Zero-tillage (ZT) and crop residue retention on the soil surface—two components of Conservation Agriculture (CA)—have been identified as promising management practices for sustainable agri- cultural intensification for some time. However, CA technology uptake by farmers in the dry areas of West Asia and North Africa (WANA) has yet to happen large-scale, even though the positive im- pact on yield has been demonstrated repeatedly. To explain the observed consistent increases in crop yield under ZT, a range of soil quality indicators were regularly monitored during 2008 to 2012 in the long-term ZT/CA trials at the headquarters of the International Center for Agricultural Research in the Dry Areas (ICARDA) in northern Syria. Results showed that CA had a positive im- pact on soil fertility. This was measurable by higher soil organic matter (SOM) and microbial bio- mass contents, increased levels of extractable phosphate, sometimes (but not always) higher amounts of larger water-stable soil aggregates, increased soil infiltration capacity and soil water retention. The buildup of SOM and associated carbon (C) sequestration was in the range of 0.29 Mg C/ha/yr, i.e. rather modest. High amounts of surface residues delayed the desiccation of the topsoil during the fallow period, but could not diminish the overall longer-term drying of the topsoil. The observed positive changes in soil quality were little, but nevertheless, in combination with the economic savings that ZT offers, this type of agricultural intensification provides an attractive op- tion for farmers in WANA, from the standpoints of economy and ecological efficiency.
The Mediterranean region has been cultivated for millennia and is the site of the evolution of settled agriculture and the center of origin of a range of important crops [
In recent decades, agricultural production has intensified in the Mediterranean region, especially the increase with continuous cropping instead of fallow and application of mineral fertilizer. To mitigate the negative impact of such intensification, various management practices have been developed, such as rotations as an alternative to mono-cropping [
Uptake of CA by farmers in the Mediterranean region is lagging behind uptake elsewhere in the world, even in countries such as Morocco and Tunisia where considerable research has been done [
To address some of these issues, the International Center for Agricultural Research in the Dry Areas (ICARDA) has carried out experiments on-station or demonstrations with farmers, to test ZT and residue reten- tion and their impact on crop yield [
The experimental fields were located at ICARDA headquarters, Tel Hadya, 30 km south of Aleppo in northern Syria (36.011˚N 36.931˚E, 285 m above sea level). The climate is semiarid with an average annual temperature of 17.8˚C. Details of the station and its role in technology development and transfer in the WANA region have been recently described by [
The main soils of the region have been classified as Vertisols, Inceptisols, and Aridisols [
ICARDA has maintained three long-term CA experiments at Tel Hadya, termed B4, B7 and C16 trials (
In the B4 trial, the effects of tillage (CT vs ZT) and early and late planting were tested in a four-course rotation of wheat, chickpea, barley, and lentil. The trial was initiated in 2006/07 in a split plot design with tillage as main plots and crops as sub plots; the trial was further split into time of sowing (early vs late) sub-sub plots in 2008/ 09; there were four replications and a plot size of 780 m2.
The barley-lentil-wheat-chickpea rotation treatment was duplicated so as to have each crop present in the field each year. CT comprised moldboard ploughing after cereals and disk ploughing after legumes soon after harvest in June or July, and a second tillage by a tine cultivator before sowing in the autumn. ZT comprised sowing di- rectly into uncultivated stubble, with pre-sowing control of weeds (Glyphosate) usually only necessary on late-sown plots. All plots were sown with a ZT seeder. Early planting was undertaken after the first germinating rains (early-mid November) and late planting about 4 weeks later (early-mid December). Crops received basal fertilizer, i.e. 150 kg/ha di-ammonium phosphate (18% N, 46% P2O5) for cereals and 150 kg/ha triple super- phosphate (46% P2O5). Cereals received an additional 40 kg/ha urea (46% N) around tillering 2 - 3 months after sowing. Post-emergence weeds were controlled with selective herbicides as necessary.
Google Earth image of ICARDA headquarters in northern Syria and location of the three main conservation agriculture trials (C16, B4 and B7); dotted sub-plots in C16 indicate the three blocks of this trial (1-3 from west to east); the network of research station fields and roads is clearly visible
The B7 trial was laid out in a split-split plot design with four replications and a plot size of 120 m2 and a total of 112 plots. Established in 1998, the trial had three treatments: wheat-legume crop rotation, different levels of land preparation by tillage, and plus/minus N-fertilizer application. Seven different tillage practices were tested, which furthermore differed depending on whether wheat or a legume had been the preceding crop. Only four tillage treatments were considered in this study, mainly to reduce sample size but also because for the remaining three tillage treatments a different drilling machine—a conventional seeder with different row spacing rather than a direct drill/ZT seeder—was used, which could have confounded the results. In the chosen Tillage treat- ment No. 2, tillage and seed bed preparation for wheat and the legumes were done by a cultivator and 1 - 2 passes with a sweeper. Crop residues were chopped and spread after harvest, i.e. retained on the plots. Tillage treatment No. 4 was equal to No. 2, except that wheat residues were incorporated into the soil with a moldboard plough and the plots were then rolled and levelled with a sweeper once, before planting of legumes. Tillage treatment No. 6 was ZT plus residue retention (full CA) and No. 7 was ZT without residue retention. Planting and fertilizer application was done as in B4.
This experiment was established in 2006/07 to study the impact of tillage (CT vs ZT) on growth and yield in a flexible rotation of the common cereal and legume crops (wheat, barley, lentil, chickpea) in the region. Tillage main plots were split into time of planting (early vs late) sub plots in 2007/08 and these were further split into residue (R) management (R+ = residue retained; R− = residue partly removed) sub-sub plots in 2008/09. The experiment was thus a split-split plot design with three replications, with a total of 24 plots each 0.5 ha in area. The CT Late R− treatment was similar to what farmers practice in the region, whereas the ZT Early R+ treat- ment represented the land-conserving land management practice recommended by the ICARDA agronomists and soil scientist.
The crop sequence from 2006/07 to 2011/12 was wheat, lentil, barley, wheat, lentil and wheat. Crops were established and managed as described under B4 above. The cereal residue management was as follows: on the R+ plots at harvest, cereal plants were harvested as high as possible and stubble and loose residues were fully retained. On the R− plots cereals were harvested as low as possible and loose residues were baled and removed from the field, with short stubbles and the non-baled fraction of loose straw retained in the field, which made R− a partial residue removal treatment. In lentil years (2010/11) all straw was removed in the hand harvesting oper- ation.
To assess changes in soil quality in response to tillage, residue retention, N-fertilizer application and crop rota- tion, a range of soil quality indicators were measured. A comprehensive description of the applied methods for chemical and physical analyses is given in [
The first soil sampling of treatments was carried out on 28 October 2008, a few days before (early) planting. A total of n = 10, n = 4 and n = 3 sub-samples was taken at 0 - 25, 25 - 60 and 60 - 100 cm depth respectively and mixed to get one composite sample per depth. Samples were taken in the CT and ZT late-planted wheat plots of B4, all considered plots of B7, and all plots with full residue retention (R+) of plot C16. These samples were subject to analysis of soil texture, (1 N ammonium acetate) extractable potassium (K), plant-available min- eral N distinguished into ammonium-N and nitrate-N, 0.5 M sodium bicarbonate extractable (Olsen) P, and SOM.
This soil sampling was repeated on 11 March, 29 April, 13 July 2009 and January 2010 (C16 only) to assess seasonal trends of mineral N. This was repeated in the early planted plots of C16 in March 2010 with composite soil samples (n = 3 sub-samples) taken at depth increments of 0 - 5, 5 - 10 and 10 - 25 cm, to detect possible differences with depth. Similarly, extractable P analysis was repeated in December 2009 in C16 at 0 - 25, 25 - 60 and 60 - 100 cm depths.
SOM sampling was undertaken in February 2012 in the early-planted plots of B4, with soil samples taken at 0 - 10 and 10 - 30 cm depths. In March 2010, SOM at very fine depth increments (0 - 2, 2 - 5, 5 - 10 cm depth), was studied by [
As bacterial biomass is considered a more sensitive indicator of soil change than gross chemical measure- ments, the C and N contents in the soil microbial biomass were measured in April 2009 in soil samples from 0 - 25 cm depth in the late planted plots of B4 and the Late R+ plots of C16.
Water stable soil aggregates (WSAs) were measured in April 2009 at 0 - 25 cm depth from ZT and CT plots of B4 and C16 (early planting only), and in February 2012 in the early-planted plots of B4, with samples taken at 0 - 10 cm and 10 - 25 cm depth. Field fresh soil (50 g) was put on two stacked sieves with mesh-widths of 0.5 mm and 0.2 mm. The sieves were then inundated in water, wet-sieved (motorized equipment with constant number of strokes/min.) for 30 sec., and the remaining soil in each sieve, the WSA, was weighted. Results were expressed in per cent abundance of WSAs in the different size classes.
Soil water infiltration was measured in C16 eight times over a period of 1.5 years, i.e. 28 October and 28 De- cember 2010, 2 February, 15 March, 3 April, 6 October and 28 November 2011 and 18 March 2012. In studying infiltration, we were interested in comparing ZT with CT plots, assuming that planting date would have little or no effect on soil physical properties. At each time, three infiltration measurements were made on each of the late-planted plots following [
In February 2012, undisturbed, 100 cm3 ring cores samples (n = 24) were taken at 0 - 5 cm depth in the late-planted plots of B4 for analysis of soil water retention (SWR) curves. The SWR method involved stepwise water drainage imposed via a (max. 1 m) hanging water column attached to a sandbox on top of which the sam- ples are placed, repeated measurement of the weight of the samples, and subsequent calculation of correspond- ing water content using the weight of the soil after oven-drying and its known volume.
To study the impact of different levels of surface residue retention on soil water evaporation and topsoil water content under wheat after conventional tillage land preparation, a one-factorial field trial (randomized plot de- sign with four reps) was initiated in 2010. Surface residue in the form of loose wheat straw was applied at the beginning of April into the standing wheat crop at six different levels, i.e. 0 (bare soil), 1.5, 3, 4.5, 6 and 7.5 t/ha. On 20 April 2010, the whole experiment was irrigated with a boom sprinkler at a rate of 30 mm; two days later, soil moisture determination started. Measuring intervals were gradually increased, from daily to weekly. Wheat on all plots was harvested at the end of May. On 28 June, the harvested plots were irrigated once again with 30 mm and soil moisture monitored under fallow conditions, i.e., eliminating transpiration and focusing on evapo- ration only. Soil moisture was determined either by TDR probe (TRIME; 15 cm rod length; factory calibrated) at 0 - 15 cm depth (five replications per plot) or manually at 0 - 15 and 15 - 30 cm.
Statistical analysis of the results was undertaken with GenStat 14. To account for the interdependency of soil properties, the different soil depths were considered as repeat measures in the ANOVA.
The texture of the soils of the B4, B7 and C16 sites was clay throughout. With depths the clay content signifi- cantly increased by approximately 10% units (least significant difference [LSD] = 1.5%). Sites differed signifi- cantly in clay (LSD = 4%), partly in silt (LSD = 3%) and sand content (LSD = 2%).
The extractable K concentration was high in all depths and sites (
. Average texture and extractable potassium at 0 - 25, 25 - 60 and 60 - 100 cm depth in the B4, B7 and C16 trials (soil sampling Oct. 2008).
Site | 0 - 25 cm | 25 - 60 cm | 60 - 100 cm | |
---|---|---|---|---|
Clay (%) | B4 | 65 | 70 | 76 |
B7 | 58 | 62 | 64 | |
C16 | 62 | 66 | 71 | |
Silt (%) | B4 | 31 | 26 | 22 |
B7 | 34 | 29 | 30 | |
C16 | 25 | 23 | 22 | |
Sand (%) | B4 | 4 | 4 | 2 |
B7 | 9 | 9 | 7 | |
C16 | 12 | 11 | 7 | |
Extractable K (mg/kg) | B4 | 272 | 157 | 116 |
B7 | 290 | 180 | 145 | |
C16 | 339 | 195 | 162 |
Soil bulk density varied amongsites, with the top 50 cm significantly lower in C16 than in B4 and B7 (LSD site = 0.1 g/cm3;
Ammonium-N concentrations at 0 - 25, 25 - 60 and 60 - 100 cm depth, with on average 5 - 12 mg/kg in 0 - 25 cm and less than 3 mg/kg in 25 - 100 cm, were highest at the beginning of the 2008/09 cropping season in Oc- tober 2008 in all plots and depths, as expected in response to rapid mineralization of organic matter after some first light rains. This ammonium was then quickly transformed into nitrate, with nitrate concentrations between on average 8 and 20 mg/kg, but highly variable throughout. A translocation of nitrate from the top to the bottom layer towards March was visible in all plots. Subsequently, nitrate-N was depleted in response to plant uptake, most pronounced in C16 (nitrate-N < 4 mg/kg in all depth and treatments). CT treatments could not be signifi- cantly distinguished from ZT treatments, as variation of results was very high. Likewise, none of the other treatments had any significant impact on mineral N concentrations.
January 2010 sampling revealed that none of the factors tillage, residue management and planting date had a significant effect on ammonium or nitrate content in C16. Ammonium-N was less than 1.5 mg/kg at all depths, and nitrate concentrations slightly reduced with depth from 6 to 4 mg/kg.
Sampling in March 2010 at depths of 0 - 5, 5 - 10 and 10 - 25 cm also revealed no significant differences among treatments.
Topsoil (0 - 25 cm) extractable P was significantly higher under the corresponding ZT treatment in B4, B7 and C16 compared to treatments 2 and 4 of B7, or the CT treatments of the B4 and C16, repeatedly during the mea- suring period (
The increases in extractable P in response to ZT were not large, and sometimes not significant (B4 and C16, Oct. 2008) due to the considerable variability of results, but nevertheless pointed towards a basic improvement of fertility under ZT, reflected in higher extractable P contents. Surprisingly, in B7, straw residue removal (Treatment No. 6) did not negatively affect extractable P in comparison to Treatment No. 7 where straw was re- tained. Likewise, the N-treatment (plus and minus N fertilizer application) in B7 did not affect the results sig- nificantly. The positive impact of ZT in B4 was only measurable in the top 10 cm of the soil.
Differences in SOM in 0 - 25 cm and deeper soil depths were small in October 2008 and could not be attributed to any of the applied treatments in B4 and C16. However, C16 had significantly higher SOM contents in 0 - 25 cm than B4. In B7, N fertilizer application significantly (p = 0.048) increased SOM contents averaged across the
Soil bulk density at various depths in the B4, B7 and C16 trials; soil sampling Sept. 2008
three depths (N+: 0.92 %, N−: 0.88 %, LSD: 0.04 %;
The assessment of topsoil SOM in C16 in very fine depth-increments in March 2010 revealed that ZT signifi- cantly increased SOM contents in 0 - 2 cm depth as compared to CT (LSD tillage = 0.3%;
The intensive soil sampling in September 2011 in C16 comparing CT Late R− and ZT Early R+ was supposed to take care of the variability of SOM in the field. The results, surprisingly, revealed that the SOM in C16 had a pronounced west-east gradient (
The microbial C and N contents in soil samples from 0-25 cm depth in the Late R+ plots of C16 were signifi- cantly higher than in the soil samples from ZT and CT late-planted plots of B4 (LSD site, C-content = 19 mg/kg, LSD site, N-content = 4 mg/kg). The microbial C content varied greatly, and no differences among tillage treat- ments could be observed. However, tillage had a significant impact on microbial N in B4, with ZT increasing microbial N content (
In April 2009, ZT significantly increased the percent WSA > 0.5 mm in B4 (
Extractable (Olsen) phosphate measured in the B4, B7 and C16 trials, in Oct. 2008, Dec. 2009 and Feb. 2012; LSDs Tillage × Depth: B4, Feb. 2012 = 2.7 mg/kg, B7, Oct. 2008 = 1.3 mg/kg, C16, Dec. 2009 = 2.7 mg/kg)
C16, had significantly less WSA in the 0.2 - 0.5 mm class under ZT. C16 had significantly more large (>0.5 mm) WSAs than B4, irrespectively of the method of soil tillage.
In February 2012, WSA determination in B4 was repeated, at depths of 0 - 10 cm and 10 - 25 cm. There was a tendency towards a higher percentage of large WSA under ZT at 0 - 10 cm, but the impact was not significant (
C16 ZT plots consistently had a higher hydraulic conductivity than CT plots, with the exception of December 2010 (
Soil organic matter (%) measured in October 2008 in the B7 (top left, LSD Nitrogen = 0.04 %), in Mar. 2010 (top right, LSD Tillage = 0.3%) and in September 2011 in the C16 trial distinguished by trial blocks (bottom)
Microbial C and N content measured in April 2009 under CT and ZT plots of ICARDA’s long-term trials B4 (left) and C16 (right); LSD, B4, Microbial N = 3 mg/kg
Water-stable aggregates by aggregate size class of soil taken April 2009 at 0 - 25 cm depth, and February 2012 at 0 - 10 cm and 10 - 25 cm depth in ZT and CT plots of the B4 and C16 trial
At all applied soil water pressure heads, soil samples from the ZT plots of B4 at 0 - 5 cm depth had higher water contents than samples from the CT plots. The overall tillage effect was significant because of the significant differences between ZT and CT water contents at −1 cm (pF 0 = saturation) and −2.5 cm (pF 0.4) pressure head (
Soil moisture determination during the late part of the wheat cropping season and following the application of 30 mm of irrigation water did not reveal a significant influence of the varying residue levels on moisture deple- tion by transpiration and evaporation. Different levels of residue application also did not have an effect on grain yields—as expected, because the residue layers were only applied in April. However, comparing yields averaged across all residue treatments with yields of a neighboring non-irrigated field with the same wheat crop, managed in the same way until the experiment started, the impact of 30 mm of irrigation on wheat grain yield was signifi- cant (LSD = 0.371 t/ha), with 1.642 t/ha under non-irrigated conditions and 2.015 t/ha under irrigated conditions average across the all residue levels.
Topsoil hydraulic conductivity (cm/h) of CT and ZT plots of the C16 trial over the period Oct. 2010 to March 2012
Soil water retention curves of undisturbed soil sam- ples from ZT and CT plots of the B4 trial
As was the case for the in-season observations, soil moisture dynamics measured after harvest under fallow conditions at 0 - 15 cm depth were not affected by the residue treatments (
On 5 August, when soil moisture was measured by gravimetric sampling at 0 - 15 cm and 15 - 30 cm depth, surface residue retention amounts had a significant impact on total water amounts (mm/30 cm depth) when combining the two sampling depths (
High extractable K concentrations are common for the calcaric soils of the Mediterranean region and K is gen- erally not yield limiting and also not applied as mineral fertilizer. The “baseline” assessment in October 2008 confirmed K was not limiting in the trial sites; the observed significant differences among the three experiments were considered to be most likely due to edaphic reasons. Higher topsoil bulk densities in the B4 and B7 trials
Soil moisture dynamics at 0 - 15 cm soil depth under fallow plots with vary- ing levels of surface residue retention in 2010 and application of 30 mm of irrigation 28 June; all observations by TDR except gravimetric sampling 5 Augus
located in the south of the station as compared to C16 in the north may in part be due to the slightly higher sand content in C16.
Soil organic carbon contents could be increased by reducing tillage and by retention of crop residues and or- ganic matter recycling. In a previous study carried out at Tel Hadya, it was shown that SOC was significantly increased by application of large amounts (10 Mg/ha/2 yr) of compost and less intensive tillage [
In a similar study from Spain [
Higher extractable P-contents in the topsoil under NT compared to CT were found by [
Soil mineral N contents varied greatly over time and depth and no significant impact of any of the applied treatments on mineral N could be shown. It is generally assumed that SOM mineralization and nitrate release are stimulated by tillage [
Determination of the impact of tillage on WSA did not result in a clear-cut trend suggesting that reduced til- lage and residue retention would always increase aggregate stability, as has been proven elsewhere [
It was somewhat surprising that soil water infiltration capacity varied significantly over time, which basically meant, at least for the soils at ICARDA headquarters, that the applied method [
We tried to answer the question whether in addition to higher water infiltration capacity by reduced tillage and (surface) residue management a surface residue layer would conserve soil moisture. During the late crop- ping season, high surface residue levels did not contribute to moisture conservation in the top 15 cm of soil, contradicting what has been reported elsewhere [
Corresponding yield data [
In this study of rainfed cropping systems typical of Mediterranean northern Syria, reducing tillage and retaining crop residues promoted significant improvements in soil quality, as reflected by an increase in SOC, higher ex- tractable P contents and improved soil water infiltration capacity, all of which are suitable indicators of soil quality changes under CA. However the positive changes in soil quality were not pronounced, similar to the ac- companying increases in crop yields. This is not surprising, as productivity in the semiarid agro-ecosystems in northern Syria is constrained primarily by the limited and erratic rainfall. However, in combination with the economic savings—by reducing tillage, farmers save labor and money that would otherwise be invested in im- plements and tractor power—this type of conservation agriculture provides an attractive option for farmers in the West Asia and North Africa region, from the standpoints of both economy and ecological efficiency.
The authors wish to thank all ICARDA staff who contributed to this research, especially Jürgen Diekmann, Shukri Ismail, Shereen Baddour, Zainab Hamou, Fatima Hussein and Nezha Merjaneh. The mentioned studies in part were funded by ACIAR and AUSAID (Project no. CIM/2008/027), an internship of Lona van Delden with ICARDA was sponsored by The Agriculture and Environment Foundation (c/o Marlene and Jürgen Diek- mann), Muhammad Ansar received funds from Rawalpindi Arid Agriculture University, Pakistan, and Kayoko Shimonaka was funded by the Tottori University and UNU, all of which is greatly appreciated.