Previous PageTable Of ContentsNext Page

MODELLING CHANGES IN SOIL MICROBIAL BIOMASS IN RESPONSE TO
ADDED CROP RESIDUES

M.E. Probert and B.A. Keating

CSIRO Division of Tropical Crops and Pastures, 306 Carmody Road, St Lucia, Qld 4067

Summary. Soil microbial biomass (SMB) was measured through the growth of two wheat crops following application of three residue treatments (none and lucerne hay or barley straw incorporated into soil). The residue treatments affected SMB in the 0-15 cm soil layer, but not in deeper soil layers; no effects of fertiliser-N application on SMB were found. Inadequacies of current models to simulate changes in SMB, in response to added crop residues and through time, are discussed. The algorithm in CERES models that determines the effect of C:N ratio on the decomposition of residues is shown to give poor prediction of changes in SMB with respect to the type of residue added and the effects of fertiliser-N.

INTRODUCTION

The benefit of legume leys in terms of the additional nitrogen contributed to subsequent nitrophillous crops is invariably greater than the relatively small effects on soil organic matter or total soil nitrogen. This has major implications when attempts are made to model such systems. It is obvious that any model which endeavours to describe the dynamics of soil nitrogen by treating the soil organic matter as a single pool must be inadequate for capturing the effects of a ley on the nitrogen supply to following crops. Models need to reflect the ‘quality’ of the soil organic matter in as much as some parts of it are more susceptible to decomposition and nitrogen is mineralised more rapidly. The dilemma for modellers is how this might be achieved, especially against the background that there is no real consensus on how the effects can be quantified experimentally. Soil microbial biomass (SMB), light-fraction carbon, particulate carbon and more readily oxidizable carbon have been investigated as measures of the more labile carbon.

In the SoilN module of APSIM (5), the soil organic matter is treated as two pools which are referred to as BIOM (being the more active component) and HUM (comprising the remainder of the soil organic matter). The addition of this one extra pool, which is similar to the representation used in other models (e.g. (2)), greatly increases the number of parameters needed to define the flows of carbon and nitrogen in the system. Selection of appropriate parameters and validation of model performance requires datasets where pertinent soil properties have been measured through time. In the experiment reported here, we determined how SMB responded to inputs of crop residues and changed through time to provide a basis for testing the adequacy of APSIM for modelling such effects.

MATERIALS AND METHODS

The experiment was carried out at Gatton, Qld on an alluvial, clay soil that had been cropped with forage sorghum in the previous season to create a nitrogen responsive situation. Treatments imposed were three residue treatments (none, and either lucerne hay or barley straw incorporated into the surface soil with a rotary hoe on 3 June 1993) in factorial combination with four rates of application of nitrogen fertiliser (0, 40, 80 and 200 kg N/ha as ammonium nitrate). Rates of application of residues were: lucerne 3560 kg/ha of dry matter containing 109 kg N; straw 3584 kg/ha, 29 kg N. There were three replicates. Wheat (cv Hartog) was sown on 29 June 1993 and grown with supplementary irrigation. After harvest, wheat residues were left in situ and a second wheat crop sown on 27 May 1994 using zero tillage practices.

Soils were sampled on four occasions during the first crop and twice (before planting and after harvest) in the second season. At all samplings, subsamples were obtained from the 0-15 and 15-30 cm depths. Deeper layers were sampled from the no residue, no fertiliser treatment at the first and final samplings.

Analytical

SMB was determined as ninhydrin-positive compounds (NPC) released and extracted after 10 days fumigation with CHCl3 (1), with a factor of 3.5 to convert NPC to microbial-N (6). This methodology was chosen because, unlike the fumigation-incubation methods for soil biomass determination, it is applicable in the presence of fresh plant residues. The irrigation schedule ensured that the soils were moist at the time of sampling, so no pre-incubations were necessary and fumigations were imposed on the field fresh samples. Other determinations at the site provide estimates of total soil carbon, the C:N ratio of the soil (14.5), and soil bulk density to calculate SMB-N on a kg/ha basis.

Modelling

Simulations were carried out using APSIM v1.

RESULTS AND DISCUSSION

Measured soil microbial biomass

In the surface 0-15 cm layer, SMB-N increased in response to the added residues with initial response being greater to the lucerne than to the barley straw (Fig. 1), but no significant effects were found from application of fertiliser-N. The increase after adding lucerne happened rapidly with the greatest effect being observed by the time of the first sampling (19 days after the residues were added). In the absence of added residues, there was little indication that SMB-N varied during the growth of the first wheat crop or through the subsequent fallow and the next crop. No significant effects of the treatments were found in the 15-30 cm soil layer at any sampling. SMB-N decreased with soil depth, though the proportion of total soil N measured as biomass-N was rather constant in the top 60 cm of soil and only declined in the 60-90 cm layer (Table 1).

Table 1. Soil microbial biomass-N (mg/kg) and as proportion of total soil N in different soil layers for the no residue, no fertiliser treatment.

Depth (cm)

Soil biomass-N

 

(mg/kg)

(% of total N)

0-15

30.0

2.0

15-30

18.4

1.7

30-60

16.2

2.0

60-90

8.4

1.2

Modelling the soil biomass pool

The concepts built into the APSIM SoilN module involve three pools of carbon in soil, viz. the fresh organic matter (FOM), a labile pool (BIOM) which is notionally the microbial biomass, and the remainder of the soil organic matter (HUM). When each of these pools decompose, there is a loss of CO2 to the atmosphere and a transfer of carbon and nitrogen to the BIOM and HUM pools, specified in terms of the proportion of carbon that is retained in the system and the proportion of the retained carbon that goes to the BIOM pool. Corresponding nitrogen transfers are calculated based on the C:N ratios of the receiving pools (C:N of the BIOM pool is set at 8; C:N of HUM pool is calculated at initialisation from the soil C:N ratio). Any immobilisation demand (e.g. when high C:N material is added to soil and undergoes decomposition) is met from the soil mineral N. Much of the code for the SoilN module has been derived from the CERES family of models, and the description of FOM in SoilN v1 is identical to CERES-Maize (4). Briefly FOM is considered to comprise three fractions (carbohydrate-like, cellulose-like and lignin-like) that have decreasing rates of decomposition. The relative amounts of each fraction (0.2:0.7:0.1) are assumed to be always the same for new material added to soil. Decomposition of all soil organic pools depends on temperature and moisture factors, but that of FOM also depends on a zero-to-unity C:N ratio factor, CNRF, which in CERES is calculated using the nitrogen available to support the decomposition process and assumed to be the nitrogen in the decomposing FOM plus the mineral-N in the soil layer. A modified C:N ratio, CNR, is calculated as

CNR = FOM-C / (FOM-N + MIN-N)

from which

CNRF = exp {-0.693 * (CNR - 25) / 25}

where FOM-C and FOM-N are the carbon and nitrogen in the FOM pool and MIN-N is the sum of ammonium- and nitrate-N in any soil layer. Thus decomposition of FOM is reduced at values of CNR > 25. On this basis, the model predicts that where sufficient fertiliser-N is applied, thereby increasing MIN-N, added residues would decompose at the same rate irrespective of their C:N ratios. The experimental data do not support this. Using the increase in SMB-N as an indicator of the decomposition process, barley straw decomposed more slowly than lucerne, and no significant effect of fertiliser application was found on the change in SMB-N following application of either residue.

To conform with the absence of any fertiliser effect in the observed data, it could be assumed that the rate of decomposition of FOM depends simply on its C:N ratio (CN = FOM-C / FOM-N). Using

CNRF = exp {-0.015 * (CN - 15)}

simulated changes in BIOM-N in the top soil layer during the first wheat crop are as shown in Fig. 2. Effects of the treatments or of time on predicted BIOM-N in the 15-30 cm layer were very small and are not presented. The assumption causes a difference in the predicted BIOM-N between the lucerne and straw treatments, but comparison with the observed data (Fig. 1) indicates that the model is decomposing the straw too rapidly, increases BIOM-N to a greater extent than the experimentally measured SMB-N, and is not able to reproduce the time trend in the data for the lucerne treatment. Other aspects of model performance, such as soil mineral-N, crop N uptake and grain yield (data not shown), were consistent with an overprediction of decomposition the straw, resulting in immobilisation of N and underprediction of crop N uptake by the straw treatment when no fertiliser-N was applied.

CONCLUSIONS

Experimentally it was found that soil microbial biomass responded to the nature of the crop residues incorporated into the soil. Modelling of the changes in soil microbial biomass depends on accurate predictions of the decomposition of crop residues (roots and incorporated material). The results from the simulations suggest that the algorithms in CERES, and which have been incorporated into the APSIM SoilN module, do not adequately represent the decomposition of different crop residues of varying C:N ratio. Decomposition rates of FOM may depend on factors other than C:N ratio; for example, lignin and polyphenol contents have been shown to be important (3).

REFERENCES

1. Amato, M. and Ladd, J.N. 1988. Soil Biol. Biochem. 20, 107-114.

2. Bradbury, N.J., Whitmore, A.P., Hart, P.B.S. and Jenkinson, D.S. 1993. J. Agric. Sci. 121, 363-379.

3. Fox, R.H., Myers, R.J.K. and Vallis, I. 1990. Plant Soil 129, 251-259.

4. Jones, C.A. and Kiniry, J.R. (Eds) 1986. CERES-Maize: a simulation model of maize growth and development. (Texas A&M Univ. Press: College Station).

5. McCown, R.L., Hammer, G.L., Hargreaves, J.N.G., Holzworth, D.P. and Freebairn, D.M. 1995. Agric. Systems 48, (in press).

6. Sparling, G. and Zhu, C. 1993. Soil Biol. Biochem. 25,1793-1801.

Previous PageTop Of PageNext Page

...

Search Site