There's lots of talk about this accident, and turns, with varying reference to wind. I'm not a Mallard pilot, nor am I familiar with Perth, but to inform myself a little, I did some Google Earth measuring of the flying display area before commenting here. Presuming that is was the intention of the pilot to avoid flying over land (built up areas) as depicted in the flying display graphic kindly posted, let's consider dimensions:
The water area depicted on the graphic equates roughly to the runway and infield dimensions of an airport with one 8000 foot runway, and a bisecting crossing 4500 foot runway. Let alone an aircraft the size [and lesser maneuverability] of the Mallard, would it be considered normal flight operations to fly a circuit inside the infield of an airport of these dimensions? I think not - that would be unusual flying. Possible, but unusual.
Perhaps this pilot practiced the turn at a place of similar dimensions, to get used to what most anyone would consider tight maneuvering. I have no idea. I know that water flying can dramatically alter perception of space for maneuvering - it can look like there's room, but there's not that much. This is a major teaching point when I train flying boat pilots.
I train that a water landing is preceded with one or more reconnaissance passes, with consideration for dimensions, the approach, hazards, and the possible affect of wind. I wonder if this pilot had the opportunity to fly some practice or reconnaissance passes before landing.
It is certain that he, as any of us would, maneuvered so as to avoid overflight of crowds or even shore. Once it started going bad, tightening the turn was not going to fix it. He had already run out of room for maneuvering within the normal flying characteristics of that 'plane. I fly light floatplanes and flying boats, and what I see in that video would worry me for maneuvering a light seaplane, let alone the larger Mallard.
I can imagine that this pilot felt pressure to perform with his aircraft, and further would not want to disappoint his passenger. He pressed further into a maneuver than perhaps he had the practice, performance and maneuverability with which to complete it.
The facts will come out, but in the mean time, discussion of what is evident so far can benefit all pilots in reminding that when maneuvering close to the ground/landing in a non aerodrome environment, extra awareness and caution are needed to assure that the required minimum space is available, and that there is an escape plan if the maneuver is going wrong. This video will become an element of my training, with the obvious aspects discussed, even before a formal investigation report is written. Anything we do to make pilots think is good.