Matlab App for exploring the skill score between timeseries data sets or gridded data sets.
The code is provided as Open Source code (issued under a BSD 3-clause License).
ModelSkill is written in Matlab(TM) and requires v2016b, or later. In addition, ModelSkill requires the dstoolbox and the muitoolbox.
The ModelSkill App enables data to be loaded and then compared on a Taylor diagram. This form of plot was originally proposed for the comparison of model timeseries output (Taylor, 2001) and has subsequently been adapted for the comparison of morphological model outputs (Bosboom and Reniers, 2014; Bosboom et al., 2014). In this implementation the Taylor approach is used but modified in line with the mehtod proposed by Bosboom for the analysis of grid and mesh data. The options include the ability to create and add points to a Taylor diagram, output the results to the Clipboard, and estimate global and local skill scores.
Bosboom J and Reniers A, 2018, The Deceptive Simplicity of the Brier Skill Score, In: Handbook of Coastal and Ocean Engineering, Series, pp. 1639-1663.
Bosboom J and Reniers A J H M, 2014, Scale-selective validation of morphodynamic models, 34th International Conference on Coastal Engineering, pp. 1911–1920, Seoul, South-Korea.
Bosboom J, Reniers A J H M and Luijendijk A P, 2014, On the perception of morphodynamic model skill. Coastal Engineering, 94, 112-125, https://doi.org/10.1016/j.coastaleng.2014.08.008.
Taylor K E, 2001, Summarizing multiple aspects of model performance in a single diagram. Journal of Geophysical Research - Atmospheres, 106 (D7), 7183-7192, 10.1029/2000JD900719.
The ModelSkill manual in the app/doc folder provides further details of setup and configuration of the model. The files for the example use case can be found in the app/example folder.
The repositories for dstoolbox, muitoolbox and muiAppLIb.