The existence of detached/isolated auroral structures in the subauroral ionosphere has been recognized and studied for decades. One major subset of such detached auroras is the so-called “isolated proton aurora” (IPA). IPA is characterized by substantial hydrogen emissions and thus inferred to be proton aurora, but is also accompanied by other emission lines. In particular, IPA is usually dominated by 557.7 nm green-line emissions in optical luminosity. To date, there is still a lack of dedicated spectrographic study and detailed comparison among structures in different emission lines of IPA. The intensity ratios between the 557.7 nm and hydrogen emissions in IPA have not been well established from concurrent observations or theoretical models. In this study, we report an IPA event during ∼0245–0345 UT on 12 October 2021, the Canada Thanksgiving storm night. Using multi-station, multi-wavelength optical instruments, including the Transition Region Explorer (TREx) spectrograph, we investigate the evolution and spectrographic properties of the IPA. In-situ and ground magnetometer data show evidence of electromagnetic ion cyclotron (EMIC) waves associated with the passage of IPA, supporting a causal link between the EMIC wave and the proton precipitation. The precipitating proton energies are estimated to range between a few keV and a few tens of keV according to the IPA emission heights inferred from triangulation analyses. Via careful examination of the spectral intensities and the elevation-angle profiles of the 557.7, 427.8, and 486.1 nm emissions based on the spectrograph data, we conclude that the 557.7 nm emissions contained in the IPA were unlikely to owe their source to energetic electron precipitation from the magnetosphere, but were the byproduct of the proton precipitation. The intensity ratio between the 557.7 nm (427.8 nm) and the 486.1 nm emissions of the IPA are confined within a relatively narrow range around ∼26 (∼4), which may serve as validation tests for existing and developing proton transport models.