Since the end of WWII, weather radar has been used to observe and measure precipitation, including rain, hail and snow. Traditionally, weather radars operate in the S-band frequency range (about 3 GHz) with a corresponding wavelength of about 10 cm. The National Weather Service (NWS), Airforce Weather Agency (AFWA), and Federal Aviation Administration (FAA) currently operate an network of over 150 weather radars throughout the U.S. called NEXRAD (Next Generation Weather Radar System). The current generation of weather radars does not directly measure precipitation, only the energy reflected back to the radar from whatever the radar is looking at (rain, hail, snow, insects, etc). The presence of precipitation must be inferred by the amount of returned energy to the radar and, for hail in particular, the location of strong radar echoes in storms. Forecasters look for areas of high reflectivity in middle to upper levels of storms as a sign of large hail, and also look for signs of sustained updrafts that could keep hailstones aloft long enough to become large.
For some back ground on how weather radar works, check out:
Weather Radar Observations of Hail
To a weather radar, hail can look very big compared to a raindrop (typical raindrops are 1-3 mm in diameter whereas large hail can be more than 10 times that size). Large hail returns a lot of energy back to the radar (in radar meteorology language, hail has a large "backscatter cross section"), especially if it's exterior has a liquid coating. Radar reflectivity values exceeding 60 dBZ are not uncommon for large hail.
NWS forecasters use computer algorithms to examine the radar data and calculate statistical forecasts of the probability of hail and severe hail (greater than 3/4 inch in diameter), and a probable maximum hail size. These calculations are based on the strength of echoes above the freezing and -20 degree C levels in thunderstorms. The numbers are used as guidance, but their accuracy varies greatly. For very large hail, there is not so much of a problem as the extremely strong echoes are unmistakably large hail signatures. There is still much uncertainty for hailstones smaller than about two inches in diameter. One reason for the uncertainty is the size distribution of the particles. The same echo could be produced by a small number of large stones or a very large number of small stones, or even large raindrops. Another factor is the extent of melting below the part of the storm the radar beam is passing through. At distances of more than about 100 km from the radar, this can also be a major factor.
Detecting hail with conventional NEXRAD radar becomes problematic when the hail is small and/or mixed with rain. This is where additional information from storm spotters and CoCoRaHS volunteers (see link below) becomes very important so that forecasters can determine if a storm is likely to produce flash flooding or damaging hail.
Observing Precipitation With Polarimetric Radar
Polarimetirc radars can detect the presence of hail more accurately than conventional NEXRAD radar. To do this, polarimetric radars take advantage of the fact that raindrops larger than about 1mm are oblate - not spherical - when they fall. The reason they are not spherical is due to aeordymanic drop on the base of the drop surface which tends to flatten them out
into "hamburger bun" shapes. Oblate raindrops have a bigger "cross section" for horizontaly polarized radiation compared to the vertical direction. Thus, the ratio pf the received horizontal to vertical backscattered power(Differential Reflectivity -ZDR) is usually greater than one (click here to see how ZDR changes with drop shape).
In contrast, hail tends to be spherical in shape. Even if its not spherical (conical is also a common shape for hail), hail tumbles when it falls so it looks spherical to the radar. That means that the ratio of back scattered power at horizontal and vertical polarization is typically near zero (smilar to small raindrops).
CoCoRaHS has worked closely with the CSU-CHILL radar facility near Greeley, Colorado to provide ground validation of rain, hail, and snow precipitation measurements.