↖ CPA Weather Lab

Rain Gauge Forensics
Tipping Bucket vs. Piezo

Backyard PWS · Camp Hill, PA
Ecowitt GW3000B · 5-min resolution
Comparison start: Apr 14 16:56
(tipping bucket first online)
⚠ Dataset trimmed to 2026-04-14 16:56 — first tipping bucket observation. Pre-cutoff piezo-only data excluded from all metrics.
Tipping Bucket
Piezoelectric
49 events · 18 co-observed · 26 piezo-only · 5 trad-only
0.900" Trad. Period Total
0.920" Piezo Period Total
+0.020" Piezo Surplus (+2%)
26 Piezo-only events
5 Trad-only events
Section 01
Daily Accumulation
Section 02
Mean Rate Divergence by Intensity
Section 03
Co-observed Event Detail — Timing, Totals & Peaks
Event Window Onset Lead/Lag Cessation Lead/Lag Trad Total Piezo Total Δ (P−T) Peak Trad (in/hr) Peak Piezo (in/hr) Peak Winner
Section 04
Observation Agreement by Rate Bin
Section 05
Diagnostic Verdict
👻

Ghost Signals Below 0.05 in/hr

225 of 457 rainy obs in the lightest bin were piezo-only. The tipping bucket was silent. These are likely vibration artifacts, surface tension events, or heavy mist — not accumulating rain.

📉

Piezo Undercounts Above 0.1 in/hr

Divergence flips negative once rain intensifies. At 0.1–0.2 in/hr, piezo runs −0.039 in/hr below trad on average. At 0.2–0.5 in/hr, the gap widens to −0.058 in/hr. Apr 25 peak: trad 0.26 vs. piezo 0.17 in/hr.

⏱️

Earlier Onset, Earlier Dropout

Piezo leads onset by 5–15 min in most events — the tipping bucket needs ~0.01" to register a tip. Piezo also tends to stop 2–10 min sooner, dropping out before the drizzle tail-off fills another tip.

← CPA Weather Lab · Learn