Legend
Top 15 costliest hail events, 1980–2026 (2026 dollars) (15)
Single-city hailstorms with an insured-loss figure (5)
Storm days checked: NWS warnings and lead times (25)
SPC hail reports ≥1 in, top-15 events before 2019 (1302)
NWS severe thunderstorm warnings (SV.W) on the 25 storm days, 2016–2025 (246)
NWS tornado warnings (TO.W) on the same storm days (9)
Hail reports ≥1 in on the 25 storm days, inside an NWS warning (673): measured size
… estimated size
… qualifier unknown
Hail reports ≥1 in on the 25 storm days, outside any warning (46): measured size
… estimated size
… qualifier unknown
SPC hail reports ≥1 in, the seven also-hail-driven events (1,132)
SPC wind reports, the seven also-hail-driven events (321): measured gust
… estimated gust
… not stated (2026 preliminary reports)
Counties named in the published insured-loss accounts, 4 city storms (6)
Dark red: the 15 costliest hail-named billion-dollar events (total cost, 2026 dollars; marker at the area hit hardest). Blue: single-city storms with an insured-loss figure (dollars as published; not comparable with the red ranking); where one shares a spot with a red marker it shows as a blue ring. Where red markers share a spot they are offset slightly; a thin line leads to the true location. Teal: 25 storm days checked against NWS warnings (IEM), at the largest hail report; click for lead times. Grey: SPC reports for older events. Orange shading: MRMS MESH radar estimate (MODELED), for 10 events and six more storm days (2021–2025). Click any dot. As of 2026-10-09. Off at load (tick them in Layers), by era: for the 25 storm days (2016–2025), NWS severe thunderstorm and tornado warning polygons (ALERT, IEM) and the 719 hail reports the warning table counts (OBSERVED; measured or estimated, inside or outside a warning); NEXRAD reflectivity at the peak of six city storms (1995–2023, OBSERVED radar); SPC hail and wind reports for the seven also-hail-driven events (2010–2026; 2026 preliminary) and MESH for the three since 2019; and the counties named in the published insured-loss accounts.
Radar-estimated hail size (MRMS MESH, MODELED)
1–1.5 in
1.5–2 in
2–2.5 in
2.5–3 in
3 in +
MRMS MESH_Max_1440min, 24 h to 12 UTC, maximum over each event's days, 1 km grid. A radar estimate of maximum expected hail size, not measured stones.
NEXRAD reflectivity, Mayfest, Fort Worth, May 5, 1995, 1995-05-06 00:15 UTC (OBSERVED radar)
20–30
30–40
40–50
50–55
55–60
60–65
65+ dBZ
IEM national NEXRAD base-reflectivity composite (N0R, 5 dBZ steps). Frame: of the 5-minute frames from 20 min before to 5 min after the largest SPC report within 40 km, the one with the strongest echo within 10 km of it (65 dBZ). 60+ dBZ cores often mean large hail, but reflectivity is not a stone size.
NEXRAD reflectivity, Tri-State storm at St. Louis, Apr 10, 2001, 2001-04-11 02:15 UTC (OBSERVED radar)
20–30
30–40
40–50
50–55
55–60
60–65
65+ dBZ
IEM national NEXRAD base-reflectivity composite (N0R, 5 dBZ steps). Frame: of the 5-minute frames from 20 min before to 5 min after the largest SPC report within 40 km, the one with the strongest echo within 10 km of it (65 dBZ). 60+ dBZ cores often mean large hail, but reflectivity is not a stone size.
NEXRAD reflectivity, Dallas, Jun 13, 2012, 2012-06-14 01:15 UTC (OBSERVED radar)
20–30
30–40
40–50
50–55
55–60
60–65
65+ dBZ
IEM national NEXRAD base-reflectivity composite (N0Q). Frame: of the 5-minute frames from 20 min before to 5 min after the largest SPC report within 40 km, the one with the strongest echo within 10 km of it (70 dBZ). 60+ dBZ cores often mean large hail, but reflectivity is not a stone size.
NEXRAD reflectivity, San Antonio, Apr 12, 2016, 2016-04-13 02:15 UTC (OBSERVED radar)
20–30
30–40
40–50
50–55
55–60
60–65
65+ dBZ
IEM national NEXRAD base-reflectivity composite (N0Q). Frame: of the 5-minute frames from 20 min before to 5 min after the largest hail report of the storm day (article table), the one with the strongest echo within 10 km of it (70 dBZ). 60+ dBZ cores often mean large hail, but reflectivity is not a stone size.
NEXRAD reflectivity, Denver, May 8, 2017, 2017-05-08 21:10 UTC (OBSERVED radar)
20–30
30–40
40–50
50–55
55–60
60–65
65+ dBZ
IEM national NEXRAD base-reflectivity composite (N0Q). Frame: of the 5-minute frames from 20 min before to 5 min after the largest hail report of the storm day (article table), the one with the strongest echo within 10 km of it (71 dBZ). 60+ dBZ cores often mean large hail, but reflectivity is not a stone size.
NEXRAD reflectivity, Red Rocks, Denver area, Jun 21, 2023, 2023-06-22 03:15 UTC (OBSERVED radar)
20–30
30–40
40–50
50–55
55–60
60–65
65+ dBZ
IEM national NEXRAD base-reflectivity composite (N0Q). Frame: of the 5-minute frames from 20 min before to 5 min after the largest hail report of the storm day (article table), the one with the strongest echo within 10 km of it (78.5 dBZ). 60+ dBZ cores often mean large hail, but reflectivity is not a stone size.
MRMS MESH, also hail-driven Mar 12–14, 2024 (KS, OK, MO) (MODELED)
1–1.5 in
1.5–2 in
2–2.5 in
2.5–3 in
3 in +
MESH_Max_1440min, max over 3 convective days, 1 km grid, spike-masked; peak 3.35 in in the box. A radar estimate of maximum expected hail size, not measured stones.
MRMS MESH, also hail-driven Apr 27–28, 2021 (TX, OK) (MODELED)
1–1.5 in
1.5–2 in
2–2.5 in
2.5–3 in
3 in +
MESH_Max_1440min, max over 2 convective days, 1 km grid, spike-masked; peak 3.9 in in the box. A radar estimate of maximum expected hail size, not measured stones.
MRMS MESH, also hail-driven Apr 13–16, 2026 (MN, WI, IA) (MODELED)
1–1.5 in
1.5–2 in
2–2.5 in
2.5–3 in
3 in +
MESH_Max_1440min, max over 4 convective days, 1 km grid, spike-masked; peak 3.52 in in the box. A radar estimate of maximum expected hail size, not measured stones.
MRMS MESH, storm day Austin–Georgetown, Apr 15, 2021 (MODELED)
1–1.5 in
1.5–2 in
2–2.5 in
2.5–3 in
3 in +
MESH_Max_1440min for the convective day (24 h to 12 UTC), 1 km grid; peak within 60 km 2.9 in (the table's Peak MESH). A radar estimate of maximum expected hail size, not measured stones.
MRMS MESH, storm day Denton, Collin and Wise counties, Feb 21–22, 2022 (night) (MODELED)
1–1.5 in
1.5–2 in
2–2.5 in
2.5–3 in
3 in +
MESH_Max_1440min for the convective day (24 h to 12 UTC), 1 km grid; peak within 60 km 3.0 in (the table's Peak MESH). A radar estimate of maximum expected hail size, not measured stones.
MRMS MESH, storm day Collin County, Texas, May 19, 2023 (MODELED)
1–1.5 in
1.5–2 in
2–2.5 in
2.5–3 in
3 in +
MESH_Max_1440min for the convective day (24 h to 12 UTC), 1 km grid; peak within 60 km 3.1 in (the table's Peak MESH). A radar estimate of maximum expected hail size, not measured stones.
MRMS MESH, storm day south-central Minnesota, Aug 11, 2023 (MODELED)
1–1.5 in
1.5–2 in
2–2.5 in
2.5–3 in
3 in +
MESH_Max_1440min for the convective day (24 h to 12 UTC), 1 km grid; peak within 60 km 2.8 in (the table's Peak MESH). A radar estimate of maximum expected hail size, not measured stones.
MRMS MESH, storm day Austin–Georgetown, Sep 24, 2023 (MODELED)
1–1.5 in
1.5–2 in
2–2.5 in
2.5–3 in
3 in +
MESH_Max_1440min for the convective day (24 h to 12 UTC), 1 km grid; peak within 60 km 5.3 in (the table's Peak MESH). A radar estimate of maximum expected hail size, not measured stones.
MRMS MESH, storm day Dallas–Fort Worth, Mar 8, 2025 (after midnight) (MODELED)
1–1.5 in
1.5–2 in
2–2.5 in
2.5–3 in
3 in +
MESH_Max_1440min for the convective day (24 h to 12 UTC), 1 km grid; peak within 60 km 2.1 in (the table's Peak MESH). A radar estimate of maximum expected hail size, not measured stones.
