This page displays images of echo statistics created using program package by Donaldson and Patterson. Once a month a script runs through all raw radar data file for each radar and makes URP meta files for selected products. The ones that are done every month are PPIs of Doppler corrected reflectivity (DBZ) and uncorrected reflectivity (DBT). Occassionally the CLOGZ product is run for the S-Band radars. For old C-band radars the products were CLOGZ-Dopvol1A, PRECIP, and CONVPPI-lowest PPI. A followup program collects the meta files in 5 day long collections and builds distributions of observed reflectivities at every pixel. The analysis program gathers up distributions for a particular year/month and derives multiple statistics: mean reflectivity, median reflectivity, mean rainfall rate (MP), percent of NULL observations, percentage of observations above a threshold, etc. Here a small fraction of the possibilities are shown. Most products have "CLOSE" and "LONG" options that display to 110km and 220km respectively. However "LONG" is often not run.

Data is displayed in polar coordinates: x - azimuth, y - range. For example points in a single azimuth direction form a vertical line, so blockage by a tower would be a vertical line of low values. Radars are roughly organized east to west, but are also grouped by S-band and C-band.

Thinking in polar coordinates is hard if you are trying to decide where something is geographically. One solution to change that from hard to merely difficult is to choose a product and then used the pulldown to display "feature map". Then toggle back and forth. On the feature map: red = roads and airport runways, blue = geographic feature (mainly water boundaries but things like Niagara escarpment are in the category too!) white=borders, orange=railways, green = power corridors, and magenta dots = wind turbines. Some maps have dotted lines, which are extensions of airport runways, shown since air traffic on those lines appears in radar statistics.

The units of the distribution have no simple interpretation. The total counts of each DBZ byte value at each range are divided by the number of files and an empirical scaling factor is applied to give values that run in the range 0 to 100, as a convenience for plotting. Intercomparison of plots shows meaningful changes, even if individual plots do not have meaningful values.