DE0 used Euler timestepping (and so is first order accurate in time).
DE1 uses the average of two Euler steps to get a second order time stepping method.
DE2 uses the average of three Euler steps to get a third order time stepping method.

So DE2 should take 3 times as long as DE0 and DE1 takes 2 times as long as DE0.

DE0 is very diffusive, so if you have say a periodic forcing, the height of the wave will decrease  quite markedly. So not so good for complicated tsunami forcing problem. 
But DE0 works quite well for flood modelling.