GFCI Holds Dry But Trips Only After Rain On Outdoor Circuit Decision Tree

Why this matters

A GFCI that holds for days then trips within hours of a rain event is reporting moisture-driven leakage above 5 mA somewhere on its protected circuit, and the leak path closes as water bridges a conductor to ground or to neutral. The temptation is to call it a defective device and swap it, but a weather-correlated trip almost never originates in the GFCI itself. The real fault lives in a flooded box, a failed in-use cover, a buried splice, or a damaged direct-burial cable that only conducts when wet. Chasing it dry is futile; the circuit has to be diagnosed in the failure condition or simulated wet. Getting this wrong leaves a customer with no outdoor power every storm and exposes the contractor to a shock-injury claim if the GFCI is bypassed to stop the nuisance.

Symptom presentation

The customer reports outdoor receptacles, landscape lighting, a well pump, or a pond pump that work fine in dry weather and lose power during or shortly after rain. The GFCI (face device or breaker) is found tripped after the weather clears. Ask: Does it trip during the rain, or hours later as water wicks into a box? Is the affected load on the LOAD side of a single GFCI feeding multiple outdoor outlets? Has any landscaping, irrigation trenching, or hardscape work happened recently? Direct-burial UF cable nicked by a shovel is the single most common slow-leak source on circuits more than a few years old.

Quick checks before isolation

Press TEST then RESET on the GFCI to confirm it functions; a self-testing UL 943 device that will not trip on TEST is end-of-life and gets replaced before anything else. Walk every outdoor box on the circuit and confirm each has a listed in-use (bubble) cover per NEC 406.9(B)(1), gaskets intact, and is mounted so water sheds away. Open each box lid and look for standing water, corrosion green on terminals, and back-stab connections weeping. Confirm the circuit map by tripping the breaker and verifying which receptacles go dead. Note whether the GFCI is the breaker in the panel or a face device with downstream LOAD wiring.

Isolation tree

Step 1 - Reproduce the fault. Disconnect all LOAD-side conductors at the GFCI and cap them. Reset the GFCI. If it now holds through a rain event (or through a hose-test simulation, described below), the fault is downstream and you proceed to Step 2. If it trips with LOAD disconnected, the fault is on the LINE side or inside the device; megger the home run and replace the GFCI.

Step 2 - Divide the downstream run. At the midpoint box, separate the circuit into upstream and downstream halves and cap the downstream conductors. Reset and wet-test. If it holds, the fault is in the second half; if it trips, the fault is in the first half. Repeat the split at the next box in the faulted half. This binary search converges on the offending box or cable segment in two or three iterations on a typical 4-to-6-outlet outdoor run.

Step 3 - Box versus cable. With the suspect segment isolated, disconnect the device in the box and cap the conductors inside the box. Wet-test the box alone (spray the exterior). If it now holds, the leak is the device, the cover, or the in-box connection - water entering a non-listed cover or a cracked device is the culprit. If it still trips with the box conductors capped, the cable feeding that box is leaking; suspect direct-burial UF damage, a buried unapproved splice, or conduit holding water that wicks into a nick.

Step 4 - Confirm the cable. Megger the suspect cable segment at 500 V from each conductor to the equipment grounding conductor (EGC) and conductor-to-conductor. A sound cable reads above 100 megohms. Pour water at the suspected entry point or run an irrigation cycle, then re-megger. A reading collapsing below 1 megohm under wet conditions confirms insulation breach. If the cable is in conduit, blow it out and check for standing water at low points, a missing weep hole, or a failed conduit seal at the building penetration.

Confirming the diagnosis

A confirmed moisture fault is repeatable on demand: a controlled hose-test or irrigation cycle on the isolated segment trips the GFCI within minutes, and the same segment holds indefinitely when dry. A confirmed cover or box leak clears the instant the box is sealed with a listed in-use cover and gasket and re-tested wet. A confirmed direct-burial cable breach shows the megger reading crashing only when water reaches the damaged point, and the location can be narrowed by wetting along the trench in sections.

Remediation

Replace failed or non-listed covers with listed in-use covers per NEC 406.9(B)(1); every 15A and 20A 125V and 250V receptacle in a wet location requires one whether or not anything is plugged in. Seal box penetrations and add weep holes at conduit low points. For a damaged UF run, the correct repair is a new cable in conduit or a listed underground splice kit (direct-burial rated, gel-filled or resin), buried at code depth per NEC 300.5 Table; do not wrap-and-tape a buried splice. Restore the in-use cover and re-megger before closing. If the GFCI failed its self-test, install a current self-testing UL 943 listed device and note the install date on the cover.

Do not defeat GFCI protection on an outdoor circuit to stop weather-driven nuisance trips. NEC 210.8(A) requires GFCI protection at all 125V through 250V, 15A and 20A outdoor receptacles, and a wet-leakage path that trips the GFCI is exactly the shock hazard the device exists to interrupt. Bypassing it transfers full liability for any rain-day electrocution to the installer. Document any customer refusal of the cable or cover repair in writing.

References

  • NEC 2023 Article 210.8(A) - GFCI protection, outdoor and wet locations
  • NEC 2023 Article 406.9(B)(1) - Listed in-use covers for receptacles in wet locations
  • NEC 2023 Article 300.5 - Underground installations, minimum burial depths
  • UL 943 - Standard for Ground-Fault Circuit-Interrupters, end-of-life self-test
  • NEC 2023 Article 250.114 - Equipment required to be grounded