MATLAB: How to use fitdist or gevfit and force the GEV distribution to a specific type (Type I)

Question

I am using annual maxima wind speed data and am trying to model the distribution using GEV Type 1. When I use either fitdist() with 'GeneralizedExtremeValue' or gevfit(), I get a negative k value for my data, but I want to fit the distributing using k = 0 for GEV Type 1. Is it possible to specify the GEV Type or to force k to 0 using either of these functions?

Some sample code is here:

pd_wind_annualmax_GEV = fitdist(WNDSPD_hub_annualmax, 'GeneralizedExtremeValue'); %WNDSPD_hub_annualmax contains 35 years of annual max wind speeds

pd_wind_annualmax_GEV_v2 = gevfit(WNDSPD_hub_annualmax)

These yield the same result but I tried both to see if it is possible to force k to 0 for Type 1 with one of the functions.

The first line yields:

Generalized Extreme Value distribution

k = -0.0646859 [-0.209939, 0.0805676]

sigma = 2.65585 [2.08858, 3.3772]

mu = 31.2358 [30.2828, 32.1888]

The second line yields:

-0.0647 2.6559 31.2358

Thank you for your help!

Answer 1

I think you want the extreme value type I or Gumbel distribution, which MATLAB seems to call the "ExtremeValue" distribution. So, the command would simply be:

pd_wind_annualmax_GEV = fitdist(WNDSPD_hub_annualmax, 'ExtremeValue');