September 1, 2026

Setting a Soil Moisture Baseline: The Number Each Bed Should Be Watered At

A soil moisture reading on its own tells you almost nothing. Is 30% wet or dry? It depends entirely on the bed — a sandy bed and a compost-rich raised bed can read the same number and be in completely different states. The reading only becomes a decision once you know that bed's dry line: the baseline below which this bed actually needs water. That's the piece every sensor page and forum thread skips.

Why one number can't cover every bed

Capacitive probes report a percentage, but that percentage isn't an absolute measure of "how much water a plant can drink." It's relative to the soil around the probe. A fast-draining sandy bed holds less water and dries quickly, so it crosses into "thirsty" at a different reading than a compost-rich bed that stays damp for days. Push the same probe into both and the same number means two different things.

This is why generic advice — "water when it reads below X" — falls apart the moment you have more than one bed. The threshold that keeps a heavy-feeding tomato bed happy will leave a lean, well-drained herb bed soggy, and vice versa. You don't need one perfect number. You need a per-bed number.

What a baseline actually is

A baseline is the reading at which a given bed should be watered — its personal dry line. You find it by watching how that bed behaves: water it thoroughly, then watch the probe fall over the following days and note where the plants start looking thirsty. That reading is your trigger for that bed. Do it once per bed and you've turned a meaningless live number into a real "water this one, skip that one" call.

Siltwell gives every bed a starting point of 25% capacitive moisture as its default threshold, then lets you override it per bed. That default is a sensible place to begin, not a claim about your soil — you tune each bed's moisture_threshold_pct down or up once you've watched how it actually dries. A sandy bed might settle at a higher trigger because it drops fast; a moisture-retentive bed might sit lower. The point is that each bed carries its own number.

Two readings, not one — how false alarms get filtered

A single reading can lie. A probe can catch a momentary dry pocket, a settling error, or a blip right after you've walked past. If a sensor fired an alert on every dip below the line, you'd learn to ignore it fast.

Siltwell only calls a bed dry when the two most recent readings are both below that bed's threshold. One dip isn't enough; the bed has to actually be sitting below its dry line. That two-reading confirmation is the difference between a probe that cries wolf and one you trust. When it does fire, you get a "this bed is low" warning and a push notification — for the specific bed, using the specific threshold you set.

Setting it up, and what it costs

The workflow is straightforward once a probe is mapped to a bed:

• Start at the default. Every bed begins at 25%. Leave it there while you watch a full dry-down cycle.
• Watch how the bed dries. Water deeply, then track the reading over the next few days alongside how the plants look.
• Set the per-bed threshold. Move that bed's number to the point where it genuinely wants water — higher for fast-draining soil, lower for retentive soil.
• Repeat per bed. Each bed gets its own line, so the same probe network makes independent calls.

Per-bed soil-moisture alerts, and connecting a station in the first place, are part of Siltwell Pro ($22/year, with a 14-day free trial). Siltwell supports Ecowitt, Ambient, Tempest, and Davis. If you're running a Tempest, note it has no soil sensor — you'd add an Ecowitt WH51 capacitive probe alongside it, which is the best-value probe and pairs to any Ecowitt gateway or Wittboy console.

Where a baseline stops helping

A per-bed baseline narrows the guesswork; it doesn't repeal it. A probe reads one spot in one bed, so a dry pocket a foot away can go unseen, and capacitive percentages drift with soil type — which is exactly why the number is only meaningful once it's calibrated to that bed rather than trusted straight out of the box. Use the baseline to kill the obvious mistakes: watering a bed that's still wet, or letting a fast-draining bed go dry while you weren't looking. Fewer bad guesses is the win. Your eyes and the weather still get a vote.