Keyboard shortcuts

Press ← or → to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

Zone water needs and duration

Open a zone’s detail view to see its governing model, water need, planned minutes, and limiting factors.

From depth to time

At its simplest:

runtime in minutes = required gross depth in mm / application rate in mm/hour × 60

The gross depth accounts for the applicable capture efficiency. The final runtime also reflects configured adjustments, zone caps, restrictions, and scheduling capacity.

For example, 5 mm applied at 10 mm/hour needs 30 minutes before other limits. A wrong application rate produces a wrong duration even when the soil model is otherwise correct.

Measure the rate with catch cups where practical. A catalog head rate is a starting estimate.

What sets the required depth?

ModelBasis
SoilThe zone’s reconstructed water need and refill plan
WeeklyRemaining target after credited rain and applied water, spread across eligible sessions
Explicit manual runRequested duration, subject to the manual dispatch policy

Check the model on the zone. Sessions per week affects weekly scheduling; it is not the soil model’s primary trigger.

Without a soil sensor

The deficit is modeled water missing from the root zone, not a measured moisture percentage. LocalSky replays recent weather and watering history:

deficit = previous deficit + plant water use - effective rain - retained irrigation
plant water use = reference ET × the plant's seasonal crop coefficient
retained irrigation = valve-open hours × application rate × retained fraction

The result stays between zero and the soil’s available storage. Soil texture and root depth set that storage; the plant’s allowed depletion sets the watering trigger. Uncertain starting conditions remain a range until the evidence resolves them. Forecast rain and the next permitted watering opportunity shape the refill.

Area does not multiply runtime. Application rate is already a water depth across the zone. Sun/shade is saved but does not currently apply a separate soil-demand adjustment. A sensor can supply moisture checks; the modeled deficit should still be read as an estimate.

A 10 mm/hour rotor with 70% retained water supplies only 7 mm to the roots in an hour. A 38 mm/hour spray with the same setting supplies 26.6 mm in an hour. A long capped run can therefore leave a rotor zone with a remaining deficit. Check actual application rate with catch cups and verify soil and plant settings before treating a modeled deficit as proof that the yard needs more water.

Why minutes can be capped

A requested refill may exceed the zone’s run limit, a restriction’s cap, or the available watering window. The displayed plan reflects the limit.

Before raising a cap, check application rate, soil, roots, supply recovery, and available time. A longer run can create runoff or exceed a well’s recovery capacity.

Cycle-and-soak divides valve-open time into segments and adds soak waits. Total elapsed time can therefore exceed the watering minutes.

Read uncertain values

A missing deficit is not zero water need. The soil reconstruction may lack enough evidence or retain a range of possible starting conditions.

A probe percentage is not automatically an absolute measurement of plant-available water. Without calibration, LocalSky should not present a precise future percentage curve.

The separate no-watering soil outlook illustrates drying without future irrigation; the progressive water plan includes projected watering and rain. They answer different questions.

Seasonal water budget

The engine’s seasonal adjustment scales the plan before the final limits. At 100%, it leaves that adjustment unchanged. Check the effective result after changing it; a cap can prevent an increase from producing longer watering.

Watering decisions · Weekly scheduling · Tuning