Every competition image of a fountain is rendered on a still day. Almost no site is still. We now put an anemometer on any exterior site for a minimum of eight weeks before we fix a single nozzle height, and the data has changed the scheme on four of the last six projects.
The numbers that matter
Not the mean. The mean is comforting and useless. What sets the design is the ninetieth-percentile gust and how often it occurs during opening hours. A site with a 3 m/s mean and 11 m/s gusts on forty afternoons a year is a much harder site than one with a steady 5.
The working rule from the rig: a clear vertical jet loses its shape at about 5 m/s, and by 7 m/s the droplet cloud has drifted roughly a third of the jet height downwind. A laminar rod is finished at 3. A water bell is finished at 2. A film wall, where the water never leaves the surface, is untroubled at 8. Choosing the effect is therefore mostly choosing what the site will allow.
Designing the fallback first
The mistake is to design the still-day scheme and then bolt on a wind cut-out, because the result is an installation that spends thirty days a year switched off and looking derelict. At Millrace Court we designed the windy-day condition first — a 34 m rill that reads perfectly well on its own — and treated the jets as the thing that gets added when conditions allow. The courtyard has never looked broken.