36 37 A fast method is presented for deriving the tropospheric CO concentrations in the Venus 38atmosphere from near-infrared spectra using the night side 2.3 µm window. This is validated using 39 the spectral fitting techniques of Tsang et al. (2008a) to show that monitoring CO in the deep 40 atmosphere can be done quickly using large numbers of observations, with minimal effect from 41 cloud and temperature variations. The new method is applied to produce some 1,450 zonal mean 42 CO profiles using data from the first eighteen months of operation from the Visible and Infrared 43Thermal Imaging Spectrometer infrared mapping subsystem (VIRTIS-M-IR) on Venus Express. 44 These results show many significant long and short-term variations from the mean equator-to-pole 45
A C C E P T E D M A N U S C R I P T ACCEPTED MANUSCRIPT2 increasing trend previously found from earlier Earth-and space-based observations, including a 46 possible North-South dichotomy, with interesting implications for the dynamics and chemistry of 47 the lower atmosphere of Venus. 48 491. Introduction 50 51The concentration of carbon monoxide (CO) in the troposphere (surface to 40km) of Venus was 52 first proposed to be retrievable through modeling by Kamp et al. (1988). This is possible through 53 the observation of the thermal emission window at 2.3 µm, where the radiation is escaping from the 54 deep atmosphere. For a thorough review of this topic, see Taylor et al. (1997) and Tsang et al. 55 (2008b). The first measurements of CO using this window were made by Bézard et al. (1990) from 56 CFHT observations. Pollack et al. (1993) also conducted ground-based observations to measure the 57 mean abundance of CO in the troposphere. However, the first attempt to measure spatial variations 58 in CO at 35 km was by Collard et al. (1993), using 2.3 µm spectra obtained by the Near Infrared 59Mapping Spectrometer (NIMS) on the Galileo spacecaft. 60 611.1 Collard et al. 1993 Method 62 63 These authors used Galileo/NIMS spectra at 2.3 µm from the fly-by of Venus in 1990. Rather than 64 using a spectral fitting technique, Collard et al. (1993) used ratios of radiances at two different 65 wavelengths. Between 2.20 and 2.30 µm, the absorption is purely due to CO 2 and cloud opacity, 66 whilst at 2.30 to 2.43 µm, the absorption is due to strong vibrational-rotation CO bands as well. The 67 wavelengths chosen by the authors were 2.252 µm, outside the CO band, and 2.330 µm, with strong 68 CO absorption. A distinct correlation is observed between these two wavelengths, which is due to 69 the optical depth of the cloud layer, as one might expect. However, an off-branching set of points 70 was also seen beneath the main branch. This was interpreted as being due to the increase in the CO 71 abundance at these locations on the planet, causing the increased absorption at 2. The premise for this analysis was to repeat the experiment of Collard et al. 1993 described in 111 Section 1.1, by taking radiation emitted at 2.30µm, which should be sensitiv...