So, I have residuals from GJR-GARCH-t(1,1) process, then I want to fit the left tail of residuals with GEV distribution, then after I get GEV parameters, I generate random data with the GEV parameter which I got before..

Then I put the random data (GEV) on the left tail , then residuals from GJR-GARCH-t(1,1) on the other side for calculating VaR 

I am confused on fitting GEV Distribution for left tail using block maxima because block maxima is the maximum value which are located on right tail, not left tail

This is for my final project here. Thanks..