fuzzycorr

Logo

A Python package for correlating simulated with observed datasets via fuzzy map comparison.

View the Project on GitHub beatriznegreiros/fuzzycorr

Fuzzy Map Comparison for the Evaluation of Hydro-morphodynamic Numerical Models


This repository contains the work developed for a Master Thesis on fuzzy map comparison methods to evaluate the performance of hydro-morphodynamic numerical models.

The goal of the study

Sediment transport and river flow processes are simulated with numerical models such as SSIIMM, Hydro_AS_2D and others. The applicability of models built with such softwares is done comparing the simulated and the observed data sets, in a step called validation. With the purpose of analysing simulated and observed bed elevation change, two main methods of comparison arise: comparison via statistical methods such as RMSE (Root Mean Squared Error) or visual human comparison. Local measures of similarity (or asimilarity) like the RMSE are very sensible to locational and value uncertainty and may indicate low agreement even when the overall patterns were adequately simulated.

In contrast, visual comparison is able to capture global similarity, which is one of the reasons why modelers still use it for model validation. Humans are capable of finding patterns without deliberately trying and therefore this comparison offers substantial advantages over local similarity measures. Nevertheless, more research has to be done to implement automated validation tools with some features of human thinking. That is necessary because human comparison is not transparent and prone to subjective interpretations; moreover, it is time consuming and hardly reproducible.

In this context, the concept of fuzzy set theory has proven to be remarkable in capturing spatial patterns similar to human thinking. By introducing fuzziness of location, one introduces tolerance to spatial uncertainties in the outputs of hydro-morphodynamic models. Fuzzy logic is, therefore, highly applicable to these models because the last hold considerable uncertainties in model’s structure, parameters and input data.

This work addresses such necessity in evaluating model performance through the use of fuzzy map comparison. The tools developed here could be applied to assess hydro-morphodynamic model’s validity and aid modelers on designing calibration procedures.

Package description

This package contains the following modules:

Usage

Code description

The repository is coded in Python 3

Dependencies and Environment

The necessary modules for running this repo are specified in the environment.yml file, to install all packages in the environment simply:

References

Ross Kushnereit    |    Chris Wills

Developed in:

Image