Previous PageTable Of ContentsNext Page

The development of a model in APSIM for the simulation of grazing oats and oaten hay

Allan Peake1, Anthony Whitbread2, Bill Davoren2, Jeff Braun3 and Sarah Limpus4

1CSIRO Sustainable Ecosystems/Agricultural Production Systems Research Unit (APSRU), P.O. Box 102, Toowoomba, QLD, 4350
2
CSIRO Sustainable Ecosystems, PMB2, Glen Osmond, SA, 5064
3
Agrilink Agricultural Consultants Pty Ltd, PO Box 118, Watervale, SA, 5452
4
The University of Queensland, Faculty of Natural Resources, Agriculture & Veterinary Science, Gatton, QLD, 4343

Abstract

A prototype oats model based on the wheat model within the APSIM simulation framework was developed and subsequently applied to simulate the effect of starting soil water content on oaten hay production in the Mid-North region of South Australia. Model parameters such as radiation use efficiency, canopy extinction coefficient, maximum rates of water extraction and rates of leaf death were derived from field experiments conducted in 2007 at sites in Queensland (QLD) and South Australia (SA). In general, simulation of the field experiments showed good agreement between the observed and simulated biomass of three different cultivars. In the simulation of the SA experiments, oaten hay yield was predicted satisfactorily, although initial biomass growth rates were under-predicted at Tarlee and over-predicted at Pinery. Long term simulations revealed substantial differences in production potential between Pinery and Tarlee, attributable to the 90mm higher annual rainfall received at Tarlee. Increased availability of soil water at sowing had a greater impact on biomass yield at Pinery, also attributable to the lower average rainfall at Pinery. Further testing of the oats model is required to generate greater confidence in its predictive abilities. Additional model development is required in order to dynamically predict biomass left in the paddock underneath the harvesting height, and how oaten hay quality parameters respond to agronomic management.

Introduction

Oats fodder is an important forage source for the beef and dairy cattle industries across Australia (Armstrong et al., 2004). In Western Australia, 650,000 tonnes of oaten hay were produced in the year 2005/2006, of which 85% was exported to international feed markets (Winfield et al., 2007). In Queensland and New South Wales approximately 500,000 hectares of oats are planted annually, which is predominantly used as an on farm standing forage or hay for gazing animals including sheep and in feedlots to supplement the diet of meat and dairy cattle (QDPI&F, 2008). The importance of oat crops to the red meat industry in QLD and NSW alone is estimated to be worth $97 million annually.

In Southern Australia, the online crop management tool Yield ProphetTM has been used increasingly by farmers to manage winter crops, and forecast grain yields (Hunt et al., 2006). Based on crop models within the Agricultural Production Systems Simulator (APSIM; Keating et al., 2003), Yield ProphetTM has not previously possessed an oats simulating capability. However, agronomists have inquired about the possibility of integrating an oaten hay simulation capability into Yield Prophet to assist in forecasting oaten hay production, and to explore how top-dressed nitrogen (N) is likely to affect the yield and quality of oats crops. This paper reports on the development of an oats and oaten hay simulation capacity within APSIM that will be available for demonstration purposes in Yield ProphetTM from 2008.

Materials and Methods

Two field experiments were conducted to obtain essential data for the development of a simulation model for grazing oats and oaten hay production within APSM. In the first experiment, physiological parameters were measured on QLD grazing oats varieties. In the second experiment, an oaten hay variety was monitored in the Mid-North agricultural zone of South Australia to assess the suitability of the oats model for predicting oaten hay yields. After testing the models ability to simulate oats with these data sets, a simulation experiment was conducted on oaten hay to determine the effect of starting soil water and sowing date on oaten hay yields in the Mid-North.

Field experiments

The first field experiment was conducted at Gatton (QLD) (770mm mean annual rainfall) using early (Coolibah) and late maturing (Taipan) oats varieties in a randomised complete block design with three replicates. Sown on 24th May 2007, the trial was initially fully irrigated and fertilised with 350 kg N/ha to ensure unlimited N supply. Biomass cuts were taken regularly through the early part of the trial. Samples were partitioned into green leaf, dead leaf, stem and floral parts, and the green leaf area index was determined on a subsample. Light interception was measured with three solarimeters placed in each variety, one in each replicate, and light extinction coefficient was then determined using the Beer-Lambert equation. Once enough data had been generated to determine the light extinction coefficient, irrigation ceased and maximum rates of crop water extraction were determined as the crop moved into a water supply limitation. In each plot, soil water content was assessed using a neutron moisture meter three times weekly.

Experiments were also conducted in the Mid North agricultural zone of South Australia using the oaten hay variety Wintaroo. At Tarlee (470 mm mean annual rainfall), the experiment was sown on 11/5/07 on a red-brown earth with PAWC of 245mm, and two treatments were imposed: ‘Best Practice Dryland’ (50 kg N/ha applied at sowing, no irrigation) and ‘High Input’ (100 kg N/ha applied at sowing, 100 mm of irrigation applied at anthesis). At Pinery (380mm mean annual rainfall), Wintaroo was sown on 5/6/07 on a red-brown earth with PAWC of 190mm, and grown under dryland conditions with 50 kg N/ha applied at sowing. Regular biomass cuts were taken at both sites and partitioned as described for the Gatton experiment. Soil water and nitrogen were measured at each of the field sites from soil cores taken at or near sowing. At each site, soil characterisations were carried out to measure drained upper limit, the lower limit of plant extractable soil water, bulk density and subsoil constraints.

Simulation experiments

Experimental results from Gatton, Pinery and Tarlee were compared with APSIM simulations of each experimental treatment. Measured soil water and nitrogen at sowing, soil characterisations and meteorological data from each field site were used to parameterise each simulation as appropriate. After testing the models ability to simulate the field experiments, long term simulation experiments were conducted using 100+ years of SILO weather data (Jeffrey et al. 2001) from the nearest representative localities (Tarlee and Balaklava). Long term simulations were conducted for oaten hay production at both Pinery and Tarlee, to determine the potential oaten hay production across the range of historical weather data in a monocropping situation with soil mineral N and surface organic matter reset each year at sowing, while soil moisture at sowing varied as simulated from the fallow conditions between crops. Long term simulations were then run at Pinery for three different levels of plant available water at sowing: 20, 60 and 100mm. In this simulation, soil water was reset to the same PAW each year at sowing as were surface organic matter and soil N. The long term simulations were run assuming a planting density of 200 plants/m2, with 110 kg/ha of nitrate-N available in the soil at sowing, and an additional 50 kg N/ha of fertiliser added at sowing. The simulation planting dates were 7th May at Pinery and 25th May at Tarlee. The harvest of oaten hay harvest date was simulated to occur on the last day of the flowering period as predicted by APSIM.

Results and Discussion

Field experiment simulations

Measurement of physiological parameters at Gatton allowed the APSIM oats model to be parameterised with a canopy extinction coefficient of 0.45, and maximum rate of water extraction of 0.15 mm/mm/day at the soil surface (results not shown). Radiation use efficiency of 1.28 g/MJ was measured in an oats experiment grown at Gatton in 2002 (Huth, unpublished data). The rate of leaf death in response to water stress was reduced to 0.01/day in order to accurately simulate dead leaf accumulation.

Using the derived parameterisation, good agreement was found between the simulated and observed oats biomass yield for all experimental sites (Fig. 1.) At Gatton, the model showed excellent prediction of biomass accumulation for both cultivars (data not shown for Taipan), although the experimental measurements ceased prematurely when the experiment was affected by lodging and a severe leaf rust infestation prior to anthesis.

At Tarlee, the model predicted anthesis biomass production well in both the High Input and Best Practice Dryland treatments (Fig. 1b & 1c), but tended to under-predict biomass production at the earlier growth stages. When simulating partitioned biomass at Tarlee, simulated leaf death was substantially slower than the observed rate of leaf death which was accelerated due to the effect of bacterial leaf blight Pseudomonas syringae (results not shown). At Pinery, the model over-predicted biomass accumulation throughout the experiment but was only 0.7t/ha higher than the observed biomass at anthesis (Fig. 1d). This is a good result given the severe water stress that the oats experienced at Pinery in 2007.

Long-term simulations

Substantial differences in production potential were simulated between Pinery and Tarlee (Fig. 2a) attributable to the 90mm higher annual rainfall received at Tarlee, although in high rainfall years Pinery was simulated to produce comparable anthesis above-ground biomass to Tarlee.

Median anthesis above-ground biomass yields of 9.1, 10.6 and 12.1 t/ha were simulated for Pinery (Fig. 2b) from simulations starting with 20, 60 and 100mm PAW at sowing, respectively. The effect of PAW at sowing was more pronounced at Pinery than Tarlee (results not shown), probably because of the lower annual rainfall experienced at Pinery. In the 20mm PAW simulations, 8t/ha or more of above-ground anthesis biomass was produced in 60% of years at Pinery (Fig. 2b), while the same biomass yield was achieved in 80% of years at Tarlee (results not shown). It is important to note that total above-ground biomass production at anthesis is likely to be 10-20% higher than actual oaten hay yields, given that oaten hay production leaves approximately 10cm of stem in the paddock after cutting.

Future work

While the comparisons between simulated and observed biomass yield at anthesis (Fig. 1) are encouraging, it must be remembered that this represents a very small data set for testing the model. Considerable further evaluation of the oats model must be carried out before it could be used with confidence to assist with managing and forecasting commercial oaten hay production. Future work will require the development of a dynamic biomass cutting function in APSIM that will allow the prediction of biomass left in the paddock after hay has been removed. Additionally, successful oaten hay production involves not only the production of maximum biomass, but also the management of several quality parameters that are affected by soil N supply and water content of the soil. Simulation of the response of these quality parameters to agronomic management would be highly beneficial to the oaten hay industry.

Conclusions

A prototype oats crop simulation model has been developed within the APSIM framework, and it is available for demonstration purposes for oaten hay producers in 2008. Widespread testing of its ability to accurately predict biomass production needs to occur to generate greater confidence in the predictive ability of this simulation model. Further development is required to develop a dynamic cutting function which accurately predicts biomass left on the paddock after hay has been removed, and to simulate the oaten hay quality parameters that are important to the industry.

Acknowledgements

We would like to acknowledge the assistance of the oats breeding team at SARDI for planting and managing the Pinery Site. We would also like to thank Tod Eadie and Greg Roberts from The University of QLD and CSIRO Plant Industry, for their assistance in managing the Gatton trial. We also thank Dr Gurjeet Gill at the University of Adelaide for the use of a laboratory and leaf area meter. Financial assistance from Balco Australia, Grain & Graze, and the Agricultural Excellence Alliance was gratefully appreciated.

References

Armstrong, K. De Ruiter, J. & Bezar, H., 2004, ‘Chapter X: Fodder oats in New Zealand and Australia- History, production and potential’, In Suttie & Reynolds (Eds) Fodder Oats: A World Review,FAO Plant and Protection Series 33

Hunt J, vanRees H, Hochman Z, Carberry PS, Holzworth D, Dalgliesh NP, Brennan LE, Poulton PL, van Rees S, Huth NI, Peake AS (2006). Yield Prophet: An online crop simulation service. In ‘Ground Breaking Stuff – Proceedings of the 13th Australian Society of Agronomy Conference, Perth, Western Australia, 10-14 September 2006.

Jeffrey SJ, Carter JO, Moodie KM, Beswick AR (2001) Using spatial interpolation to construct a comprehensive archive of Australian climate data. Environmental Modelling and Software, 16, 309-330.

Keating BA, Carberry PC, Hammer GL, Probert ME, Robertson MJ, Holzworth D, Huth NI, Hargreaves JNG, Meinke H, Hochman Z, McLean G, Verburg K, Snow V, Dimes JP. Silburn M, Wang E, Brown S, Bristow KL, Asseng S, Chapman S, McCown RL, Freebairn DM, Smith CJ (2003) An overview of APSIM, A model designed for farming systems simulation. European Journal of Agronomy, 18, 267-288.

Queensland Department of Primary Industries & Fisheries (QDPI&F), 2008, ‘Forage oat breeding program’, viewed 14/4/2008, http://www.dpi.qld.gov.au/cps/rde/xchg/dpi/hs.xsl/4791_5038_ENA_HTML.htm

Winfield, K., Paynter, B., Malik, R. (2007) Oat hay quality for export and domestic markets. Farmnote 209, Department of Agriculture and Food, WA. http://www.agric.wa.gov.au/pls/portal30/docs/FOLDER/IKMP/FCP/CER/EXPORT_HAY_QUALITY_FN.PDF

Previous PageTop Of PageNext Page