OR/14/007 Simulation modes: Difference between revisions
| Line 10: | Line 10: | ||
! ID !! Name !! Description | ! ID !! Name !! Description | ||
|- | |- | ||
| 1 || Nash-Sutcliffe Efficiency (NSE) || [[Image:table4.1maths.png|frameless| | | 1 || Nash-Sutcliffe Efficiency (NSE) || [[Image:table4.1maths.png|frameless|200px| ]] | ||
|- | |- | ||
| 2 || Root Mean Squared Error (RMSE) || [[Image:table4.2maths.png|frameless| | | 2 || Root Mean Squared Error (RMSE) || [[Image:table4.2maths.png|frameless|200px| ]] | ||
|- | |- | ||
| 3 || Mean Absolute Percentage Error (MAPE) || [[Image:table4.3maths.png|frameless| | | 3 || Mean Absolute Percentage Error (MAPE) || [[Image:table4.3maths.png|frameless|200px| ]] | ||
|} | |} | ||
Revision as of 13:45, 26 March 2015
| Mackay, J D, Jackson, C R, and Wang, L. 2014. AquiMod User Manual (v1.0). Nottingham, UK, British geological Survey. (OR/14/007). |
AquiMod has two different simulation modes, calibration and evaluation, which are described below.
Calibration
The calibration mode is used to determine the most efficient model structures and parameter sets for the chosen application, i.e. those that produce groundwater level time-series that best replicate available observational data. In calibration mode, AquiMod has an in-built Monte Carlo parameter sampling algorithm that runs a specified model structure using multiple unique parameter sets (up to ~107) which are randomly sampled from a user-defined parameter space. For each parameter set, the model is run and a measure of simulation efficiency is calculated by comparing the simulated groundwater level time-series to available observed data. Three different objective functions can be used in AquiMod to determine model efficiency (Table 4.1).
| ID | Name | Description |
|---|---|---|
| 1 | Nash-Sutcliffe Efficiency (NSE) | 
|
| 2 | Root Mean Squared Error (RMSE) | 
|
| 3 | Mean Absolute Percentage Error (MAPE) | 
|
Evalutaion
When using the evaluation mode, the user is required to specify one or more fixed parameter sets. AquiMod then determines the model efficiency for each of these and also generates time- series output files for each model component (e.g. groundwater level time-series for the saturated zone).