We propose a framework for model calibration based on niching genetic algorithms.
It consists of interactive multimodal search, visualization, and parameters analysis.
The output of the framework is a set of alternative model calibration solutions.
The multimodal framework is applied to two different applications.
Experiments showed the framework benefits when interactively validating models.