this production system shift, fertilizer. The yield gap caused by soil degradation is exacerbated by a
despite its short-term economic justification (Ernst et al., 2018), follows costly yield quality gap. This form of soil degradation is almost im-
a costly intensification pathway that, on a soil quality basis cannot be perceptible to farmers in production conditions, as soybean dominates
qualified as sustainable, because productivity and nutrient use effi- the summer phase of the rotations in the Pampas region. The results of
ciency progressively and sharply degrade. our experiment clearly show that high soybean frequency in the crop
Continuous annual cropping systems with high frequency of soy- rotation can cause a yield penalty on wheat crops. Rotations with
bean, even under no-till, showed degradation of soil aggregation which pastures, or a less drastic change towards increased frequency of maize
influence soil biological and physical processes such as root growth and and sorghum in the rotation, would prevent the expression of this

8
O.R. Ernst, et al. Field Crops Research 257 (2020) 107934

difficult-to-diagnose soil degradation, most likely reducing environ- Droogers, P., Bouma, J., 1997. Soil survey input in exploratory modeling of sustainable
mental pollution and increasing both wheat yield and nitrogen use ef- soil management practices. Soil Sci. Soc. Am. J. 61, 1704–1710.
Ernst, O., Siri-Prieto, G., 2009. Impact of perennial pasture and tillage systems on carbon
ficiency. input and soil quality indicators. Soil Tillage Res. 105 (2), 260–268. https://doi.org/
10.1016/j.still.2009.08.001.
Author contribution Ernst, O., Kemanian, A., Mazzilli, S., Cadenazzi, M., Dogliotti, S., 2016. Depressed at-
tainable wheat yields under continuous annual no-till agriculture suggest declining
soil productivity. Field Crops Res. 186, 107–116.
Oswaldo Ernst: conceived and established the long-term experi- Ernst, O., Dogliotti, S., Cadenazzi, M., Kemanian, A., 2018. Shifting crop-pasture rotations
ment, built the analytical framework and statistical models for the to no-till annual cropping reduces soil quality and wheat yield. Field Crop
Res.Research 218, 180–187.
analysis, and wrote the manuscript. Armén Kemanian and Santiago FAOSTAT, 2016. Agricultural Data: Crops and Livestock Primary and Processed. FAO,
Dogliotti: contributed to the conception of the manuscript, discussed Rome [Online] Available at http://faostat.fao.org. (Accessed May 2018).
the results with OE and edited manuscript. Guillermo Siri-Prieto: con- Franzluebbers, A.J., Sawchik, J., Taboada, M.A., 2014. Agronomic and environmental
impacts of pasture and crop rotations in temperate North and South America. Agric.
tributed to the design of the changes to the long-term experiment and
Ecosyst. Environ. 190, 18–26. https://doi.org/10.1016/j.agee.2013.09.017.
edited the manuscript. Sebastián Mazzilli: did the simulations with García-Préchac, F., Ernst, O., Siri-Prieto, G., Terra, J.A., 2004. Integrating no-till into crop
CropSyst and contributed to the data discussion. and pasture rotations in Uruguay. Soil Tillage Res. 77, 1–13. https://doi.org/10.
1016/j.eja.2008.10.005.
Garnett, T., Appleby, M.C., Balmford, A., Bateman, I.J., Benton, T.G., Bloomer, P.,
Declaration of Competing Interest Burlingame, B., Dawkins, M., Dolan, L., Fraser, D., et al., 2013. Sustainable in-
tensification in agriculture: premises and policies. Science 341, 33–34.
Grahmann, K., Verhulst, N., Peña, R.J., Buerkert, A., Vargas-Rojas, L., Govaerts, B., 2014.
The authors declare that they have no known competing financial Durum wheat (Triticum durum L.) quality and yield as affected by tillage-straw
interests or personal relationships that could have appeared to influ- management and nitrogen fertilization practice under furrow-irrigated conditions.
ence the work reported in this paper. Field Crops Res. 164, 166–177.
Huggins, D.R., Pan, W.L., 1993. Nitrogen efficiency components analysis: an evaluation of
cropping system differences in productivity. Agr. J. 85, 898–905.
Acknowledgments Huggins, D.R., Allmaras, R.R., Clapp, C.E., Lamb, J.A., Randall, G.W., 2007. Corn-soybean
sequence and tillage effects on soil carbon dynamics and storage. Soil Sci. Soc. Am. J.
71, 145–154.
Funding for this research was provided by the Instituto Nacional de
INASE, 2015. Evaluación De Cultivares. http://www.inia.org.uy/convenio_inase_inia/
Investigación Agropecuaria FPTA #303. Support was also provided by Evaluacion_CI/Ano2015/JornadaInvierno2016.pdf. (Last cheked May 2018). .
Hatch Appropriations under Project #PEN04571 and Accession Jackson, R.B., Lajtha, K., Crow, S.E., Hugelius, G., Kramer, M.K., Piñeiro, G., 2017. The
ecology of soil carbon: pools, vulnerabilities, and biotic and abiotic controls. Annu.
#1003346. The long-term experiment was funding by World Bank
Rev. Ecol. Evol. Syst. 2017 (48), 419–445.
(PRENADER 036) (1993-1998); Comisión Sectorial Investigación Kasper, M., Buchan, G.D., Mentler, A., Blum, W.E.H., 2009. Influence of soil tillage sys-
Científica (CSIC I+D) (1997-1999; 2006-2008); CSIC Programa Grupos tems on aggregate stability and the distribution of C and N in different aggregate
de Investigación No 774, Proyecto No 139 (2011-2014). This research is fractions. Soil Tillage Res. 105, 192–199.
Kemper, W.D., Rosenau, R.C., 1986. Aggregate Stability and Size Distribution.
part of the doctoral dissertation of Oswaldo Ernst at the Universidad de Kumar, S., Kadono, A., Lal, R., Dick, W., 2012. Long-Term Tillage and Crop Rotations for
la República, Uruguay. 47–49 Years. Influences Hydrological Properties of Two Soils in Ohio. Soil Sci. Soc.
Am. J. 76, 2195–2207.
Mazzilli, S., Ferreira, G., Ernst, O., 2013. Zonas agrícolas en expansión: desarrollo de
Appendix A. Supplementary data herramientas para un manejo sustentable de los sistemas de producción. In: Ernst, O.,
Mazzilli, S., Kemanian, A. (Eds.), Zonas agrícolas en expansión: desarrollo de her-
Supplementary material related to this article can be found, in the ramientas para un manejo sustentable de los sistemas de producción. Calibración,
validación y aplicación de un modelo modelos de simulación de rendimientos de
online version, at doi:https://doi.org/10.1016/j.fcr.2020.107934. cultivos. Instituto Nacional de Investigación Agropecuaria. No. 50. Serie FPTA-INIA-
Uruguay, pp. 9–32.
References Mazzilli, S.R., Kemanian, A., Ernst, O., Jackson, R.B., Piñeiro, G., 2015. Greater humifi-
cation of belowground than aboveground biomass carbon into particulate soil organic
matter in no-till corn and soybean crops. Soil Biol. Biochem. 85, 22–30.
Alvarez, C.R., Taboada, M.A., Perelman, D., Morrás, J.M., 2014. Topsoil structure in no- Mazzilli, S., Ernst, O., de Mello, V., Pérez, C., 2016. Yield losses on wheat crops associated
tilled soils in the Rolling Pampa. Argentina. Soil Res. 52, 533–542. https://doi.org/ to the previous winter crop: impact of agronomic practices based on on-farm analysis.
10.1071/SR13281. Eur. J. Agron. 75, 99–104.
Alvarez, R., Steinbach, H.S., 2009. A review of the effects of tillage systems on some soil McLellan, E.L., Cassman, K.G., Eagle, A.J., Woodbury, P.B., Sela, S., Tonitto, C.,
physical properties, water content, nitrate availability and crops yield in the Marjerison, R.D., van Es, H.M., 2018. The nitrogen balancing act: tracking the en-
Argentine Pampas. Soil Tillage Res. 104, 1–15. vironmental performance of food production. BioScience 68 (3), 194–203.
Aparicio, V., Costa, J.L., 2007. Soil quality indicators under continuous cropping systems Modernell, P., Rossing, W.A.H., Corbeels, M., Dogliotti, S., Picasso, V., Tittonell, P., 2016.
in the Argentinean Pampas. Soil Tillage Res. 96, 155–165. https://doi.org/10.1016/j. Land use change and ecosystem service provision in Pampas and Campos grasslands
still.2009.02.005. of southern South America. Environ. Res. Lett. 11 (11) http://iopscience.iop.org/
Baeza, S., Paruelo, J.M., 2020. Land Use/Land cover change (2000–2014) in the Rio De La article/10.1088/1748-9326/11/11/113002. Available 27/6/2018.
Plata Grasslands: an analysis based on MODIS NDVI time series. Remote Sens. (Basel) Moll, R.H., Kamprath, E.J., Jackson, W.A., 1982. Analysis and interpretation of factors
2020 (12), 381. https://doi.org/10.3390/rs12030381. which contribute to efficiency to nitrogen utilization. Agron. J. 74, 562–564.
Blum, W.E.H., 2005. Functions of soil for society and the environment. Rev. Environ. Sci. Müller, N.D., Gerber, J.S., Johnston, M., Ray, D.K., Ramankutty, N., Foley, J., 2012.
Biotechnol. 4, 75–79. https://doi.org/10.1007/s11157-005-2236-x. Closing yield gaps through nutrient and water management. Nature 490, 254–257.
Boeni, M., Bayer, C., Dieckowc, J., Conceição, P.C., Pinheiro Dick, D., Knicker, H., Salton, Munkholm, L.J., Heck, R.J., Deen, B., 2013. Long-term rotation and tillage effects on soil
J.C., Motta Macedo, M.C., 2014. Organic matter composition in density fractions of structure and crop yield. Soil Tillage Res. 127, 85–91.
Cerrado Ferralsols as revealed by CPMAS 13C NMR: influence of pastureland, crop- Nelson, D.W., Sommers, L.E., et al., 1996. Total carbon, organic carbon, and organic
land and integrated crop-livestock. Agric. Ecosyst. Environ. 190, 80–86. https://doi. matter. In: Page, A.L. (Ed.), Methods of Soil Analysis, Part 2, 2nd ed. Agronomy. 9:
org/10.1016/j.agee.2013.09.024. 961-1010. Am. Soc. of Agron., Inc..
Carignano, M., Staggenborg, S.A., Shroyer, J.P., 2008. Management practices to minimize Newman, K., Verburg, P.H., Stehfest, E., Müller, Ch, 2010. The yield gap of global grain
tan spot in a continuous wheat rotation. Agron. J. 100, 145–153. production: a spatial analysis. Agric. Syst. 103, 316–326.
Cassman, K.G., 1999. Ecological intensification of cereal production systems: yield po- Novelli, L.E., Caviglia, O.P., Wilson, M.G., Sasal, M.C., 2013. Land use intensity and
tential. Soil quality. And precision agriculture. Proc. Natl. Acad. Sci. U.S.A. 96, cropping sequence effects on aggregate stability and C storage in a Vertisol and a
5952–5959. Mollisol. Geoderma 195, 260–267. https://doi.org/10.1016/j.geoderma.2012.12.
da Silva, F.D., Carneiro Amado, T.J., Oliveira Ferreira, A., Assmann, J.M., Anghinoni, I., 013.
de Faccio Carvalho, P.C., 2014. Soil carbon indices as affected by 10 years of in- Novelli, L.E., Caviglia, O.P., Piñeiro, G., 2017. Increased cropping intensity improves crop
tegrated crop–livestock production with different pasture grazing intensities in residue inputs to the soil and aggregate-associated soil organic carbon stocks. Soil
Southern Brazil. Agric. Ecosyst. Environ. 190 (2014), 60–69. Tillage Res. 165, 128–136.
DIEA, 2019. Annual Statistics From the Agricultural Sector in Uruguay. DIEA-MGAP. Peiretti, R.A., Dumanski, J., 2014. The transformation of agriculture in Argentina through
http://www.mgap.gub.uy. soil conservation. Int. Soil Water Conserv. Res. 2, 14–20. https://doi.org/10.1016/
Doran, J.W., Parkin, T.B., et al., 1994. Defining and assessing soil quality. P. 3-21. In: S2095-6339(15)30010-1.
Doran, J.W. (Ed.), Defining Soil Quality for a Sustainable Environment. SSSA Spec. Pravia, M.V., Kemanian, A.R., Terra, J.A., Shi, Y., Macedo, I., Goslee, S., 2019. Soil carbon
Publ. No. 35, Soil Sci. Soc. Am., Inc. and Am. Soc. Agron., Inc, Madison, WI. saturation, productivity, and carbon and nitrogen cycling in crop-pasture rotations.

9
O.R. Ernst, et al. Field Crops Research 257 (2020) 107934

Agric. Syst. 171, 13–22. AL. Soil Conservation Service. 1972. Soil Survey Laboratory Methods Procedures for
Redo, D.J., Aide, T.M., Clark, M.L., Andrade-Núñez, M.J., 2012. Impacts of internal and Collecting Soils and Samples. Soil Surv Report 1. USDA, Washington. EE.UU.
external policies on land change in Uruguay, 2001–2009. Environ. Conserv. 39, Stöckle, C.O., Donatelli, M., Nelosn, R., 2003. CropSyst, a cropping systems simulation
122–131. https://doi.org/10.1017/S0376892911000658. model. Eur. J. Agron. 18, 289–307.
Rossiter, D.G., Bouma, J., 2018. A new look at soil phenoforms – definition, identification, Tourn, S.N., Videla, C.C., Studdert, G.A., 2019. Ecological agriculture intensification
mapping. Geoderma 314, 113–121. through crop-pasture rotations does improve aggregation of Southeastern-Pampas
Russelle, M.P., Hester man, O.B., Sheaffer, C.C., Heichel, G.H., 1987. Estimating nitrogen Mollisols. Soil Tillage Res. 195, 104411. https://doi.org/10.1016/j.still.2019.
and rotation effects in legume-corn rotations. In: Power, J.F. (Ed.), The Role of 104411.
Legumes in Conservation Tillage Systems. Soil Conservation Society, Washington, van Ittersum, M.K., Rabbinge, R., 1997. Concepts in production ecology for analysis and
D.C, pp. 41. quantification of agricultural input-output combination. Field Crops Res. 52,
Salton, J.C., Mercante, F.M., Tomazi, M., Zanatta, J.A., Concenc, G., Silva, W.M., Retore, 197–208.
M., 2014. Integrated crop-livestock system in tropical Brazil: toward a sustainable van Ittersum, M.K., Cassman, K.G., Grassini, P., Wolf, J., Tittonel, P., Hochmanna, Z.,
production system. Agric. Ecosyst. Environ. 190 (2014), 70–79. 2013. Yield gap analysis with local to global relevance: a review. Field Crops Res.
Salvo, L., Hernández, J., Ernst, O., 2010. Distribution of soil organic carbon in different 143, 4–17.
size fractions, under pasture and crop rotations with conventional tillage and no-till Varvel, G.E., 2006. Soil organic carbon changes in diversified rotations of the western
systems. Soil Tillage Res. 109, 116–122. Corn Belt. Soil Sci. Soc. Am. J. 70, 426–433.
Sasal, M.C., Andriulo, A.E., Taboada, M.A., 2006. Soil porosity characteristics and water Waring, S.A., Bremner, J.M., 1964. Ammonium production in soil under water logged
movement under zero tillage in silty soils in Argentinian Pampas. Soil Tillage Res. 87 conditions as an index of nitrogen availability. Natura (Landon) 201, 951–952.
(1), 9–18. Wingeyer, A.B., Amado, T., Pérez-Bidegain, M., Studdert, G.A., Perdomo Varela, C.H.,
Sasal, M.C., Boizard, H., Andriulo, A.E., Wilson, M.G., Léonard, J., 2016. Platy structure Garcia, F.O., Karlen, D.L., 2015. Soil quality impacts of current south american.
development under no-tillage in the northern humid Pampas of Argentina and its Sustainability 7, 2213–2242. https://doi.org/10.3390/su7022213.
impact on runoff. Soil Tillage Res In Press, Corrected Proof, Available online 21 Yoder, R.E., 1936. A direct method of aggregate analysis of soil and study of the physical
March 2017. nature of erosion losses. J. Am. Soc. 28, 337–351.
Saxton, K.E., Rawls, W.J., 2006. Soil water characteristic estimates by texture and organic Zadoks, J.C., Chang, T.T., Konzak, C.F., 1974. A decimal code for the growth stages. Weed
matter for hydrologic solutions. Soil Sci. Soc. Am. J. 70 (5), 1569–1578. Res. 14, 415–421.
Soil Quality Institute, 1999. Soil Quality Test Kit Guide. NRCS Soil Quality Inst. Auburn.

10