The paper proposes a new model for the statistical characterization of extreme values of wind speed based on the Inverse Burr distribution.

The paper also provides a theoretical derivation of the Inverse Burr model as a mixture of Inverse Weibull models.

Inverse Burr parameters were estimated through maximum likelihood and moment estimations and through a new proposed quantile estimation tool.

Numerical applications were performed on several real wind speed datasets to validate the proposed model as an alternative to existing models.

A comparison between estimation techniques was performed also on synthetic samples of wind speed, generated through known distributions.