Difference between revisions of "Modelling tool user interfaces"
Line 32: | Line 32: | ||
|style='background: #F5FCFF' |Long (hours) | |style='background: #F5FCFF' |Long (hours) | ||
|} | |} | ||
+ | |||
+ | Run times | ||
+ | Run times for models built with a given tool can vary considerably with the size and complexity of the set-up, the number of timesteps in the run, and the computer being used. This makes is difficult to generalise, however, due to their relative simplicity WRSM, SPATSIM, and ACRU4 models in these exercises generally ran in a matter of a few minutes for a 30 year run, while the ArcSWAT and MIKE-SHE models could range from 15 minutes to several hours to complete this. | ||
+ | |||
+ | The run times have import particularly for parameter adjustment testing, limiting how much can be done in the time available for a project. However, the tools that have long run times, as well as SPATSIM, have made the parameter adjustment process quick and allow batch run facilities. Once set-up time-consuming testing runs can be ongoing without manual intervention needed from the modeller, freeing their time for other work. Although WRSM and ACRU4 run quickly, the process of making parameter adjustments in the model set-up to be tested takes a long time and a lot of manual effort as described above. |
Revision as of 21:12, 31 May 2021
The table below gives a comparison of the user interface for each of the modelling tools.
Characteristic | WRSM-Pitman | SPATSIM-Pitman | ACRU | SWAT | MIKE-SHE |
---|---|---|---|---|---|
General layout | GUI; have to manually add complexity | GUI | GUI; easy to add complexity | ||
Formats for input-output data | .dfs0 (excel file which can be converted to dfs0 using the software); .dfs2 (ASCII which can be converted); .shp | ||||
Computing resources | Light | Light | Light | Intensive. Need a good processor. | |
Run times | Quick (minutes) | Quick (minutes) | Quick (minutes) | Long (hours) | Long (hours) |
Run times Run times for models built with a given tool can vary considerably with the size and complexity of the set-up, the number of timesteps in the run, and the computer being used. This makes is difficult to generalise, however, due to their relative simplicity WRSM, SPATSIM, and ACRU4 models in these exercises generally ran in a matter of a few minutes for a 30 year run, while the ArcSWAT and MIKE-SHE models could range from 15 minutes to several hours to complete this.
The run times have import particularly for parameter adjustment testing, limiting how much can be done in the time available for a project. However, the tools that have long run times, as well as SPATSIM, have made the parameter adjustment process quick and allow batch run facilities. Once set-up time-consuming testing runs can be ongoing without manual intervention needed from the modeller, freeing their time for other work. Although WRSM and ACRU4 run quickly, the process of making parameter adjustments in the model set-up to be tested takes a long time and a lot of manual effort as described above.