Winter storms can cause significant societal impacts in the densely-populated regions of the Northeast United States. Mesoscale snow bands embedded within winter storms are often the main focus of snowfall forecasts and analyses. This study investigates the relationship between observed surface snowfall rates and local enhancements in radar reflectivity (i.e. mesoscale snow bands) using data from 264 storm days over 11 winter seasons (2012–2023). We compare hourly surface snowfall rates obtained by ASOS weather stations with the area × time fractions of locally-enhanced reflectivity features and of all echo passing over the 25 km radius vicinity of the surface observation. Our analysis focuses on non-orographic snow storms with surface winds < 5 m s<sup>−1</sup>.</p> <p>Our findings show that most of the time snow rates are low (75 % of hours had liquid equivalent snow rates less than 1 mm hr<sup>−1</sup>). Heavy snow rates (> 2.5 mm hr<sup>−1</sup> liquid equivalent) are rare (< 4 % of observations). When enhanced reflectivity features pass over a location, only 1 out of 4 hours have heavy surface snow rates. High spatial resolution vertical cross sections from airborne radar obtained during the NASA IMPACTS field campaign and rapid update RHIs from ground-based radar demonstrate that enhanced reflectivity features in snow aloft usually lack the vertical column continuity characteristic of reflectivity structures in rain. Ice streamers with higher reflectivities are tilted and smeared on their way to the surface as their constituent snow particles are dispersed laterally by the horizontal winds within the storm.