MRMS, Explained

In short: MRMS is NOAA’s blend of every radar in the network into one national grid, updated about every two minutes. It is the right product for the big picture, for rainfall totals, and for the history of rotation and hail along a storm’s path. For the shape of one storm, a single radar’s own scan is still better.

Every limitation of a single radar (the rising beam, the cone of silence, the blind spots between sites) has the same fix: use all the radars. That’s MRMS, the Multi-Radar Multi-Sensor system developed by NOAA’s National Severe Storms Laboratory and run operationally since 2014.

MRMS ingests roughly 180 U.S. and Canadian radars, plus surface observations, lightning detection, rain gauges, and model analyses, and fuses them into a seamless grid covering the country at about 1 km resolution, updating every ~2 minutes. Forecasters use it constantly. Few consumer weather apps show you more than its reflectivity mosaic, which is a shame, because several MRMS products answer questions single-site radar simply cannot.

How it differs from the radar you know

• No cone of silence, and far fewer gaps. Where one radar’s beam is too high or blocked, a neighboring radar fills in, and the mosaic always uses the best available view of each grid cell. Across the Intermountain West the radars are far enough apart that much of the ground between them is sampled only well above 10,000 ft, so the mosaic inherits the overshoot rather than curing it.
• It’s an analysis, not a sweep. MRMS grids are computed from quality-controlled ingredients: clutter and biology largely scrubbed, beam blockage corrected. The tradeoff is a touch of latency and smoothing versus raw Level 2 data from your closest site.
• Rule of thumb: storm structure (hooks, couplets) → single-site base data. Storm history, totals, and wide-area awareness → MRMS.

The products worth knowing

Seamless composite reflectivity

The national precipitation picture with each pixel drawn from the optimal radar. This is the product to leave on for general situational awareness, especially if you live 80+ miles from your nearest radar site, where single-site data quietly degrades (see What Radar Can’t See).

Rain or snow

MRMS labels each pixel rain or snow from the surface air and wet-bulb temperatures in the RAP and HRRR model analyses: snow where the wet-bulb temperature is below 0 °C (32 °F) and the air temperature below about 2 °C (36 °F), rain otherwise. It is a temperature rule applied to radar echo, not a radar measurement of type, so the line can sit a few miles off near the changeover. MRMS has no freezing rain or sleet class. MesoRadar’s Rain / Snow / Ice view adds one Ice class (freezing rain or sleet) from the HRRR model: its freezing rain and sleet flags, plus a temperature-profile check on its air aloft, and only where MRMS itself sees precipitation. The legend marks Ice as model-assisted. How the warm layer aloft decides between snow, sleet and freezing rain is in Snow, Sleet & Freezing Rain.

In MesoRadar: Precipitation Type (rain / snow) and Rain / Snow / Ice, on the National tab.

Rotation tracks

MRMS continuously computes azimuthal shear, the strength of rotation in the velocity field, at low levels (roughly 0 to 2 km) and mid levels (3 to 6 km), then accumulates the maximum over trailing windows of 30, 60, 120, 240 and 360 minutes, plus a 24-hour version.

The result is a map of where rotation has been: dark, narrow streaks tracing every mesocyclone’s path. Uses that make this genuinely practical:

• “Did that rotation pass over my house?” After a warned storm, the track answers it to within a kilometer or two near a radar, and less precisely far from one.
• Post-event analysis. NWS damage-survey teams use rotation tracks to decide where to look. You can replay an outbreak the next morning and see every circulation’s full path.
• Live line-storm awareness. During a QLCS event with brief spin-ups, tracks reveal which segments of the line have been producing circulations.

A caution: a rotation track records rotation aloft, not confirmed tornadoes. Strong track ≠ tornado path; it’s a where-to-pay-attention product.

MESH, and which way it is wrong

Maximum Estimated Size of Hail, computed across multiple radars from the reflectivity above the freezing level, weighted so that echo colder than minus 20 degrees counts fully and echo between 0 and minus 20 counts in part (the Witt et al. 1998 method). MESH paints swaths showing estimated maximum hailstone size along a storm’s entire path.

Two uses:

• Ahead of time: which storms are the hailers, and how big.
• Afterward: hail swath maps showing which neighborhoods took 1.5″ stones. Roofers, adjusters, and homeowners filing claims use MESH swaths for precisely this. It is a statistical fit between reflectivity aloft and reported hail, and it depends on a model temperature profile. It is an estimate, and by design a high one. Witt and colleagues fit MESH so that about three quarters of observed maximum hailstones are smaller than the estimate. Comparisons with hail reports find every version least reliable as stones approach 4 inches (Wendt and Jirak 2021). Verify with ground truth. It is still a strong first answer to “was my street hit?”

Precipitation rate and QPE

MRMS quantitative precipitation estimation blends dual-pol rainfall algorithms (leaning on KDP) with gauge corrections into rolling rainfall totals: 1, 3, 6, 24, 48, 72 hours. During flash-flood events this is the product: reflectivity shows intensity right now, but flooding is about accumulation, and QPE shows where 3 to 6″ has already fallen. Pair it with the training-cells concept from How to Read Weather Radar. Like the other MRMS analysis products, the accumulation totals are part of Pro in MesoRadar; the live mosaic and precipitation rate are free.

Echo tops and heights

18 dBZ echo-top height approximates storm depth: a quick proxy for updraft strength when comparing cells. Pilots watch it to route around a storm rather than through it.

A practical MRMS workflow

1. Quiet day: seamless composite for the big picture; QPE if anything is training.
2. Severe watch: composite for awareness; jump to single-site base reflectivity and SRM on any cell that organizes (structure needs the raw data; see Hook Echoes).
3. During warnings: rotation tracks to see each circulation’s history and heading; MESH to size up the hail threat along its path.
4. Next morning: 24-hour rotation tracks and MESH swaths to review what happened. And whether it happened to your roof.

In MesoRadar the mosaic is the National tab; rotation tracks, MESH and the rainfall accumulations are in the same product picker. Where to find them.