large spike of IO write ( io_written_kbytes_per_second)

We happen to draw system resources usage graph for 24 window, where in we noticed spike of IO write and specific point in time.  to reason out we dig out more  we moved to  Tuple_mover_operations to see if any mergout happened for the particular timeframe, but no success .

My questions ?

1) Is there a way to know what cause this IO spike ?
2) Can we  know, how much of IO do a perticular query take ? no system table gives this info @ query level.