Progress and Challenges of Ozone Satellite Remote Sensing Inversion

Fig. 1. Schematic diagram of ozone satellite remote sensing detection technology^［65］. (a) Nadir observation; (b) occultation observations; (c) limb-viewing

Download full size

Fig. 2. Main ozone satellite detection sensors (ultraviolet sensors and infrared sensors)

Download full size

Fig. 3. Detailed explanation of retrieval algorithm for total ozone column

Download full size

Fig. 4. Main sensors and applications of ozone profile retrieval in recent years

Download full size

Fig. 5. Analysis of VOC_s and NO_x control zones in typical areas (Diagonal line in figure is ‘ridge line’)

Download full size

Table 1. Specific parameters of commonly used satellite sensors for ozone detection

View table

Table 1. Specific parameters of commonly used satellite sensors for ozone detection

Sensor	Satellite platform	Research institution	Satellite transit time	Temporal resolution /d	Spatial resolution /km	Spectra range /μm	Service period
GOME	ERS-2	ESA	10：30	3	40×320	0.24-0.79	1995-2003
GOME-2	METOP-A	ESA	9：30	1.5	40×80	0.25-0.79	2012-2023
SCIAMACHY	ENVISAT	ESA	10：00	6	30×60	0.24-2.40	2002-2012
OMI	AURA	NASA	13：45	1	13×24	0.27-0.50	2004-2023
OMPS	Suomi-NPP	NASA	13：30	1	50×50	0.25-0.42	2011-2023
EMI	GaoFen5	China	13：30	1	13×48	0.3-0.5	2018-2023
TROPOMI	Sentinel-5P	ESA	13：30	1	3.5×5.5	0.27-0.32， 0.310-0.495， 0.675-0.775， 2.305-2.385	2017-2023
TOU	FY-3A	China Meteorological Administration	13：40	1	50×50	0.30-0.36	2008-2018
	FY-3B						2010-2023
	FY-3C						2013-2023
AIRS	AQUA	NOAA	13：30	1	14×14	3.74-15.4	2002-2023
CrIS	Suomi-NPP	NASA	13：30	1	14×14	3.9-15.4	2011-2023
TES	AURA	NOAA	13：45	16	0.5×5	3.3-15.4	2004-2023
IASI	Metop-A	ESA	21：30	1.5	80×40	3.63-15.50	2006-2023

Table 2. Application analysis of accuracy verification of satellite ozone observation data

View table

Table 2. Application analysis of accuracy verification of satellite ozone observation data

Source of validation data	Application analysis
Surface	Bian et al^［108］ used observation data of Dobson and TOMS to analyze the long-term trend of total ozone in Beijing and Kunming from 1979 to 2000，and found that the two data had high consistency in the measurement of total ozone.
	Chen et al^［109］ analyzed the ozone concentration from ground-based stations and TOMS ozone observation data in Shanghai. The results show that their trends are similar（correlation coefficient is 0.81），but the ground-based results are lower than the results of TOMS data.
	Hong et al^［110］ compared and analyzed the total column ozone obtained from OMI-DOAS algorithm with ground-based observations in Seoul，and found that the seasonal average of ground-based observations was underestimated by 2.68%.
Radiosonde	Cai et al^［111］ used ozone sounding data to verify the ozone profile and tropospheric ozone column concentration of GOME satellite in Lhasa，Xining and Beijing. The results show that the monthly average ozone concentration in the lowest layer（0-2.5 km）of the GOME satellite has a good correlation with the ground observations. In addition，the average deviation of the tropospheric column concentration is within 10%.
	At Zhongshan Station，Amundesen-cott South Pole Station and Neumayer Station，Zhang^［112］ verified the correlation between the data of ozone sounding and the ozone vertical profile of the sixth edition of AIRS in 2008，and found that the two data showed similar trends after analysis.
	Chen et al^［113］ verified OMPS ozone profile using ozone sonde data from 2016 to 2018 in Beijing. The results show that the relative deviation between the ozone vertical distribution of OMPS and the ozone sonde data is less than 10% in the middle and upper stratosphere，but larger（15%-40%）in the middle and upper troposphere.
Satellite observation	Wang et al^［114］ used DOAS and multiplicative algebraic reconstruction to extract vertical ozone profiles from atmospheric limb scattering measurements. Comparing the extracted ozone profile with the ozone measurement of SCIAMACHY，the deviation of the two data is less than ± 10%.
	Van Peet et al^［115］ improved the ozone profile retrieval algorithm（OPERA）and applied it to GOME and GOME-2 satellites. The ozone retrieval results of the two satellites are analyzed and verified with the data of the ultraviolet radiation data center（WOUDC）. It is suggested that the relative difference in the troposphere is 20%，while it in the stratosphere is 15%.
	Zawada et al^［116］ developed a two-dimensional tomography algorithm based on OMPS-LP，and compared the dataset of the developed algorithm with the MLS observations. Within 20 minutes，the difference in the stratosphere is less than 5%.
	Ma^［8］ developed a global tropospheric ozone dataset with long time series for OMI/MLS and TOMS/SBV，which improved the consistency of original ozone data and ground-based data by 44.89%.