Simulating growth and development of lowland rice in APSIM Xike Zhang1,2, Holger Meinke1,2, Peter DeVoil1,2, Gon van

Laar3, Bas A.M. Bouman4 and Yahya Abawi1 1 Predictive, Precision Systems, Queensland Department of Primary Industries, P.O. Box 102, Toowoomba, Qld 4350. www.dpi.qld.gov.au/ Emailxike.zhang@dpi.qld.gov.au; holger.meinke@dpi.qld.gov.au
2

Agricultural Production System Research Unit, Queensland Department of Primary Industries, P.O. Box 102, Toowoomba, Qld 4350. www.apsru.gov.au/Email peter.devoil@dpi.qld.gov.au; yahya.abawi@dpi.qld.gov.au 3 Plant Science, Wageningen University, P.O Box 430, 6700 AK Wageninigen, The Netherlands.www.dpw.wageningen-ur.nl/cwe/ Email gon.vanlaar@wur.nl 4Crop, Soil and Water Sciences Division, International Rice Research Institute (IRRI), DAPO Box 7777, Metro Manila, Philippines, Email b.bouman@cgiar.org Abstract A lowland rice module operating in the APSIM framework was developed based on ORYZA2000. This required rewriting the code to simulate rice physiological processes as intended in ORYZA while using the existing APSIM suite of modules for water and nitrogen. The model can now be used to simulate rotational systems. This paper describes the key features of APSIM-ORYZA, including its input requirements, interactions with other components of APSIM (e.g. evapotranspiration, soil water balance, soil nitrogen dynamics, etc) and output. We also test the models performance against some original ORYZA test data. The future development needs are highlighted. Media summary Based on a stand-alone rice model (ORYZA-2000), APSIM-ORYZA was developed to allow the simulation of cropping systems including lowland rice. The new module can now be used for cropping system simulations (e.g. intercropping, crop residue management, interaction between crop and soil fertility status). The module has been tested using limited data sets and will be fully validated under different environmental conditions. Key Words Crop modelling, rice, APSIM, ORYZA, cropping systems

Introduction
There have been intensive efforts on studying rice production system resulting in the development of several rice simulation models (Bouman et al., 2001; Ritchie et al., 1986). ORYZA2000 is one of the most advanced and has been intensively tested (Bouman et al., 2001, and references therein). However, it is a single-crop model addressing crop growth, development and crop response to environmental constraints. It does not simulate crop rotations or carry-over effects although there is an increasing demand for the ability to simulate rice-based cropping systems (especially the intensive agriculture system in China and southeast Asia countries). Such a systems capability will allow investigation of N dynamics, crop sequence, intercropping, crop residue management and soil management (eg. soil erosion and fertility). The cropping system model, APSIM (Agricultural Production System Simulator, Keating et al., 2003) is designed for the simulation of complex cropping systems and their management. System-related processes are available to any crop module through APSIMs infrastructure and generic crop libraries (Wang et al., 2002), greatly increasing the efficiency of model development by reducing redundancy and potential for error.

Collaborative research efforts between Wageningen University, IRRI (International Rice Research Institute) and APSRU (Agricultural Production System Research Unit) have now resulted in the incorporation of ORYZA into the APSIM modelling framework. This paper describes features of APSIM-ORYZA, its requirements and compatibility with other components of APSIM library of modules. Testing on the module continues.

Model description
Crop modules in APSIM simulate crop phenology, leaf area development, biomass production, yield and nitrogen accumulation in response to environmental variables such as temperature, solar radiation, soil water and nitrogen supply. Rice-specific processes as simulated in ORYZA were implemented within APSIM. Table 1 summarises the input requirements and key parameter values for APSIM-ORYZA. Table 1. Input variables for key physiological processes simulated in APSIMORYZA and their links to other APSIM-modules.
Simulation Input requirement Value Source Output

Phenology

Meteorological data daily max & min temp, rainfall, solar radiation;

Time series data

met module in APSIM

report module in APSIM

Base, optimum and maximum temperature;

8, 30, 42C

Oryza

Transplanting shock index.

0.4

Oryza

CO2 assimilation and yield formation

Extinction coefficients for visible light and for N distribution in the canopy;

function of development stage

Oryza

report module in APSIM

Initial light-use efficiency of a single leaf;

function of temperature

Oryza

Leaf N content;

calculated

Oryza

Spikelet formation factor.

65 spikelets/g DM

Oryza

Water and nitrogen uptake by plant

Potential evapotranspiration;

calculated

eo module in APSIM

Water availability in soil profile;

calculated

soilwat2 module in APSIM

Deltas sent back to soilwat2 and soiln2 modules in APSIM for water and N balance calculation

Soil nitrogen supply.

calculated

soiln2 module in APSIM

Temperature is the main driving force for phenological development (Bouman et al., 2001, and reference therein) and is used to calculate the timing of the start of photoperiod sensitive phase, panicle initiation, anthesis and physiological maturity. The development rate is calculated from hourly temperature (which is in turn estimated from daily maximum and minimum temperature, Bouman et al., 2001) using three cardinal temperatures: base temperature (8C), optimum temperature (30C) and maximum temperature (42C). Transplanting shock reduces this rate and so does water limitation. Crop development ceases (ie. death) when daily average temperature falls below 12C for three consecutive days. The model calculates the effective leaf area index by taking account of leaf rolling under stress and light interception from stems. The instantaneous assimilate rate of a single leaf depends on leaf nitrogen content. The assimilation rate of whole canopy is obtained by integrating the instantaneous value over canopy depth and sinusoidal course of the day. Growth estimates of different organs are achieved via dry matter partitioning, leaf death and loss, maintenance respiration and growth respiration. The number of spikelets at flowering is calculated from an empirical index of spikelet formation factor which was derived from field experiment (Bouman et al., 2001). The final spikelet number for yield formation is adjusted by a spikelet sterility which is a function of high or low temperature and water limitation. Water availability is determined via the APSIM-Soilwat2 (note that while water ponding is implemented, simulation of paddy N chemistry is not simulated). Crop water uptake is based on potential evapotranspiration which is computed in APSIM-Eo. Soil nitrogen supply

is calculated by APSIM-Soiln2 and APSIM-Fertiliz if nitrogen fertilizer is applied. APSIM maintains mass balance for all processes.

Model testing
The performance of APSIM-ORYZA under potential production conditions (non-water and non-nitrogen limiting) was tested on a field experiment conducted in IRRI farm in Philippines in 1991 by simulating the dry weight of the whole panicle, total dry weight of above-ground biomass and leaf area index (Figure 1). Simulated vs. measured data showed good agreement and was the same as results obtained when using ORYZA-2000. Overall, simulated LAI matched the pattern of growth, with the model slightly under-simulating the maximum leaf area. This is probably due to the inaccuracy of the parameter of specific leaf area in the late period of leaf growth. The calibration of that parameter would produce close simulation of LAI to the measurement.

Figure 1. Measured versus simulated biomass production and leaf area index of Oryza sativa cv. IR72. Measured data were obtained from a potential production experiment conducted by IRRI in Philippines in 1991

Figure 2. Actual versus predicted yield production and nitrogen concentration in leaves of Oryza sativa cv. IR-72. Measured data were obtained a field experiment with three levels of nitrogen fertilizer application conducted by IRRI in Philippines in 1992

Model performance was further tested under nitrogen-limited conditions (Figure 2). The test data were derived from a lowland field experiment with three levels of nitrogen fertilizer application conducted by IRRI in Philippines in 1992. Results showed that the simulated leaf nitrogen concentrations are lower than measured values in 0 kg N/ha and 180 kg N/ha treatments and higher/lower in 225 kg N/ha treatment. Although Figure 2 showed good agreement between simulated and measured dry weight of total panicles under 0 and 225 kg N/ha conditions, the model did not reproduce the high yield in 180 kg N treatment. Neither ORYZA-2000 nor APSIM is able to simulate the nitrogen dynamics of a paddy soil, ie. the transformation of nitrate and ammonium in an anaerobic environment. Thus the simulation of soil nitrogen supply using the dryland nitrogen module in both models is inadequate and further development must address this issue.

Conclusion
APSIM-ORYZA aims to simulate growth and development of rice with the capacity to deal with other important features of rice cropping system such as N dynamics, crop sequence, intercropping, crop residue management and soil management (eg. soil erosion and fertility). Preliminary testing of the modules performance using two independent data sets without water and nitrogen-limiting conditions shows reasonable agreement between predicted and measured values for crop biomass production and yield of storage organ. However there are some discrepancies between the simulated leaf nitrogen concentration and yield of storage organ and the measured values under different nitrogen application conditions. The results highlight the need to develop module to simulate nitrogen dynamics in paddy soil, which is underway. At the same time, APSIM-ORYZA needs to be further tested under different field-management and cropping systems.

References
Bouman BAM, Kropff MJ, Tuong TP, Wopereis MCS, ten Berge HFM and van Laar HH (2001). ORYZA2000: modelling lowland rice. Los Banos (Philippines): International Rice Research Institute, and Wageningen University. 235p. Keating BA, Carberry PS, Hammer GL, Probert ME, Robertson MJ, Holzworth D, Huth NL, 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 (2203). An overview of APSIM, a model designed for farming systems simulation. European Journal of Agronomy 18, 267-288. Ritchie JT, Alocilja EC, Singh U and Uehara G (1986). IBSNAT and the CERES-Rice model. Proceedings of the Workshop on Impact of Weather Parameters on Growth and Yield of Rice. April 1986. International Rice Research Institute, Manila, Philippines. Wang E, Robertson MJ, Hammer GL, Carberry PS, Holzworth D, Meinke H, Chapman SC, Hargreaves JNG, Huth NI, McLean G (2002). Development of a generic crop model template in the cropping. European Journal of Agronomy 18,

Using Notepad++ to compile and run APSIM

Print RSS

This page demonstates how to use the freely available Notepad++ to compile APSIM source code and run APSIM simulations.

Installing Notepad++

Install Notepad++ from the Notepad++ web site. Ensure the NppExec plugin is installed by going to the Plugins | Plugin Mangaer | Show Plugin Manger menu and selecting NppExec:

Setting up Notepad++ to run APSIM simulations

To configure Notepad++ to execute external programs go to the Plugins | NppExec | Execute (F6) menu and type the command: cmd /c C:\Apsim\Model\ApsimRun.exe "$(FULL_CURRENT_PATH)", replacing c:\Apsim with your APSIM development path if you have it located in a different directory. Click Save and give it the name Run APSIM.

To assign this RunAPSIM command to a menu, go to PlugIns | NppExec | Advanced options. Enter a menu item name Run. Select the associated script Run APSIM. Click Add. This will add a new menu to the Macro menu in Notepad ++.

To assign this menu item to a key press (so that you can simply press F5 to run APSIM, go to Settings | Shortcut mapper, find the Run menu item under the Plugin commandstab and click Modify to give it the key press of your choice. F5 is a fairly standard run key press.

Also because F5 is already mapped to the generic Run feature of Notepad++ you will have to disable this other F5 shortcut. To do this, under the Main menu tab in Settings | Shortcut mapper find Run (usually number 184) and change the shortcut to "None". So now we only have one F5 shortcut which is our Apsim Run shortcut. You can still access the generic Run feature of Notepadd++ from the main menu in Notepad++. You just don't have the F5 shortcut to it anymore.

Now whenever you have a simulation file loaded in Notepad++ (e.g. .apsim file), you can press F5 to

run it.

Setting up Notepad++ to compile APSIM source code

To compile APSIM source code follow the steps above setting up an Execute command of: C:\Apsim\Model\Build\RunMake.bat $(CURRENT_DIRECTORY) with a name of Compile APSIM and a menu command Compile. I use a shortcut key of CTRL SHIFT B which is the same as Microsoft Visual Studio. To get Notepad++ to highlight error messages from the compile, go to PlugIns | NppExec | Console Output Filters and enter these 2 filter lines:
*"%FILE%", line %LINE%:*

%FILE%(%LINE%) : *

Now you can compile the file that you have open in Notepad++ by pressing CTRL SHIFT B. Any

error message will be displayed in the window at the bottom. Double clicking an error message will take the editor to that line.

Building Apsim
Print RSS

Obtaining a working development tree from scratch

The Apsim source is stored in a version control system called subversion . You can find a quick tutorial on its use here . The current APSIM repository URL is http://apsrunet.apsim.info/svn/apsim/trunk, pre 7.0 versions of apsim usehttp://apsrunet.apsim.info/svn/development/trunk. You will need to install the CollabNet Subversion Command-Line Client v1.6.17 (for Windows). The APSIM development team also recommend TortoiseSVN. Once installed, create a directory on your hard disk to contain the development tree (e.g. c:\Apsim), right click on it and select SVN Checkout. Enter http://apsrunet.apsim.info/svn/apsim/trunk as the URL of repository and click OK.

Updating a development tree to the latest revision

To update a development tree to the most recent revision, right click on it and select SVN Update. This will bring down all changes made by the APSIM development team to your development tree. Where you've made changes, SVN will try and merge your changes into what it brings down from the repository. If you don't want to keep your changes, then right click on the development tree and select Tortoise SVN|Revert following an update. Once you have done an update and optionally a revert, you need to either build the entire tree, or download the appropriate binaries from the APSIM Build System.

Obtaining binaries from the APSIM Build System

Determine the Revision Number of your development tree by right clicking on it and selection TortoiseSVN|Show log. The entry in bold (at the top if you've just done a checkout) will tell you your revision number. Once you know your revision number, you can obtain all binaries for your development tree by going to the APSIM Build Systemand clicking the binaries link for your revision. This will download a binaries zip file which you can unzip to your Model (e.g. c:\Apsim\Model) directory. For cluster installations, here is a self extracting linux executable of version 7.4.

Obtaining third party tools and support libraries

You will need to install a range of open source build libraries (117Mb) (boost, libxml2, and iconv precompiled win32 binaries). Unzip this file alongside the apsim source tree, not underneath (e.g. c:\BuildLibraries). Microsoft Visual Studio 2010 is required for the infrastructure, .NET and C++ modules. Lahey Fortran 95 version 5.6 is required to build the fortran modules. To build the GUI, you will need: TeeChart for .NET component Quantum Whale editor component You'll also need a good text editor. The APSIM development team recommend NotePad++. You can then configure the text editor to compile the module by following the instructions here.

If you want to build everything

After you've installed compilers and support libraries, a set of batch files (or shell scripts on unix systems) build part or all of apsim. They are in Apsim/Model/Build.
CleanAll.bat deletes all binary files in the source tree. BuildAll.bat cleans and builds everything from scratch. The compiler output from each module is inModel/Build/BuildAllOutput.xml RunAll.bat runs every simulation in the example and test directories. This can take some

time. These scripts can be run by double clicking them in Windows Explorer, or running them within the Visual Studio Command Prompt DOS window. The build and run system operates in parallel - an xml description of individual tasks describes their order. It will use all available processors, and your system may become unresponsive for a few minutes. Individual apsim modules (eg SoilWat.dll) can be built with the script Model/Build/RunMake2.bat. A description of its use with a text editor is here. The output of the compilation appears on stdout. By default, apsim is built in "Release Mode" with no debugging information. The switch to build a debuggable version is inModel/Build/Win32VSCPP.make.

If you want to compile a FORTRAN module on Windows

To compile an APSIM FORTRAN module, you'll need to have LF95 installed. Open the module's source file in your text editor, and use Model/Build/RunMake2.bat to build the module (details here).

If you want to compile a C++ or .NET module on Windows

Building C++ or .NET modules requires Microsoft Visual Studio 2010. You can use a text editor to edit and compile the source file, or open the projects solution file in the VS2010 IDE.

If you want to compile the APSIM user interface

The APSIM user interface uses many third party components that will need to be purchased. If you're interested in doing this then post a message to the APSIM Forum.

Compiling under LINUX

The linux build process is identical to the windows build; though the build scripts have .sh extensions. Non-GUI .NET components build under mono 2.10 or later. Currently, there are no pre-built packages for Ubuntu installs, so the package must be built from source, the mono interpreter added to binfmt_misc, and the mono shared library path added told.so.conf.d. You will need the following packages installed:
gcc gfortran libboost1.40-all-dev libxml2-dev tcl8.5-dev tcllib

Currently, no GUI components can be built under linux.

Running under LINUX

Apsim is a mix of native and .NET applications, each runtime has a few peculiarities that must be addressed. The native applications are linked against several shared libraries (eg libApsimShared.so) which are not installed in a standard shared library path, forcing the use of the LD_LIBRARY_PATH environment variable. If you are working in a login session (eg a console on your desktop) you can add this to your .profile (not /etc/environment - see Launchpad bug#366728). Alternatively, follow the logic in Model/ApsimRun.sh to convert .apsim/.con files and run apsim. .NET applications require a working mono installation using binfmt_misc as described in the previous section. For cluster installations, here is a self extracting linux executable of version 7.4.