When the Pond Breathes Better Than Expected

The funny thing about a garden system is that it teaches you through its problems. This past week, we discovered something quietly important while chasing mosquitoes and wrestling with plumbing: our water is disappearing faster than the San Diego sun alone should take it, and that's actually telling us something good.

Let me back up. We've been measuring the reservoir carefully—it dropped from 3.6 inches down to 1.5 inches over eight days, losing about 4 gallons a day. On paper, that sounds alarming. But when we compared it against the official evaporation rates for coastal San Diego (the "pan-evaporation" standard that water managers use), our reservoir is losing water roughly 40% faster than open water would. That extra loss? It's the plants drinking. It's the sand wicking moisture up from below. It's the whole system working—which means our wicking bed design is actually pulling water where we need it. That's a small victory wrapped inside what looked like a leak.

The mosquito problem, though, is more straightforward. Dense pond plants create perfect nurseries for mosquito larvae—still water, vegetation to cling to, shelter from fish. We're going to thin the plants significantly and probably introduce some mosquito fish (gambusia, if we can source them locally). It's an old-fashioned solution, but it works because it respects how ecosystems actually organize themselves.

The real constraint we're bumping against right now is feeding twice a day. We've been managing only once-a-day feedings in the mornings, which isn't pushing enough nitrogen through the system to keep everything thriving. Our fish and plants are getting by, but not flourishing. The oxygen levels held steady at five milligrams per liter last week even without the air pump running—that was encouraging—but we noticed some dead zones forming in a few spots in the sand, which tells us water isn't flowing uniformly through it all.

This points toward a redesign. The manifold pipe we rebuilt recently is sitting in the sand, trying to distribute water through it, but sand doesn't wet evenly. Instead, we're thinking of moving the manifold so it sprays directly over the water surface, then letting water percolate sideways through the sand layers naturally. It turns out making sand moist is easy; making it stay wet uniformly is the real puzzle.

August is teaching us to listen closely to what the system is actually doing versus what we hoped it would do. Next week we tackle the manifold redesign and figure out how to automate that second feeding. One step at a time.