| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Department of Soil Chemistry and Plant Nutrition, Institute of Soil, Water and Environmental Sciences, The Volcani Center, Agricultural Research Organization, P.O.B. 6, Bet Dagan 50250, Israel
b Agricultural Research Organization, Gilat Research Center, D.N. Negev 85280, Israel
c Present address: Réseau des Missions Déchets, APCA-Chambres d'Agriculture 9, Avenue George V, F-75008 Paris, France
* Corresponding author (abartal{at}volcani.agri.gov.il)
Received for publication April 26, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: CMC, cattle manure compost • NCSOIL, a model of nitrogen and carbon dynamic in soil • Q10, coefficient of increase in the decomposition rate constant per temperature increase of 10°C • SSC, sewage sludge compost
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The overall objective of the present study was to examine the possibility of using the NCSOIL model with parameters of composts that were determined according to results of an incubation study, to establish guidelines on the optimal amount of composts and fertilizer N, which should be applied annually to wheat under specific conditions. The specific objectives were to (i) determine the parameters of the composts that control the decomposition rate of the organic waste and the rate of inorganic N release and (ii) predict the available N content in soil that is cropped with wheat for three consecutive years in the greenhouse and receives annual applications of composts.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Soil and compost materials for the incubation experiments were the same as those used in the wheat pot experiment (Bar-Tal et al., 2004). The soil was a Typic Rhodoxeralf, very sandy, 4% clay, 2% silt, and 94% sand, with 1.7% calcium carbonate and pH 7.5. Total organic C and organic matter contents were 3.25 and 5.6 g kg–1, respectively, as determined by the dichromate oxidation–titration method. The total N content was 96 mg kg–1, as determined after digestion with H2SO4 and H2O2 and analysis by a colorimetric method in an autoanalyzer (Lachat Instruments, Milwaukee, WI).
The two types of compost were a commercial cattle manure compost (CMC) and a domestic sewage sludge compost (SSC) from the Netanya sewage treatment plant. Chemical analyses of the two composts are given in Table 1. The pH and EC were determined in a 1:5 water extract. Organic matter and total C and N contents in the compost were determined by the same method as that used for the soil. Total C was 58% of the organic matter. The P content was determined with the Lachat autoanalyzer in the same digestion as the N. Inorganic N as NH4 and as NO3 was determined by steam distillation in a 1:8 soil to 1 M KCl extract.
|
Mineral N concentration was determined periodically throughout the 140 d of incubation by sampling 5 g of soil. Soil samples were extracted in 1:5 1 M KCl and tested for ammonium and nitrate with the Lachat autoanalyzer. For measurements of the CO2 emission rate, a trap of 4 mL of alkaline solution (1 M NaOH) was installed in a cup above the soil before each jar was tightly closed for 48 h. The surface area of the solution in the cup was 13 cm2, compared with the 150 cm2 of the soil surface in the jars. Excess NaOH was titrated with 0.2 M HCl after carbonates had been precipitated with BaCl2, and the result was subtracted from that obtained from a control without soil. Both mineral N and CO2 were sampled twice a week during the first few weeks of incubation or after rewetting, and once a month later, when mineralization had slowed down. During the first two weeks, NH3 volatilization was measured by using a trap containing 4 mL of boric acid–indicator solution (Keeney and Nelson, 1982) in a similar cup to that used in the alkaline trap. Ammonium was titrated with 0.01 M H2SO4.
Greenhouse Experiment
The average weekly minimum and maximum air temperature ranged from 12 to 18°C and from 20 to 30°C, respectively. The soil and composts were air-dried and ground to pass through a 2-mm sieve. At the beginning of the first year, December 1997, thirty-two 100-L plastic containers (40-cm diameter, 60-cm height) were each filled with 80 kg soil. The composts were applied to two successive years of wheat crops, then a third year of wheat crops after one year of break without planting, irrigation, and compost application. Six treatments included three application rates of each of the two composts of 325, 650, and 1300 g (dry weight) of compost per container, or an equivalent of 3, 6, and 12 kg m–2, respectively. Thus, a wide range of composts' load was obtained according to the main goal of the research. Two additional treatments were controls without composts: one was fertilized and one was unfertilized. A randomized design of the eight treatments in four blocks was applied: one container for each treatment in each block. The upper 15-cm soil layer of each container was taken out and compost was added to it and thoroughly mixed, according to one of the eight treatments (including the controls without composts) and then replaced on top of the remaining soil. The containers were irrigated until leaching began, which indicated the attainment of water-holding capacity. Wheat (cv. Ayalon) was then sown at 100 seeds per container (926 seeds m–2). This rate is much higher than typical field seeding rates to ensure that there were adequate plants for several samplings during the growth until final harvesting. Soil moisture was controlled according to weight loss and adjusted by daily manual irrigation with deionized water in the first two years, and by drip irrigation with tap water of electrical conductivity of 0.9 to 1.0 dS m–1 in the last year. The quantity of drainage was negligible, less than 1% of the applied water. The fertilized control was irrigated with nutrient solutions. The wheat was harvested 135 to 148 d after seeding, and the soil was left to dry.
At the beginning of the second and third years, the top 15-cm soil layer was removed and mixed with an aliquot of compost identical to the one it had received. The soil mixed with compost was returned to the container and irrigated until the water holding capacity was attained. Seeds were sown at a density of 60 seeds in each container (556 seeds m–2).
Soil and plants were sampled during the course of the growing period. Plants were sampled in the first year on Days 32, 57, 85, and 135 (harvest) after seeding; in the second year on Days 32, 57, and 140 (harvest) after seeding; and in the third year of cropping on Days 36, 69, 97, and 148 (harvest) after seeding. The aboveground biomass of 10 plants was removed from each container, weighed, dried at 65°C, and analyzed for dry weight and for N concentrations by wet digestion, as described above for the compost. Four replicate soil cores were collected from each container during the first year on Days 32 and 135 (harvest) after seeding; in the second year on Days 48, 64, and 140 (harvest) after seeding; and in the third year of cropping on Days 36, 69, 97, and 148 (harvest) after seeding. The initial soil and the soil samples from the cores, divided into layers of 0 to 15, 15 to 35, and 35 to 55 cm deep, were analyzed for ammonium and nitrate, following extraction in 1 M KCl at a 1:5 soil to solution ratio.
Modeling
The model framework we used was NCSOIL (Molina et al., 1983; Hadas and Molina, 1993); it comprises several organic pools that differ in their functions. Organic residues represent materials added to the soil (i.e., the compost in the present case). This pool can be divided into several components, differing in biochemical stability: the microbial biomass consists of a zymogenous microbial pool that feeds on the residue pools, and the soil autochthonous microbial pool that feeds on decayed microbial biomass and on the mineralizable soil organic matter that is not derived from microbial biomass. All organic pools are defined according to their C content, first-order decomposition rate constant, microbial use efficiency factor (i.e., the fraction of decomposed C assimilated by the heterotrophic microbial biomass), and the C to N ratio. The rate of C flow from one pool to another depends on the first three pool characteristics, and that of N flow is determined by the C flow, in accordance with the C to N ratio of the organic pools. Further details on pool characteristics and processes included in the model were presented by Molina et al. (1983)(1990) and Hadas and Molina (1993).
Optimization of Compost Pool Characteristics
Mineralization results obtained with the soil–compost mixtures were transformed into values appropriate to the compost itself, by subtracting the results obtained with the untreated soil control from those obtained with the mixture, to remove the effect of soil organic matter; in fact, the latter was very low in this soil. For each compost the data of two variables (N and CO2) in three treatments (application rates), measured at each sampling date were used as input to optimize the compost decomposition parameters as defined by NCSOIL. Each data point was the average of the two experimental replicates. The model was adjusted to simulate the mineralization of the two composts by searching for optimum values of residue pool parameters, by means of the Marquardt algorithm, which had been introduced as a module in NCSOIL (Barak et al., 1990). The best fit of simulated to measured data was based on the weighted least sum of squares of residuals, represented by a 
2 value as follows:
 |
[1;2], mineral N and CO2); j is the sampling index (sampling time); k is the index of the experimental treatment (here k [1;3], three application rates for each compost); Yijk are the measured values; Yi(jkA) are the simulated values with a set of parameters called A; SDi is the standard deviation of the Yi measurements; and DF is the degree of freedom. We assumed that each compost could be considered as comprising two pools (labile and resistant), representing two phases in decomposition. The objective was to determine the characteristics of the two components of the residue pools (carbon content, C to N ratio, decomposition rate constant) from the data obtained in the first 140 d of the experiment. Optimization for each of the composts addressed four parameters, that is, C and N partitioning between labile and resistant pools (total C and N were known), and their decomposition rate constants. There is a difficulty in the simultaneous optimization of the labile fraction degradation rate and initial size because they are correlated as the CO2 flux and mineralized N are products of the two parameters. In the optimization process initial values of the parameters are introduced. The initial value of the labile fraction was based on the accumulative amount of CO2 produced during the first 8 weeks of high flux rates. Varying the initial size of the labile pool in the range of the amounts of CO2 produced during the first 4 to 12 weeks of high flux rates did not change considerably the obtained optimized parameters. The optimization of the decomposition rate constant of the resistant pool was not possible during the 140-d incubation period because the estimated decomposition was not significant during that period. Therefore, in the first step of optimization we assumed that the decomposition rate constant was 0.0001 d–1, based on published data for stabilized soil organic matter as listed by Molina et al. (1994), then only this parameter was optimized using the data obtained in the longer incubation experiment, which lasted 370 d. Then, the optimization of the other parameters with the data of the short-term incubation experiment was repeated. The new parameters were again used in optimization of the decomposition rate constant of the resistant fraction with the long-term experiment and this process continued until the change in the obtained parameters was smaller than 5%.
During the first stage of the modeling work (the incubation experiment), the model was only slightly modified to account for the inorganic N loss observed in the first few days. The daily rate of ammonia volatilization following application of compost in the sandy soil was 0.5% of the NH+4–N concentration in the soil.
The effect of drying and rewetting on C and N mineralization could have been partly due to microbial biomass decay (van Veen et al., 1985), and partly due to the breakdown of resistant organic matter fractions (van Gestel et al., 1991). In the present study we assumed that in the drying period the zymogenous microbes would die and become labile, and that part of the resistant component of the compost would break down and become part of the labile component. The amount of this component was optimized with incubation data following a period of drying and employing the parameters obtained in the previous short and long incubation experiments (140 and 370 d) described above.
During the second stage of the modeling work (the greenhouse wheat experiments), two modifications were introduced into the framework model. The first modification was the activation of the temperature reduction factor option. This function exists in NCSOIL but is constant during the course of the cycle, whereas in the real greenhouse the temperature varies with time. To take account of this variation, we introduced a weekly temperature reduction factor, based on temperature measurements. The average weekly soil temperatures were calculated from the greenhouse air and soil temperatures, which were measured daily throughout the experiments. We assumed that the optimal reference temperature was 30°C, and that the Q10 coefficient (increase in the decomposition rate constant per temperature increase of 10°C) is in the range of 1.6 (Clay et al., 1985) to 2.0 (Stanford et al., 1975); both values were tested.
The second modification concerned the effect of living plants on the C and N economy in soil, and was added as a new module called "N-uptake and root deposition," following the work of Hadas et al. (2002) on corn (Zea mays L.). The general idea is that the plant affects the N balance as a sink for mineral N, and the C balance as a source of organic C; the proportions of N taken up as NH+4 and as NO–3, respectively, depend on the ratio between their concentrations; and the organic C input comprises the exudations of living roots, and the dead roots left by the sampling of the aboveground plant material (of which the latter relates specifically to our greenhouse experiment). The plant effect in the model is not computed by means of growth equations but is obtained from plant sampling and estimated based on values in the plant physiology literature. The input data to this new module comprise three variables determined at four sampling dates: the cumulative dry weight production of the aboveground plant biomass, the cumulative aboveground N uptake, and the amount of dead roots left after the aboveground sampling. The first two are calculated directly from measurements, whereas the last is estimated according to the first of the following plant physiology assumptions; the other assumptions are used to estimate the total N uptake and root depositions in the soil.
From the change in the measured aboveground dry weight between two consecutive sampling dates, one is able to calculate by linear interpolation the daily aerial plant growth for that period and, by using Assumption 1, similarly to calculate the daily root growth and thus the daily total growth.
Assumption 4 implies that the partitioning of carbon in the direction of roots has to be doubled based on the quoted references. However, there is a lack of information on the dynamic of exudates deposition with plant age; thus, we used a constant value that allocated more C to the roots than to the shoot in the first 20 d of growth. However, as the total weight of the plant was small at this age the effect on the rhizosphere was relatively small.
In our greenhouse experiment, aboveground sampling introduced another input of organic C into the soil: the dead roots. At each sampling date the amount of C that enters the system from the dead roots is calculated by applying Assumptions 1 and 3 to the measured dry weight of the aerial biomass sample. This amount is divided into two pools according to the following assumption:
Finally, when it detects N deficiency in the soil, the root system reacts according to the following assumption:
Consequently, root deposits also increase by 50% when N is deficient.
To represent greenhouse conditions, the N contributed through tap-water irrigation was also added to the model, as a new daily source of mineral N, according to the NO–3–N concentration in the water (5 mg L–1) and the amounts of water applied.
Mineral N concentrations in the soil, as measured during the plant growth period, were used to evaluate the results of simulations that used the compost parameters in the modified NCSOIL model.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
Optimization of Compost Parameters
The finding of two phases of decomposition justified the assumption that composts can be considered to comprise two major components: one labile and one resistant. The simulations of the incubation results with the optimized compost parameters are shown in Fig. 1, and the parameters are presented in Table 2. Both composts showed some similarities: first, the labile component represented 16 to 20% of the total organic carbon in the compost; its C to N ratio was much higher than that of the resistant component; and lastly, the decomposition rate constants of the components of SSC and CMC were very close. The decomposition rate constants of the labile component were optimized with the short-term incubation experiment (Fig. 1), whereas those of the resistant component were optimized on the basis of the long-term N mineralization data (Fig. 2)
. The latter values are in the range of decomposition rate constants for stabilized soil organic matter as listed by Molina et al. (1994). Differences between the composts are mainly in the total amounts of C and the C to N ratios in each component of the residue pool, which could be estimated from the initial total C and N analyses.
|
|
Greenhouse Experiment
By using the C and N mineralization kinetics of both composts that had been determined from the parameters optimized in the incubation study, and applying the modified version of NCSOIL (that included the temperature reduction factor and plant effect), we successively simulated soil C and N turnover in the three sequential years of wheat grown in the greenhouse. Between successive years, the pools were recalculated and included in new input files to take into account changes caused by the drying effect and the new addition of compost.
Simulated mineral N turnover during the last year is shown in Fig. 3 . The effect of the type of compost (SSC or CMC) on the N economy seems very mild compared with the effect of the application rate. The plant N uptake curve is the linear integration of measured values obtained from plant shoot samplings, and estimates of the root contribution, both in the third year. For the high dose of compost (12 kg m–2) the uptake rate (represented by the slope of the curve in Fig. 3) was almost constant for the first 70 d after seeding of the last year (Days 730–800), after which it gradually fell; whereas for the lower doses the N uptake rate was greatly reduced after 40 d (Day 770). The pattern of soil mineral N content variation was complementary to that of plant N uptake. The soil mineral N content was the result of the activities of sinks (plant uptake) and sources (net N mineralization and N in the irrigation water); it initially comprised that left over from the second growing year plus the initial mineral N content of the added compost, and decreased as a result of plant uptake. The soil mineral N content approached a minimum of 5 g m–2 after 40 d (Day 770) in the low- and medium-dose treatments, but only after 100 d (Day 830) in the high-dose treatments, and after that it gradually increased, especially in the high application rate treatments. Soil mineral N fluctuations were simulated successfully and showed close agreement with measured data. The net N mineralization was low at the beginning, even negative in the SSC treatments (gross immobilization exceeded gross mineralization). Around Day 830, net N mineralization increased slightly, and as much N was mineralized during the last 50 d as during the first 100 d of the last year (Days 730–830). The simulated mineral N contents in the low-rate compost treatments were below measurable values, otherwise the simulations fitted the measured mineral N concentrations very well.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Nitrogen dynamics in compost-amended soils ranges from drastic immobilization to net N mineralization (Hue and Sobieszczyk, 1999), depending on the maturity and the C to N ratio (Bernal et al., 1998). Concerning the release of mineral N in the present study, the major part of the compost (comprising 80–85% C) remained in the soil after 140 d of incubation at 30°C. Because of the high C to N ratio of the labile fraction, immobilization occurred, especially with the SSC, and the net mineralization of the organic N was just 11 to 14%. One may assume that under field conditions, with lower temperatures and slower mineralization, mineralized N would account for a negligible proportion of plant N nutrition requirements. Regarding the effect of the soil, this experiment was performed in a much depleted sand that contained little microbial biomass and little organic matter. Because of the small contribution from the soil, microbial biomass buildup was based only on the C and N released by the compost, and the results of the incubation experiment precisely reflected the decomposition of the compost. This could be used as a standard method to determine the mineralization properties of organic waste.
The negligible effect of a drying period on organic C and N dynamics in soil that we observed and optimized confirms the findings of van Gestel et al. (1991). If attention is restricted to the three weeks following rewetting of the soil, the C flush and mineral N decrease may appear important, but when the long-term stabilization of the curve is considered, these phenomena can be neglected.
Greenhouse Experiment
The dynamics of mineral N in the six treatments in the greenhouse experiment was strongly affected by plant uptake, and showed evidence of N deficiency in the low- and medium-application rate treatments (Fig. 3). When N uptake terminated, around Day 40 after seeding for the low- and medium-application rate treatments, the soil mineral N content was very low (approximately 5 mg m–2), whereas in the high-rate treatment, N uptake slowed down around Day 100 after seeding (Day 830), which, again, was when the soil mineral N content approached 5 mg m–2.
The shape of the net N mineralization curve obtained in the greenhouse experiments was rather different from that obtained during the incubation study (Fig. 1). During the latter, a period of 50 d with no accumulation preceded a period of steadily increasing mineral N. In the greenhouse there was no "no-accumulation period," but a 100-d period (Days 730–830) with little net mineralization, followed by increased accumulation until harvest. The relative amount of N mineralized in the greenhouse was also smaller than in the incubation. We assume that there are three main reasons for these differences.
|
|
Field Simulation Scenarios
To extrapolate our findings and estimate the amount of compost required to meet wheat demand for N, field scenarios of compost and fertilizer application were simulated. The optimal total N uptake by wheat was taken to be 35 g m–2, a value derived from data collected for wheat in a permanent-plots-fertilization experiment (Kafkafi and Halevy, 1974), of which 5 g m–2 was provided in the irrigation water. The N uptake was not predicted in this simulation but was given as input data as long as the amount of mineral N in the soil is sufficient to meet the target. In the first scenario, we assumed application of compost at 6 kg m–2 of compost to two different soils: the depleted sandy soil that was used in our incubation study (mineralizable C content, 60 mg kg–1; C to N ratio, 12; microbial biomass C content, 3 mg kg–1) and an agricultural soil (mineralizable C content, 1200 mg kg–1; C to N ratio, 12; microbial biomass C content, 60 mg kg–1). NCSOIL stopped the simulation whenever the mineral N was exhausted. This compost rate was not sufficient to support the reference plant N uptake; therefore, two equal fertilizer top dressings, applied on Days 14 and 35, were gradually adjusted to avoid soil inorganic N depletion. In the second scenario, we adjusted the minimal rate of compost necessary to run the full cycle of wheat in the sandy soil with the above-described wheat N uptake (Kafkafi and Halevy, 1974).
The simulated results are presented in Fig. 6 . Sewage sludge compost or CMC applied to the sandy soil at 6 kg m–2 without N fertilizer contributed only 15 or 12 g N m–2, respectively, toward the total N consumption of 35 g m–2. In the agricultural soil, endogenous organic N also contributed to net N mineralization. Therefore, in agricultural soil 12 or 13 g N m–2 had to be applied as fertilizers in addition to application of 6 kg m–2 of SSC or CMC, respectively; whereas in the depleted sandy soil 20 or 23 g N m–2 had to be applied as fertilizers in addition to application of 6 kg m–2 of SSC or CMC, respectively. Consequently, in the agricultural soil 8 and 10 g m–2 of N fertilizer were saved, respectively, in the SSC and CMC treatments, relative to the amounts needed in the sandy soil. In the sandy soil there was a steep increase in mineral N concentration after fertilizer N application on Days 14 and 35 (Line 4 in Fig. 6), and the soil was almost completely depleted after 90 d, whereas in the agricultural soil the changes in mineral N concentration after the fertilizer N application on Days 14 and 35 (Line 3 in Fig. 6) were smaller and the concentration remaining at harvest was higher. For the second scenario, in which only compost, without N fertilizer, was applied to the sandy soil (Line 2 in Fig. 6), the amounts of SSC and CMC required to meet the target N uptake were 17 and 20 kg m–2, respectively. Application of such enormous amounts of compost would be both impracticable and a threat to the environment because of the large initial concentrations of mineral N present before the main period of N uptake by plants, and because the large quantities of mineral N produced by late N mineralization after harvest (Fig. 6) would be subject to leaching toward the ground water.
|
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Cropping conditions affected C and N dynamics in the soil mainly through: (i) soil temperature (which affected the rates of biochemical processes compared with standard conditions), (ii) mineral N uptake by plants, and (iii) the release of very labile organic C in root exudates. The modified NCSOIL model, which included these effects, successfully simulated compost mineralization under a wheat crop, using mineralization characteristics optimized in an incubation experiment. The contribution of composts to plant N nutrition was limited because of the immobilization of N by the soil microbial biomass, which was enhanced by the available C contained in root exudates.
In the short term, large amounts of compost are required to supply the inorganic N that it contains, because the rate of mineralization of organic N from compost falls short of the uptake by wheat. In the long term, the compost will have a residual effect on soil fertility and crop N availability. The scenarios in sandy and fertile soil enable one to estimate the appropriate combination of compost and mineral N to meet crop demand without leaving residual inorganic N at the end of the cropping season.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Related articles in JEQ:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
