Describes how to compile the source code, input requirements, and miscellaneous issues related to running the model.

• Hardware requirements
• Compiling and running the source code
• Input requirements

Hardware requirements

The hardware requirements for DHSVM depend on the size of the application, that is, the number of pixels within the watershed. However, most modern workstations with a large amount of memory (512Mb RAM or more) and a fast processor (Pentium IV, PowerPC G4, or equivalent) will be amply sufficient.

DHSVM does not make use of multiple processors, so a fast single processor machine will result in a shorter runtime than a slower multi-processor machine (assuming no other processes are active)

Compiling and running the source code

DHSVM is written in ANSI-C, and will run under most operating systems. It has been successfully implemented on Pentium PC based systems (under FreeBSD, Linux, and Microsoft Windows), Apple PowerPC based systems (OS X), SUN workstations, and HP workstations.

The Gnu C-Compiler (gcc) is available for all these platforms and operating systems and is the compiler of choice. This means that this is the only compiler used in testing the model codes, and no attempts will be made to support other compilers.

You will need to modify the makefile that comes with the DHSVM source code to reflect the location of the necessary header and library files on your system. DHSVM has an option to display its output in an X-window while the model is running. This requires the availability of the X11 library on your system. To compile the code for the X-window display, uncomment the line in the makefile with DEFS=-DHAVE_X11. 

Similarly, DHSVM can read and write NetCDF files. This option is only available if you have the netcdf libraries on your system. To download the netcdf software and obtain more information, visit the Unidata website. To compile the code for reading and writing NetCDF files, uncomment the line in the makefile with DEFS=-DHAVE_NETCDF.

Input requirements

As a spatially distributed hydrological model, DHSVM is input intensive. In broad terms, the following input is needed for the implementation of the model in a specific area:

• Digital Elevation Model (DEM) of the basin
• Soil textural and hydraulic information
• Vegetation information
• Meteorological conditions at a subdaily timestep, in particular precipitation, air temperature, humidity, wind speed, incoming shortwave radiation and incoming longwave radiation
• Information about the stream and road network (location, width, etc.)

Soil and vegetation information is needed at the same resolution as the resolution at which the model is implemented. For each pixel a soil and vegetation type is specified, and a lookup table is used to store the associated soil and vegetation properties.

The specific content and format of each of the input files is specified on the Model Input page.