Two (or even three?) reasons -- in that particular case the Bernoulli effect is indeed bigger because the outlets are directly underneath the hull, but a bigger reason is likely to be that the incoming water has quite a lot of air bubbles in it which lowers the water density under the whole front of the boat, and the boat bows sink as a result, and the boat then slides "down the slope" and hits the gates. I've seen this happen in deep locks with a lot of bubbles, the bows dropped several inches so water came flooding in through the scuppers which were normally above water -- and ships have been sunk this way due to underwater gas eruptions. There's also the question of where the water coming into the lock ends up when it travels down the lock and bounces off the bottom gates, which raises the level there -- if there's a gap between the boat and the bottom gates this water comes up to the surface as a bulge and pushes the boat forwards, if the boat is against the bottom gates there's no space for this to happen. As usual there's no simple answer to a complicated question, there are all sorts of things going on when you fill a lock, some obvious and some not so obvious. Bernoulli is one of them, but given the relatively low water velocities (except right where the paddle culvert enters the lock) it's not usually the biggest one, changes in water level or just water flows against the hull have a bigger effect. But you can make some rough engineering guesstimates, along with numbers for Bernoulli pressure drop (0.07psi at 1m/s, proportional to square of speed) and water flow vs. gradient and likely water level changes, as well as observing what actually happens to the water level and boat trim as a lock fills, and how long this takes (to give water flow rates). These suggest that Bernoulli does indeed have an effect, but other things usually have a considerably bigger effect.
