Soil & Tillage Research

Transcription

1 Soil & Tillage Research 109 (2010) 1 8 Contents lists available at ScienceDirect Soil & Tillage Research journal homepage: Can non-inversion tillage and straw retainment reduce N leaching in cereal-based crop rotations? E.M. Hansen a, *, L.J. Munkholm a, B. Melander b, J.E. Olesen a a Aarhus University, Faculty of Agricultural Sciences, Department of Agroecology and Environment, P.O. Box 50, DK-8830 Tjele, Denmark b Aarhus University, Faculty of Agricultural Sciences, Department of Integrated Pest Management, Forsøgsvej 1, DK-4200 Slagelse, Denmark ARTICLE INFO ABSTRACT Article history: Received 2 June 2009 Received in revised form 9 October 2009 Accepted 2 April 2010 Keywords: Reduced tillage Non-inversion tillage Direct drilling Ploughing Straw Nitrate leaching Finding ways of reducing nitrate leaching in Northern Europe has become an extremely important task, especially under the projected climate changes that are expected to exacerbate the problem. To this end, twofield experimentswere established under temperate coastal climate conditions to evaluate the effect of tillage, straw retainment and cropping sequences, including cover crops, on nitrate leaching. The experiments were established in autumn 2002 on a loamy sand with 92 g clay kg 1 and a sandy loam with 147 g clay kg 1. The tillage treatments were stubble cultivation to 8 10 cm or 3 4 cm, direct drilling, or ploughing to 20 cm. The hypothesis was that (i) decreasing soil tillage intensity would decrease leaching compared to ploughing, (ii) leaving straw in the field would decrease leaching compared to removing straw, and (iii) a spring/winter crop rotation with catch crops would be more efficient in reducing nitrate leaching than a winter crop rotation. Overall, we were not able to confirm the three hypotheses. The effect of soil tillage on leaching might be blurred because the studied crop rotations had a high proportion of winter crops and because catch crops were grown whenever the alternative would have been bare soil in autumn and winter. The spring/winter crop rotation with catch crops was not found to be more efficient in reducing nitrate leaching than the winter crop rotation. In contrast, in a single year the winter crop rotation showed significantly lower leaching than the spring/winter crop rotations, probably due to the spring/winter crop rotation including peas, which may be considered a high-risk crop. Our study highlights that management practices that improve biomass production throughout the year are crucial in order to tighten the nitrogen cycle and thereby reduce nitrate leaching. ß 2010 Elsevier B.V. All rights reserved. 1. Introduction Nitrate leaching has been of major concern in Northern Europe during recent decades due to its effect on human health and the environment. This has led to the introduction of EU (e.g. European Community, 1998, 2000, 2006) and national legislation, e.g. Danish Action Plans I, II and III for the Aquatic Environment (Dalgaard et al., 2005). In Northern Europe climate changes can increase the risk of nitrate leaching unless preventative measures are taken. Warmer temperatures and higher CO 2 concentrations may lead to higher demands for nitrogen (N) fertiliser (Olesen and Bindi, 2002; Olesen et al., 2007), but extreme weather events, such as heavy storms or droughts, will make fertiliser recommendations less reliable than in a stable climate. Besides, a warmer climate may result in increased turnover of soil organic matter especially during winter (Olesen et al., 2004b), which may further increase the risk of nitrate * Corresponding author. Tel.: /1760; fax: address: Elly.M.Hansen@agrsci.dk (E.M. Hansen). leaching. Cost-efficient strategies to tighten the N cycle are therefore urgently needed. Such strategies should aim to minimise N availability when there is minimal root uptake and risk of percolation. Viable strategies may include non-inversion tillage, straw retainment and cropping sequences that include cover crops (Christensen, 2004). Surprisingly few studies have been published on the effect of tillage on N mineralisation and nitrate leaching in temperate coastal climates. A number of studies have shown that over-winter losses of nitrate can be reduced by up to 25% when intensive autumn tillage is omitted before sowing a spring crop (Hansen and Djurhuus, 1997; Stenberg et al., 1999; Thomsen, 2005). Mixed results have been found when comparing different intensities of tillage before sowing a spring crop (literature quoted by Hansen and Djurhuus, 1997; Hooker et al., 2008). It is remarkable that the quoted studies were carried out in cropping systems based on spring-sown crops when currently autumnsown crops dominate in many temperate coastal climate areas such as Northwestern Europe. Postponement of intensive tillage to the spring is not an option when growing autumn-sown crops. Thus, it is highly relevant to investigate the effect of tillage on /$ see front matter ß 2010 Elsevier B.V. All rights reserved. doi: /j.still

2 2 E.M. Hansen et al. / Soil & Tillage Research 109 (2010) 1 8 nitrate leaching in cropping systems with a high proportion of winter crops. Retainment of straw, i.e. plant material with a high C:N ratio, is expected to immobilise a significant amount of N in the autumn. Experiments under controlled conditions have shown that 1 3 kg N per Mg straw can be immobilised (Christensen, 1985). The immobilisation of N was also found to depend on the placement of the straw. Straw left on the soil surface decomposes more slowly than when incorporated (Christensen, 1986). This means that there is a slower decomposition of straw in untilled than in tilled soil. In practice, immobilisation of N is only one aspect of the complex effect of straw retainment on crop growth and N utilisation in different tillage systems. There is a lack of studies investigating the effects of straw retainment at cropping system level. In order to examine N leaching in ploughed and non-inversion tillage systems in Denmark an experiment was started in autumn 2002 on a sandy loam and a loamy sand soil. The experiment was placed on a previously ploughed soil in order to compare the transition phase from ploughing to non-ploughing with continuous ploughing. The overall aim was to study the effect of soil tillage intensity and crop rotation on several aspects of crop development, environment and economics. This paper deals with the aspect of N leaching as affected by tillage, straw management and crop rotation in the years We hypothesized that decreasing soil tillage intensity would decrease mineralisation and thus decrease leaching compared to ploughing. We also expected leaching to be lower where straw was retained in the field due to immobilisation of N. In addition, we wanted to evaluate differences in leaching between a winter crop-based rotation and a spring/ winter crop rotation. 2. Materials and methods 2.1. Experimental sites The experiment was established on sandy loams at Foulum ( N, E) and at Flakkebjerg ( N, E), Denmark. Both soils are based on ground morainic deposits from the last glaciation. The soil at Foulum is classified as a Mollic Luvisol and the soil at Flakkebjerg as a Glossic Phaeozem according to the WRB (FAO) system (Krogh and Greve, 1999). The clay (<2 mm), silt (2 20 mm), fine sand ( mm) and coarse sand ( mm) contents of the soil (0 25 cm) at Foulum were 92, 126, 444, 307 g kg 1 and at Flakkebjerg 147, 137, 426 and 270 g kg 1, respectively. The organic carbon content at Foulum was 18 g kg 1 and at Flakkebjerg 12 g kg 1. The mean annual temperature ( ) at Foulum and Flakkebjerg is 7.3 8C and 7.7 8C, respectively, and annual precipitation is 626 and 558 mm, respectively (Olesen, 1993). Precipitation and average temperatures for 3-month periods are shown in Table 1. Soil physical properties (e.g. bulk density and penetration resistance) in the different tillage systems were reported by Munkholm et al. (2008). In autumn 2005 the bulk density at Table 2 Crop rotations (R2 R4) and straw management. R2 R3 R W. barley W. wheat/cc W. wheat/cc 2004 W. rape S. barley/cc S. barley/cc 2005 W. wheat Pea Pea 2006 W. wheat W. wheat W. wheat Straw Left Removed Left CC, catch crop of undersown perennial ryegrass. Foulum and Flakkebjerg was 1.25 and 1.31 g cm 3, respectively in ploughed plots. In short, it was found that the former ploughed layer in plots with non-inversion tillage was noticeably compacted as indicated by increasing bulk density and penetration resistance s The experiment was established in autumn 2002 as part of a larger experiment. The actual design was a randomized complete split-plot design in four replications with two factors: crop rotation as main plot and soil tillage as sub-plots. In this study we used rotations R2, R3 and R4 (Table 2). In crop rotation R2 and R4, straw was cut and retained after harvest, in R3 straw was removed. The catch crop in winter wheat and spring barley (Table 2) was perennial ryegrass (Lolium perenne L.), which was undersown in spring. Before the experiment was established in autumn 2002 the field had been cropped and cultivated according to normal agricultural practices. In 2002 the crop was oat (Avena sativa L.). The tillage systems were direct drilling (D), harrowing to 3 4cm(H 3 4 ) or 8 10 cm (H 8 10 ) and ploughing (P). The H 8 10 and H 3 4 treatments were only stubble cultivated. A rotary harrow (Bomford Dyna Drive) was used in H 8 10 and H 3 4, but in Foulum a spring-tine harrow (CMN maskintec A/S) was used in H 3 4 when the soil was not too dry. The crops were sown with a single-disc drill (Gaspardo Scan-Seeder DP300) in D, H 3 4 and H 8 10 and with a traditional seed drill (Nordsten Lift-o-matic CLH300) in P. In all treatments crops were sown at the same row distance of 17.5 cm. At Flakkebjerg the P treatment was ploughed in late autumn prior to the sowing of spring crops. Catch crops in the nonploughed treatments were killed by herbicide around 1 November and 30 kg N ha 1 of the fertiliser recommendation for winter rape was applied in autumn. All other crops received the total amount of manure/fertiliser in spring. At Foulum the catch crops in all treatments were killed by herbicides around 1 November, and the soil was ploughed in spring. All crops except peas were fertilised with 100 kg NH 4 N ha 1 in pig slurry and the rest of the fertiliser recommendation was supplied as mineral fertiliser (Table 3). Analyses of slurry were carried out prior to application and the N content was used to calculate an application rate to meet a target of 100 kg NH 4 N ha 1. At spreading, slurry samples were collected to determine the actual N content. The slurry was applied with Table 1 Precipitation (P) and average temperature (T) for 3-month periods at Foulum and Flakkebjerg meteorological stations. Long-term mean Foulum Flakkebjerg Foulum Flakkebjerg Foulum Flakkebjerg Foulum Flakkebjerg P T P T P T P T P T P T P T P T December February March May June August September November All years, December November Precipitation measured at 1.5-m height.

3 E.M. Hansen et al. / Soil & Tillage Research 109 (2010) Table 3 Fertiliser and manure application to crops in crop rotations R2 R4, for winter barley (WB), winter wheat (WW), winter oilseed rape (WR) and spring barley (SB). Slurry Mineral fertiliser N Total inorganic N Mg ha 1 Ammonium N kg ha 1 Org. N Foulum 2003 R2, WB R3, WW/CC R4, WW/CC R2, WR R3, SB/CC R4, SB/CC R2, WW a R2, WW R3, WW b R4, WW b Flakkebjerg 2003 R2, WB R3, WW/CC nd R4, WW/CC nd R2, WR c 163 R3, SB/CC R4, SB/CC R2, WW a R2, WW R3, WW b R4, WW b CC, catch crop of undersown perennial ryegrass; nd, not determined. a Winter wheat after winter oilseed rape. b Winter wheat after pea. c Including 30 kg N ha 1 applied 21 August trailing hoses to winter crops and injected after ploughing but before sowing to spring barley. In plots with spring barley and noninversion tillage, the slurry was injected before sowing. Numbers of germinating seeds per square meter were calculated from counting four randomly chosen rows of 1 m. Some crops were in some years characterised by poor growth in certain patches, which were assessed visually as the percentage of the plot covered by patches ( percent patches ). Generally, the plants in areas with poor growth had no recognisable attack of diseases, pests or nutrients deficiencies, but the plants were much smaller than normal and resembled bonsai plants (Passioura, 2002), which are plants growing in small containers. Passioura (2002) drew attention to the fact that roots emerging from germinating seeds in non-inverted soil may have access to a very small volume of disturbed soil. The phenomenon is described in more detail by Hansen et al. (submitted for publication). Each plot was harvested for grain yield with a plot combine. For winter wheat, barley and winter rape the dry matter content was determined by a near-infrared spectroscopy analyzer (Infratec TM 1241 Grain Analyzer, Foss A/S; Buchmann et al., 2001) on which also protein content in the cereals was determined. The N content in grains was calculated using a protein factor of 5.70 for wheat and 6.25 for barley. The N content in winter rape from Foulum was determined by the Dumas method (Hansen, 1989), while the winter rape from Flakkebjerg was determined on the Infratec TM 1241 Grain Analyzer, and a protein factor of 6.25 was used (personal communication, Johannes Ravn Jørgensen). Samples of total above-ground biomass were taken in 1-m 2 in each plot approximately 1 week after BBCH 85 (Lancashire et al., 1991). The amount of straw and other residues left after harvest was estimated from the sample of total above-ground biomass subtracting the grain dry matter yield. The N content in straw was determined by the Dumas method (Hansen, 1989) Soil nitrate and percolation For calculation of nitrate leaching, soil water samples were taken using porous ceramic cups permanently installed in the autumn of 2002 at a depth of 1 m. The sampling system consisted of suction cups (655x01-B1M1, 1 bar, standard, Soilmoisture Equipment Corporation, Goleta, CA) mounted on PVC pipes (Hansen et al., 2000). Suction and collection tubes were protected by strong plastic tubes, which were taken outside the experimental field in furrows made by equipment on a tractor. Two samplers were installed per plot (i.e. eight per treatment giving a total of 96 samplers per location). A suction of approximately kpa was imposed 2 3 days before sampling. During this period the suction decreased as a result of water sampling. The soil water samples from each plot were bulked before analysis, frozen within a few hours and later analyzed for nitrate-n (Best, 1976). Generally, sampling was carried out once every other week, except in periods of drought or frost. Percolation was calculated using the model Evacrop (Olesen and Heidmann, 1990), which is a programme for calculating daily

4 4 E.M. Hansen et al. / Soil & Tillage Research 109 (2010) 1 8 Table 4 Nitrogen leaching (kg N ha 1 ) and percolation (mm) at Foulum during three leaching periods (1 July 30 June). Crop rotation R2 R3 R4 Straw Left Removed Left Crops W. rape W. wheat W. wheat S. barley/cc Peas W. wheat S. barley/cc Peas W. wheat Leaching period Leaching P c H ab H bc D a LSD.95 ns 15.9 ns ns ns ns ns ns ns Percolation All treatments CC, catch crop of undersown perennial ryegrass. Values within a column followed by the same letter are not significantly different. actual evapotranspiration and runoff from the root zone using a simple soil vegetation atmosphere transfer (SVAT) model. Daily values of precipitation, temperature and potential evapotranspiration are required. Evacrop does not include tillage system or straw handling. In two previous soil tillage experiments on coarse sand and sandy loam the soil water content was measured by neutron moisture meters (Hansen and Djurhuus, 1997). Based on this study and a study by Djurhuus (1985), it was assumed that soil tillage and straw handling had only a small and insignificant effect when evaluating nitrate leaching (Hansen and Djurhuus, 1997). This is in agreement with Stenberg et al. (1999) who on the basis of measurements in individually tile-drained plots on a sandy loam assumed that drainage did not differ in treatments with different soil tillage. Under the actual soil and climatic conditions it was therefore decided to use the percolation calculated by Evacrop for all treatments Calculations and statistics Leaching was calculated per day from daily values of percolation and percolation-weighted daily nitrate concentrations. Annual leaching was calculated from 1 July to 30 June. In order to diminish interactions from the installation of the suction cups, the measurements predating 30 June 2003 were rejected. Analyses of variance for the effect of tillage in each crop rotation and leaching period were carried out according to a randomized block design. Analyses of variance for the effect of crop rotation and tillage were carried out according to a split-plot design with rotation as the main plot factor and tillage as the subplot factor. In both cases the SAS general linear models procedure GLM were used (SAS Institute, 1988). 3. Results and discussion At Foulum the effect of soil tillage on nitrate leaching during the three leaching periods , and was not significant in all crop rotations except in R (Table 4). At Flakkebjerg there were significant effects of soil tillage on nitrate leaching in R2 during the period and in R3 during the period (Table 5). However, as it appears from the discussion below it is a key issue to bear in mind that yields of precrops and actual crops are of utmost importance when evaluating crop and soil management effects on nitrate leaching Foulum Leaching from the three rotations at Foulum during the leaching period showed no significant effect of soil tillage. Leaching was low in all cases, varying between 11 and 24 kg N ha 1 1(Table 4). The crop in R2 was winter rape during Although the establishment of winter rape with non-inversion tillage was poorer than with ploughing, the number of plants in each treatment was sufficient (>60 per m 2 ) to take up most of the plant-available N and thereby keep nitrate leaching at a low level (11 24 kg N ha 1 ), which is in agreement with results from Table 5 Nitrogen leaching (kg N ha 1 ) and percolation (mm) at Flakkebjerg during three leaching periods (1 July 30 June). Crop rotation R2 R3 R4 Straw Left Removed Left Crops W. rape W. wheat W. wheat S. barley/cc Peas W. wheat S. barley/cc CC/Peas W. wheat Leaching period Leaching P 12 b a H ab b H a b D 38 a b LSD ns ns ns ns 7.7 ns ns ns * Percolation All treatments CC, catch crop of undersown perennial ryegrass. Values within a column followed by the same letter are not significantly different. * P = 0.08.

5 E.M. Hansen et al. / Soil & Tillage Research 109 (2010) Johnson et al. (2002). There were no significant differences between tillage treatments, although leaching tended to be highest with direct drilling. Generally, the yields of winter rape were high and not significantly different in the four soil tillage treatments (Table 6). In R3 and R4, a ryegrass catch crop was undersown in winter wheat on 8 May 2003 before the leaching period. The uptake in above-ground plant material at the end of October was between 16 and 24 kg N ha 1 with no significant difference between tillage treatments. The average uptake in R3 and R4 was 20 and 22 kg N ha 1, respectively (Table 7), similar to the levels found by Hansen and Djurhuus (1997), where leaching was significantly reduced by the growth of the catch crop. In summary, winter rape and a catch crop of perennial ryegrass were able to keep leaching at a low level during the leaching period Foulum R3 and R4 also had a ryegrass catch crop during this period undersown in the previous spring barley crop. Leaching was generally higher than in the first period, due to the weather conditions and a poorly developed catch crop. The above-ground catch crop N uptake averaged only 11 and 16 kg N ha 1 in R3 and R4, respectively (Table 7). The catch crop was established late (13 May 2004) and the barley main crop harvested late (3 September 2004) due to wet weather conditions. Late sowing means that the catch crop had little time to establish before the barley covered the ground, and late harvest means that the catch crop was exposed to competition from the barley crop for a long time before the barley was harvested. Higher than normal precipitation in August and September may also have leached nitrate below the root zone of the poorly developed catch crop. In R2, where winter wheat succeeded winter rape, there was significantly higher leaching in D than in P and H 3 4. This does not confirm our hypothesis that decreasing soil tillage intensity would decrease leaching compared to ploughing. Winter rape has been Table 6 Dry matter (DM) grain yield (kg ha 1 ) and total-n uptake in grain and straw (kg ha 1 ) at Foulum DM grain Total-N DM grain Total-N DM grain Total-N R2 Crop W. rape W. wheat W. wheat P a a 152 a H b b 135 ab H ab b 105 b D ab b 121 ab LSD.95 ns ns ns 1 ns 1568 ns 2 R3 Crop S. barley Peas W. wheat P 5410 a ab H ab b H ab a D 4590 b ab LSD.95 ns 2 ns ns ns ns ns 3 R4 Crop S. barley Peas W. wheat P 5555 a 190 a a H ab 164 ab ab H b 146 bc ab D 4115 b 132 c b LSD ns 51 ns ns 1 P = P = P = Only three replications. Table 7 Dry matter (DM) (kg ha 1 ) and total-n uptake in catch crop (kg N ha 1 ) at Foulum and Flakkebjerg approximately 1 November DM Total-N DM Total-N R3 Foulum P H H D LSD.95 ns ns * ns ns R4 Foulum P H H D LSD.95 ns ns ns ns R3 Flakkebjerg P 792 a 10 a H b1 3 b H a 9 a D 883 a 9 a LSD ns ns R4 Flakkebjerg P 829 a 9 a H b1 4 b H a 7 ab D 890 a 9 a LSD ns ns * P = The plots were by mistake stubble cultivated with rotary harrowing after harvest of winter wheat. found to leave more soil mineral N residues in the autumn after harvest than cereals (Johnson et al., 2002). Apparently the winter wheat sown on 24 September 2004 was not able to reduce the soil mineral N after winter rape to the same degree in all four tillage treatments. The explanation could be that reduced tillage and especially direct drilling resulted in slower deep rooting during autumn, winter and early spring as measured by Munkholm et al. (2008), who found that the formerly ploughed layer in reduced tillage treatments was noticeably compacted as indicated by increases in both penetration resistance and bulk density Foulum In the leaching period , winter wheat was grown in all three crop rotations. In R2 it followed winter wheat and in R3 and R4 peas. No significant differences were found between tillage treatments (Table 4). Like winter rape, peas have been found to leave more soil mineral N residues in the autumn after harvest than cereals (Johnson et al., 2002). In 2005 the total-n content in pea seeds and straw in the ploughed treatment in R4 was 150 kg N ha 1, which was significantly higher than with direct drilling (Table 6). As the straw was left in R4, this could explain the tendency to higher leaching in P than in D (discussed below for Flakkebjerg ) Flakkebjerg In contrast to Foulum, there was a significant effect on leaching in R2 in at Flakkebjerg with winter rape (Table 5). Leaching in H 3 4 and D was significantly higher than in P. This is another example of results that do not confirm our hypothesis that

6 6 E.M. Hansen et al. / Soil & Tillage Research 109 (2010) 1 8 Table 8 Dry matter (DM) grain yield (kg ha 1 ) and total-n uptake in grain and straw (kg N ha 1 ) at Flakkebjerg DM grain Total-N DM grain Total-N DM grain Total-N R2 Crop W. rape W. wheat W. wheat P 2453 a 144 b H a 173 b H b 144 b D 1197 b 233 b LSD ns ns ns ns R3 Crop S. barley Peas W. wheat P a 145 a H b 89 b H c 60 b D c 62 b LSD.95 ns ns ns ns R4 Crop S. barley Peas W. wheat P 2987 a a 150 a 7074 a 136 a H b a 121 ab 6028 b 107 b H a b 74 b 6951 a 124 ab D 3211 a b 79 b 6621 ab 137 a LSD ns 1061 ns 874 ns * 25 ns ** * P = ** P = decreasing soil tillage intensity would decrease leaching compared to ploughing. As leaching only took place from December 2003 to May 2004, the extra leaching with non-inversion tillage is probably a result of a poor establishment of winter rape in the dry autumn of 2003, especially in H 3 4 and D. From August to September only 48 mm of rain fell compared with the long-term average of 114 mm. In P, H 8 10,H 3 4 and D there were 100, 79, 17 and 5 plants per m 2, respectively. Yields in H 3-4 and D were significantly lower than in P and H 8 10 (Table 8), but generally the low yields in winter rape were also due to bird damage. Fertilisation in the autumn with 30 kg N ha 1 most likely increased leaching in the poorly established winter rape plots. If the winter rape had been well established, there would probably have been negligible leaching, as seen at Foulum in R2 in (Table 4). In R3 and R4 the ryegrass catch crop undersown on 9 May 2003 had an above-ground N uptake ranging between only 3 and 10 kg N ha 1 at the end of October (Table 7). An uptake below 10 kg N ha 1 is less than half of the uptake at Foulum. Since the catch crops were sown at the same time (8 9 May) in both locations and the winter wheat main crop was harvested at the same time (11 12 August), the different development of the catch crop after harvest could be because the winter wheat harvested in 2003 was more competitive at Flakkebjerg than at Foulum. This is implied by the average winter wheat yields at Flakkebjerg being almost 1300 kg DM ha 1 higher than at Foulum (data not shown). Early competition, in particular, greatly affects the growth of undersown grass (Olesen et al., 2004a) Flakkebjerg In R2, where winter wheat succeeded winter rape, there was no significant difference in leaching between tillage treatments. Although yields in winter rape were poor, especially in D, high N levels were found in straw (Table 8) due to a high weed density. In R3 and R4 the ryegrass catch crops were undersown in spring barley on 29 April days after the sowing of spring barley. Leaching varied between 24 and 34 kg N ha 1 in R3 and between 21 and 26 kg N ha 1 in R4 (Table 5). The catch crop N uptake was between 13 and 16 kg N ha 1 in R3 and 15 and 19 kg N ha 1 in R4 (Table 7). Observations in the field showed that the ryegrass germinated satisfactorily and was well established at harvest in R3 and R4. Probably the almost concurrent sowing of spring barley and ryegrass resulted in less competition from spring barley than from winter wheat the year before. Besides, the spring barley yields were generally low (Table 8), which may have further lessened the competition against the catch crop Flakkebjerg In R2, winter wheat succeeded winter wheat and no significant difference in N leaching was found between tillage treatments. In R3 and R4, winter wheat succeeded peas and, in contrast to all other crop rotations, leaching in R3 was significantly higher in P than in non-inverted treatments. The same trend was found in R4. Pre-crop yields (peas) were almost non-existent in H 3 4 and D due to poor germination under the dry soil conditions in April The peas were sown on 4 April 2005 and only 10 mm rain during April compared with the long-term average of 34 mm. But also damage caused by hares and deer in the early stage of development contributed to the poor yields. At Flakkebjerg, leaching after other crops than peas was generally either higher or not significantly different when comparing reduced tillage with ploughing. It is, therefore, a reasonable assumption that the superior growth of peas and thus the higher N fixation in ploughed treatments explains the higher leaching with these treatments than with noninverting treatments. Although the results confirm our hypothesis that decreasing soil tillage intensity would decrease leaching compared to ploughing, it was an implicit prerequisite that crops in non-inverted treatments were satisfactorily established. The result emphasizes the importance of including the type, establishment and development of the pre-crop in leaching studies with different tillage treatments Tillage Overall we were not able to confirm our hypothesis of reduced nitrate leaching with decreasing tillage intensity. A significant effect of tillage was found in only three out of 18 cases (crop rotations locations years). Furthermore, leaching was higher for reduced tillage than for ploughing in two of the cases. Our

7 E.M. Hansen et al. / Soil & Tillage Research 109 (2010) inconsistent results illustrate that tillage affects N mineralisation, uptake and loss in a complex manner. The summarised leaching for the whole period showed no significant effect of tillage at Flakkebjerg and Foulum (data not shown). At Flakkebjerg, there was a significant interaction between soil tillage and crop rotation with increased leaching following the unsuccessful attempt to establish winter rape in R2 and peas in R3 and R4. The literature indicates a high risk of nitrate leaching when practising intensive autumn tillage before a spring crop. Hansen and Djurhuus (1997) found that leaching significantly increased with tillage intensity in the autumn on sandy loam. Annual average nitrate leaching levels across 5 years ranged from 76 to 35 kg N ha 1 with the highest level reported for autumn ploughing with autumn stubble cultivation and lowest for reduced tillage without autumn stubble cultivation. Compared with autumn ploughing, reduced tillage also tended to result in lower nitrate leaching on a coarse sandy soil, although differences were not significant. A study on a sandy loam in Sweden (Stenberg et al., 1999) showed a less pronounced effect than the Hansen and Djurhuus s (1997) study for the sandy loam of the omission of autumn tillage before a springsown crop. In the Swedish study there was a substantial growth of weeds and volunteers on the unploughed plots, while in the Danish study the unploughed plots were kept bare by herbicide spraying. The Swedish study therefore actually implies a somewhat smaller effect of tillage than the Danish study. In a lysimeter experiment, Thomsen (2005) found a significant effect of carrying out tillage in the autumn rather than the spring, but the effect was likewise smaller than that showed by Hansen and Djurhuus (1997). The lack of significant effects of tillage in our study may be because we did not include high-risk intensive autumn tillage before spring crops. Catch crops were always grown before spring crops and tillage before spring crops was postponed until late autumn (Flakkebjerg) or early spring (Foulum). But why did we not see a difference for the autumn-sown crops? This may be related to the effect of tillage on early season root and shoot growth. Munkholm et al. (2008) showed that intensive tillage before winter wheat resulted in faster early season growth. These studies were carried out in the same tillage experiment as described in this study. Consequently, we expect an increased uptake of N during the autumn and winter for intensive tillage, especially compared to direct drilling. In addition, N uptake in grain and straw was in some cases significantly higher in P than in D. Besides, in some crops at Foulum there was a larger number of patches with poor growth in plots with non-inversion tillage than in ploughed plots (Hansen et al., submitted for publication). These observations imply that more mineral N could have been present in D than in P after harvest. This will counterbalance the expected, higher N mineralisation for intensive tillage. For winter wheat following winter rape (high amount of available N) at Foulum our results indicate that the effect of intensive tillage on early season development was more important than the effect on N mineralisation as nitrate leaching decreased for this crop with tillage intensity (Table 4). This study highlights the need to take the risk of poor crop establishment into consideration when environmental effects of non-inversion tillage are evaluated. At Flakkebjerg, the occasional very dry spell in spring and autumn led to poor establishment for direct drilling and shallow tillage. Our results show that poor establishment of autumn-sown crops with reduced tillage can result in higher leaching than in ploughed treatments. In contrast, poor establishment of spring-sown, N-fixing crops with reduced tillage can result in lower leaching than in ploughed treatments, but this is indeed an unsound way to reduce leaching. If reduced tillage causes poor establishment of a fertilised spring-sown crop, leaching may be higher with reduced tillage than with ploughing due to the poorer N utilisation with reduced tillage. Nitrate leaching in all treatments were estimated by combining calculated percolation for ploughed treatments with values of measured nitrate-n concentration in actual treatments. If percolation actually were higher in plots with non-inversion tillage, due to reduced evapotranspiration in plots with reduced yields, then leaching in non-inverted plots would be underestimated. Hence, if percolation estimates were biased the real difference between non-inversion tillage and ploughing would be more pronounced and thereby strengthen the conclusions. Overall, these results demonstrate that (i) management practices that improve crop growth and yields are crucial to minimise nitrate leaching and (ii) leaching after high-risk crops like peas and rape (Johnson et al., 2002) may need measures other than growing winter wheat to reduce nitrate leaching Straw The general similarity in leaching between R3 and R4 (where the only difference was that straw was removed in R3 and left in R4) suggests that leaving straw in the field did not result in significant net N immobilisation in this short-term experiment. This finding is in agreement with Stenberg et al. (1999) and Àlvarez et al. (2008). Results from continuous spring barley cropping summarised by Schjønning (1983) showed slightly lower concentrations of nitrate in plots with straw incorporation than in plots without, but the difference was not significant. Nyborg et al. (1995) found that immobilisation of N by straw occurred during the first few years of their experiment, but disappeared thereafter. This is in agreement with Catt et al. (1998) who suggested that the effect of straw incorporation on nitrate leaching is small and shortlived. It was expected that incorporation of straw by ploughing would immobilise more N than straw left on the soil surface in the directly drilled plots. Ambus and Jensen (2001) found that deep incorporation of ground residues in the autumn immobilised about twice as much N as shallow incorporation of the same amounts of residues. But the crop residue particle size is important. Ambus and Jensen (2001) found that with incorporation of ground residues, the immobilisation was 30% higher than with residues cut in 25- mm pieces. In field experiments residues are typically larger than in laboratory experiments, which may explain some inconsistency between results from field experiments and laboratory experiments. Only very sparse information exists on the importance of crop residue particle size for decomposition and soil N dynamics in the field (Ambus and Jensen, 2001). As summarised by Jin et al. (2008), also higher heterogeneity in residue distribution and less soil-residue contact may influence the result under field conditions. If significant immobilisation took place in our experiment or if immobilisation was more pronounced in one treatment than the other, the net effect was not large enough to be detected. However, in an experiment with 15 N-labelled fertiliser, Àlvarez et al. (2008) did not find net immobilisation in either ploughed or no-tilled plots Crop rotation At Foulum the overall effect of crop rotation and the interaction between soil tillage and crop rotation on nitrate leaching during the leaching periods , and was not significant. At Flakkebjerg there was a significant interaction between soil tillage and crop rotation in the leaching period This was probably due to the tendency for leaching in R2 to decrease with increasing soil tillage, while this was less pronounced in R3 and R4 (Table 7). In there was an almost significant effect of crop rotation (P = 0.06) and a significant

8 8 E.M. Hansen et al. / Soil & Tillage Research 109 (2010) 1 8 interaction between crop rotation and soil tillage. The interaction was probably due to the different effect of soil tillage in R3 and R4 (after peas) compared with R2 (after winter wheat). Concerning differences between crop rotations, leaching in R2 (29 kg N ha 1 ) tended to be lower (P = 0.06) than leaching in R4 (35 kg N ha 1 ), but was not significantly different from R3 (34 kg N ha 1 ). Thus leaching in plots after a high-risk crop (Johnson et al., 2002) such as peas with straw left on the field as in R4 ( at Foulum) tended to be higher than after winter wheat. Apparently the winter wheat sown on 19 September 2005 in R4 was not able to reduce soil mineral N after peas to the same degree as winter wheat after winter wheat in R2. However, the difference was only 6 kg N ha 1. The summarised leaching for the whole period showed no significant effect of crop rotation at Flakkebjerg and Foulum (data not shown). 4. Conclusions It was hypothesised that decreasing soil tillage intensity would decrease mineralisation and thus leaching compared to ploughing. Overall, we have not been able to confirm this in the studied crop rotations that have a high proportion of winter crops, and where catch crops are grown whenever the alternative would have been bare soil in autumn and winter. It was hypothesised that leaving straw in the field would increase immobilisation of N and thus decrease leaching compared to the removal of straw. Overall, we were not able to demonstrate this difference between two crop rotations where straw was either left or removed from the soil. It was hypothesised that a spring/winter crop rotation with catch crops (R3 and R4) would be more efficient in reducing nitrate leaching than a winter crop rotation (R2). We have not been able to demonstrate this. On the contrary, in a single year ( ) the winter crop rotation (R2) showed a significantly lower leaching level than the spring/winter crop rotations (R4), probably because the spring/winter crop rotation included peas, which similar to oilseed rape may be considered a high-risk crop. Our study highlights that management practices that improve active plant growth and biomass production throughout the year (i.e. optimize N uptake) are crucial to tightening the N circle and thereby reducing nitrate leaching. Acknowledgements The staffs at the Research Centre Flakkebjerg and Research Station Foulumgård are gratefully acknowledged for their technical assistance in the field. The work was financed by the Danish Ministry of Food, Agriculture and Fisheries under the research programme Agriculture from a holistic resource perspective. References Ambus, P., Jensen, E.S., Crop residue management strategies to reduce N- losses interaction with crop N supply. Commun. Soil Sci. Plant Anal. 32, Àlvarez, C.R., Àlvarez, R., Sarquis, A., Residue decomposition and fate of nitrogen-15 in a wheat crop under different previous crops and tillage systems. Commun. Soil Sci. Plant Anal. 39, Best, E.K., An automated method for determining nitrate-nitrogen in soil extracts. Queensland J. Agric. Anim. Sci. 33, Buchmann, N.B., Josefsson, H., Cowe, I.A., Performance of European artificial neural network (ANN) calibration for moisture and protein in cereals using the Danish near-infrared transmission (NIT) network. Cereal Chem. 78, Catt, J.A., Howse, K.R., Christian, D.G., Lane, P.W., Harris, G.L., Goss, M.J., Strategies to decrease nitrate leaching in the Brimstone Farm Experiment, Oxfordshire, UK, : the effect of straw incorporation. J. Agric. Sci., Camb. 131, Christensen, B.T., Tightening the nitrogen cycle. In: Schjønning, P., Elmholt, S., Christensen, B.T. (Eds.), Managing Soil Quality: Challenges in Modern Agriculture. CAB International, Wallingford UK, pp Christensen, B.T., Wheat and barley straw decomposition under field conditions: effect of soil type and plant cover on weight loss, nitrogen and potassium content. Soil Biol. Biochem. 17, Christensen, B.T., Barley straw decomposition under field conditions: effect of placement and initial nitrogen content on weight loss and nitrogen dynamics. Soil Biol. Biochem. 18, Dalgaard, T., Børgesen, C.D., Hansen, J.F., Hutchings, N.J., Jørgensen, U., Kyllingsbæk, A., How to half N-losses, improve N-efficiencies and maintain yields? The Danish Case. In: Zhu, Z., Minami, K., Xing, G. (Eds.), 3rd International Nitrogen Conference, Contributed Papers, Science Press, NJ, USA, pp Djurhuus, J., Vandforbrug i vårbyg i relation til jordbearbejdning og halmnedmuldning (Actual evapotranspiration in spring barley in relation to soil tillage and straw incorporation). Tidskr. Planteavl 89, (in Danish with English summary). European Community, Council Directive 98/83/EC of 3 November 1998 on the quality of water intended for human consumption. European Community, Directive 2000/60/EC of the European Parliament and the Council of 23 October 2000 establishing a framework for Community action in the field of water policy. European Community, Directive 2006/118/EC of the European Parliament and the Council of 12 December 2006 on the protection of groundwater against pollution and deterioration. Hansen, B., Determination of nitrogen as elementary N, as alternative to Kjeldahl. Acta Agric. Scand. 39, Hansen, E.M., Djurhuus, J., Nitrate leaching as influenced by soil tillage and catch crop. Soil Tillage Res. 41, Hansen, E.M., Djurhuus, J., Kristensen, K., Nitrate leaching as affected by introduction or discontinuation of cover crop use. J. Environ. Qual. 29, Hansen, E.M., Munkholm, L.J., Olesen, J.E., submitted for publication. N-utilization in non-inversion tillage systems. Hooker, K.V., Coxon, C.E., Hackett, R., Kirwan, L.E., O Keeffe, E., Richards, K.G., Evaluation of cover crop and reduced cultivation for reducing nitrate leaching in Ireland. Technical reports: Vadose zone processes and chemical transport. J. Environ. Qual. 37, Jin, K., Sleutel, S., De Neve, S., Gabriels, D., Cai, D., Jin, J., Hofman, G., Nitrogen and carbon mineralization of surface-applied and incorporated winter wheat and peanut residues. Biol. Fertil. Soils 44, Johnson, P.A., Shepherd, M.A., Hatley, D.J., Smith, P.N., Nitrate leaching from a shallow limestone soil growing a five course combinable crop rotation: the effects of crop husbandry and nitrogen fertilizer rate on losses from the second complete rotation. Soil Use Manage. 18, Krogh, L., Greve, M.H., Evaluation of World Reference Base for Soil Resources and FAO Soil Map of the World using nationwide grid soil data from Denmark. Soil Use Manage. 15, Lancashire, P.D., Bleiholder, H., van den Boom, T., Langelüddeke, P., Strauss, R., Weber, E., Witzenberger, A., A uniform decimal code for the growth stages of crops and weeds. Ann. Appl. Biol. 119, Munkholm, L.J., Hansen, E.M., Olesen, J.E., The effect of tillage intensity on soil structure and winter wheat root/shoot growth. Soil Use Manage. 24, Nyborg, M., Solberg, E.D., Izaurralde, R.C., Malhi, S.S., Molina-Ayala, M., Influence of long-term tillage, straw and N fertilizer on barley yield, plant-n uptake and soil-n balance. Soil Tillage Res. 36, Olesen, J.E., Annual Agrometerological Overview 1992, SP rapport no. 10. Danish Institute of Plant and Soil Science (in Danish with English summary). Olesen, J.E., Bindi, M., Consequences of climate change for European agricultural productivity, land use and policy. Eur. J. Agron. 16, Olesen, J.E., Heidmann, T., EVACROP. A Programme for Calculating Daily Actual Evapotranspiration and Runoff from the Root Zone. Version Research Note. Research Centre Foulum, 65 pp. (in Danish with English summary). Olesen, J.E., Hansen, P.K., Berntsen, J., Christensen, S., 2004a. Simulation of aboveground suppression of competing species and competition tolerance in winter wheat varieties. Field Crops Res. 89, Olesen, J.E., Rubæk, G., Heidmann, T., Hansen, S., Børgesen, C.D., 2004b. Effect of climate change on greenhouse gas emission from arable crop rotations. Nutr. Cycl. Agroecosyst. 70, Olesen, J.E., Carter, T.R., Diaz-Ambrona, C.H., Fronzek, S., Heidmann, T., Hickler, T., Holt, T., Minguez, M.I., Morales, P., Palutikof, J., Quemada, M., Ruiz-Ramos, M., Rubæk, G., Sau, F., Smith, B., Sykes, M., Uncertainties in projected impacts of climate change on European agriculture and ecosystems based on scenarios from regional climate models. Climatic Change 81 (Suppl. 1), Passioura, J.B., Soil conditions and plant growth. Plant Cell Environ. 25, SAS Institute Inc., SAS/STAT. User s Guide, Release Edition. SAS Institute Inc., Cary, NC. Schjønning, P., Halmnedmuldning. In: Hansen, J., Kyllingsbæk, A. (Eds.), Kvælstof og planteproduktion. Tidsskrift for Planteavls Specialserie. Beretning nr. S 1669, pp Stenberg, M., Aronsson, H., Lindén, B., Rydberg, T., Gustafson, A., Soil mineral nitrogen and nitrate leaching losses in soil tillage systems combined with a catch crop. Soil Tillage Res. 50, Thomsen, I.K., Crop N utilization and leaching losses as affected by time and method of application of farmyard manure. Eur. J. Agron. 22, 1 9.