Parameters of response distribution , response expectations What is the meaning of this sentence?
"In extreme value analysis, the parameters of the response distribution instead 
of the response expectation are linked to the covariates."

I understand that response expectation is E(Y) , but not sure what parameters of response expectations being linked to covariates is.
 A: It's difficult to be absolutely certain given the limited context, but it looks like an extreme-value model is being contrasted with an ordinary regression model, or perhaps with a generalized linear model. (However, the contrast is not quite so clear-cut as I'll explain.)
In an ordinary regression model (or indeed, in a GLM), the (conditional) expected value of the response is written as a function of (i.e. "linked to") the covariates. 
By contrast, "in extreme value analysis, the parameters of the response distribution [...] are linked to the covariates."
That is, some response distribution is specified (presumably one of the extreme value distributions), and the parameters of that distribution are written as functions of the covariates. Presumably (and again, absence of context makes it hard to be sure) there's some link function to a linear predictor.
[But in fact even with a GLM or a regression model, the expected value is a parameter of the distribution (possibly after a suitable reparameterization), so again the parameters of the distribution are written as functions of the covariates]
