Javad,
I am very much surprised by what I have just revealed to myself.
I can not imagine that accuracy would be un affected by a bird on the base.
I can now claim that a bird on the base, will not affect your ability to obtain a fix though, or the speed at which the fix is obtained.
My first test was with a real bird, of which I will not mention the suspected species. This bird was not going anywhere in a hurry, and had zero detectible affect on the ability of the LS to obtain a fix.
Because this bird was a little low on the "meat index", I looked around for something else to simulate a bird. The only thing I could come up with was a muddy pair of shoes. I figured, that this pair of shoes was much more of an obstacle for signal reception, than any bird ever would be.
With this pair of shoes sitting on the base, the LS was obtaining a four engine fix in less than one second. The color orange for "float" was not even being displayed when the engines were reset. Straight to all green.
Unless we get a real bird for testing, I think this puts the issue to bed.