Pushing precipitation to the extremes in distributed experiments: recommendations for simulating wet and dry years