The concentration of hydroxyl radicals is estimated from the degradation of NOx and selected volatile organic compounds (VOCs) during the transport of polluted air masses from the city of Freiburg to the Schauinsland mountain through a steep valley called Gro¿es Tal. The approach is based upon chemical and meteorological measurements made during the Schauinsland Ozone Precursor Experiment (SLOPE96) at two ground-based sites and aboard a small aircraft. Separation of chemistry and transport is achieved through the degradation of chemical compounds with significantly different reactivity towards OH. The transport time of the air between the two measurements (90¿5 min) and the influence of mixing with background air on the calculated OH concentration is quantified with the help of a dispersion experiment with SF6 and from airborne chemical measurements. The OH concentration (7¿106 cm-3) is almost a factor of 2 larger than what is calculated with a chemical box model constrained by the measured trace gas concentrations and photolysis rates. The radical budget can be closed within the experimental uncertainties, when an upper limit is adopted for the photolysis of nitrous acid, in addition to the radical production from photolysis of ozone, H2O2, and carbonyl compounds as well as ozonolysis of unsaturated VOCs. Biogenic VOCs (i.e., isoprene, terpenes, and a number of oxidated compounds) comprise about half of the total VOC reactivity in the transported plume. The results from SLOPE96 confirm the assumptions made in the analysis of an earlier experiment conducted in 1992, when noontime OH concentrations of 6--8¿106 cm-3 were derived in the presence of NOx mixing ratios between 70 ppb at the entrance of the valley and 15 ppb at Schauinsland. Comparison with direct measurements from different studies qualitatively reveals the expected dependence of on the NO2 mixing ratio with a maximum around 1--2 ppb of NO2. ¿ 2000 American Geophysical Union