| Nitrous oxide(N2O)is an important anthropogenic greenhouse gas(GHG)in the atmosphere,except for carbon dioxide(CO2)and methane(CH4).The Total Carbon Column Observatory Network(TCCON)offers long-term global distributions of N2O molar fractions,but it suffers from comparatively expensive maintenance,mobility inconvenience,and a lack of infrastructure.In contrast,the portable ground-based Fourier transform infrared spectrometer EM27/SUN has a relatively higher sensitivity for near-surface observations,allowing continuous measurements in typical areas and serving as a valuable data supplement.In this thesis,the total columns of N2O in Hefei are overserved with EM27/SUN.The daily variation pattern of N2O column concentration was obtained by the established PyQt inversion algorithm system,which combines the optimal estimation method.Moreover,we analyzed the factors that influence the inversion results and compared the observed data results with the TCCON site for verification.The main work and conclusions are as follows:(1)The line-by-line integrated radiative transfer model(LBLRTM)was used to pick a better N2O absorption window in the near-infrared band,which was 4682.5～4756.1 cm-1.We investigated the fundamental process of the PROFFAST inversion algorithm,developed the PyQt inversion algorithm system to process the spectral data,and got the daily variation pattern of XN2O from September 2021 to December 2022.The transport characteristics analysis of the gas stream was carried out in conjunction with the backward trajectory model HYSPLIT for column concentration variation.(2)On the XN2O variation,the effects of a-priori profiles(temperature,pressure and molecular concentration),near-surface pressure variation,and other extraneous factors were studied.When the a-priori temperature profile increase by 1 K,XN2O decreases by around 0.09 ppb.XN2O increases by no more than 0.008 ppb with a priori atmospheric pressure profile increase by 1 hPa.For both values obtained by using constant pressure and daily variable pressure,the XN2O change error range is between-0.20 and 0.27 ppb,respectively.Near-surface pressure loss of 1 hPa causes a 0.24 ppb drop in XN2O.Particularly,the long short-term memory network LSTM approach in deep learning was used to fit the data for prediction and analysis,and superior prediction results were obtained.(3)The column concentration results were verified using the high-resolution TCCON site observations.It demonstrated the correlation between EM27/SUN(resolution is 0.5 cm-1)and TCCON retrievals from the co-located IFS 125HR spectrometer(highest resolution is 0.00096 cm-1).Further,we extracted empirical correction factors to improve the accuracy of the EM27/SUN inversion.The highest relative error of the EM27/SUN inversion results decreased from 2.41%to 1.36%.(4)The seasonal variation of XH2O,XCO2,XCH4 and XCO in Hefei was analyzed.We investigated the relationship between them and XN2O as well as their sensitivity to the changes in pressure,temperature,and molecular a-priori profiles.The findings show that CO and H2O increase positively as the temperature profile rises by 1 K,while CH4 and N2O rise negatively.When adopting an average temperature profile,CO is more sensitive to changes in column concentration than other greenhouse gases.N2O is most sensitive to pressure fluctuations as the near-surface pressure rises by 1 hPa.In addition,XN2O has a lower correlation with changes in column concentrations of other GHGs in the examination of correlations. |
PDF Full Text Request