Technical details

Programming language

The Hurley Pasture Model is written in Advanced Continuous Simulation Language (ACSL, pronounced “axle”). This is a high level language created in the 1970s for applications ranging across pharmaceutical, engineering, and aerospace. It was developed substantially by Mitchell and Gauthier Associates (USA) and remains very advanced and capable despite its age. We use the latest stable version ACSL 11.8.4.

Statements are written in the form of largely differential equations, in a very intuitive style and can be written in almost any sequence. The main programme (*.csl) is written in the ‘system level’ language of ACSL. When prompted, the ACSL software translates the ACSL system statements, re-writing them into a recognised programming language (eg Fortran or C), which is then compiled. The compiler is installed along with ACSL. We use Compaq Visual Fortran 6.6 that was originally packaged with ACSL 11.8.4. ACSL then begins a ‘real time’ simulation which is ‘run’ and controlled using ACSL ‘runtime’ language and commands input from the keyboard, or more conveniently, input from pre-determined ‘batch’ files of commands (such as the examples in the ‘HPM.cmd’ file). By collating any number of levels of ACSL runtime command batch files, the user can assemble their own library/catalogues of ever higher level commands (eg devising ones called ‘do climate change’, or ‘rerun paper 2012’). So, at any time, after say altering one aspect of the model, a whole re-analyses of historic predictions, can be re-produced, and re-analysed confidently.

The entire state of the model can be saved to a file (extension .SAV) and archived, which allows the researcher to return to an exact state at any time in the future and re-run from there.

Reference Manuals for ACSL, essential for any further development using the ACSL language, were produced by Mitchell and Gauthier. Some remain available from Amazon, or Google/books, and may be available from libraries. The ISBN number for reference manual for ACSL 10.1 is 0-925649-04-X

Running the model

Unfortunately the commercial software ACSL is no longer supported by the most recent vendor (Aegis Technologies, USA). To compile and run the model requires a specific operating system (Microsoft Windows XP Professional, 32bit) which is no longer supported by Microsoft, and various software packages (eg Compaq Visual Fortran 6.6) which is also no longer supported by Microsoft/Compaq. Obtaining licenced copies of these comes down to approaching those suppliers.

However, if the software can be obtained, the models can be run highly effectively, and securely, on a modern computer by using a Mac (iMac or MacBook) running any recent Mac OS (eg Yosemite or Sierra), as the host operating system, and while running a Windows Professional 32 bit operating system as a ‘virtual PC’ contained within eg VMWare Fusion (for Sierra this is Fusion 8). Any legitimate ‘real’ and licenced XP PC can be ‘migrated’ onto a Mac, and into VMWare Fusion, to create a ‘virtual’ PC, so allowing it to run on a computer that remains by all measures, state of the art. Files can be moved seamlessly between the two operating systems (by swiping with the mouse).

We had some limited success in running the models in early versions of Windows 7, but these appear to use an emulator for XP, and run very slowly.

While the use of a Mac sustains complete functionality for the model, it is unknown to what extent the original software vendors will continue to distribute licences for ACSL, Visual Fortran, or XP. Recent postings on the Aegis Technologies website state: “Please contact Jtuck@aegistg.com in regards to any of these matters”

A full set of manuals and all required software will be assembled here, if agreement can be gained from the vendors to allow us to sustain the legacy of our more than 25 years commitment to this programming language.

For further advice contact Professor Tony Parsons directly.

On Insight and Validation

The model is conceived, and parameterised at the level of fundamental processes, wherever possible, and so is ‘validated’ scientifically at that level, which is at least one, preferably two, levels below the level and scale at which predictions are being made. This allows the model predictions to be independent of, and so can be assessed independently from, any available field scale measurements. This also allows the model to be used to explore combinations of circumstances that have as yet not arisen, and so no field data exists! This is distinctly different from where, in many ‘system’ models, notably in pastoral/practical agriculture, components of models would have been parameterised ‘backwards’ – from field scale empirical data. Independence from already observed field scale information has given rise to many counter-intuitive predictions, which when subsequently tested were found to be valid, and so has progressed our understanding. Despite this distinction, the HPM currently produces predictions of numerous components of the whole grassland ecosystem, that are in a very realistic ballpark to what later searches have found had been measured. Differences in how models are conceived and used, and the relevance of this, many find challenging.  The topic needs careful consideration, before assuming what kind of model suits the purpose in hand.