A tall tower flux measurement setup was established in metropolitan Houston, Texas, to measure trace gas fluxes from both anthropogenic and biogenic emission sources in the urban surface layer. We describe a new relaxed eddy accumulation system combined with a dual-channel gas chromatography - flame ionization detection used for volatile organic compound (VOC) flux measurements in the urban area, focusing on the results of selected anthropogenic VOCs, including benzene, toluene, ethylbenzene and xylenes (BTEX), and biogenic VOCs including isoprene and its oxidation products, methacrolein (MACR) and methyl vinyl ketone (MVK). We present diurnal variations of concentrations and fluxes of BTEX, and isoprene and its oxidation products during summer time (May 22 - July 22, 2008) and winter time (January 1 - February 28). The measured BTEX values exhibited diurnal cycles with a morning peak during weekdays related to rush-hour traffic and additional workday daytime flux maxima for toluene and xylenes in summer time. However, in winter time there was no additional workday daytime peaks due mainly to the different flux footprints between the two seasons. A comparison with different EPA National Emission Inventories (NEI) with our summer time flux data suggests potential underestimates in the NEI by a factor of 3 to 5. The mixing ratios and fluxes of isoprene, MACR and MVK were measured during the same time period in summer 2008. The presented results show that the isoprene was affected by both tail-pipe emission sources during the morning rush hours and biogenic emission sources in daytime. The observed daytime mixing ratios of isoprene were much lower than over forested areas, caused by a comparatively low density of isoprene emitters in the tower's footprint area. The average daytime isoprene flux agreed well with emission rates predicted by a temperature and light only emission model (Guenther et al., 1993). Our investigation of isoprene's oxidation products MACR and MVK showed that both anthropogenic and biogenic emission sources exist for MACR, while MVK was strongly dominated by a biogenic source, likely the isoprene oxidation between the emission and sampling points.