SSSAJ Grow Your Career with SSSA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (3)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Wang, H.
Right arrow Articles by Zhou, H. F.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Wang, H.
Right arrow Articles by Zhou, H. F.
Agricola
Right arrow Articles by Wang, H.
Right arrow Articles by Zhou, H. F.
Related Collections
Right arrow Other Soil Management
Right arrow Soil Models
Right arrow Nutrient Management
Soil Science Society of America Journal 66:1304-1310 (2002)
© 2002 Soil Science Society of America

DIVISION S-6—SOIL & WATER MANAGEMENT & CONSERVATION

Simulation of Soil Carbon Dioxide Flux During Plant Residue Decomposition

H. Wang*,a, D. Curtinb, Y. W. Jamea, B. G. McConkeya and H. F. Zhouc

a Semiarid Prairie Agricultural Research Centre, Agriculture and Agri-Food Canada, Box 1030, Swift Current, SK, Canada S9H 3X2
b New Zealand Institute for Crop & Food Research, Private Bag 4704, Christchurch, New Zealand
c Shenyang Agricultural Univ., 120 Dongling Road, Shengyang, PRC

* Corresponding author (wangh{at}em.agr.ca)


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 Total Carbon Dioxide Emissions
 CONCLUSION
 REFERENCES
 
Based on information obtained from recent studies, we modified the CENTURY model to improve simulation of short-term soil respiration, especially in soils with surface-applied crop residues. This involved adding N availability as a factor controlling the decomposition rate. Translocation by filamentous fungi was assumed to be the mechanism supplying mineral N to residues decomposing on the soil surface. When available N is nonlimiting, the N availability factor is 1, otherwise decomposition rates of all pools of soil surface and belowground organic matter are reduced proportionately until N supply meets demand. The modified model was evaluated using CO2 flux data from a laboratory experiment which included different wheat (Triticum aestivum L.) straw types (fresh and weathered straw), straw placements (incorporated and surface-applied) and soil water regimes (continuously moist and alternating moist-dry conditions). In general, CENTURY successfully simulated daily CO2 fluxes in these treatments, except for an underestimation in the first day after watering and an overestimation immediately after rewetting dry soil in the moist-dry water regime. For treatments with surface-applied straw, CENTURY overestimated soil respiration, while the modified version gave substantially better simulations. The correlation between measured and simulated total (in 77 d) respiration was improved by model modification. CENTURY underestimated the soil mineral N remaining in the soil at the end of the experiment. The modified model gave improved mineral N simulations for the surface straw treatments.

Abbreviations: LOFIT, lack of fit • NT, no-tillage • SOM, soil organic matter


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 Total Carbon Dioxide Emissions
 CONCLUSION
 REFERENCES
 
THE CONCENTRATION OF CO2 in the atmosphere has increased substantially since the beginning of the industrial revolution. There are concerns that continuing increases in levels of CO2 and other greenhouse gases will contribute to climate change. Soils contain about three times as much C as the atmosphere, and they have the potential to store additional C (Campbell and Zentner, 1993). Research shows that, if properly managed, agricultural lands could be an important sink for C (Janzen et al., 1998). Reduction in tillage intensity, especially the adoption of no-tillage (NT) cropping, is widely recognized as an effective management technique to enhance C storage in soil (Kern and Johnson, 1993; Lal and Kimble, 1997; Dao, 1998). Many long-term experiments have shown that adoption of NT may increase soil organic matter (SOM) in temperate regions and subhumid and humid tropics (Paustian et al., 1997a). The effect of management on C sequestration is strongly affected by weather, soil conditions (e.g., texture), and agronomic factors, such as crop rotation, fertilizer application, and residue treatment (Janzen et al., 1998). As a result, the SOM gain in response to NT can vary widely within a region such as the Canadian prairies (Campbell et al., 1995, 1996; Janzen, et al., 1998).

Given the complexity and multifaceted interactions between the factors listed above, simulation models that describe SOM turnover and N dynamics in soils are useful for estimating management-induced SOM changes (Paustian et al., 1997b). The CENTURY model (Parton et al., 1987) is one such model. It was originally developed for simulating long-term management effects on SOM in Great Plains grasslands. It is widely used, and it has been extensively evaluated, in a variety of ecosystems and locations (Paustian et al., 1992; Kelly et al., 1997; Scholes et al., 1997; Smith et al., 1997). Recently, CENTURY has been used to simulate short-term changes in SOM. Del Grosso et al. (2001) used the DAYCENT model (Parton et al., 1998), which is the daily time-step version of CENTURY, to simulate daily C dynamics and trace gas fluxes. The System Approach to Land Use Sustainability program (SALUS), which combines a daily based crop-growth model and CENTURY, was developed to simulate continuous crop, soil, water, and nutrient conditions under different managements (Schulthess and Ritchie, 1996). Paul et al. (1999) used SALUS to predict the intra and interyear differences in field CO2 fluxes. Recently, the CENTURY SOM module was incorporated into the Decision Support System for Agrotechnology Transfer software program (DSSAT) and this improved simulations of residue decomposition (Gijsman and Bowen, 1999; Gijsman, 2000).

Because crop residues have a short turnover time in soil compared with other SOM pools (Parton et al., 1987), their decomposition strongly affects the short-term dynamics of soil C. Curtin et al. (2000) indicated that the lower CO2 flux under NT than under conventional tillage could be attributed to slower decomposition of crop residues on the surface of NT soil than when they were incorporated by tillage. Duiker and Lal (2000) concluded that lack of significant difference in soil CO2 flux between treatments receiving different rates of surface-applied residue was partly because of undecomposed surface residue which did not contribute to the CO2 flux. Although CENTURY contains above and belowground C pools, decomposition of above-ground pools is not well-described because quantitative studies on the mechanism of soil surface decomposition are limited (Frey et al., 2000). The objective of this study was to use information provided by recent research to modify the CENTURY model to improve the simulation of short-term soil CO2 emissions, especially when plant residue is applied on the soil surface.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 Total Carbon Dioxide Emissions
 CONCLUSION
 REFERENCES
 
The Experiment
Details of this laboratory experiment have been reported by Curtin et al. (1998). A bulk sample of a Swinton silt loam (Typic Haploboroll) was taken in the fall of 1994 from the top 15 cm of a field at the experimental farm of Semiarid Prairie Agricultural Research Centre, Agriculture and Agri-Food Canada, Swift Current, SK. The field had been summer fallowed in the preceding growing season. The soil contained 200 g clay kg-1, 500 g silt kg-1, and 300 g sand kg-1 and had a pH of ~6. The C and N concentrations of the soil were 15 and 1.5 g kg-1, respectively. The soil was sieved (<4 mm) while still field moist (145 g H2O kg-1). Visible plant residues were removed by hand. Subsamples (equivalent to 13 kg dry soil) were placed in wooden boxes (39.5 by 34.5 by 10 cm) lined with black polythene sheets and packed to a bulk density of 1.0 Mg m-3. The boxes were placed in a greenhouse for 3 wk to allow the effects of soil disturbance on CO2 emissions to subside. Distilled water was added (as a fine mist) every 2 to 3 d to return the soil to field moisture content.

A factorial, randomized complete block design with three replicates was used in this experiment. There were three factors: straw type (fresh vs. weathered straw), straw placement (surface-applied vs. incorporated), and soil moisture (continuously moist vs. alternating moist-dry conditions). There were two controls: no-straw soil with either continuously moist or moist-dry moisture regimes.

Wheat straws were collected in November, 1994 from fields adjacent to that from which the soil was sampled. Samples were taken from the current year's stubble (referred to as fresh straw) and from standing stubble of the previous year's crop (weathered straw) (the field with weathered straw had been chemically fallowed in 1994). Both straw types, which were from the same variety of hard red spring wheat (Lancer) were collected by cutting upright stubble 4 to 5 cm above the soil surface. The straws were air-dried in the laboratory and cut into 2.5-cm lengths using a scissors. Straws were analyzed for ash, cellulose, hemicellulose, and lignin (Goering and van Soest, 1970). Total N and C were measured using an automated elemental analyzer (Carlo Erba, Milan, Italy).

Straw was either applied evenly on the soil surface or incorporated uniformly into the soil at a rate of 38.2 g per box, equivalent to 2800 kg straw ha-1, which is about the average straw yield in semiarid areas of Saskatchewan (Campbell and Zentner, 1997). After straw application, the boxes were transferred to a growth chamber that was maintained at 20°C and 60% relative humidity throughout the experiment. The next day, distilled water was applied to all boxes to increase soil moisture content to 220 g H2O kg-1 (90% of field capacity, determined as described by Peters [1965]). To minimize damage to soil aggregates because of droplet impact, watering was always done by misting with distilled water from a hand-operated spray can. The soils were either watered every 2 to 3 d to compensate for evaporative losses and to keep soil moisture content to >85% of field capacity (continuously moist treatment) or watered only once when soil moisture content declined to 7% (moist-dry treatment). The growth chamber was in darkness except during watering and measuring.

An aluminum collar (15-cm i.d.) was inserted to a depth of 4 cm in each soil box to facilitate measurement of CO2 emissions by the chamber method. The collar remained in place throughout the experiment. Emissions of CO2 were measured every 2 or 3 d during the 77 d experiment using a closed-path infrared analyzer (LI-COR Model LI-6000, LI-COR, Lincoln, NE) with a flow rate of 1.2 L min-1 in a vented chamber. Measurements of soil respiration using similar equipment have been reported by Rochette et al. (1992 a,b). The chamber, which was attached to the collar using quick-fit clamps (a rubber gasket was placed between the collar and the chamber to provide an air-tight seal), enclosed an area of 177 cm2 and had a volume of 2 L. Concentrations of CO2 in the chamber were recorded every second, and fluxes (in units of µmol CO2 m-2 s-1) were calculated from the rate of increase in CO2 concentration during a 200-s deployment period. Accumulation of CO2 in the chamber was linear over 200 s. Daily output of CO2-C (g C m-2 d-1) was estimated assuming the respiration rate was constant during the 24 h period. Total CO2 production during the experiment was calculated by linear interpolation of the mean fluxes and integration over time. Soil mineral N at commencement and at end of experiment was determined using the method described by Hamm et al. (1970).

The CENTURY Model and Modification
CENTURY (Parton et al., 1987) is a computer simulation model of plant-soil ecosystems that simulates the dynamics of grasslands, forest, and crops. The grassland, crop, and forest systems have different growth submodels which are linked to a SOM submodel that includes three SOM pools (active, slow, and passive), above and belowground plant residue pools, and a surface microbial pool. The active pool represents microbes and microbial products and it has a turnover time of 1 to 5 yr. The slow pool includes resistant plant material and soil-stabilized microbial products and it has a turnover time of 20 to 50 yr. The passive pool includes physically and chemically stabilized SOM with a turnover time of 400 to 2000 yr. Above and belowground plant residues are partitioned into structural and metabolic pools as a function of the lignin/N ratio in the residue. Turnover times for structural and metabolic pools are 1 to 5 yr and 0.1 to 1 yr, respectively. Decomposition of each pool is assumed to be microbially mediated with an associated microbial respiration. Each pool has a potential decomposition rate which is reduced by multiplicative functions of soil moisture and soil temperature. Each SOM pool has an allowable range for the ratio of C to other elements. Flows of N and P between SOM pools are related to the C flows. The quantity of each element flowing out of a particular pool equals the product of the C flow and the C/element ratio of the pool. Mineralization or immobilization of N and P occurs as is necessary to maintain the ratios. The decomposition rate is reduced if the quantity of any element is insufficient to meet the immobilization demand. Detailed descriptions of the model are given by Parton et al. (1987), Paustian et al. (1992), and Metherell et al. (1993).

In the CENTURY model, potential decomposition rates of the surface plant residue pools and surface microbial active pool are arbitrarily set at 20% lower than those of in-soil pools, because it is assumed that moisture conditions are less optimal on the surface than in the soil (Parton et al., 1987). However, it is recognized that, in addition to the moisture factor, the availability of N also influences the decomposition rate (Parr and Papendick, 1978; Henriksen and Breland, 1999). It is relatively easy to simulate the N availability for SOM pools because mineral N in the same soil layer should be readily available for use in decomposition. The main N source for decomposing surface C pools is considered to be soil N obtained by fungal translocation (Hendrix et al., 1986; Beare et al., 1992). Recently, Frey et al. (2000) found that the annual N flux by filamentous fungi from the soil to surface-applied wheat straw in a NT field in Colorado was 2.4 g m-2. They also showed that the N flux mediated by fungi was constant over the period of their test (185 d after the placement of straw). The N flux would be expected to increase under conditions of high soil N availability (Holland and Coleman, 1987) and low initial residue N concentration (Frey et al., 2000). This information provided an opportunity for us to modify the CENTURY model to improve the simulation of soil CO2 emissions, especially where crop residues are placed on the soil surface, as under NT management.

In addition to soil moisture and soil temperature, we therefore included N availability as a multiplicative factor to reduce potential decomposition rates. We assumed that mineral N present in the soil layer to which residues are incorporated is readily available for use in decomposition, while only the limited amount of N translocated from the soil by fungi will be available for decomposition of surface residues. Further, it was assumed that the N flux rate from soil to the surface residue is constant. Available N was estimated on a daily basis for the soil and soil surface separately. When the available N is sufficient for decomposition of soil or surface pools, the N availability factor is 1, otherwise decomposition rates of all pools of soil surface and belowground organic matter are reduced proportionately until N supply meets demand. As the initial concentrations of soil mineral N and straw N in the study by Frey et al. (2000) were close to those in our experiment (Curtin et al., 1998), we assumed a daily rate of potential N supply from soil to surface straw via fungal translocation of 6.6 mg N m-2 (i.e., 2.4 g m-2 yr-1).

The source code for CENTURY (version 4) used in this analysis was downloaded directly from the website (http://www.nrel.colostate.edu/projects/century/) (this version is the updated version and may differ slightly from the program originally used). We used only the SOM submodel for simulations of soil CO2 flux during plant residue decomposition under the controlled conditions. The time step was changed from a weekly to daily basis.

The proportions of SOM in the active, slow, and passive pools were assumed to be 3, 45 and 52%, respectively. Monreal et al. (1997) used this pool-size allocation as the initial (1967) distribution of organic matter when simulating long-term crop rotation effects on SOM in soils of the Canadian prairies. The simulated changes in soil organic C and N were close to the observed values and they were superior to those achieved using CENTURY in Sweden (Parton et al., 1982) and on the Great Plains (Cole et al., 1989). Effects of soil moisture on decomposition were calculated (using relative water content data) using the equation provided by CENTURY. The temperature effect was calculated by an equation obtained from the CENTURY Tutorial (Parton et al., 2001).

The performances of CENTURY and the modified model were evaluated and compared using the following statistics. The association between simulated daily CO2-C fluxes and measured values was assessed by Pearson's correlation (r). Systemic errors were determined by the sum of squares attributable to lack of fit (LOFIT) (Whitmore, 1991). This method enables the experimental errors to be distinguished from the failure of the model. The statistical significance of LOFIT was obtained by comparing the ratio between the mean square due to lack of fit and the mean square because of the random error to tabulated F values with the appropriate degrees of freedom (Smith et al., 1996). The mean difference between measurement and simulation (M) was calculated to evaluate consistent errors (Addiscott and Whitmore, 1987). Pearson's correlation was calculated between measured and simulated values of total CO2-C evolution and soil mineral N at the end of the experiment.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
Total Carbon Dioxide Emissions
 CONCLUSION
 REFERENCES
 
Daily Carbon Dioxide Fluxes
The soil had a mineral N content of 18.4 mg N kg-1 soil at the beginning of the experiment. According to CENTURY, decomposition of incorporated straw should not be limited by N availability. Consequently, simulated values of daily CO2 flux for no-straw controls and treatments with incorporated straw were the same using the modified model as when the original CENTURY model was used.

In the no-straw control under continuously moist conditions (Fig. 1a) , small fluctuations of daily CO2 evolution occurred throughout the experiment. This appeared to be due mainly to day-to-day drift of the equipment (LI-COR LI-6000 analyzer) as the measurement accuracy of the LI-6000 system is about ±10%. Towards the end of the experiment, soil respiration increased slightly, possibly because of growth of moulds on the soil surface as a result of the high moisture and humidity. This was also observed for other continuously moist treatments (Fig. 1b, c, d, and e). For the reasons given above, the correlation between measured and simulated values of daily CO2 flux was poor in the no-straw (continuously moist) treatment and both systemic errors (LOFIT) and consistent errors (M) were high (Table 1). However, the simulation line was within, or very close to, the standard error of most of the measured data indicating that the simulated values were reasonably acceptable.



View larger version (35K):
[in this window]
[in a new window]
  		Fig. 1. Measured and predicted (...CENTURY, —modified model) daily CO2-C fluxes. Bars are standard errors of measurements.

 

View this table:
[in this window]
[in a new window]
  		Table 1. Assessments for simulations on daily CO2-C fluxes by Pearson's correlation (r), the lack of fit (LOFIT) and the mean difference between measurement and simulation (M).

 
The two straw types (fresh and weathered straw) had similar composition [i.e., lignin, cellulose and hemicellulose contents, and C/N ratios (Curtin et al., 1998)]; as a result, CO2 flux from fresh straw treatments was very similar to that of corresponding weathered straw treatments (Fig. 1). For straws incorporated into continuously moist soil, there were good correlations between simulated and measured CO2 fluxes (Table 1), indicating that the pattern of CO2 production was generally matched by the CENTURY simulation. However, the model tended to overestimate the CO2 flux 1 d after the first water application to the incorporated straw treatments (Fig. 1b and c). Paul et al. (1999) attributed an overprediction of CO2 evolution by the SALUS model to the fact that the model does not consider the lag period prior to residues being colonized and comminuted. In our study, predicted CO2 fluxes were in good agreement with measured values by Day 4 after the first watering, suggesting that colonization occurred rapidly compared with the field situation. This may be because the straw used in our experiment was cut into small (2.5 cm) pieces and it was mixed well into the soil. The overestimation on Day 1 after the first watering and the increases of CO2 flux before the end of the experiment were main reasons for the significant lack of fit between measured and simulated values. However, the t tests for M indicate that there was no significant bias in the simulations for either straw type.

Under moist-dry conditions, CENTURY successfully simulated the soil respiration pattern of the no-straw control and treatments with incorporated straws (Fig. 1f, g, and h), as shown by highly significant r values (Table 1). As in the continuously moist environment, the model overestimated the CO2 flux on Day 1 after the initial watering. Large flushes of CO2 production were observed immediately after the rewatering, which may be attributed to fast mineralization of labile organic substances (Dao, 1998) and release of trapped CO2 by displacement with water. CENTURY does not take these process into account and it underestimated respiration on the first day (incorporated straw treatments) or two (no-straw control) after rewatering. The F tests for LOFIT indicate that the error in the simulated values was significantly greater than the error inherent in the measured values for all three treatments. This was mainly because of the overestimation of respiration after the first watering and underestimation after the second watering. However, as shown by M values (Table 1), the consistent error was only significant for the no-straw control.

In general, CENTURY overestimated CO2 fluxes from treatments with surface-placed straw (Fig. 1d, e, i, and j). The modified model greatly improved the simulation in these cases. According to calculations using the modified model, decomposition rates of C pools on the soil surface were only ~15 to 30% (continuously moist soil) and 13 to 20% (moist-dry conditions) of those of the corresponding in-soil pools. These results suggest that N availability may be a serious limitation to the decomposition of surface-placed straw. All statistics used to evaluate the simulations, except the r for treatments under the continuously moist conditions (Table 1), were better for the modified vs. the original CENTURY model. Respiration rates simulated by the modified model were relatively constant during the period of the experiment (0.4–0.5 g CO2-C m-2 d-1). In contrast, the original model predicted a significant decline with time in the CO2 flux. Measured values showed little, if any, trend with time.


   Total Carbon Dioxide Emissions
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 Total Carbon Dioxide Emissions
CONCLUSION
 REFERENCES
 
CENTURY satisfactorily simulated total CO2-C production during the experimental period for the no-straw controls and the incorporated straw treatments under the two water regimes (Fig. 2) . Simulated values of total respiration in the incorporated straw treatments were within, or very close to, the standard error of the measured values. The model slightly underestimated the total respiration for the two controls. CENTURY overestimated the total respiration for all treatments with straw applied on the soil surface, while the modified model greatly improved the simulation for these treatments. Overall, the correlation between measured and simulated total soil respiration was improved by modification of CENTURY.



View larger version (17K):
[in this window]
[in a new window]
  		Fig. 2. Measured and predicted total CO2-C emissions (•,{circ} no straw controls, {blacktriangleup},{triangleup} fresh straw incorporated, {blacktriangledown},{triangledown} weathered straw incorporated, {blacksquare},{square} fresh straw on the surface, {diamondsuit},{diamond} weathered straw on the surface under continuously moist (closed symbols) and moist-dry (open symbols) conditions. Bars are standard errors of measurements.

 
Soil Mineral Nitrogen
CENTURY underestimated soil mineral N remaining at the end of the experiment in all treatments and, especially, in the treatments with straw applied on the soil surface under the moist-dry conditions (Fig. 3) . However, the linear relationship between measured and simulated values (P < 0.001) suggests that the structure of CENTURY is generally sound, though it could be further improved. The modified model improved simulation for treatments with surface-applied straw and therefore improved the relationship for mineral N (Fig. 3).



View larger version (19K):
[in this window]
[in a new window]
  		Fig. 3. Measured and predicted soil mineral N at the end of the experiment (•,{circ} no straw controls, {blacktriangleup},{triangleup} fresh straw incorporated, {blacktriangledown},{triangledown} weathered straw incorporated, {blacksquare},{square} fresh straw on the surface, {diamondsuit},{diamond} weathered straw on the surface under continuously moist (closed symbols) and moist-dry (open symbols) conditions. Bars are standard errors of measurements.

 

   CONCLUSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 Total Carbon Dioxide Emissions
 CONCLUSION
REFERENCES
 
The CENTURY model appeared to be structurally sound in predicting the trend of soil respiration under different straw managements and soil moisture conditions in a controlled chamber experiment. Modifications of the model were made to include an N availability factor and this resulted in improved simulations of soil respiration when plant residue was applied on the soil surface. In this study, decomposition of incorporated straw was not limited by N availability. Further studies to quantify the interrelationships between N availability, surface residue decomposition, and environmental factors are needed to improve models of soil C cycling, particularly under no-till conditions.


   ACKNOWLEDGMENTS
 
We thank Dr. Con Campbell for his advice and constructive comments during the study. Technical assistance in computer programming by E. Chan and J.Y. Lou is greatly appreciated.

Received for publication April 2, 2001.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 Total Carbon Dioxide Emissions
 CONCLUSION
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (3)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Wang, H.
Right arrow Articles by Zhou, H. F.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Wang, H.
Right arrow Articles by Zhou, H. F.
Agricola
Right arrow Articles by Wang, H.
Right arrow Articles by Zhou, H. F.
Related Collections
Right arrow Other Soil Management
Right arrow Soil Models
Right arrow Nutrient Management

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome