Sign inRegisterCheck connection

NEXRAD Level III Products and GRLevel3 as a Display Client

NEXRAD Level III products are generated radar images, not a full WSR-88D volume. What GRLevel3 displays as a client, and what it omits.

Back to Historical weather observations and forecasts

GRLevel3 is a Windows viewer for NEXRAD Level III products. It is not a radar. It does not sit on a tower in Florence or Norman. It does not replace the WSR-88D volume scan. It is a display client: software that requests already-generated products, draws them on a map, and can overlay warnings and reports. The historical TNET Weather URL /radar/grlevel3 pointed visitors at that workflow—Level III images, often of the Phoenix area, produced on a schedule and shown on a station site. This article explains what Level III includes, what it does not, and how a display client fits a personal weather page. It is not a GRLevel3 license page, download mirror, or key-recovery service.

TNET Weather does not rehost GRLevel3 output, NEXRAD files, or Gibson Ridge binaries. Current official imagery belongs on NWS Radar. TNET is not a warning service.

Historical context

In the mid-2000s, PWS operators who wanted more than a national GIF used Level III feeds. NOAA’s Radar Product Generator turns each volume into a catalog of products. Gibson Ridge Software’s GRLevel3 became a common hobby tool for pulling those products from RPCCDS and rendering them locally. Operators then exported stills or loops to FTP. Those files are not current weather and are not restored here. What remains useful is the product distinction the filename was trying to name.

Level II is the volume; Level III is the product

NOAA NCEI’s NEXRAD documentation splits the archive the way a scientist should split a PWS caption:

Level-II (base) data are the meteorological base quantities from the volume: reflectivity, mean radial velocity, and spectrum width, plus dual-polarization base fields (differential reflectivity, correlation coefficient, differential phase). Files typically hold several minutes of base data depending on the volume coverage pattern. That is the three-dimensional sample the radar actually collected.

Level-III products are generated from that base data by the Radar Product Generator. The NEXRAD Radar Operations Center describes Level III as the output product data of that generator, provided in near real time to RPCCDS and archived. NCEI states that over 75 Level-III products are routinely available. Most are already mapped, color-scaled, and suited to a two-dimensional display. They assist analysis, forecasts, warnings, and tracking. They are not a substitute for opening the volume.

A GRLevel3 window that shows “BR” or “N0Q” is showing a product. A mosaic on a consumer website is usually a further derived composite of many products. Neither is the Level-II volume. If you need every tilt, dual-pol moment, and gate, you are in Level II (and in a different Gibson Ridge program, GRLevel2 / GR2Analyst, which this URL is not about).

What Level III typically includes

The following are the families a PWS page of this era actually published, described from NCEI product notes rather than from a recovered screenshot.

Base reflectivity. Echo intensity in dBZ at a stated elevation. Classic products include 124 nmi and 248 nmi ranges (N0R/N0Z-class and their digital successors N0Q-class). Four low elevation angles are commonly available; exact angles depend on the scan strategy. This is the “rain hitting the ground” thumbnail only in a loose sense: it is returned power on one tilt, not a gauge.

Composite reflectivity. The maximum reflectivity from all scanned heights in the volume (NCR/NCZ-class). It answers “how strong is the strongest echo in the column?” It does not tell you at which height that maximum lived. A composite can look more filled-in than the lowest tilt because it can pull elevated cores into the picture.

Echo tops. Height of the echo top (NET/EET-class). Useful for a quick look at deep convection. Echo top is a radar-derived height of detectable return, not the thermodynamic tropopause and not a cloud-top infrared temperature from GOES.

Base velocity and storm-relative velocity. Radial wind toward or away from the radar; storm-relative versions subtract estimated storm motion. These are Doppler products. They are still Level III images, not a full dual-Doppler analysis.

Precipitation accumulations and VIL. One-hour and storm-total precipitation, and vertically integrated liquid, are algorithm outputs. Caption them as derived. They are not the station’s rain.

Hydrometeor classification. Dual-pol products compare polarimetric moments to categories (rain, hail, biological, clutter). NCEI describes them as most-likely echo sources, not a laboratory ID of each drop.

GRLevel3, as Gibson Ridge documents it, displays high-resolution base products, dual-polarization products, and derived products, and can show local storm reports and severe-weather warnings. Warnings on that map are NWS polygons. Drawing them in a hobby viewer does not make the hobby site a forecast office.

What Level III does not include

State these limits in any caption that says “Level III” or “GRLevel3”:

  • Not the raw volume. You do not have every gate of every moment at every tilt unless you pulled Level II.
  • Not a national mosaic. A single-site product is polar, centered on one antenna. Stitching many radars is a different chain; see third-party mosaics.
  • Not satellite. Infrared and water-vapor imagery are GOES radiance. A lightning overlay in GRLevel3 is a display layer, not a new sensor. See lightning and satellite imagery.
  • Not ground truth at the funnel. Reflectivity is not millimeters. VIL is not a bucket.
  • Not a warning decision. Official alerts stay on weather.gov.
  • Not TNET’s imagery. This restored URL does not generate or refresh Level III frames.

Range matters inside a single product. A 124 nmi base-reflectivity image and a 248 nmi image are not the same map with extra margin. The long-range product trades resolution and samples a higher beam at the edge. Operators who exported “CR248” (composite, 248 nmi) and “BR1” (base reflectivity, lowest tilt) were showing two products, not two cameras on one storm.

GRLevel3 as software, not as a data source

Treat the program the way you treat a plotting library. Provenance is: WSR-88D → Radar Product Generator → Level III file → GRLevel3 (or another client) → PNG on a website. The scientific source is the NWS product, identified by site ID, product code, and time. “Generated with GRLevel3” is a processing credit, not an independent observation.

Gibson Ridge publishes current GRLevel3 information at grlevelx.com. TNET is not affiliated with Gibson Ridge, does not sell licenses, and does not host the installer. Current radar belongs on NWS; archives belong on NCEI NEXRAD or NOAA Open Data Dissemination, not a 2005 station thumbnail.

Export cadence is a publishing choice. The radar’s volume time is the observation time. FTP upload time is the page time. A 48-frame loop is a historical sequence of products. Animating them does not create a continuous field between volumes.

Practical reading rules for a Level III panel

Read the product code and site ID before the colors. Separate base (one tilt) from composite (column maximum) from algorithm (VIL, accumulations, class). Keep UTC on the image. Do not mix a Phoenix KIWA product with a national mosaic in one unlabeled stack. If clutter filtering is on, say so; it is derived. Link the live view to radar.weather.gov instead of a personal cache that will go stale.

The physics of reflectivity and beam height are on radar as remote sensing. How KIWA’s beam sits over the Phoenix metro is on Phoenix-area NEXRAD.

Modern relevance

Level III is a lesson in evidence type. A generated product is still observed radar in origin, but it is already processed. TNET’s public vocabulary is how the service works. Precipitation structure belongs among the atmospheric evidence families, not as a GRLevel3 screenshot and not as a promised connection speed. The Horizon study review may be read for how independent work relates weather to Starlink conceptually; it does not turn a Level III color table into a throughput forecast.

Historical observation pages: observations hub.

Sources