This gallery documents the coral farm I built as an addition to my home, from the first earthmoving to the frag racks that held the corals. It was my second coral farm: the first was a series of grow-out systems running off a main tank. This one was a purpose-built room for aquaculture, and the build took several years, a fight with the township and more lessons than I expected. The photos follow the project in order, and each stage has its own write-up in the My Coral Farm series.

I built it as an insulated room rather than a traditional greenhouse, so it could hold heat in winter and stay cooler in summer at a manageable cost. Every material choice, from the floor tile to the ceiling paint, was made for a space that would be warm, wet and humid for years on end.

The site and the permit fight

This was my first construction project as an addition to an existing structure. I had done more than my share of remodelling, but I had no idea what I was in for with the township. At the same time as the farm, I was putting in an in-ground swimming pool, and most of my backyard was effectively a cliff, dropping away at about 45 degrees a few feet past the bottom of the driveway.

In 1999, to prepare for both projects, I had the area for the coral farm graded down to the level of the house foundation and a large pit dug for the pool. When I met with the building inspector for South Park Township in 2000, the project came to a halt. I was told I had 30 days to return the backyard to its original grade or face a fine of $2,000 a day. On the way out, the inspector mentioned that a storm sewer line ran through the back of the yard and that I was not going to be building anything.

The contractor who had levelled the yard sent a backhoe and two bobcats, and over two days they moved about 150 tons of soil, returning the yard close to its original grade while preserving the platform for the farm. The surveyor marked an easement of 7.5 feet on either side of the sewer line. The good news was that the line started on my property and ran across the neighbour’s, rather than crossing mine. After about two months, the township dismissed the charges and granted the permits, but the process took the whole summer and construction was pushed back to 2001.

Footer, foundation and shell

In the summer of 2001 we broke ground as soon as the weather allowed. The crew dug the footer, and a drain stack in the garage had to be extended, an unplanned but short job. Because the pool excavation blocked access, a concrete truck could not get down behind the house, so the cement was poured by hand and wheeled over a board several blocks high.

The original design fell short of the end of the house, so I changed it on the fly and tapered the end to follow the variance in the backyard, which gained about 30 square feet. I had an extra six inches of gravel laid and added reinforcement in the concrete to be safe. One setback was that there was no drain in the middle of the room; adding one would have meant another permit, and after the previous year I did not want to go back. The pad was poured level with the house so I could cut through the basement wall, where a window was cut down to the floor to make the doorway.

The frame used standard 2x4s, with pressure-treated wood wherever the walls touched concrete, and oversized roof timbers to keep roofing nails from coming through. I used R13 insulation in the walls and R19 in the ceiling, with baffles installed first, because insulating a vaulted roof without them leads to rotting shingles within a few years in a humid building. Outside, taupe vinyl siding was matched to the painted brick of the house. In one summer the farm structure, the back of the house, the pool and a pool fence were all finished.

Lighting shelf, wiring and walls

Inside, I built a shelf around the perimeter of the vaulted room for recessed lighting and ballasts, rather than hanging shop lights. In the evening, fluorescent lamps in that shelf gave soft, indirect light that would not shock the corals or interrupt their evening polyp extension, and the shelf hid ballasts for supplementary lighting. Power ran on several circuits: one for the heater and exhaust fan, one for the perimeter lighting shelf and three more for equipment, all tied into a main GFI, which is essential in a wet room.

For the walls, I wanted something more durable than green board. In a smaller enclosed system I had run for about six years, I had seen even green board break down in the humidity. I looked at what zoos and laboratories use and found FRP, fibreglass-reinforced panel, which is white, reflective, waterproof and sealed with plastic channels, but at around $26 a sheet and more than 25 sheets needed, it was expensive. Within two months I found poly board, a polymer-based alternative without the fibreglass, still water and chemical resistant, and saved about $12 a panel by driving an hour to a distributor.

Floor, ceiling and air

The floor is ceramic tile, chosen because it was cheap, reflective and durable and could continue into the house. The tiles cost roughly 60 cents each, and I had the first three courses of block tiled as well. The surface is slick when humid, but it brightened the room dramatically compared with bare concrete and conducts the coolness of the slab in summer. I ran electrical sockets every three feet along the walls.

The ceiling is water-resistant green board, sealed with a two-part epoxy paint from PPG designed for indoor pools and spas. The two parts cost about $70 and covered 250 square feet; the mix needs 45 minutes before the first coat and 24 hours between coats. I chose a sky blue, not ideal for reflection but reminiscent of water and sky. For air movement, I installed reversible outdoor ceiling fans on timers, about $45 each, alongside a ventilation fan. Heat comes from a ceiling-mounted, vented natural-gas heater by Modine. A vented heater matters: unvented heaters release carbon dioxide into a closed room, enough to lower the pH of the aquarium water.

Stock tanks and sand

After months of measuring, I was sure the room could hold six independent systems with sumps. I first bought Rubbermaid stock tanks, but once they were laid out the stepped sides meant the usable surface area was barely more than my old glass tanks, and the platforms to raise them would not leave room for sumps underneath. I switched to Tuff Stuff tanks of 110 gallons, which are tapered rather than stepped, hold more water per floor area and have a lower profile, at the cost of walls that bow slightly when filled and a black residue that took a long time to clean off. The sumps were dropped from the plan.

For substrate I used deep sand beds, which I had found to be an excellent way of controlling nutrients and feeding the fish and corals on my first farm. Starting in 2002, I could use Southdown play sand at two or three dollars a bag, compared with about $16 for 20 pounds of aquarium sand in the late 1990s. Each 110-gallon tank took about 300 pounds of sand for a bed around five and a half inches deep. I also photographed the right and wrong ways to fill a tank: pour sand into water and it clouds for about a week, but lay the sand first and trickle the water in slowly from a hose and the tank is clear the same afternoon.

Water movement and wave makers

Water movement was the biggest weakness of my old grow-out system, which relied on powerheads I was never satisfied with. Toward the end of that system I had tried an Ocean Currents Rotator, which clogged until I added foam pre-filters, and a Sea Swirl. In the new farm I could match the method to each tank, starting with two 360-degree rotating pulsers driven by Mag pumps mounted on the tank wall. One limit quickly became clear: in tanks only about 20 inches high with a six-inch sand bed, strong current would disturb the bed.

The most exciting device was the Wave2k, which pushes a surge across the surface and another along the bottom of the tank. In early tests it kept detritus in suspension for hours, and in a 110-gallon tank it made the surface look like the chop in a tide pool, thrashing an unsecured gorgonian back and forth as it would be on a reef. I planned to use it for a bare-bottom, species-focused tank of gorgonians and Dendronephthya, both of which need food kept in suspension.

Light: from skylights to metal halide

My goal from the start was a farm run on natural light. It did not work out. The house does not run east to west, the farm roof sits below the main roof and falls into shadow early and late in the day, and the three skylights, kept modest so the addition would blend with the house and the roof stay strong, did not admit enough light. I raised the roof partly to catch more light and partly so heat could rise away from the tanks, and the vaulted ceiling and the home’s orientation gave direct sun for only part of the day.

Supplemental lighting went in during December 2002 and changed by February 2003. A fluorescent strip had to sit within about three inches of the water to give enough intensity, which meant constantly moving it, so I bought a 20,000 K metal halide for that tank. For the large-polyp stony coral tank I tried a 96-watt power-compact quad, claimed to match a 175-watt halide. Colour and growth were good for three months, but at about $200 per fixture and $50 per bulb, and with two fixtures needed per tank, I replaced it with metal halide too. I never quite matched the colour I had once seen in zoanthids under 175-watt halides on my previous system.

Frag shelves and racks

How you hold the frags matters as much as the tank. On my first farm I started by wedging fragments into floral picks pushed into the sand, which were always being knocked over and kept everything at the bottom of the tank. I tried Plexiglas shelves bent with a heat gun and drilled for the picks, then deli containers filled first with crushed coral, which trapped detritus and grew algae, and later with aragonite sand.

For this farm, a local acrylic fabricator built racks to my design, cut to fit the oval Tuff Stuff tanks with three support strips glued underneath to stop them bowing. Holes for coral plugs were drilled about two inches apart on the SPS racks and about an inch apart on the zoanthid racks, since zoanthids tolerate close neighbours. The racks sat on simple boxes made from cut PVC pipe, which kept them off the deep sand beds; I avoided egg crate, which I had found to be an algae trap.

About the photos

Most of these pictures date from the early 2000s, and they show it. In the early days of my reef keeping, most photos were taken on film, and the first consumer digital cameras produced low-resolution images; I actually traded coral frags for my first digital camera. During my years photographing aquacultured coral, nearly everything was shot from above, looking down into the shallow grow-out tanks, which made white balance far easier than shooting a display tank through the front glass. The construction photos were simply a record of the work, and I regret not having taken more of the building and roofing stage.