Previous PageTable Of ContentsNext Page

EM38 and crop-soil simulation modelling can identify differences in potential crop performance on typical soil zones in the Mallee.

Anthony Whitbread1, Rick Llewellyn, David Gobbett and Bill Davoren.

1CSIRO Sustainable Ecosystems, PMB2 Glen Osmond 5064, South Australia. Email anthony.whitbread@csiro.au

Abstract

In the Southern Australian Mallee, performance within paddocks shows enormous spatial variation. Since these differences are largely related to elevation and soil variation, the use of EM38 surveys have been shown to be useful in sensibly zoning cropping paddocks into areas of like yield potential. This paper outlines the use of soil characterisation, and crop-soil modelling to predict the long term crop performance or yield potential of paddock zones with low, moderate or high subsoil constraints at four sites in the Mallee. Long term simulations of wheat production at two of these sites (Carwarp and Loxton) showed that the median wheat yield on most unconstrained zones was 60 to 70% higher than that of the constrained zones (i.e. 1.5-1.7 t/ha of grain yield on an unconstrained sandy loam compared with 0.9-1 t/ha on a constrained heavy clay loam). At Carwarp however, there were six seasons (none at Loxton) from the period 1956 to 2007 in which the most constrained zones out-yielded the other zones, and this occurred in seasons with wet finishes. At the other two sites, Bimbie and Pinnaroo, crop performance was not simulated to be very different in the EM38 defined zones. At Pinnaroo this was due to smaller relative differences in the rooting depth and crop lower limit (related to soil texture) between the three zones than the other sites. At Bimbie, large observed yield differences do not support this result and may indicate that the parameterisation of the APSIM water balance model was not adequate to simulate the differential effects that soil texture may have on the infiltration and redistribution of water as well as the influence of texture (or osmotic potential) on soil water uptake.

Key Words

Spatial soil variation, temporal yield variation, zones, crop modelling, APSIM.

Introduction

Farmers have long been aware that crop performance within paddocks shows enormous spatial variation, especially in the cropping regions of the Southern Australian Mallee. These differences in yield are driven predominantly by soil variation and are often as great as season to season variability (Sadras et al. 2002). With the advent of readily available tools to detect soil spatial variability (e.g. Electromagnetic Induction: Jones et al. 2008; Llewellyn et al. 2008) and equipment to manage and monitor its effects (e.g. variable rate fertiliser spreaders and harvester yield monitors), farmers are now in a position to manage and possibly take advantage of variation. Because relative yield differences between zones delineated on the basis of subsoil constraints (inferred from electromagnetic induction) are not constant, a combination of field results and modelling may assist to assess the likely longer-term economics of zone management. This paper outlines an approach where the soil within the zones were characterised for their water holding capacity and chemical sub-soil constraints and the Agricultural Productions Systems sIMulator (APSIM) (Keating et al. 2003) was used to simulate the long term yield performance on the zones.

Methods

Four sites in the Victorian and SA Mallee were selected from Llewellyn et al. (2008) where yield mapping information was available for cereal crops in 2006 and 2007. These were Bimbie, 80 km SE of Mildura (34°27’ S, 142°58’ E), Carwarp (34°27’ S, 142°12’ E), 30 km S of Mildura and Pinnaroo (35°20’ S, 140°54’ E) 170km S of Mildura and Loxton (34O 29’ S, 140O 34’ E), 150 km W of Mildura. From each site, nine soil cores were taken in 2006 on a transect across the paddock, divided into increments of 0-20, 20-40, 40-60, 60-80, 80-110 cm. Sub-samples of the soils were analysed for pH, electrical conductivity (EC) and chloride, measured in a 1:5 soil:water extract. Boron was determined using 0.01M CaCl2 extracting solution and immersion in a 98oC water bath (Rayment and Higginson 1992). Soil sand, silt and clay fractions were determined after sieving to remove gravel. An EM38 survey was undertaken in April 2005 and used to create 3 EM-based soil classifications for each paddock (Table 1). The soil cores were assigned to the soil classifications in which they were located and the results presented are averages of the cores falling into these zones. Plant available water capacity (PAWC) of each zone was characterises from the drained upper limit (DUL), crop lower limit (CLL) and rooting depth. DUL was determined at a point within each zone using the pond technique of Dalgliesh and Foale (1998). CLL was determined for each zone using the soil moisture measured at the harvest of wheat crops in 2006 (nine cores across the three soil classes) and in 2007 (27 cores across the three soil classes). CLL was taken as the lowest soil moisture value measured in the 2 seasons. For each paddock and within each zone, the rooting was assumed to cease at the soil layer where the concentration of B exceeded 19 mg/kg (Nuttall et al. 2003), Cl exceeded 1000 mg/kg and/or EC exceeded 1 dS/m (regarded as very high to extreme soil salinity rating for clay contents 10-40 % by Shaw 1999 Table 8.5 p 136), or at the limit sampling depth which was usually to a physical barrier such as rock.

Simulation

Wheat (cv. Yitpi) growth in each year for the period 1957 to 2006 was simulated within in each zone at each location with APSIM version 5.3. Climate data were sourced from a SILO sourced weather station (Jeffrey et al. 2001) located closest to each site. The simulations were reset each year so that starting soil N and organic matter remain the same in all years. Starting soil mineral N was assumed to be the same (52 kg N/ha to 110cm) for each soil class in the paddock. Wheat was sown between April 25 and July 15 on the first date when there had been 10 mm rain over the preceding 5 days and the soil profile contained at least 10 mm of available soil water. A sowing application of 30 kg of N was included each year.

Results

Soil profiles for each zone

The use of EM38 to differentiate soil boundaries based on the sensing of subsoil characteristics such as clay content and salt concentration has been shown by Llewellyn et al. (these proceedings) as an effective tool for zoning Mallee paddocks into management zones of similar yield potential. Average zonal cereal yields measured in the dry 2006 and 2007 seasons do not vary much between the zones with the exception of Carwarp (Table 1). The estimated depth of rooting was consistently deepest in the low subsoil constraints zone (Table 1). However, PAWC wasn’t necessarily highest in these zones because the finer textured soils most common in the moderate and highly constrained zones hold more moisture between DUL and CLL. At each location, available P (0-10cm) was similar across the 3 zones and ranged from 22 to 41 mg/kg across the locations. Total organic carbon (TOC) in the topsoil (0-10cm) were consistently lowest (<0.9 %) in the low subsoil constraints (sandier) zones and between 1.14 and 1.30 % in the high zones, reflecting the influence that clay protection can have on soil organic matter.

Table 1. The soil textural and chemical characteristics of soil profiles in the 3 zones defined by the EM38 survey and the plant available water capacity (PAWC) to the depth of constraint, plant available P (0-10 cm) and total organic carbon (TOC) (0-10cm).

 

Zone/soil constraint

Estimated rooting depth

aB

Cl

EC
1:5

1ECa

PAWC

Mean zone yield
2006/07

2TOC

   

(cm)

(mg/kg)

(mg/kg)

dS/m

dS/m

mm

t/ha

%

Bimbie

Low

110

-

-

-

0.3-1.1

101

0.9/1.3

0.84

 

Moderate

80

18-20

-

-

1.1-1.5

87

0.8/1.5

0.97

 

High

50

18-20

1400-1600

1.5

1.5-2.2

91

0.7/1.5

1.30

Carwarp

Low

110

-

-

-

0.0-0.6

107

0.6/1.2

0.74

 

Moderate

110

-

-

-

0.6-0.9

91

0.2/1.1

0.82

 

High

60

-

1050

1.1

0.9-1.6

65

0/0.8

1.14

Loxton

Low

110

-

-

-

0.1-0.4

66

0.9/0.5

0.71

 

Moderate

110

-

-

-

0.4-0.7

58

0.9/0.9

0.97

 

High

50

12-19

1000

1.0

0.7-1.7

80

0.5/0.5

1.14

Pinnaroo

Low

80

-

-

-

0.4-1.0

61

0.9/1.3

0.86

 

Moderate

60

20

-

-

1.0-1.4

64

0.8/1.5

1.15

 

High

50

23-30

500

0.9

1.4-2.6

56

0.7/1.5

1.21

1ECa = apparent electrical conductivity. 2TOC based on a Walkley-Black extraction multiplied by 1.3. aNote, Soil pH >8.5 in all cases

For the zones containing low sub-soil constraints, low values of CLL indicate that roots were able to extract water from the soil profile (data shown for Carwarp, Figure 1a-c and Loxton, Figure 1d-f). The shape of the PAWC diagram, influences the dynamics of the water balance with the shape of a PAWC diagram for a sandy, unconstrained soil being narrow and deep (Figure 1d), whilst a soil higher in clay, but with subsoil constraints will be wider at the top and potentially holding more soil moisture near the surface (Figure 1c, 1f).

Figure 1. The characterisation of the crop lower limit (CLL) and the drained upper limit (DUL) for Carwarp containing low (Fig. 1a), moderate (Fig. 1b) and severe (Fig. 1c) subsoil chemical constraints and Loxton containing low (Fig. 1d), moderate (Fig. 1e) and severe (Fig. 1f) subsoil chemical constraints.

Table 2. Simulated wheat grain yield (with a sowing N application of 30 kg/ha) within zones surveyed as containing low, moderate and high subsoil constraints for all sites over the period 1957 to 2007.

     

Probability of yield

 

Constraints

Median
Yield (t/ha)

< 1 t

1 - 1.5 t

> 1.5 t

Bimbie

Low

1.48

0.29

0.23

0.48

 

Moderate

1.60

0.31

0.17

0.52

 

High

1.46

0.33

0.19

0.48

Carwarp

Low

1.51

0.40

0.10

0.50

 

Moderate

1.25

0.44

0.13

0.42

 

High

0.94

0.52

0.13

0.35

Loxton

Low

1.69

0.23

0.21

0.56

 

Moderate

1.29

0.35

0.33

0.33

 

High

1.01

0.50

0.27

0.23

Pinnaroo

Low

1.63

0.27

0.15

0.58

 

Moderate

1.65

0.29

0.13

0.58

 

High

1.59

0.31

0.15

0.54

The simulation of wheat growth over the long term.

Crop modelling at Carwarp and Loxton showed large differences in median yield between zones and a much greater probability of achieving grain yield in excess of 1.5 t/ha in the low subsoil constraint zone (Table 2). Interestingly at Loxton despite the high subsoil constraint zone having a higher PAWC of 80 mm compared to 66 mm in the low zone (Table 1) wheat performance was much poorer. This result reflects both the combination of potentially higher soil evaporation losses and less accessible mineral N in the shallow, but heavier textured ‘high’ zone soil. At Bimbie and Pinnaroo, the simulated wheat growth on the 3 zones resulted in similar yields (Table 2). This result was expected at Pinnaroo because the soil differences (rooting depth, PAWC and texture) between zones were small(note narrow ECa range and little difference in zone yields, Table 1). At Bimbie where the estimated rooting depth of the high zone was 50 cm, larger effects on the growth of wheat, were expected despite similar PAWC’s. While these expected differences were not measured in 2006 and 2007 (Table 1) the farmer confirmed that there were large yield variations across the paddock related to the sub-soil constraints. Not simulating differences in crop growth between different zones probably indicates that APSIM has not been parameterised correctly to simulate the differential effects that soil texture may have on the surface evaporation and the redistribution of water within the profile as well as the influence of texture (or osmotic potential) on soil water uptake. In the seasons with good rainfall, simulated yields in the high zone at Bimbie and Pinnaroo outperformed the low and moderate zones due to higher TOC (Table 1) resulting in higher N mineralisation.

Discussion

At the four sites studied in this paper, the three zones identified using apparent EC readings from the EM38 survey reflected soil classes that contain low, moderate and high sub-soil constraints. At Carwarp and Loxton, the large differences in rooting depth, the shape of the PAWC diagram and the effect of texture on soil water dynamics (i.e. the effect of texture on the storage of water in the profile) resulted in large and consistent differences in simulated yield. At Pinnaroo, these soil differences were not large enough to result in differences in yield potential between zones, however at Bimbie, despite large soil differences simulated yield potential was similar. This result is suggested as being due to the APSIM water balance model not being parameterised to simulate plant water uptake in soils of high salt levels. Differences in the potential grain yield of wheat growing in management zones delineated with EM38 soil mapping has major implications for the use of site specific management of inputs, leading to reduced inputs in zones with high subsoil constraints and visa versa in zones with few subsoil constraints.

Acknowledgements

This work is part of a Mallee Sustainable Farming Inc project and Training Growers to Manage Soil Water Project (GRDC Research Code CSA00011) supported by the Grains Research and Development Corporation. The support of the participating farmers at each site is gratefully acknowledged.

References

Jeffrey SJ Carter JO Moodie KB Beswick AR (2001). Using spatial interpolation to construct a comprehensive archive of Australian climate data. Environ. Modelling Software. 16, 309–330. (www.bom.gov.au/silo/, accessed 30/04/2008)

Jones B Llewellyn R and O’Leary G (2008). Sodium, chloride, clay and conductivity: consistent relationships help to make EM surveys useful for fertiliser and crop choice decisions in the Mallee. Proceedings of the 14th Australian Agronomy Conference, Adelaide, Australian Society of Agronomy, In press.

Keating BA Carberry PS Hammer GL Probert ME Robertson MJ Holzworth D Huth NI Hargreaves JNG Meinke H Hochman Z McLean G Verberg K Snow V Dimes J Silburn M Wang E Brown S Bristow K Asseng S Chapman SC McCown RL Freebairn DM Smith CJ (2003) An overview of APSIM, a model designed for farming systems simulation. Eur. J. Agron, 18, 267-288.

Llewellyn R Jones B Whitbread A and Davoren B (2008) The role for EM mapping in precision agriculture in the Mallee Proceedings of the 14th Australian Agronomy Conference, Adelaide, Australian Society of Agronomy, In press

Nuttall JG Armstrong RD Connor DJ (2003) Evaluating physiochemical constraints of Calcarosols on wheat yeidl in the Victorian southern Mallee. Australian Journal of Agricultural Research 54: 487-497.

Rayment GE and Higginson FR (1992). Australian Laboratory Handbook of Soil and Water Chemical Methods’. Inkata Press, Melbourne.

Sadras VO Roget DK O'Leary GJ (2002) On-farm assessment of environmental and management constraints to wheat yield and rainfall use efficiency in the Mallee. Australian Journal of Agricultural Research 53:587-598.

Shaw RJ (1999). Soil salinity – electrical conductivity and chloride. In (Eds Pervill KI Sparrow LA and Reuter DJ) Soil Analysis an Interpretation Manual. (CSIRO publishing Australia) pp 129-145.

Previous PageTop Of PageNext Page