Meteorologist: May 17–19 storm dropped about 4.25 inches on Sanford watershed — NOAA products escalated from watch to flash‑flood emergency

Brian K. Rappold, a forensic meteorologist retained by the plaintiffs, told the court he reconstructed rainfall across the Sanford watershed using Level‑2 Doppler radar (Gaylord, MI) calibrated against five rain gauges (COOP/CoCoRaHS/Mesowest). Rappold described two linked steps: (1) convert radar reflectivity (dBZ) to rate using published Z–R algorithms, and (2) ‘ground‑truth’ or calibrate radar estimates to observed gauge totals to reduce bias.

For the main May 17–19 event, Rappold said the radar‑grid average across the engineered watershed was about 4.25 inches over roughly a 40‑hour period; this average corresponds to roughly a 28‑year rainfall (≈3.5% annual exceedance probability) under NOAA Atlas 14 frequency tables. He presented incremental 5‑minute radar time series and a cumulative rainfall curve for the watershed centroid; the largest 5‑minute radar increment in the center reached roughly 0.21 inches and the heaviest rainfall rates occurred late evening on May 18.

Antecedent events: Rappold reconstructed four antecedent storms (Apr 29, Apr 30, May 14, May 15) and provided their average depths and durations; these prior rains matter because hydrologists use antecedent rainfall to infer watershed soil moisture and runoff responsiveness.

NWS communication timeline: Rappold also compiled National Weather Service products that were issued during the period: early hazardous outlooks and forecasts (May 14–16), a flood watch (May 17), forecast updates (May 18) projecting 1–4 inches of additional rain, and a sequence of flood warnings and flash‑flood warnings; at least one flash‑flood emergency cited reported releases from upstream dams (Smallwood and Secord) and warned of life‑threatening rises. Rappold said multiple area rivers and several dams were reported at high flows as the storm progressed.

Why it matters: Rappold said the watershed average and the NWS escalation show the storm was an unusually strong multi‑day event that created severe hydrologic stress across the local river system, explaining why hydrology experts treated antecedent moisture and calibration carefully in their HMS modeling inputs.