Classical studies of urban meteorology have concentrated on the interpretation of data from single sites, investigated exhaustively. Here, surface roughness layer (SRL) data from six major US cities in a nation-wide network of urban sites are examined. It is found that the velocity variances in time, derived by averaging the conventional variances from a network of n stations, are nearly equal to the velocity variances in space, derived as the variances among the n average velocities. This near equality (referred to here as “quasi-ergodicity”) is modified during sunlit hours, when convection preferentially elevates the former due, presumably, to enhanced coupling with winds above the SRL. Although the data show no significant overall dependence of the ratio of these two variances (expressed here as Rst, the ratio of the corresponding velocity standard deviations) on wind speed, this is not the case when data nearest local noon are considered. In such conditions, when convection should be most active, the data show a dependence of Rst on the reciprocal of the wind speed that is statistically significant on most occasions. Analysis of dependencies on the average rate of change of temperature with time is also carried out. It is concluded that the average state of the SRL in urban and suburban areas tends towards ergodicity, in much the same way has been observed elsewhere for the case of forests.