This thesis presents the development of four different remote-sensing instruments dedicated to atmospheric research and their use in field campaigns between 2008 and 2012. The instruments are based on uv-visible spectrometers and installed respectively on a scientific aircraft, ultralight aircraft, and cars. One of the instruments is targeted to operate from an Unmanned Aerial Vehicle (UAV). The Differential Optical Absorption Spectroscopy (DOAS) technique is used to quantify the molecular absorption in the spectra of scattered sky light. These absorptions are then interpreted by modeling the transfer of radiation in the atmosphere. Airborne platforms enable new measurement geometries, leading for instance to a high sensitivity in the free troposphere. On the other hand, a miniaturization effort is required, especially for the instruments onboard ultralight aircraft and UAV. Reaching the limited size, weight, and power consumption is possible through the use of compact spectrometers and computers, together with custom built electronics circuits and housings. A common target of the different experiments is to quantify tropospheric nitrogen dioxide (NO2). Regarding this trace gas, the developed instruments provide complementary findings, such as the vertical distribution in the pristine Arctic or the levels in the exhaust plumes of large cities like Riyadh. Car-borne measurements in North-West Europe reveal the horizontal gradients of surface NO2 at various scales. The UAV payload is intended to produce high spatial resolution maps of tropospheric NO2 columns.
Acknowledgements iii
Abstract v
List of Figures xiii
List of Tables xv
1 Introduction 1
2 Some aspects of the Earth atmosphere 5
2.1 Atmospheric composition and dynamics 5
2.1.1 Trace gases and aerosols 5
2.1.2 General vertical structure 9
2.1.3 A zoom in the troposphere 11
2.1.4 Transport in the atmosphere 12
2.2 Nitrogen oxides in the troposphere 14
2.2.1 NO2 in air quality 15
2.2.2 NO2 and tropospheric ozone 16
2.2.3 Sources and sinks of NOx 19
2.2.4 Time and space patterns in the NO2 _eld 20
3 Scattered-light DOAS 23
3.1 Di_erential Optical Absorption spectroscopy 23
3.1.1 Light absorption by molecules 23
3.1.2 Spectroscopy of the uv-visible scattered-light sky 27
3.1.3 Principle of DOAS 29
3.2 DOAS in practice 31
3.2.1 Ring E_ect 31
3.2.2 Resolution of the linear problem 32
3.2.3 Nonlinearities due to o_set and shift 34
3.3 Radiative transfer models 35
vii
viii CONTENTS
3.3.1 The radiative transfer equation 35
3.3.2 The discrete ordinates method 38
3.3.3 UVSPEC-DISORT 39
3.3.4 Air mass factors and weighting functions 40
4 Instrumental aspects in DOAS measurements 43
4.1 Components of DOAS instruments 43
4.1.1 Input optics 43
4.1.2 Grating spectrometer 46
4.1.3 CCD and CMOS detectors 50
4.2 Characterization 52
4.2.1 Calibration, instrument function, sampling ratio 52
4.2.2 Dark current, o_set, and spectral stray light 55
4.2.3 Polarization response 56
4.3 Detection limit and optimal spectral resolution 58
5 Airborne DOAS in Arctic 61
5.1 Geophysical context 62
5.2 ALS-DOAS instrument and POLARCAT campaign 64
5.2.1 Instrumental description 64
5.2.2 The POLARCAT-France spring campaign 65
5.3 Spectral analysis and pro_ling method 67
5.3.1 DOAS analysis 67
5.3.2 Radiative transfer modeling 69
5.3.3 Retrievals with a maximum a posteriori 72
5.3.4 Error analysis 75
5.4 Results 77
5.4.1 Residual columns and O4 DSCD scaling factor 77
5.4.2 Measured versus simulated slant columns 79
5.4.3 NO2: stratospheric e_ect and detection limit 81
5.4.4 Retrievals of aerosol extinction and NO2 82
5.5 Interpretation 88
5.6 Conclusions on POLARCAT 95
6 DOAS measurements from an ultralight aircraft 97
6.1 Motivation 97
6.2 The ULM-DOAS instrument and Earth Challenge 99
6.2.1 Instrument description 99
6.2.2 The Earth Challenge expedition 101
6.3 Spectral analysis and NO2 column retrieval 103
6.3.1 DOAS analysis 103
6.3.2 Air mass factors calculation 104
CONTENTS ix
6.4 Sensitivity studies and error analysis 109
6.5 Results 112
6.5.1 Comparisons with satellites 112
6.5.2 Other interesting measurements 117
6.5.3 Soil signature above desert 119
6.6 Conclusions on Earth Challengenge 120
7 Mobile-DOAS measurements 123
7.1 State-of-the-art of mobile trace gases measurements 123
7.2 Description of the Mobile-DOAS 125
7.3 Retrieval scheme for NO2 vertical columns 126
7.3.1 Case of a homogeneous NO2 _eld 126
7.3.2 Realism of the geometrical approximation? 127
7.3.3 Solution for the NO2 _eld inhomogeneities 131
7.3.4 Error budget 133
7.4 Participation to the CINDI campaign 133
7.5 Routine measurements in Belgium 135
7.6 Comparison with Chimere Model 136
7.7 Conclusion on the Mobile-DOAS measurements 138
8 SWING: trace gases imaging from an UAV 141
8.1 Interest of UAV measurements 141
8.2 Whiskbroom imaging from a UAV 142
8.2.1 Geometry of whiskbroom imaging 142
8.2.2 Choice of the spectrometer and simulations 144
8.3 Instrument and platform description 147
8.4 Test _ights from an ultralight aircraft in Belgium 150
8.5 Conclusions and perspectives 152
8.6 Acknowledgments 152
9 Conclusions and perspectives 155
A Logarithmic weighting functions for LIDORT 163
A.1 Motivation 163
A.2 Jacobian inputs for an absorber 164
A.2.1 Optical depth derivative 164
A.2.2 Single scattering albedo derivative 165
A.2.3 Phase function coe_cients derivative 165
A.3 Jacobian inputs for the aerosol extinction 165
A.3.1 Optical depth derivative 166
A.3.2 Single scattering albedo derivative 166
A.3.3 Phase function coe_cients derivative 167
A.4 Perspectives 167
x CONTENTS
B Two remarks on the inverse problem 169
B.1 Impact of the unretrieved forward model parameters 170
B.2 Inequality-constrained maximum a posteriori solution 171
C Contributions to the scienti_c literature 175
C.1 Peer-reviewed 175
C.2 Proceedings and technical documents 176
C.3 Popular science 176
Acronyms 179
Bibliography 205