The only input file that the model needs is the daily climate file in .csv format (file name: InputClimate.csv). All remaining files are created through the GUI. The input climate file can be found in the folder ?C:\Program Files\Toolkit\CLASS\Class CGM\Project name?. Format of a sample climate file is shown in "Appendix A.1 ? Input data file" in the CLASS CGM Users Guide 1.0.1.

The model generates 6 output files which are described in "Appendix A.2 ? Output data files" in the CLASS CGM Users Guide 1.0.1. All the output files will be stored in the folder ?C:\Program Files\Toolkit\CLASS\Class CGM\Project name?. All output files are in ?.csv? format and therefore you can open them directly in Excel. The user can view the numerical values of the outputs by clicking on the data tab at the top of the GUI or you can view the same outputs as a plot by clicking on the plot tab.

Algorithms for the crop growth model are described in a detailed technical report (Johnson, 2003). Algorithms for the water balance component are detailed in the CRC for Catchment Hydrology Technical Report 04/12 by Tuteja et al., 2004 (available under 'Publications' below). The technical report has been peer reviewed by international and national experts in distributed hydrologic modelling. Testing of the CLASS CGM has been done at two levels. The first level includes testing of the growth component wherein hydrological variations (eg. soil moisture, rainfall and evaporative demand etc.) are switched off and known climate and hydrological input were used to check growth computations from CGM against results from Johnson (2003) and manual calculations from the algorithms. A second level of checking was done against the water balance computations from HYDRUS-2D (Simunek et al., 1999). The pasture growth component was switched off and results from the water balance component U3M-1D on soil moisture and plant transpiration across the soil profile were checked against results from the HYDRUS-2D model. These tests have been conducted for a 2.3m deep homogeneous as well as a 2.3m deep heterogenous soil profile. Both models were used with same climate and soils data and the evaporative demand was forced as potential plant transpiration. In the case of homogeneous soil profile, simulations were done for 183 days with no plant water stress in both models. The results on soil moisture variation for each 10cm layer across the soil profile, actual plant transpiration and leakage from the soil profile were within 0.5%. For the case of heterogeneous soil profile, simulations were done for daily climate data from 1975-2000 (26 years) and a soil profile from the Little River catchment (Macquarie River basin, New South Wales). Simulations were done with plant water stress in both models. Results on soil moisture within each of the four soil materials were within 0-2% and at the interface of the material were within 0-5%. These differences are largely due to differences in numerical architecture of the two models and plant water stress function implementation in the two models (soil moisture based plant water stress function in CLASS U3M-1D as against pressure based soil moisture stress function in HYDRUS). The water balance component U3M-1D in CLASS uses an explicit solution of the Richards? equation for water balance calculations. As the time step and the layer thickness are decreased, the implicit solution is approached. The default soil layer thickness value of 10cm is found to be optimum for speed as well as accuracy. A value smaller than 10cm is likely to increase the run time without any appreciable increase in the accuracy. A value greater than 20cm will affect the accuracy of the solution of the Richards? equation.