Abstract: Radio-frequency (RF) noise has typically been measured at a handful of fixed, representative locations within the urban environment (Achatz, Lo, Papazian, Dalke, & Hufford, 1998; Achatz & Dalke, 2001; Dalke, Achatz, Lo, Papazian, & Hufford, 1997; Wepman & Sanders, 2011; Wagstaff & Merricks, 2005; Spaulding & Disney, 1974). In this work, we discuss the development of a mobile RF noise measurement system and the necessary geospatial and statistical post-processing techniques required to characterize the variations in noise on the street-scale in the VHF sections (60 – 300 MHz) of the spectrum. We discuss the design of our mobile noise measurement system, with special focus on the choice and calibration of preselection filters and preamplifiers necessary to reliably measure low RF noise levels while avoiding intermodulation distortion problems that arise in an environment with many strong emitters. Additionally, we describe post-processing techniques developed to reliably merge and interpolate the RF data with geolocation data which are collected on two very different (microsecond and multisecond, respectively) timescales. We use a preliminary urban dataset from Boston, MA to show that the geo-statistical properties of RF noise power can vary appreciably over street-scale distances, and that these spatial variations are repeatable over tactically relevant times.