Some Problems of the total ozone of ultraviolet scattering inversion
|School||Graduate School, Chinese Academy of Sciences|
|Keywords||total atmospheric ozone the effective reflectivity of surface solar ultraviolet radiation back-scattered radiation forward scattered radiation|
The scientists in many fields are currently giving more and more attention to ozone and its continuous monitoring on a global scale. Now total atmospheric ozone derived from back-scattered ultraviolet radiation measured by satellite has been a main method to obtain total ozone data. Based on simulating calculations, several hyposthesises and parameters in current retrieval algorithms which may have important influence to the retrieval of total ozone are studied and discussed. At last the feasibility study of determining total atmospheric ozone by forward scattered ultraviolet radiation from the clear zenith blue sky is made. The main results are summarized as follows:(1) Within the limitation of TOMRADE model, the paper makes total ozone retrieval error analysis induced by ozone vertical distribution, variations of temperature profile, the reflectivity of effective reflecting surface, the height of cloud top and stray light et al. The results obtained: if the measuring wavelength pair is unchangeable the error induced by ozone vertical distribution is dependent on the solar zenith angle, total atmospheric ozone, the effective reflectivity of surface and the height of differences between the actual and the assumed ozone profile. If the solar zenith angle is not larger than 60°the influence of stratospheric ozone profile is less and the error of retrieval of total ozone is less than 1.5%. With the solar zenith angle larger and larger the error can be larger than 4%. The influences of stratospheric and tropospheric ozone profile are less and less when the effective reflectivity is larger and larger. The influence of temperature is weak and only when the temperature of the level of maximum ozone concentration is different between the actual and assumed temperature profile, the error is large. When the solar zenith angle is larger and larger the error is larger and larger. The error induced by temperature profile can be larger than 9%. If the assumed effective reflectivity of surface is higher or lower than the true one, the retrieval total ozone value is higher or lower than the true one accordingly. The error is less and less with the larger solar zenith angle. If the assumed effective pressure of surface is higher or lower than the true one, the retrieval total ozone value is lower or higher than the true one accordingly. The error is less and less with the larger solar zenith angle. Ozone errors due to accurate cloud height in retrieval include radiation interpolation error and ozone retrieval error above cloud and below cloud. The last two errors are larger comparing to the first. The ozone retrieval error above cloud and below cloud could not cancel each other, which will induce large errors. The largest error of retrieval of total ozone can be as large as 20% because of the influence of cloud top height. The error induced by two different models is larger and larger with the solar zenith angle and can be as large as 15%. The errors induced by the solar zenith angle and satellite zenith angle are larger and larger with secθ0 + secθ(θ0the solar zenith angle,θthe satellite zenith angle). The stray light can always make the retrieval total ozone value lower and the influence is stronger and stronger when the solar zenith angle is larger and larger. If the stray light can be not larger than times 0.1% then the error is not larger than 0.5%. Because of the departure of measuring wavelength center the error of retrieval of total ozone is from 0.84% to 1.11%. The change of the width of filter function can lead to 2% error. Because of the radiometric calibration the error is not larger than 2%. Im(360)(2) Current all TOMS ozone retrieval algorithms treat cloud as opaque Lambert surface and consider the effective reflectivity of cloud top height is independent on the wavelength. On the basis of atmospheric radiative transfer simulating calculations the paper shows that cloud is not opaque and radiation can penetrate into the cloud even though the optical thick cloud. In addition, the reflectivity of cloud top height is dependent on the wavelength because of the compositive action of cloud scattering, rayleigh scattering and ozone absorption. Simulating calculations using the V7 algorithm show the multiple scattering below the cloud top height, especially in cloud can enhance the ozone absorption path length. The effect can make the retrieval value of total ozone amount larger than the actual one. We call the phenomena“cloud absorption effect”and make a discussion on the influence parameters. At last, a new retrieval algorithm on the basis of simulating calculations is developed. The new algorithm can reduce the error induced by“cloud absorption effect”. (3) Within the limitation of TOMRADE model, it has been shown that useful determinations of total atmospheric ozone can be made from the forward scattered ultraviolet radiation from the clear zenith blue sky. Because of the complex interplay between penetration and sensitivity to ozone, there is an optimum wavelength for maximum sensitivity to ozone for a given ozone amount and path length. The wavelength based on measurements at a single wavelength should be selected from short to long as increases. Based on the accuracy of the intensity ratio (diffusely transmitted to extraterrestrial input flux) within 1%, reflectivity within 0.05,and surface pressure within 10 hPa, low total ozone and short would not produce better than 5% retrieval accuracy even though sΩ0sλ1 =308nm is selected. Measurements on a pair of wavelengths are found to reduce the effects of surface reflectivity and pressure and the retrieval results are in better accuracy. The wavelengthsλ1 =308 nm,λ2=3125 nm,λ3=3175 nm associate withλ4=3300 nm respectively and produce three pairs denoted by p1, p2, p3 respectively. P1 is selected when sΩ0≤1.5, p2 is selected when 1. 5< s ?0≤3.0, p3 is selected when s ?0 >3.0. Based on the same accuracy of the intensity ratio, reflectivity and surface pressure the accuracy of retrieval results is within 4% at least under the selection principle.