Replicating "natural" flow patterns is the chief goal, but closed systems also require the following features.
1) Detritus suspension for feeding corals and export to the protein skimmer and or mechanical filter. Poor detritus suspension will allow it to settle in the substrate where it is harder to reduce or export. Upward laminar flow or strong circular flow will achieve this.
2) Move water from the substrate to the surface for gas exchange. Upward, laminar flow or strong circular flow is ideal.
3) Removal of coral mucous and sloughed off "skin". Captive reefs are missing the silent army of cleaner invertebrates and fish that keep corals healthy. While butterflies clean away dead coral tissue (necrosis) in the wild, they linger too long in aquarium conditions, and don't know when to stop "cleaning". Laminar and circular flow are less efficient in this aspect. Direct flow directed down at the rock formation is best, but at the cost of losing the more significant features. Although direct flow, aimed at corals, is the most common configuration in reef keeping, it promotes detritus settling, interferes with surface skimming and oxygenation, and turns your live rock into a mechanical filter as water is pushed through it.
4) Variable flow patterns to discourage corals from growing in one direction toward the flow of water and subsequent food. This can also be achieved with good laminar flow, as food is available from all directions. Oceans Motions makes a few devices that serve this purpose very well.
5) To provide the proper conditions for efficient surface skimming. Cross-flow and up-flow are best. An effluent line working against another, near the overflow box will break surface tension, causing the overflow to draw water from below the surface, allowing surfactants to form a film as they float around the tank in limbo.
It's easy to forget that the overflow box and closed loop influents are moving just as much water per hour as the effluent lines are. Circular and cross-flow will allow the two to work together for a wave effect.
Flow is a matter of quality, rather than quantity. It would be a lot easier to fine tune if we could see how it was acting and reacting. You can use dyes or fish food to trace flow patterns, but success is limited.
I find that a good system will skim all floating flake food within 15 seconds. The flake food that sinks should be kept suspended indefinitely, with little or no "dead spots".
Another test is to make sure you have good surface movement with just the closed loop on, with no return lines at or near the surface. Good circular or upward laminar flow will achieve this.
Open rock work will aid in achieving many of these goals, but live rock still remains a limiting factor.
I use a combination of circular and upward laminar flow, with Oceans Motions 4-Way wave makers. This tank has two Dart pumps and a Blueline 100 return (from sump) pump. Each Dart has its' own 4-Way with effluent lines moving water to the opposite end of the tank. The return lines on the bottom have 45 degree elbows and are arranged in a circle for a spiral ascent to the surface.
There are two clusters of three closed loop return lines located in the bottom of the tank in a circular pattern. The two closed loop intakes re lcated just off of the bottom to avoid sand intake.
This is the end opposite from the overflow box. There are two sump returns located at the surface and two closed loop returns (from the pump at the opposing end), located at the bottom of the end panel.
The top effluent is from the Blueline 100 pump on the sump and has a 45 degree elbow added. Before I added the elbow, the tank wouldn't surface skim properly as surface tension was broken by opposing flow. The bottom effluent is from the closd loop pump at the opposite end of the tank.
