| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Soil Science, Swedish Univ. of Agricultural Sciences, P.O. Box 7014, SE-750 07 Uppsala, Sweden
b Enbärsväg 8, SE-761 63 Norrtälje, Sweden
* Corresponding author (holger.kirchmann{at}mv.slu.se)
Received for publication February 23, 2006.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
60% higher in the conventional system than the organic system. The agronomic efficiency of N was 9 to 10 kg grain yield kg–1 N in the organic system compared with 16–18 kg grain yield kg–1 in the conventional system. The long-term use efficiency of P was lower in the organic system (7%) than in the conventional system (36%). These results show that yield and soil fertility are superior in conventional cropping systems under cold-temperate conditions.
Abbreviations: AL, ammonium acetate-lactate • ICP, inductively coupled plasma spectrophotometer • SOC, soil organic carbon
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Several European long-term field studies have shown that yields of arable crops are typically lower in organic systems than in conventional systems (Ivarson and Gunnarsson, 2001; Eltun et al., 2002; Mäder et al., 2002; Torstensson et al., 2006). In Sweden, Ivarson and Gunnarsson (2001) showed significant yield reductions for all types of crops, averaging 25% in systems that included livestock and 45% in those that did not.
Most of the results available on organic systems were obtained from arable soils with a history of inorganic fertilizer application for one to several decades before organic cropping principles were applied and were, therefore, influenced by soil fertility improvements from previous management. Few, if any, organic farming experiments have been performed on arable soils that had never received any inorganic fertilizer. In our study, no fertilization had occurred for the 40 yr before the initiation of the experiment. These conditions made the site particularly suitable to investigate the impact of different nutrient application strategies used in organic and conventional production systems on soil fertility characteristics.
The objective of the present study was to compare the performance of organic and conventional crop–livestock systems with respect to (i) crop yield and nutrient uptake; (ii) changes in pH, plant-available N, P, and K; (iii) changes in soil C; (iv) crop nutrient use efficiencies; and (v) leaching of N to surface and ground waters.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Experimental Design and Nutrient Inputs
Until 1940, the farm had dairy cows in combination with a cropping system of cereals, legumes, potatoes, and sugarbeet (Beta vulgaris L.), with an occasional input of mineral fertilizers. In 1942, sheep were introduced and more land was converted into perennial grassland such that by 1978 most of the land was perennial grassland. In 1979, a grassland field on the farm was plowed, left fallow, repeatedly chisel plowed to reduce weeds, and drainage was installed. In 1980, the current experiment started. The cropping systems were laid out in plots with a net size of 35 by 90 m each, surrounded by a buffer zone of 5-m width. One organic plot and one conventional plot, both aimed at supporting dairy production but with different crop rotations, were established (Fig. 1
). Two unfertilized and nonlimed plots from a different study with a similar crop rotation as that in the conventional system (forage crop replaced by barley, Hordeum vulgare L.) and a net size of 8 by 40 m were used as the control treatment. The control was used as a comparison to determine the impact of the organic and conventional treatments on soil characteristics and crop production.
|
All plots, except those that had an in-sown forage or cover crop, were moldboard plowed in the autumn each year. Before moldboard plowing, crops were chisel plowed except clover (Trifolium spp.) and grass crops, which were disc harrowed and chisel plowed. Cover crops in the organic system were disc harrowed and chisel plowed before plowing in spring. Weeds in the control and conventional system were treated with herbicides, mainly phenoxy acids. In the organic system, tillage was used for weed control, resulting in four cultivations per crop compared with 3 in the conventional system and control (Table 1). The number of cultivations (Table 1) does not include seeding operations, shallow plowing after manure or slurry application, or cultivation after harvesting.
|
2 x 15 Mg ha–1 across a 6-yr rotation; mean dry matter 20.8%) was equivalent to an animal density of 0.56 dairy cows ha–1, which is 50% greater than the animal density that could be supported from yield levels. In the conventional system, the mean annual input was 147 kg N ha–1, of which 113 kg N was from inorganic fertilizer, 23 kg N from liquid animal manure (slurry), and 11 kg N from N2–fixation. The amount of cattle slurry (2 x 30 Mg ha–1 across a 6-yr rotation; mean dry matter 10%) added to the conventional system was equivalent to only 0.34 dairy cows ha–1, which is less slurry than could be produced from yields of the conventional system supporting an animal density of 0.8 dairy cows ha–1. Nitrogen fixation by cover crops and by residues of the main legume was estimated using aboveground crop data and the STANK in MIND model (Version 4:1, Swedish Board of Agriculture, 2006) including a grassland submodel (Fagerberg et al., 1990). In 1985, 100 kg P ha–1 was applied to the organic system as 2.5 Mg basic slag ha–1 (CaO x P2O5 x SiO2, 4 CaO x P2O3; Thomas phosphate) containing 40 g P kg–1. In 1988, apatite (160 g P kg–1) was applied to the organic system at a rate that applied 646 kg P ha–1. Potassium was applied to the organic system in 1991 as potassium aluminum silicate (adularia) at a rate that applied 66 kg K ha–1, and in 1996 potassium yeast waste (potassium sulfate) was used to apply 100 kg K ha–1. In the conventional system, P and K were supplied as super phosphate [Ca(H2PO4)2], potassium chloride (KCl), and through cattle slurry resulting in mean annual inputs of 29 kg P and 82 kg K ha–1 yr–1 (Table 1). Solid manure was applied in the organic system twice during the rotation cycle in the autumn, before winter wheat (Triticum aestivum L.) and potato production (Fig. 1). The organic system received a one-time animal manure slurry application in 1983 and solid manures were applied thereafter. During the 18 yr of this study, the mean annual input of P was higher in the organic system than in the conventional system, but K applications were higher in the conventional system than in the organic system (Table 1). Before field application, manure and slurry were analyzed for dry matter content, total-N, and NH4–N using nondried material (Kjeldahl method), and total P and K.
At the initiation of the experiment in 1980, the conventional system was limed with 4.5 Mg ha–1 calcium carbonate (CaCO3), equivalent to 2070 kg CaO ha–1, but the pH of the organically managed soil was not adjusted due to the recommendation of the Swedish organic farming organization Naturenlig Odling. However, the basic slag applied to the organic system in 1985 had a liming effect equivalent to 1125 kg CaO ha–1.
Lysimeters, similar to those described by Bergström (1987), were installed in each cropping system at the initiation of the experiment. The lysimeters were placed in pits 9 m long by 3 m wide by 1.2 m deep (covering a surface area of 27 m2) and consisted of rubber sheets with vertical sides extending up to the topsoil (0.4 m below the soil surface). A drainage pipe was placed at the bottom of the rubber sheet and covered with a 50-mm layer of sand. This drainage pipe was connected via a watertight plastic pipe to a measuring station, where water flow rates were recorded and samples were collected. The soil was repacked in the lysimeters layer by layer, in accordance with the original stratification of the soil profile.
Soil Carbon Balances
The carbon balance for the topsoil (0–30 cm depth) of the two systems were calculated using the ICBM model (Andrén and Kätterer, 1997), which was adapted to handle discontinuous annual carbon input (Kätterer et al., 2004). The model, which consists of two C pools, one being recently applied organic material and the other soil organic matter, was initialized by a steady-state assumption, that is, the initial carbon mass was in equilibrium with previous land use. The measured C concentrations were converted into C mass using a pedotransfer function for estimating soil bulk density (Kätterer et al., 2006). The C inputs to the soil were from manure or slurry, crop residues left in the field, and weeds and cover crops in the organic system. The input of crop shoot and root residues, including root turnover and root-derived organic compounds, was estimated from crop yield records using algometric relationships based on literature values for barley (Pettersson, 1989; Johansson, 1992); wheat (Flink et al., 1995); winter rape (Petersen et al., 1995; Gosse et al., 1999); potato (Opena and Porter, 1999; Alva et al., 2002); sugarbeet (Vamerali et al., 2003; Brown and Biscoe, 1985); and ley (annual forage crop) (Kuzyakov and Domanski, 2000). For bean (Vicia faba L.), pea (Pisum sativum L.), and oat, we used the same functions as for barley. The mean C input from cover crops was assumed to be 1.6 Mg C ha–1. Default values for model parameters originally calibrated for another Swedish field experiment were derived from Andrén and Kätterer (1997).
The model was also used to calculate C balance scenarios whereby differences in management practices exclusive to organic or conventional farming systems were eliminated. That is, we assumed the same crop residue management, a C return through animal manure proportional to biomass production, and the use of cover crops in both systems. Moreover, the initial difference in C content between the two systems was also eliminated in these simulations.
Analyses, Calculations, and Statistics
The plow layer and subsoil layers of the soil were sampled before the start of the experiment in 1979 to determine the soil fertility status, and analyzed for exchangeable P, K, Mg, and Ca according to Egnér et al. (1960) using ammonium acetate-lactate (AL-solution), which is positively correlated with plant availability, nonexchangeable P and K using hydrochloric acid (HCl), and pH (H2O). The cation exchange capacity of the soil was determined after saturation with 0.5 M sodium acetate (pH 8.2) and replacement with 0.25 M barium chloride (buffered with sodium acetate to pH 7.0). Exchangeable cations were determined after extraction with 0.25 M barium chloride and measured with an inductively coupled plasma spectrophotometer (ICP) (PerkinElmer, Optima 3000 DV, Germany). The bulk density of the topsoil was measured according to Blake and Hartge (1986). After 1979, soil samples were collected in the autumn of 13 different years during the 18-yr experimental period and analyzed for pH and AL-extractable K and P. Concentrations of NH4–N and NO3–N in soil were measured in spring and autumn each year; triplicate soil samples were taken at 0.3-m intervals down to a 0.9-m depth and extracted with 2 M KCl. Extracts were analyzed colorimetrically for NO3 and NH4 on an autoanalyzer (TrAAcs 800, Germany). Total soil C and N were measured by dry combustion (LECO CNS Analyzer, USA).
Yield determinations were made on three subplots (25 m2) in each cropping system. Nitrogen, P, and K was determined on harvested and removed products using a CNS Analyzer (LECO) and an ICP–AES (PerkinElmer, Optima 3000 DV). The presence of weeds was quantified from three or five sampling areas of 0.25 or 0.5 m2 within all crops, except forage crops, in each system each spring and summer before and after mechanical or chemical weed control operations. Number and fresh weight of the most frequently observed weed species as well as total weed fresh weight were recorded. Total dry mass of weeds after weed control were calculated from fresh weight, assuming water content of 75%.
Drainage water from the lysimeters was collected during drain-flow periods in autumn/winter/spring on a weekly basis and analyzed for NH4–N and NO3–N using the phenate method and the colorimetric Cd-reduction method, respectively (American Public Health Association, 1985). Only results from 3 yr of the experimental period (1990, 1992, and 1993) are reported in this paper, primarily because it takes several years for a repacked soil profile containing >10% clay to settle properly (Bergström, 1990).
The agronomic efficiency of N by grain crops (1981–1998) was calculated as the increase in grain yield of the organic and conventional systems compared with the control divided by the N input (Table 1). The P and K input use efficiency was calculated as (amount of the nutrient in crop yield component) – (nutrient in crop yield components of the control)/(nutrient application rate) (Ladha et al., 2005).
The main management systems were not replicated, so subsamples taken within plots were used as replicates for statistical analysis. Data were analyzed with descriptive statistics and ANOVA (single factor) using Microsoft Excel (Microsoft Corp., Redmond, WA) and Minitab (Minitab Inc., State College, PA).
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
Concentrations of plant-available K were also very low in the topsoil and subsoil, with a mean value of 57 mg K kg–1 dry soil, and were not significantly different (P > 0.05) between layers. Extractable Mg was low in the soils of both systems. Concentrations remained low down to 0.7 m and were at the threshold for Mg deficiency to occur in plants (100 mg kg–1 Mg soil). Below that depth, concentrations increased.
Calcium concentrations increased below 0.7 m depth along with an increase in soil pH, indicating that the soil is calcareous in deeper layers. A pH of 5.6 (H2O) in the plow layer indicated that the soil was acidified in the top layers, which is corroborated by a 1930 report that indicated a pH value of 6.4 in the topsoil of the same experimental area. Determinations of nonexchangeable P and K (HCl-extraction) in the plow and subsoil layers (n = 6) revealed mean values of 220 ± 48 mg P kg–1 soil and 900 ± 300 mg K kg–1 soil, with no significant difference (P > 0.05) between soil layers.
In the topsoil of the whole field, mean soil C concentrations were 25.3 ± 3.5 g kg–1 (range 18–32 g C kg–1 soil; n = 29) and total N was 2.5 ± 0.3 g kg–1 (range 2.0–3.2 g N kg–1 soil; n = 29), resulting in a C/N ratio of 10:1 ± 0.54. The organic matter content was somewhat higher than that of adjacent cultivated soils due to the earlier use of the field for perennial grassland.
Changes in Nutrient Status and Acidity in Soil
The addition CaCO3 in 1980 increased the pH value from 5.7 to 6.1 in the conventionally managed soil. The addition of basic slag to the organically managed soil in 1985 was primarily for the application of P, but it did increase soil pH from 5.6 to 5.8 (Fig. 2A
). Differences in soil pH between the organically and conventionally managed soils were significant (P < 0.0001). Year-to-year variations in soil pH values ranged from 0.3 to 0.5 pH units despite sampling at the same time each year and may be related to the specific crop that was grown in the rotation. The largest variation occurred after sugarbeet was grown. However, a comparison between the organically managed and control soils showed that the control soil, despite no lime addition, had significantly higher (P < 0.001) soil pH values (6.0) than the organically treated soil (5.7) during the 18 yr (Fig. 2A). This indicated that the organically managed soil was more susceptible to acidification.
|
Since the initial P and K fertility status of the soil was very low at this experimental site, P and K application rates were based on twice the normal crop removal rates. As a consequence, concentrations of exchangeable soil P were expected to increase in both the organic and the conventional systems. Figure 2B reveals that exchangeable P in the control remained almost constant across time (10 mg P kg–1 soil), whereas P levels in the conventionally and organically managed soil increased significantly across time (P < 0.05) to 20 and 40 mg P kg–1 soil, respectively.
Data in Fig. 2B indicate a higher P availability in the organically treated soil than in the conventional treatment (P = 0.02). The addition of 903 kg P ha–1 to the organically managed soil from cattle manure, basic slag, and apatite (Table 1) increased the levels of exchangeable soil P more than the addition of 519 kg P ha–1 from cattle slurry and super phosphate in the conventionally managed soil during the same time period. The addition of 100 kg P from basic slag to the organic system increased exchangeable P from 15 to 30 mg P kg–1 soil, which is equivalent to 120 kg P ha–1. The addition of 646 kg P ha–1 as apatite in 1988 further increased exchangeable P from 30 to 40 mg P kg–1 soil, which is equivalent to 80 kg P ha–1. The increase in the soil P after the apatite addition may, however, overestimate plant P availability because ammonium AL can partly dissolve apatite. The concentration of exchangeable P in the conventional system showed a trend from 7 to 19 mg P kg–1 soil across time, which is equivalent to 45 kg P ha–1 or 1.6 kg P ha–1 yr–1, but the trend was not significant (linear regression; R2 = 0.34, P = 0.15).
To find out the percentage of fertilizer P that became plant available, the following was calculated. Changes in amounts of plant available P in soil across time were related to the net input of P (addition of P – removal of P). The relative increase in P availability in the organic system (120 + 80 = 200 kg P ha–1) in relation to net P input (total addition of 903 kg P – crop removal of 184 kg P = 719 kg P) was 27.8%. The corresponding relative increase in P availability (45 kg P ha–1) in the conventional system (total addition of 519 kg P – crop removal of 311 kg P = 208 kg P) was 22%.
Exchangeable K in the control treatment remained relatively constant during the experimental period (50 ± 3 initial value; 47 mg K kg–1 soil final value), equivalent to 180 kg K ha–1 (Fig. 2C). Variation in plant available K in the control in different years (min. 22; max. 72 mg K kg–1 soil) could partly be explained by a higher demand for K by sugarbeet and winter wheat, and peaks could be caused by K release from crop residues. The input of 775 kg K ha–1 to the organic system did not change the level of exchangeable K (mean 54 ± 5 mg K kg–1 soil) significantly as compared with the control, although K was mainly added in plant-available form through cattle manure and yeast wastes (K mainly as K2SO4), and 66 kg K was added as potassium aluminum silicate (adularia). The application of roughly twice as much K to the conventional system (1476 kg K ha–1) as compared with the organic system through inorganic fertilizer and slurry resulted in a significantly higher (P < 0.05) soil K level of 86 ± 4 mg K kg–1 soil, corresponding to an increase of 180 to 320 kg K ha–1. The relative increase in plant-available K (140 kg K ha–1) in relation to the K input seems low (140 out of 1476 kg K = 9.5%). However, considering that the crops removed 1116 kg K over time (grain and residue was removed), 140 kg out of a net input of 360 kg K means that actually 39% of added K was present in exchangeable form.
Changes in Soil Organic Carbon
Concentrations of soil organic carbon (SOC) decreased across time in both the organic (27 to 25 g C kg–1 soil) and the conventional system (24–19 g C kg–1 soil) due to the conversion of previous grassland into arable production. The decrease of SOC was less in the organic system because it received about 40% higher C inputs (3.8 Mg C ha–1 yr–1) than did the conventional system (2.6 Mg C ha–1 yr–1). The higher C input in the organic system resulted from inclusion of cover crops, incorporation of crop residues and weeds, and use of solid manure. In the conventional system, the use of slurry, removal of crop residues and absence of cover crops led to a significantly lower C input.
Carbon dynamics were well reproduced by the ICBM model for the organic system without any model tuning (Fig. 3 ), but there was a slight deviation between modeled and measured SOC changes in the conventional system. However, the standard deviations of the SOC measurements in 1979 were high and may explain the difference. According to the model, the steady state concentrations for SOC would end at 19 and 12 g C kg–1 soil for the organic and conventional system, respectively.
|
|
In Scenario 2 (Table 3), all factors were held the same as in scenario one except now we assume that the same amount of 1.2 Mg straw ha–1 yr–1 was removed from both cropping systems. Total C inputs were 3.8 and 4.4 Mg C ha–1 yr–1 in the organic and conventional systems, respectively, and the corresponding steady state SOC concentrations were 18 and 22 g C kg–1 soil. The higher steady state SOC concentration in the conventional system is due to the limited removal of crop residues as compared with the complete removal in Scenario 1.
In Scenario 3 (Table 3), we include consideration of the yields and its relationship to potential animal stocking density. The yields obtained in the organic system can support a maximum animal density of 0.4 dairy cows ha–1 yr–1 (Agriwise, 2006) when all crops including beans and peas (Steineck et al., 2000) except potato were used as feedstuffs. Thus, five out of six crops are needed for fodder in the organic system. In the conventional system, an animal density of 0.4 units ha–1 yr–1 and manure application rate, similar to that in the organic system, is achieved when barley, rape, grassland and 50% of oats are used for animal feed, but sugarbeet, winter wheat and 50% of oats are exported to off-farm markets. Thus, 3.5 out of six crops are needed for fodder in the conventional system. Scenario 3 also assumes that both cropping systems include three cover crops in the rotations. Total annual C inputs in this scenario were 3.4 and 3.8 Mg C ha–1 yr–1 in the organic and conventional systems, respectively, and the corresponding steady state SOC concentrations were 15 and 17 g C kg–1 soil, respectively (Table 3).
In summary, when all the prerequisites for an unbiased comparison between systems (i.e., pitfalls i–iv listed above) were adjusted in the present study (Scenario 3), total C inputs became 10% higher in the conventional system than in the organic system despite considerably higher crop sales from the conventional system. Thus, we conclude that organic farming systems are not an option to sequester C in the soil under Swedish conditions. Data from other organic long-term field experiments such as the Swiss DOK (Dynamisch, Organisch-Biologisch und Konventionell) trial (Alföldi et al., 1993; Fliessbach and Mäder, 2000) and the Norwegian model farm study at Apelsvoll (Riley and Eltun, 1994; Breland and Eltun, 1999) showed that organically managed soils lost more SOC than conventionally managed soils. However, it is often claimed that organically managed soils have higher C contents and that organic farming is an option to sequester more carbon in the soil (Smith, 2004). A number of studies of organically managed soils actually reveal higher SOC contents, but, in these cases, organic C of off-farm origin was applied (e.g., Clark et al., 1998; Bulluck et al., 2002). As explained above, organic manures cannot be used in amounts independent of the management of the system to be meaningful for a scientific comparison (Faerge and Magid, 2003). The above discussion demonstrates that comparisons of agricultural systems require awareness of boundary conditions and stringent consideration of input factors.
Yields, Weeds, and Nutrient Contents in Crops
The year before the initiation of the experiment, barley was cropped to determine possible yield variations within the field and the total level of production. Barley yields averaged 1.8 Mg ha–1, varying from 1.6 and 1.9 Mg ha–1. During the 18-yr experimental period, yields tended to increase in the organic system (+33 kg dry matter ha–1 yr–1) and in the conventional system (+117 kg dry matter ha–1 yr–1), but has tended to decline in the control (–60 kg dry matter ha–1 yr–1) (Fig. 4
). None of these trends were statistically significant at P > 0.05. The comparisons of mean dry matter yields between the first and third crop rotation cycles within each cropping system (data not shown) indicated no significant changes across time. Thus, we conclude that the characteristic yield level of each system was reached within a few years without any conversion period. The nonexistence of a conversion period was due to the low initial nutrient status in the soil and not having a residual effect of earlier fertilizer (nutrient reserves) or pesticide applications (depression of weeds).
|
|
3 Mg dry matter in beans. The organic system weed biomass was significantly (P = 0.001) greater than in the conventional system, which averaged 27 ± 45 kg dry matter ha–1 yr–1 (Fig. 4). The presence of 1 Mg dry matter weed biomass in the organic system means an average withdrawal of at least 15 kg N from the main crop, assuming a N concentration of 15 g kg–1 dry matter. Furthermore, the presence of weeds in the organic system did not decrease with time despite intensive mechanical weed control, whereas the weed biomass in the conventional and control systems decreased significantly with time (–5 kg dry matter ha–1 yr–1; P = 0.019). There was also a change in weed species across time in the organic treatment. Perennial species became dominant in the organic system, which were even more difficult to control. During the first half of the experiment, the three most frequent (in respect to their mass) weed species in the organic system were charlock (Sinapis arvensis L.), spurry (Spergula arvensis L.), and red shank (Polygonum persicaria L.). During the second half of the experiment, field thistle (Cirsium arvense Scop.) and couch grass (Agropyron repens L.) became more frequent and were, together with charlock, the most abundant species. During years with clover/grass, dandelions (Taraxacum officinale F. H. Wigg. aggr.) were frequent.
The analysis of N, P, and K concentrations in barley and winter wheat grains revealed that nutrient concentrations were significantly different (P < 0.05) between the two cropping systems. Nitrogen concentrations were lower in the organically grown cereals (18–22 g N kg–1) than in the conventional (22–27 g N kg–1). The opposite was true for P and K concentrations, which were slightly higher in organic (4.1 g P kg–1 and 5.1 g K kg–1) than in conventional grains (3.8 g P kg–1 and 4.4 g K kg–1). Higher contents of N in the conventional grains and higher P in the organic grains are predictable due to higher N application to the conventional treatment and higher P application to the organic treatment. However, the organic treatment received lower K application but had higher K concentrations in grains (Table 1).
Nutrient Use Efficiency
Short-term measurements of nutrient use efficiency can be misleading if the residual effect of nutrient sources is large and not accounted for. Calculations over the long term seem more appropriate if organic nutrient sources (manures, slurries, cover crops) and untreated minerals (apatite, potassium aluminum silicate) are included. These conditions were fulfilled in the present study, as crop uptake was measured over 18 yr in the organic, conventional, and control treatments.
In general, the use efficiency of inorganic N and P fertilizers was higher than that of organic sources or untreated minerals over the long term (Table 5). In particular, crop utilization of P was extremely low in the organic system (7%) compared with the conventional (36%). It is reasonable to assume that the large application of apatite-P, which is sparingly soluble, greatly reduced the P use efficiency. However, excluding apatite-P from the calculation increased the utilization to 20% in the organic system, which is still only about half the recovery in the conventional system (Table 5).
|
What Limited Yields in the Organic System?
Lower yields in the organic system may be due to different reasons: lower inputs of N (Table 1), greater weed competition (Fig. 4), lower nutrient use efficiency (Table 5), and poorer control of pests and crop diseases. This study allowed us to identify one primary cause.
Soil analyses (Fig. 2) revealed that the amount of plant-available P was actually highest in the organic treatment. Furthermore, as mentioned above, crop analyses revealed higher P and K concentrations in grains organically grown than if conventionally grown. Thus, the P and K fertility status of the organically cropped soil in the present study, both initially and after improvement through amendments, was not a yield-limiting factor.
Data in Table 1 show that the mean yearly input of N in the organic system (108 kg ha–1) was considerably lower than in the conventional system (147 kg ha–1), followed by lower N concentrations in organically than conventionally grown grains. From this information, we conclude that available N was probably yield-limiting in the organic treatment. This conclusion must be seen in light of the initial aim, namely to choose crops in the organic rotation that enabled a maximum N input through fixation. In fact, legumes were grown as main crops every second year to supply the subsequent crop with N-rich residues (mean of 45 kg N ha–1 yr–1), and three legume cover crops provided an additional N input (yearly mean of 25 kg N ha–1 yr–1).
A central question is whether the N shortage in the organic system is limited to the present study or whether it is a general characteristic of organic cropping systems. A recent review of N input showed that lower amounts of nutrients are applied to organic systems compared with conventional systems (Kirchmann and Bergström, 2001). Even in a long-term field trial in the USA under warm, arid climatic conditions, N deficiency was pointed out as the major yield limiting factor (Clark et al., 1999). High inputs of P and K to organic systems are possible through purchase of untreated minerals (apatite, potassium magnesium sulfate, potassium aluminum silicate). Worldwide, untreated N resources (guano and Chilean nitrate) are scarce. Thus, organic farming must rely on biological N fixation or the importation of off-farm organic waste products for its N supply. Such products (e.g., fodder, meat and bone meal, food industry wastes, compost, animal manure, biogas residues) often originate from conventional production, which demonstrates a reliance on conventional production systems. Organic systems must incorporate large proportions of legumes in their crop rotations or apply N through manure or as green manure crops to maintain the N supply of crops.
It seems that only organic grassland–ruminant systems enable a high input of N to the soil. Yields of grass/clover forage crops are often similar between organic and conventional systems (Eltun et al., 2002), which was also the case in this study. In organic crop–livestock systems, however, a reduction in the number of legume forage crops in rotation and their substitution by other crops results in a lower N fixation potential (Olesen et al., 2002). The additional N inputs through legume cover crops seldom exceed 50 kg N ha–1 under cold-temperate conditions, as crop growth in autumn is limited to a maximum yield of 2 Mg ha–1 (Bergström and Jokela, 2001). Whole-year green manure crops may add amounts of N similar to legume forage crops (Olesen et al., 2002), but since green manure crops are not harvested, this practice reduces the number of harvestable crops in a rotation. In organic farming systems without livestock, on-farm N inputs are further reduced as no N is returned through manure/slurry if not purchased as approved organic fertilizers.
It appears that the lower mean N supply seems to be a major yield-limiting factor in most organic systems with the exception of grassland ruminant agriculture. The organic principal "to base organic farming on living ecological systems and cycles," relying on on-farm N input and recycling of N, can explain this (International Federation of Organic Agricultural Movements, 2006).
Inorganic Nitrogen in Soil and Leaching Water
Inorganic N (NH4 and NO3) contents in soils sampled in autumn after the crop harvest provide an indication of N leaching risk under Swedish climatic conditions (Table 5). Amounts of mineral N in the soil profile (0- to 0.9-m depth) sampled across 12 yr was 66.0 kg N ha–1 (mean of all years) in the organically managed soil, and 48.8 kg ha–1 (mean of all years) in the conventionally managed soil, but this difference was not significant (P = 0.08). There were variations between years that can probably be attributed to the type of crop grown. Highest inorganic N occurred after years when potatoes (97 kg N ha–1), oilseed rape (89 kg N ha–1), clover (80 kg N ha–1), and beans (75 kg N ha–1) were grown. Soil inorganic N was intermediate after peas (48 kg N ha–1) and cereals (44 kg N ha–1) were grown and lowest after sugarbeet (42 kg N ha–1) and legume-free grass (26 kg N ha–1) were grown. To check whether potato (only grown in the organic treatment) might affect potential leaching losses, mean inorganic N values were recalculated eliminating years with potato crops. The mean value decreased from 66 to 60 kg N ha–1 which was not significantly different (P = 0.19).
Inorganic soil N content in spring (n = 15 yr), which is an indication of N mineralization during winter, showed no significant differences (P > 0.05) between the conventional (mean 59.6 kg N) and organic treatments (mean 54.8 kg N).
The analysis of drainage water showed that mean N concentration for the experimental period was 13.1 mg N L–1 in the organic treatment and 10.7 mg N L–1 in the conventional treatment (Table 5), but were not significantly different (P = 0.7). During some periods (1992), concentrations increased in the organic treatment (P = 0.02), but during other periods (1993) they tended to increase in the conventional treatment (P = 0.25) (Fig. 5 ). Difference in drainage water N between the two cropping systems could usually be attributed to the specific crop being grown in the cropping system. Growing legumes, potatoes, sugarbeet, and cereals resulted in higher N concentrations in the drainage water. The highest N concentration (34 mg N L–1) occurred when potato was grown in the organic system. This is in line with the high levels of inorganic N left in soil after a potato crop (see above), and in good agreement with the results obtained in another Swedish study comparing organic and conventional cropping systems (Torstensson et al., 2006).
|
What are the Consequences of Conversion to Organic Cropping?
Yield data from European long-term organic field experiments reveal significantly reduced production (Eltun et al., 2002; Torstensson et al., 2006). Lower yields in organic systems than in conventional systems are often simply reported as facts without further elaboration. Lower production per unit area of land requires more land to produce the same amount of food. Agricultural production needs to be seen from the perspective of food demand. A significant reduction in food production needs to be compensated for by production elsewhere. To put it simply, substantially more land is needed to produce the same amount of food if organic agriculture is introduced on a large scale. In this context, it is important to stress that if a major proportion of agriculture were to be run according to organic principles, animal manures and other organic nutrient sources would simply not be sufficient to feed the growing world population (Chen and Wan, 2005). Furthermore, the necessary expansion of agricultural land is often restricted by various climatic factors, such as drought. Much of the world's food is produced under irrigation in semiarid areas where human population growth is diminishing water availability for agriculture (Bruinsma, 2003). As a consequence, there will be a greater dependency on food production from rain-fed agriculture in the future. In this context, it is questionable whether conversion to organic crop production really is a viable alternative to produce sufficient amounts of food.
Another important issue is that if more land is used for agriculture to produce sufficient food, leaching losses will be increased. Transformation of natural land, forests, or grazing areas to arable land results in significantly higher leaching losses. Leaching losses per unit of area are as a rule lower from ecosystems having a permanent vegetation cover (Berdén et al., 1997; Nohrstedt et al., 1996). Furthermore, even possibilities of conserving biodiversity are greatly reduced if wild nature is transformed into arable land (Green et al., 2005). In fact, we see the necessary expansion of agricultural land to maintain food production as the most serious environmental impact posed by organic farming.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Red clover cover crops were used to fix N and reduce N leaching, but were found to be insufficient to decrease the N leaching potential from the organic system. Significantly lower yields in the organic system means far greater N leaching from the organic than conventional systems if expressed per unit of product.
The use of slag and apatite as P sources and of potassium aluminum silicate [K(AlSi3O8)] and yeast wastes as K sources in organic farming instead of processed P and K fertilizers was not found to limit crop production. Instead, N was identified as being the yield-limiting element in the organic system despite excessive growth of legumes in rotation. Furthermore, the frequent growth of legumes made the organically managed soil more susceptible to acidification.
Since grass/legume forage crops are the only crops for which organic yields were similar to conventional yields, only organic grassland-ruminant systems can potentially achieve the same level of production as conventional systems under cold-temperate conditions. The much lower yields in organic than in conventional production systems require the greatest attention and must be viewed from the perspective of demand for food and land.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Crop Science | Vadose Zone Journal | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
), control (
), and conventional (
) systems over three 6-yr crop rotations (±SE, n = 3). The weed biomass in the conventional system and control (data not shown) was very low due to herbicide treatment.