On day 59 of 1987 a well calibrated set of photometric observations of auroral emissions (427.8, 630.0, 844.6, and 871.0 nm) was obtained at Sondre Stromfjord, Greenland coincident with incoherent scatter radar measurements of electron density. Ratios of these optical emissions were used to correct for atmospheric extinction and to infer the average energy of the precipitating electrons and the deviations in the atmospheric composition induced by auroral heating. Previously, the radar data taken alone had been used to infer the average energy of the auroral particles. For the first time a comparison is made of the results from this optical ratio technique with those obtained by combining the optical data with the radar data and with results obtained using the radar data alone. Specifically, we compare the average energy inferred by these techniques as well as the changes in the atomic oxygen to molecular nitrogen ratio in the lower thermosphere. This study shows the critical importance of characterizing the atmospheric scattering if reliable results are to be derived. In particular, it is shown that on occasion the optical technique may be even more reliable than using radar and optics together since, because of light scattering, the radar and the optics are sometimes observing different parts of the aurora. Magnetometer data are used to estimate local Joule heating. The combined data are also used to address the question of whether observed composition changes were produced locally. ¿ 1999 American Geophysical Union