You want a "proper study" instead of an annual time series, comparable for over a century, conducted by millions of individuals? What a goofy comment. This is gold-standard data. The paper even explains why this is superior to nerd bullshit like eBird.