Measurements of Tropospheric, Stratospheric and Mesospheric Water Vapor by Ground Based Microwave SpectroRadiometry.

Transcription

1 Measurements of Tropospheric, Stratospheric and Mesospheric Water Vapor by Ground Based Microwave SpectroRadiometry Inauguraldissertation der Philosophisch-naturwissenschaftlichen Fakultät der Universität Bern vorgelegt von Alexander Haefele von Sevelen SG Leiter der Arbeit: Prof. Dr. N. Kämpfer Institut für Angewandte Physik

2

3 Measurements of Tropospheric, Stratospheric and Mesospheric Water Vapor by Ground Based Microwave SpectroRadiometry Inauguraldissertation der Philosophisch-naturwissenschaftlichen Fakultät der Universität Bern vorgelegt von Alexander Haefele von Sevelen SG Leiter der Arbeit: Prof. Dr. N. Kämpfer Institut für Angewandte Physik Von der Philosophisch-naturwissenschaftlichen Fakultät angenommen. Der Dekan: Bern, den Prof. Dr. Urs Feller

4

5 Abstract This thesis deals with ground based remote sensing of atmospheric water vapor by microwave spectroradiometry. It is dedicated to the improvement of the calibration and inversion of the radiation measurements and to the validation of the retrieved water vapor data. The thesis is devided into 9 chapters. Chapter 5, 6 and 8 are peerreviewed articles and Chapter 7 is a research report. The impact of water vapor on the atmosphere is manifold. It aects the temperature and the composition of the atmosphere directly and indirectly through its radiative properties and plays a key role in ozone chemistry. As a favorite tracer, water vapor observations gave deep insight in middle atmospheric dynamics. In Chapter 1 the role of water vapor is discussed and past and current research is reviewed emphasizing the importance of this thesis. Chapter 2 gives a very brief overview over the dierent insitu and remote sensing techniques to measure water vapor and provides the reader with the basics of spectroscopy, radiative transfer and radiometry which are of importance to follow the calculations in the following chapters. In Chapter 3 the Middle Atmospheric Water Vapor Radiometer, MIAWARA, of the Institute of Applied Physics (University of Bern) is presented. It is in operation since the year 2002 and deployed at Bern, Switzerland. In the frame of this thesis the instrument received several innovations in hardware and the calibration process has been optimized. While details on hardware are only briey listed in Chapter 3, a detailed discussion and validation of the new calibration algorithms is given. The Seoul Water Vapor Radiometer, SWARA, is a sister instrument of MIAWARA, deployed at Seoul, South Korea. As part of this thesis the measurement software has been developed and rst measurements were performed at Bern, Switzerland, and at Seoul, South Korea. A brief description of the instrument and results are presented in Chapter 4. Diurnal cycles of trace constituents give insight in the time scales and magnitudes of chemical and dynamical processes. In validation studies a diurnal cycle in the parameter under consideration can introduce large systematic dierences due to dierences in the temporal sampling of the instruments. In Chapter 5 diurnal cycles in water vapor and ozone are derived from radiometer data and compared to those derived from climate model simulations. The diurnal cycle of ozone in the upper stratosphere and lower mesosphere has been investigated thoroughly in the past decades. However, new features in mid and lower stratospheric ozone are presented. The diurnal cycle of water vapor received much less attention. In Chapter 5 rst results are presented. A weak diurnal cycle could be found with an amplitude in the order of 1 % related to tidal advection, revealing reasonable agreement with the climate model simulations. Chapters 6 presents an intercomparison of ve ground based microwave radiometers. These instruments are part of the Network for the Detection of Atmospheric Composition Change,

6 NDACC, and are spread all over the world. Thus, global observations of middle atmospheric water vapor from a spaceborne instrument have been used as a transfer standard to derive the biases between the ground based instruments. The instruments are found to agree with each other within 510 %. The validation of the MIAWARA instrument is further elaborated in Chapter 7. In the frame of the SHOMING project (Stratospheric Humidity Observations and Monitoring) ve balloonsoundings with a FLASHB research hygrometer have been performed in the vicinity of the radiometer. These soundings provided reliable humidity proles up to 37 km and allowed an evaluation of the lowermost part of the water vapor proles retrieved from MIAWARA measurements. In Chapter 8 the ability of MIAWARA to measure in the troposphere is investigated. Spectra of 1 GHz bandwidth, as measured by MIAWARA, do in principle contain information about the vertical water vapor distribution in the troposphere. In this chapter the calibration and retrieval are presented and characterized, and a detailed validation of the obtained results is carried out revealing a mean wet bias of 1020 % compared to balloonsoundings. In Chapter 9 conclusions are drawn from this thesis and an outlook on future work is given.

7 Contents 1 The Role of Water Vapor in the Atmosphere Sources and Sinks Radiation Dynamics Chemistry Trends Measuring Atmospheric Water Vapor Microwave Remote Sensing Spectroscopy and Radiative Transfer Microwave Radiometry Middle Atmospheric Water Vapor Radiometer MIAWARA Introduction MIAWARA History Calibration Hotcold Calibration and Troposphere Correction Opacity Transmission of the Reference Absorber Optimized Observing Angle for Noise Minimization Retrieval Seoul Water Vapor Radiometer SWARA Introduction Instrument Description Results Diurnal Changes in Middle Atmospheric H 2 O and O 3 : Observations in the Alpine Region and Climate Models 45 6 Validation of Ground Based Microwave Radiometers at 22 GHz for Stratospheric and Mesospheric Water Vapor 59 3

8 7 Comparison of Insitu and Remote Sensing Measurements of Water Vapor made within the frame of the SHOMING Project over Switzerland 70 8 Tropospheric Water Vapor Proles Retrieved from Pressure Broadened Emission Spectra at 22 GHz 79 9 Conclusions and Outlook 86 Acknowledgements 88 Publication List 89 Curriculum Vitae 94

9 Chapter 1 The Role of Water Vapor in the Atmosphere 1.1 Sources and Sinks The main source of atmospheric water vapor is the evaporation at the land and sea surface. In the atmosphere the water vapor concentration is governed by the strong dependence of the saturation vapor pressure on temperature. As a consequence of the constant and negative temperature laps rate of approximately 5 K/km, the water vapor concentration shows a steep decrease in the troposphere with a scale height in the order of 2 km (see Figure 1.1). The excess water vapor condenses or sublimes to form clouds and precipitation. At the tropopause the air is typically a thousand times dryer than in the lower troposphere. Air enters the stratosphere primarily in the tropics where the tropopause is coldest and acts as a cold trap leading to a freezedried stratosphere. The zonal mean water vapor distribution as observed by the Microwave Limb Sounder (MLS) onboard the Aura satellite is shown in Figure 1.2. A lot of the features discussed in the following can be discovered in this Figure. The water vapor mixing ratio of air entering the stratosphere is governed by the temperature history of the air parcel when it travels from the troposphere to the stratosphere and lies around 3.5 ppmv [Fueglistaler et al. 2005]. In the stratosphere the BrewerDobson circulation (see Section 1.3) lets air rise and be transported poleward. A second important source of stratospheric moisture is the oxidation of methane via a series of reactions that may be summarized as follows [Remsberg et al. 1984] CH 4 + 3O 2 + hν 2H 2 O + CO + O 3. (1.1) Methane oxidation leads to a positive vertical and latitudinal gradient in water vapor as air is transported from the tropical tropopause upward and poleward (see Figure 1.2). In the thermosphere and upper mesosphere water photolyzes by absorption of shortwave ultraviolet light and in the stratosphere and lower mesosphere water vapor is primarily destroyed by the reaction with atomic oxygen: H 2 O + hν(λ < 200nm) H + OH (1.2) 5

10 6 Pressure [hpa] VMR [ppmv] Thermosphere Mesopause Mesosphere Stratopause Stratosphere Altitude [km] 10 2 Tropopause Troposphere Temperature [K] Figure 1.1: Water vapor (blue), ozone (yellow) and temperature (black) prole taken from the U.S. standard atmosphere H 2 O + O( 1 D) 2OH (1.3) Above the 1 hpa level ( 50 km) destruction dominates production leading to a decrease of water vapor in the mesosphere. Stratospheric air reaching the polar night may encounter very low temperatures such that polar stratospheric clouds (PSC) can form. The subsequent sedimentation of these ice particles leads to a dehydration of the polar stratosphere [Pan et al. 2002, and references therein]. There are large interhemispheric dierences in the extent of this dehydration process. In the antarctic vortex suciently low temperatures are reached for extended formation of PSC's. The arctic vortex, on the other hand, is disturbed by enhanced wave activity and the temperature falls only occasionally below the threshold for cloud formation. Due to large scale upwelling, adiabatic cooling makes the summer polar mesopause the coldest point in the atmosphere providing a second location in the middle atmosphere for cloud formation [Fiedler et al. 2009, and references therein]. These polar mesospheric clouds (PMC) are found in a narrow layer between 8284 km. Summers and Siskind [1999] found enhanced water vapor concentrations of 1015 ppmv in this altitude range in the arctic summer mesosphere. This water is believed to be advected from below and sequestered in or just below the PMC layer as either vapor or ice. von Zahn and Berger [2003] support this hypothesis by a modeling study and speculate about a freezedried mesopause. A second layered phenomenon is the presence of a (third) local maximum in the vertical prole of water vapor around 6570 km. This enhancement has been observed by ground based millimeter techniques [Bevilacqua et al. 1985; Nedoluha et al. 1996; Seele and Hartogh 1999] and by the HALOE experiment [e.g. Siskind and Summers 1998]. This water vapor enhancement is not in agreement with the classical water vapor chemistry (equations

11 CHAPTER 1. THE ROLE OF WATER VAPOR IN THE ATMOSPHERE 7 Zonal and monthly mean H 2 O, Aura/MLS v2.2, August Zonal and monthly mean H 2 O, Aura/MLS v2.2, February Pressure [hpa] VMR [ppmv] Pressure [hpa] VMR [ppmv] Latitude [deg] Latitude [deg] 1 Figure 1.2: Latitudeheight crosssection of monthly and zonal mean water vapor as measured by Aura/MLS (version 2.2) for August 2008 (left panel) and February 2009 (right panel). The data have been corrected for oscillations around 30 hpa according to Lambert et al. [2007]. (1.1,1.2)) and the need for additional sources of mesospheric water vapor emerged [Siskind and Summers 1998]. Summers and Siskind [1999] proposed a heterogeneous reaction on the surface of meteoric dust as possible source of H 2 O, and Sonnemann et al. [2005] suggested an interplay of an autocatalytic reaction with special dynamic conditions as encountered in the summer polar middle atmosphere as possible explanation. 1.2 Radiation Because the atmosphere is transparent for shortwave (visible) radiation and partially absorbing for infrared (thermal) radiation it acts like a greenhouse, trapping thermal energy. The dominant contributor to infrared opacity is water vapor and it is hence the dominant greenhouse gas [Soden et al. 2005]. Atmospheric water vapor increases the surface temperatrue by 21 K which accounts for 62 % of the total greenhouse eect. Carbon dioxide accounts for 22 % of the total greenhouse eect but constitutes over 50 % of the anthropogene greenhouse eect [Klose 2008]. Under the assumption that the relative humidity stays constant, any change in temperature will lead to an increase in water vapor. More water vapor in the atmosphere will in turn increase the greenhouse eect and surface temperatures will rise. Approximately half of the predicted warming in response to an anthropogenic increase in greenhouse gases such as CO 2 is due to this water vapor feedback eect [Held and Soden 2000; Soden et al. 2005]. The increase in water vapor is not expected to be uniform but will be strongest in the upper troposphere. Measurements of upper tropospheric humidity are thus important for the detection of climate change. In Haefele and Kämpfer [subm.] a new approach is presented to measure tropospheric humidity by micwrowave spectroradiometry.

12 8 Stratospheric water vapor constitutes a small positive anthropogenic radiative forcing when we consider the portion that is produced by methane oxidation, as methane concentrations are altered by human activities [Forster and Shine 2002; Forster et al. 2007]. In the stratosphere itself, however, water vapor has a cooling impact because it increases the infrared emissivity of the stratosphere. This cooling leads to an increase in midstratospheric ozone as certain reaction rates of the ozone chemistry depend on temperature [Evans et al. 1998; Haefele et al. 2008]. The extent and the importance of the contribution of stratospheric water vapor to the total anthropogenic radiative forcing and to the stratospheric cooling trend is uncertain and still under debate [Forster et al. 2007; Randel et al. 2009]. Reason for this lack of knowledge are still the quite sparse long term observations of stratospheric water vapor. In this context the validation of ground based radiometers of Haefele et al. [in press] can contribute to this discussion as long term measurements of ground based radiometers allow to derive trends and allow to intercalibrate consecutive satellite missions to compile global long term data sets. 1.3 Dynamics Middle atmospheric water vapor has a photochemical lifetime in the order of months up to 70 km and exhibits steep horizontal and vertical gradients. This makes it a valuable tracer for atmospheric dynamics. In 1949 measurements of lower stratospheric water vapor and helium over Southern England by a balloonborne frost point hygrometer gave rise to the hypothesis of a world circulation in which air enters the stratosphere through the tropical tropopause where it is dehydrated due to the low temperatures and is then transported to the midlatitudes. This hypothesis was formulated in the well known work by A. W. Brewer in 1949 [Brewer 1949]. This world circulation is called BrewerDobson circulation. Mote et al. [1995] discovered a close relation between lower stratospheric water vapor and tropical tropopause temperatures which is a consequence of this wellknown circulation. The vertical propagation of the imprint of the temperature on water vapor is called the tape recorder eect, probably because water vapor serves as kind of a memory for tropical tropopause temperatures. Indications of the tape recorder eect are also evident in Figure 1.2. The need for a downward motion above the poles that compensates for the ascending branch over the equator was already stated by Brewer [1949]. This large scale subsidence over the winter poles is evident in water vapor observations as dry mesospheric air descends to the midstratosphere and air, moistened by methane oxidation, descends from the stratopause to the lower stratosphere [Lahoz et al. 1993] (see Figure 1.2). Not only large scale dynamical features are visible in water vapor measurements but also short term variations, given that the data are suciently resolved in time. Good time resolution is one of the strengths of ground based remote sensing instruments compared to satellites, that pass one geolocation usually only twice daily. Ground based microwave radiometers have been extensively used to investigate the diurnal cycle of ozone in the stratosphere and mesosphere [Zommerfelds et al. 1989; Schneider et al. 2005; Haefele et al.

13 CHAPTER 1. THE ROLE OF WATER VAPOR IN THE ATMOSPHERE ; Palm et al. 2009] and recent work is also targeting at the mesopause region [Rogers et al. 2009]. More recently ground based water vapor measurements in the middle atmosphere have been used to study the dynamics on short time scales. Seele and Hartogh [2000] and Flury et al. [in press] investigated the evolution of water vapor during sudden stratospheric warming events using data of ground based microwave radiometers. Sonnemann et al. [2008] analyzed the 5day signal in mesospheric water vapor at high latitudes, and as part of this thesis Haefele et al. [2008] investigated the diurnal cycle of stratospheric and mesospheric water vapor, that is caused by tidal oscillations. 1.4 Chemistry Even though the water molecule in the atmosphere is inert it still plays an important role in atmospheric chemistry. First of all, water vapor is the source gas of the hydroxyl radical through reactions (1.2) and (1.3). Hydroxyl is one of the most important oxidizing agents in the atmosphere, and in the stratosphere and mesosphere OH is particularily fundamental as a natural catalyst for the chemical destruction of ozone [Summers and Conway 2000, and references therein]. In the upper stratosphere and mesosphere OH exhibits a strong diurnal cycle as its concentration is mainly governed by solar UV radiation. OH, and indirectly water vapor, has thus a strong inuence on the diurnal evolution of daytime ozone. Based on HALOE data Marsh et al. [2003] found a distinct response of sunset ozone data to water vapor revealing lower ozone values at sunset with higher water vapor concentrations. Inside the polar vortex the temperature can fall below 188 K at 20 km which allows for the formation of ice clouds (see Section 1.1). On the surface of these polar stratospheric clouds (PSC) heterogeneous reactions can take place releasing chlorine from the two reservoir gases, HCl and ClONO 2 [Brasseur and Solomon 2005]. This chlorine, which is conned to the polar vortex, is broken up by the rst sunrays in spring and starts a catalytic destruction of ozone leading to the ozone hole. As noted earlier, there is a strong asymmetry between southern and northern hemisphere. Due to the generally higher temperatures in the northern polar vortex, PSC formation is hampered and the ozone hole is less pronounced. 1.5 Trends The detection of trends requires stable long term measurements. These are particularly rare in the case of stratospheric and mesospheric water vapor. The data sets listed in Table 1.1 are the most relevant for the investigation of trends. Based on these data the following picture emerged: In the time period 1954 to 2000 water vapor increased by 1 %/year [e.g. Nedoluha et al. 1998; Evans et al. 1998; Oltmans and Hofmann 2000; Rosenlof et al. 2001]. Most intriguing is the fact that this trend is going along with a

14 10 Table 1.1: Key data sets for the investigation of long term trends of stratospheric and mesospheric water vapor. Experiment Location Time coverage NOAA frost point hygrometer Boulder, CO, USA since 1981 Halogen Occultation Experiment (HALOE) global Water Vapor Millimeter Spectrometer: WVMS1 Lauder, NZ since 1994 WVMS3 Mauna Loa, Hawaii since 1996 cooling trend in the tropical tropopause which would be likely to cause a negative trend in stratospheric water vapor [Simmons et al. 1999; Randel et al. 2000]. Secondly, increasing methane concentrations due to human activities can only account for a small portion of the increase in water vapor [Evans et al. 1998]. Thus, this positive trend in water vapor remains under debate. According to Rosenlof and Reid [2008], likely solutions to this problem are changes in the dynamics around the tropical tropopause either remotely by wave driving but more probably locally by changing sea surface temperatures. In late 2000 the positive trend in stratospheric water vapor has been interrupted by a distinct drop in the H 2 O levels (approximately -0.4 ppmv [Randel et al. 2006]). The persistently low water vapor values since late 2000 correlate with anomalously cold tropical tropopause temperatrues for the period and with a decrease in ozone near the tropopause after According to Randel et al. [2006] these changes are most likely due to an intensication of the BrewerDobson circulation.

15 Chapter 2 Measuring Atmospheric Water Vapor Water vapor in the atmosphere is extremely variable in space and time and shows a steep gradient between a very humid lower troposphere and a very dry upper troposphere and stratosphere. These facts make the measurement of water vapor a big challenge. As pointed out in Chapter 1, there are many open questions concerning water vapor, and continuous, high quality measurements of water vapor in the whole atmosphere are needed. Measurements are typically made in situ from balloon platforms or rocket sondes or remotely from the ground or from space. Big dierences arise between the dierent techniques with respect to spatial and temporal resolution and coverage. Figure 2.1 shows an overview of the common techniques with the associated altitude ranges. In Kley et al. [2000] detailed information about the dierent techniques can be found. Some of these techniques are used in the framework of the Network of the Detection of Atmospheric Composition Change, NDACC. This network is composed of more than 70 research sites that perform atmospheric soundings on a regular basis. NDACC is dedicated to long term observations of the atmospheric composition with an emphasis to the stratosphere ( The radiometers that were considered in the study of Haefele et al. [in press] are part of NDACC. While the reader is referred to the literature to get more information about other techniques [e.g. Kley et al. 2000] the physical basics of microwave remote sensing of water vapor are given in the following sections. 2.1 Microwave Remote Sensing Spectroscopy and Radiative Transfer A molecule or an atom that changes from one state of energy E n to an other state of energy E m emits or absorbs a photon of the frequency f. The energies and the frequency are related to each other according to Bohr's law: E n E m = hf (2.1) 11

16 12 in situ remote sensing Stratopause Tropopause tunable diode laser spectrometer frost point Lyman a fluorescence Lyman a absorption MOZAIC radiosonde infrared & far infrared spectrometer microwave LIDAR standard retrievals advanced retrievals operational satellites Figure 2.1: Techniques for water vapor measurements in the atmosphere. Adapted from Kley et al. [2000]. Absorption or emission does, however, not always happen at the frequency f, but there are certain broadening processes that allow photons of dierent frequencies to be emitted or absorbed. The width of the frequency distribution is reciprocal to the lifetime t of the involved energy levels: f 1 (2.2) t There is a natural broadening due to the limited lifetime of any energy state which is in the order of 10 7 MHz. In the gaseous atmosphere molecules collide with each other which leads to a further reduction of the lifetime and to a further broadening. Pressure generally governs the collision rate and this mechanism is thus called pressure broadening. At pressures encountered in the lower and middle atmosphere natural broadening is unimportant compared to pressure broadening. The probability for a photon to be emitted or absorbed by a molecule in the presence of electromagnetic radiation is given by the Einstein coecients. Einstein's considerations are the basis for the calculation of the absorption coecient, α(f), which describes the change in incident intensity I f of a ray passing through an innitesimally thin layer ds according to Lambert's law: di f = α(f)i f ds (2.3) Absorption or emission of radiation in the atmosphere is commonly described by line absorption/emission which accounts for all possible transitions between energy states for each species. Hence, the absorption coecient has to be calculated for each transition, herein after referred to as line, and the total absorption α is given by the sum of the

17 CHAPTER 2. MEASURING ATMOSPHERIC WATER VAPOR 13 individual absorption coecients α i α = no.lines i=1 α i. (2.4) The absorption coecient due to a line i depends on the number density of the absorbing gas, n, of the strength of the line, S i (T ), and of the shape of the line, F (f), which accounts for the broadening of the line α i (f) = ns i (T )F (f). (2.5) S i (T ) depends on the population of the energy levels involved in the line i and is thus a function of temperature. The line shape function F (f) accounts for the broadening of the line and depending on the underlying broadening process the line shape function looks dierent. One description of the pressure broadening is the Lorentzian line shape: F L (f) = 1 π γ L (f f 0 ) 2 + (γ L ) 2 (2.6) γ L is the line width and f 0 is the line center frequency. The third important broadening process not discussed so far is Doppler broadening. It accounts for the movement of the molecules relative to the observer and hence for the Doppler shift of the the emitted radiation. The line shape is derived from the probability distribution for the relative velocity between the gas molecules and the observer and is given by where γ D is the Doppler line width given by F D (f) = 1 ( ) 2 f f0 exp (2.7) γ D π γ D γ D = f 0 c 2kT m (2.8) where m is the mass of the molecule, k is the Boltzmann constant, T is the temperature and c is the speed of light. The combined eect of pressure and Doppler broadening can be approximated by a Voigt line shape: F V oigt (f, f 0 ) = F L (f, f )F D (f, f 0 )df. (2.9) As this is only an approximation one does not get the correct total absorption with the approach of line absorption. Instead, continuum terms have to be added to α to correct for errors in the line shape assumptions. In the case of water vapor these continuum terms are empirical corrections with a weak frequency dependence. Figure 2.2 illustrates the change of the line width as a function of pressure for the water vapor line at 22 GHz. Given a measurement of a line, the information of the vertical

18 14 Pressure [hpa] Line Width Doppler Broadening Pressure Broadening Lorentz Line Shape [MHz 1 ] km 70 km 60 km 50 km 40 km Line Width [Hz] Frequency Offset [MHz] Figure 2.2: Broadening of the water vapor line at GHz. Left panel: Line width due to pressure and Doppler broadening as function of altitude. Right panel: Lorentz line shape at dierent altitudes. distribution of an absorber is retrieved from the pressure dependence of the line width. No information about the vertical distribution, however, can be retrieved above the altitude where Doppler broadening starts to govern the total line width. In case of the 22 GHz line of water vapor this level is found at 0.02 hpa ( 80 km). Note that the Doppler line width γ D depends on the frequency f. Hence, the observation of lines at lower frequencies allows generally to reach higher altitudes. An overview of the current knowledge about the values of S and γ is given in Haefele et al. [in press, Table 4]. To calculate the transfer of radiation through the atmosphere along the path s let us rearrange equation (2.3) and add a source term accounting for atmospheric emission: di f ds = α(f, s)i f(s) + α(f, s)b f (T (s)) (2.10) B f (T (s)) is the Planck function. The incident intensity at the location s 0 is given by the integral form: I f (s 0 ) = I f (s )e τ(f,s ) + where the opacity, τ(f, s), is given by τ(f, s) = s s s 0 α(f, s)b f (T (s))e τ(f,s) ds (2.11) s 0 α(f, s )ds. (2.12) In the microwave region hf kt and the Planck function is reasonably well reproduced by the RayleighJeans approximation: B f = 2f 2 kt c 2. (2.13)

19 CHAPTER 2. MEASURING ATMOSPHERIC WATER VAPOR 15 This leads to the denition of the RayleighJeans brightness temperature T b f = c2 2f 2 k I f. (2.14) Inserting (2.13) and (2.14) into (2.11) yields the radiative transfer equation for the microwave case: T b f (s 0 ) = T b f (s )e τ(f,s ) + s s 0 α(f, s)t (s)e τ(f,s) ds. (2.15) In passive microwave radiometry the radiation T b f (s ) is equal to the cosmic background radiation: T b f (s ) = T 0 = 2.7 K Microwave Radiometry The task of a microwave radiometer is to measure T b f (s 0 ). An antenna feeds the radiation from the free space (atmosphere) into the waveguide of the radiometer. There, the signal is amplied, ltered and, in the case of a heterodyne receiver, converted to lower frequencies. The signal can be spectrally analyzed by various types of spectrometers. Common spectrometer types are: Filter banks, acoustooptical and chirp transform spectrometers, autocorrelators and digital FFT spectrometers. The spectrometer output is in good approximation linearly related to the input power. This linear relationship is determined in the calibration process. The calibration of the Middle Atmospheric Water vapor Radiometer, MIAWARA, is presented in Section 3.3. The resolution of a radiation measurement is determined by the bandwidth of individual spectrometer channels, f, by the integration time, t int, and by the total power measured by the spectrometer. The total power includes the power received by the antenna, T A, and the power generated inside the radiometer, T rec. The resolution is given by the radiometer equation: T = T A + T rec (2.16) ftint The antenna temperature, T A, is the brightness temperature of the atmosphere weighted by the sensitivity of the antenna and can be set to T b f (s 0 ) in good approximation (see Section 3.3.3). For a Dicke radiometer, that measures not total power but the dierence to a reference source, the resolution is increased by a factor of 2 1. The radiometer noise formula determines how much integration time is required to achieve a desired signal to noise ratio. This issue is further discussed in Section The need to accurately measure atmospheric water vapor is also addressed by the ground based microwave community. An overview of the currently available instruments is given in Table 2.1. The increasing number of such instruments raises the need for standards in measuring, data processing and prole retrieval to achieve better agreement between the individual instruments. Setting up and validating standards is the idea behind the work of Haefele et al. [in press]. 1 The factor of 2 applies only if the integration time of both the atmospheric and the reference measurement is t int.

20 16 Table 2.1: Inventory of existing ground based microwave instruments for the observation of middle atmospheric water vapor. The spectral resolution, f, refers to the highest resolution at line center and can be lower at the wings. The bandwidth, B, refers to the measured total bandwidth. For the retrievals a smaller bandwidth might be used. Project name Site Spectrometer f [khz] B [MHz] Trec [K] Preamplier Calibration Onsala (57 N, 12 E) Autocorr uncooled hot-cold MIAWARA Bern (47 N, 7 E) digital FFT uncooled hot-cold SWARA Seoul (37 N, 127 E) digital FFT uncooled hot-cold WVMS-1 Lauder (45 S, 170 E) Filter bank cooled noise diode WVMS-2 Table Mountain (34 N, 117 W) WVMS-3 Mauna Loa (19 N, 157 W) Filter bank cooled noise diode CAWSES1 Andoya (69 N, 16 E) CTS cooled hot-cold CAWSES3 Zugspitze (47 N, 11 E) CTS cooled hot-cold RAM Ny Ålesund (79 N, 12 E) AOS/CTS 1600/ / uncooled hot-cold MIRA5 Zugspitze (47 N, 11 E) digital FFT uncooled hot-cold MOBRA mobile AOS < 200 uncooled hot-cold MIAWARAC mobile digital FFT uncooled hot-cold

21 Chapter 3 Middle Atmospheric Water Vapor Radiometer MIAWARA 3.1 Introduction Two important steps have to be made to get from raw measurements to a water vapor prole: The calibration and the retrieval. The calibration itself is a two step procedure consisting rst of a hotcold calibration to convert the raw measurements from instrument specic units to physical units, and second of a troposphere correction to account, amongst others, for the attenuation of the middle atmospheric signal in the troposphere. The hotcold calibration includes the determination of the brightness temperatures of reference targets. In the case of MIAWARA a microwave absorber at ambient temperature serves as hot load and the sky at an elevation angle of 60 serves as cold load. While the brightness temperature of a microwave absorber is in good approximation equal to its physical temperature the brightness temperature of the sky is calculated based on the opacity and the eective temperature of the troposphere. In this thesis a new algorithm for the derivation of the opacity from a tipping curve measurement has been implemented. This new approach is discussed in Section and comparisons to the former algorithm are made. The MIAWARA instrument performs a balancing calibration scheme in order to minimize eects due to gain instability and nonlinearities. The balancing calibration was introduced by Parrish et al. [1988] and is widely applied for ground based remote sensing of middle atmospheric water vapor. This calibration scheme is related to the Dicke method including a reference measurement and a line measurement. The line measurement is a sky measurement taken under a low elevation angle and it contains the main part of the middle atmospheric signal. The reference measurement is also a sky measurement but taken in zenith direction where the beam is partly covered by a microwave absorber to compensate for the lower airmass. The line and the reference signals are balanced by adjusting the area of the beam covered by the reference absorber and by continuously adjusting the elevation angle of the line measurement. The correction of the calibrated balanced spectrum for the 17

22 18 Figure 3.1: The MIAWARA instrument on the roof of the new facilities for atmospheric remote sensing at Zimmerwald (46.88 N / 7.46 E, 907 m amsl). attenuation due to the troposphere and due to the reference absorber and for variations in the elevation angles reveals the spectrum as it would be observed at the tropopause in zenith direction. A new formulation for the equivalent transmission of the reference absorber is derived in Section and compared to a former formulation. The quest for a water vapor prole that is consistent with the calibrated measurement and with some sort of a priori water vapor climatology is called inversion or retrieval. It is, after the calibration, the second step on the way from raw measurements to a water vapor prole. The outcome of the retrieval is nothing magic but compared to an in situ measurement it may still require a little bit more of an eort to understand and to get a feeling what a retrieved prole can tell us and what it can not tell us. A characterization of the retrieval as it is applied in the case of MIAWARA and other ground based microwave instruments is thus given in Section 3.4. The organization of the chapter is as follows: A short overview of the history of the MIAWARA instrument is given in Section 3.2. Section 3.3 addresses the calibration and the troposphere correction with a special focus on the derivation of the tropospheric opacity (Section 3.3.2) and the transmission of the reference absorber (Section 3.3.3). The retrieval of water vapor proles from calibrated spectra is discussed in Section MIAWARA History The MIAWARA instrument was operated on the roof of the Institut für exakte Wissenschaften in the center of Bern from November 2002 until September 2006 when it was moved to the new facilities of the Observatory in Zimmerwald 7 km South of Bern at 905 m

23 CHAPTER 3. MIDDLE ATMOSPHERIC WATER VAPOR RADIOMETER MIAWARA Water Vapor over Bern, Switzerland, MIAWARA, v179 8 Pressure [hpa] VMR [ppmv] 10 1 Jul07 Jan08 Jul08 Jan09 Jul09 3 Figure 3.2: Water vapor record measured by MIAWARA with the digital FFT spectrometer. amsl (see Figure 3.1). Until March 2007 the spectral analysis was done by an Acousto Optical Spectrometer, AOS, and by a Chirp Transform Spectrometer, CTS [Deuber et al. 2005]. The bandwidth, B, and the spectral resolution, f, of these spectrometers are as follows: AOS: B = 1 GHz, f = 1.2 MHz; CTS: B = 40 MHz, f = 14 khz. In March 2007 MIAWARA experienced a major innovation including the following changes: Replacement of tha AOS by a digital FFT spectrometer (Acqiris, B = 1 GHz, f = 61 khz, [see Müller et al. 2009, and references therein]) New measurement PC Implementation of stepper motor controllers Replacement of the mirror suport unit The water vapor record measured with the digital FFT spectrometer since April 2007 is shown in Figure 3.2. A further extension of the backend was made in October 2008 when a second digital FFT spectrometer (BEAM, B = 25 MHz, f = 12 khz [see Müller et al. 2009, and references therein]) was included in the backend. The purpose of this spectrometer is to extend the altitude range of the instrument up to 8085 km. It is not the scope of this document to report on the details of the instrument's hardware and its operation but it is referred to the updated version of the MIAWARA operating manual [Haefele 2009].

24 Calibration Hotcold Calibration and Troposphere Correction The purpose of the calibration process is to convert the measurement of the atmospheric radiation from instrument specic units (counts) into physical units. The physical unit that is commonly used for radiant power is brightness temperature. It is the physical temperature at which a perfect blackbody would emit the same power as is measured by the instrument. As introduced in Chapter 2, the RayleighJeans brightness temperature is dened as: T b(f) = λ2 2k I f (3.1) Here, f is the frequency, λ is the wavelength, k is Boltzmann's constant and I f is the specic intensity ([W/m 2 /Hz/sterad]).The radiometer output, U(f), is given by Nyquist's law [Ulaby et al. 1981]: U(f) = G(f)k f(t A + T rec ) + U 0 (f) (3.2) Where G(f) is the radiometer gain, f is the bandwidth of one spectrometer channel, T A and T rec are the antenna and receiver temperatures, respectively, and U 0 (f) is the oset of the radiometer output. If the eld of view of the antenna is uniformly covered by an object with the brightness temperature T b, then we can set (see Section 3.3.3) T A = T b. (3.3) A calibration cycle of the MIAWARA instrument consists of a set of four measurements looking at four dierent targets as shown in Figure 3.3: 1. U hot : Hot load (microwave absorber at ambient temperature) 2. U cold : Cold load (sky at elevation angle φ cold = 60 ) 3. U line : Atmosphere (sky at elevation angle φ line = ) 4. U ref : Reference (sky at elevation angle φ ref 90 + contribution from the reference absorber) Let the brightness temperature of the hot load, T b hot, be equal to its physical temperature, T hot, which is measured with temperature sensors. The brightness temperature of the cold load, T b cold, is determined from a tipping curve measurement (see Section 3.3.2). Making use of Equation 3.2 and the four measurements of a calibration cycle the following relation can be derived: T b = T b line T b ref = (U line U ref ) T hot T b cold U hot U cold (3.4)

25 CHAPTER 3. MIDDLE ATMOSPHERIC WATER VAPOR RADIOMETER MIAWARA 21 Figure 3.3: Block diagram of the MIAWARA instrument. A calibration cycle consists of four measurements looking at the following targets: hot load, cold load, atmosphere and reference. The balanced spectrum, T b, is not yet the quantity to be inverted, instead it has to be corrected for the attenuation in the troposphere, the eects of the reference absorber and the elevation angles of the line and reference measurement to nally get the spectrum as it would be observed at tropopause height in zenith direction. Let us decompose T b line and T b ref into contributions from the middle atmosphere (stratosphere and mesosphere), T b, and into spectrally at contributions from the troposphere and the reference absorber, T b cont1 : T b line = T b (φ line )e τµ line + T b cont,line (3.5) T b ref = T b (φ ref )e τµ line A + T b cont,ref (3.6) A is the transmission due to the reference absorber that partly covers the beam (see Section 3.3.3), φ is the elevation angle of the beam, τ is the zenith opacity of the troposphere, µ ref/line is the airmass of the troposphere at the elevation angles φ ref and φ line. e τµ line is the transmission of the troposphere for a given airmass which is given by the ratio of the slant opacity τ φ at elevation angle φ and the zenith opacity: µ φ = τ φ τ (3.7) 1 T b cont must not be confused with the continuum term that is commonly used in the eld of spectroscopy accounting for generally unknown contributions to the absorption coecient with weak frequency dependence

26 22 In the stratosphere and mesosphere absorption can be neglected and the following relationship holds in good approximation: T b (φ = 90) = T b (φ)/µ (φ) (3.8) Due to the sphericity of the atmosphere the airmass factor shows a small dependence on altitude and µ (φ) refers to the airmass of the middle atmosphere (10 90 km). Subtracting (3.6) from (3.5) and inserting (3.8) yields T b (φ = 90) = T b T b cont µ line e τµ line Aµ ref e τµ ref (3.9) T b cont is the contribution from the troposphere and from the spectrally at reference absorber and is generally not known. However, the balance between the line and the reference signal allows now to set T b cont 0. The spectrum at tropopause height in zenith direction is thus in good approximation given by Opacity T b (φ = 90) = a T b (3.10) a = 1 µ line e τµ line Aµ ref e τµ ref (3.11) The brightness temperature that would be measured by a ground based instrument looking through a isothermal atmosphere under the elevation angle φ can be written as T b(φ) = T 0 e τµ φ + T eff (1 e τµ φ ). (3.12) T 0 is the cosmic background, T eff the eective temperature of the troposphere estimated from the surface temperature T amb (T eff = 0.69(T amb 273) ) [Han and Westwater 2000], τ is the zenith opacity and µ φ is the airmass factor at the elevation angel φ. For φ i 20 the airmass factor can be approximated within a 2% error as For τ 1 equation (3.12) can be linearized with respect to τ µ φ 1 sin(φ). (3.13) T b(φ) T b (φ) = T 0 (1 τµ φ ) + T eff τµ φ (3.14) A tipping curve measurement is a set of at least two measurements {U i } of the sky at dierent elevation angles {20 φ i 90}. The measurements can be related to brightness temperature as shown in Deuber et al. [2004] U i U hot U hot U offset = T 0(1 τµ φi ) + T eff τµ φi T hot T hot + T rec (3.15)

27 CHAPTER 3. MIDDLE ATMOSPHERIC WATER VAPOR RADIOMETER MIAWARA 23 τ slant airmass Figure 3.4: Illustration of the iterative search for the zenith opacity. The circles are the airmassopacity pairs {µ φi, τ φi } and the dashed line is the t whose slope is the zenith opacity. The data around a slant opacity of 0.35 to 0.4 are calculated with the initial guess for the opacity (1st iteration) and the data around a slant opacity of 0.15 are calculated with the updated value for τ (2nd iteration). The iteration is stopped if the t intersects the origin, {0, 0}. Two measurements U 1 and U 2 allow to solve (3.15) for τ τ = (T hot T 0 )(U 2 U 1 ) (T eff T 0 )(U hot (µ φ2 µ φ1 ) + U 2 µ φ1 U 1 µ φ2 ) (3.16) As e τµ φ 1 τµφ if follows that T b(φ) T b (φ). But as we assume that T b (φ) T b(φ) we are going to underestimate τ using equation (3.16). An other approach described in Han and Westwater [2000] uses an iterative method to nd τ. Again a set of at least two sky measurements is needed plus a cold load measurement U cold that is also a sky measurement at its own elevation angle, φ cold. An initial value for the opacity has to be chosen, τ 0. From here the brightness temperature of the sky T b(φ cold ) at elevation angle φ cold is calculated according to (3.12). With this value the tipping curve signals {U i } can be calibrated T b i = (U i U hot ) T hot T b(φ cold ) U hot U cold + T hot (3.17) The brightness temperatures T b i can be mapped to slant opacities solving (3.12) for µ φ τ ( ) Teff T 0 µ φ τ = τ φi = ln (3.18) T eff T b i From equation (3.7) it is apparent that the airmass-opacity pairs {µ φi, τ φi } should lie on a straight line that crosses the origin. The slope of this line is the zenith opacity. Therefore

28 24 Table 3.1: Error estimation for the opacity calculated iteratively. Input Variables Setup Variables Output Variables U hot 0.7 % U 0.7 % T eff 3.5 K T hot 2 K φ 0.5 y-axis-oset initial value (τ 0 ) 0.3 max number of iterations 5 τ (2 angles) 8% τ (4 angles) 5.7% a linear t of the {µ φi, τ φi } data gives a new value for τ. The initially chosen value for τ is replaced by the new value and the process is repeated until the y-axis-oset of the t is satisfyingly small. Figure 3.4 visualizes the process. A y-axis-oset of less than is mostly achieved after two iterations. The error of this method was assessed using a monte carlo simulation. A data set of a ctive tipping curve was constructed assuming a value of 0.1 for the opacity, a receiver temperature of 140 K and a y-factor (y = U hot /U cold ) of 2.5. Based on this data the opacity was calculated 10 5 times while all input values have been contaminated with gaussian noise. The uncertainty of the opacity is only marginally aected by the setup of the iteration process itself. However, the number of tipping angles involved has a large inuence on the uncertainty of τ. The results of this calculation are listened in table 3.1. Eect of τ on Receiver Temperature Power is not only received by the antenna but is also generated by the radiometer itself. This power is described by the receiver temperature, T rec, which corresponds to the physical temperature that a black body should have in order to emit the same power as the radiometer does (see equation (3.2)). The receiver temperature is determined with the y-factor method as described in Janssen [1993] where the cold load temperature (e.g. the sky at 60 elevation) is calculated from the opacity and the eective temperature of the troposphere using equation (3.12). To validate the tipping curve calibration the radiometer is calibrated on a monthly basis against a liquid nitrogen load which replaces the sky measurement. Figure 3.5 shows the receiver temperature of the MIAWARA radiometer derived from tipping curve calibrations and from liquid nitrogen calibrations for the years 2003 to 2006 assuming a linear relationship between T b and τ (equation (3.16)). It is apparent that T rec is subject to an annual cycle with a maximum in summer. Equation (3.16) is underestimating the opacity and as a consequence the cold load temperature T b(φ cold )

29 CHAPTER 3. MIDDLE ATMOSPHERIC WATER VAPOR RADIOMETER MIAWARA 25 Figure 3.5: Receiver Temperature derived from tipping curve calibration using equation (3.16). is also underestimated. It follows from the y-factor method that the receiver temperature is overestimated in that case and that the error in T rec is twice as large as the error in T b(φ cold ) for a y-factor of 2. Figure 3.6 shows the receiver temperature based on the opacity derived iteratively. The iterative determination of τ leads to a better estimate of T b(φ cold ) throughout the year and the annual cycle is reduced remarkably. The systematic oset in T rec between the tipping curve and the liquid nitrogen calibration is to a great extent due to uncertainty in the determination of the temperature of the liquid nitrogen load. This uncertainty is mainly related to reections on the surface of the liquid nitrogen Transmission of the Reference Absorber For the ground based detection of the weak water vapor line at 22 GHz a balancing calibration has to be performed [Parrish et al. 1988]. In contrast to the total power calibration a dierence signal U of a line and a reference target is calibrated. The line signal, U line, is taken at low elevation angles between 20 and 30. The reference measurement, U ref, is taken at a high elevation angle (φ 90 ) where a reference absorber covers a small part of the antenna beam to account for the smaller airmass. In order to retrieve the underlying water vapor distribution from the calibrated dierence spectrum T b the contribution from the reference absorber has to be determined. Let us start with the antenna temperature that is given by [Janssen 1993] T a = g(θ, φ)t b(θ, φ)dω (3.19) 4π where Θ is the azimuth, φ is the elevation angle of the beam, Ω is the solid angle and g(θ, φ) is the antenna gain normalized to one: g(θ, φ)dω = 1 (3.20) 4π

30 26 Figure 3.6: Receiver Temperature derived from tipping curve calibration using the iterative determination of the opacity. We assume that T b(θ, φ) is constant and equal to T b within the main lobe of g(θ, φ) and that the contribution from outside the main lobe can be neglected. The antenna temperature can now be approximated with T a = T b g(θ, φ)dω = T b (3.21) 4π For the reference measurement, let G sky be the solid angle, that is not covered by the reference absorber, where the antenna sees the brightness temperature T b. On the other hand, let G ref be the solid angle covered by the reference absorber, where the antenna sees its physical temperature T ref because the piece of absorber is supposed to be a perfect blackbody. The antenna temperature at reference position is thus given by We dene A as and get T a = T b g(θ, φ)dω + T ref G sky g(θ, φ)dω G ref (3.22) A = g(θ, φ)dω G sky (3.23) 1 A = g(θ, φ)dω G ref (3.24) because G sky + G ref = 4π and because of the condition (3.20). T a can now be written as T a = T ba + T ref (1 A) (3.25) and we call A the equivalent transmission of the reference absorber. The variables in the following calculations are described in Table 3.2. The transmission

31 CHAPTER 3. MIDDLE ATMOSPHERIC WATER VAPOR RADIOMETER MIAWARA 27 of the reference absorber can be estimated by solving equation (3.25) for A where now T a corresponds to T b ref and T b corresponds to T b(φ ref ): A = T ref T b ref T ref T b(φ ref ) (3.26) According to Forkman et al. [2003] A can also be estimated as follows A = T ref (T eff T 0 )(e τµ ref e τµ line) + T b T b(φref ) T ref T b(φ ref ) (3.27) Note that (T eff T 0 )(e τµ ref e τµ line ) = T b(φ line ) T b(φ ref ) (3.28) If (3.28) is used in (3.27), T b(φ ref ) cancels out. From the measurement we know that T b = T b line T b ref and if we assume to have a proper value for τ then T b line = T b(φ line ). With this we end up at equation (3.26). The opacity is critical in the determination of A. Equations (3.26) and (3.27) can only compare well if the calibrated line spectrum, T b line, and the calculated line spectrum, T b(φ line ), are consistent. Figure 3.7 shows a scatter plot of these variables, once for the linearized determination of τ and once for the iterative determination. As expected, T b(φ line ) is underestimated in the rst case, because the opacity is systematically too low. On the other hand the iteratively determined value for τ leads to a good agreement between calibration and calculation. Figure 3.8 shows, how equation (3.27) compares to (3.26), again for both methods to derive the opacity. As expected both formulations are in agreement only if the calibration according to (3.17) and the calculation according to (3.12) are consistent. Table 3.2: Description of the variables. Variable Description T ref Physical temperature of the reference absorber T eff Mean radiating temperature of the troposphere T hot Physical temperature of the hot load absorber T cold Brightness temperature of the cold load (typically sky at 60 elevation) T 0 The cosmic background temperature µ ref The airmass in reference direction µ line The airmass in line direction T b(φ ref ) = T 0 e τµ ref + Teff (1 e τµ ref ), (no reference absorber!) T b(φ line ) = T 0 e τµ line + Teff (1 e τµ line ) T b ref = (U ref U hot ) T hot T cold U hot U cold + T hot, (with reference absorber!) T b line = (U line U hot ) T hot T cold U hot U cold + T hot

32 28 March 1 21, 2005, iterative τ determination Tbline,calculated [K] Tbline,calculated [K] March 1 21, 2005, linear τ determination Tbline,calibrated [K] Tbline,calibrated [K] Figure 3.7: Calibrated values against calculated values for the sky between 20 and 30 elevation for the linearized determination of τ, left panel, and for the iterative determination of τ, right panel. reference load transmission, March 1 21, AForkman AForkman reference load transmission, March 1 21, AHaefele AHaefele Figure 3.8: Transmission due to the reference absorber calculated according to equations (3.27), abscissa, and (3.26), ordinate. Once for the linearized determination of τ, left panel, and once for the iterative determination, right panel.

33 CHAPTER 3. MIDDLE ATMOSPHERIC WATER VAPOR RADIOMETER MIAWARA Noise of calibrated and corrected single spectra empiric calculated σ [K] /24 01/03 01/13 01/23 02/02 02/12 02/22 03/03 03/13 03/23 04/ Figure 3.9: Time series of empirically determined noise and of the theoretical value for the noise Optimized Observing Angle for Noise Minimization In order to reach a certain sensitivity of the retrieval to the measurement the measurement noise, σ, has to be reduced below a certain value by averaging over several individual spectra. Averaging over n spectra reduces the measurement noise by a factor of 1/ n. We are now going to focus on the noise of a single spectrum that is calibrated and corrected for the tropospheric attenuation, the transmission of the reference absorber and for the dierent airmasses, hence, of a spectrum above the tropopause in zenith direction as given by equation (3.10). The correction is a pure scaling of the balanced spectrum with the correction factor a, which depends on the opacity, the zenith angles of the line and reference measurement and on the transmission of the reference absorber. The noise of a single, balanced and corrected spectrum is given by: at sys σ = 2 (3.29) ftint T sys is the system temperature, f the bandwidth of a single channel and t int the integration time of a single spectrum. Note that a single spectrum consists of two wobbling positions and that t int is twice the integration time of a single spectrometer scan. For the FFT of MIAWARA t int = s= 6.54 s. The factor of 2 accounts for the fact that a Dicke switching is applied. The noise also scales with the correction factor a. Given the tropospheric opacity τ, the eective temperature T eff, the zenith angle of the line measurement θ and the physical temperature of the reference absorber bar equation (3.29) can be evaluated assuming balance between the line and the reference measurement: T b line = T b ref.

34 30 (σ(θ) σ(70))/σ(70) [%] Relative increase of noise with respect to θ=70 deg zenith angle [deg] Figure 3.10: Top panel: The noise as a function of zenith angle and opacity. The contours represent the theoretic values and the colored dots are derived from measured spectra. Bottom panel: Relative increase of the noise of a corrected single spectrum as a function of zenith angle and opacity. Figure 3.9 shows that equation (3.29) is in good agreement with the standard deviation of a small frequency interval of a measured, calibrated and corrected spectrum. Figure 3.10 shows the measurement noise as a function of the zenith angle of the line measurement and of the opacity. The contour lines represent the calculated noise according to equation (3.29) and the colored dots represent the standard deviation of measured, single, corrected spectra. Figure 3.10 exhibits that for opacities below 0.2 the optimal zenith angle is higher than 75 and that lower zenith angles always lead to more noisy spectra. For opacities above 0.2 the optimal zenith angle decreases slowly to 70 for τ = Figure 3.10 shows the relative increase of the noise of a single, corrected spectrum as a function of the zenith angle of the line measurement and the opacity. One can read from this gure that the noise increases by 50 % going from a line zenith angle of 70 to 60 for an opacity of 0.1. This is a substantial increase and it implies that the line zenith angle should be kept as high as possible. The only way to inuence the line zenith angle is to change the fraction of the beam that is covered by the reference absorber during the reference measurement. However this is critical as this changes the total reectivity of the reference absorber and thus aects the amplitude and structure of the baseline. This in turn can cause unnatural variability in the retrieved proles mainly below 1 hpa. In Haefele et al. [in press] this eect is referred to several times. At the time of writing there is no suitable solution to this problem available and one has to live with the drawback of considerably longer integration times during periods of enhanced tropospheric opacity.

35 CHAPTER 3. MIDDLE ATMOSPHERIC WATER VAPOR RADIOMETER MIAWARA 31 Table 3.3: Description of the variables used in the optimal estimation retrieval of H 2 O proles. y ɛ F (x) K K b x b x a ˆx S y S a S b measured spectrum at frequencies depending on spectrometer measurement noise calculated spectrum based on an atmospheric state x derivative of F with respect to x derivative of F with respect to b the true atmospheric state forward model parameters (T, p, spectroscopic and instrumental parameters) a priori assumption of the atmospheric state, i.e. of the H 2 O distribution the retrieved atmospheric state error covariance matrix of the measured spectrum error covariance matrix of the a priori assumption error covariance matrix of the forward model parameters 3.4 Retrieval The optimal estimation (OEM) approach is discussed in great detail in Rodgers [2000] and the adaption of the general theory to the measurements of MIAWARA is presented in Deuber [2005] and Haefele et al. [in press]. The following development of the characteristics and errors of the retrieval follows closely the conventions and notations introduced in Haefele et al. [in press, Section 5] and has to be considered as complementary to this. In this Section the idea of the retrieval is just briey sketched and more weight is put on the characterization of the retrieval and on the error analysis. The variables used in the following calculations are given in Table 3.3 and all vector or matrix valued parameters are written in bold face letters. All the retrieval and modeling work is done with the Atmospheric Radiative Transfer Simulator, ARTS [Buehler et al. 2005], and with Qpack [Eriksson et al. 2005], which is a collection of matlab routines for instrument simulation and retrieval work. The spectrum as it would be measured by the instrument, y, is a function of the vertical distribution of water vapor, x, and of a set of additional parameters, b: y = F (x, b) + ɛ. (3.30) The OEM solution of the inverse problem, ˆx, is the prole that minimizes the following cost function which can be derived from Bayes theorem: c = [y F (ˆx)] T S 1 y [y F (ˆx)] + [ˆx x a ] T S 1 a [ˆx x a ]. (3.31) The variables are described in Table 3.3. Solving dc/dˆx = 0 for ˆx yields ˆx = x a + (S 1 a + K T S 1 y K) 1 K T S 1 y (y F (x a )). (3.32)

36 32 We can now calculate the sensitivity of the retrieved state to the measurement ˆx y = D y = (S 1 a + K T S 1 y K) 1 K T S 1 y (3.33) and nally the averaging kernel, A, that describes the sensitivity of the retrieved state to changes in the true state ˆx x = ˆx y y x = D yk = A. (3.34) The averaging kernel is a key quantity for the characterization of the retrieved prole. It describes how the retrieval smoothes the true state and how sensitive it is to the a priori prole. The averaging kernels of the instruments that have been considered in the validation study [Haefele et al. in press] is shown in Figure With the derivatives given in equations (3.33) and (3.34) the propagation of the errors of the measurement and of the forward model parameters can be calculated according to Gaussian error propagation. In the error analysis we consider the following error sources that are described with the corresponding covariances: Measurement noise, S y ; temperature, S T ; troposphere correction factor, S a ; line intensity, S S ; air broadening, S γ. The Covariance of the retrieved prole is now given as: S x = (A I)S a (A I) T (3.35) +D y S y D y +D y K T S T K T TD T y +D y K a S a K T a D T a +D y K S S S K T S D T y +D y K γ S γ K T γ D T y The rst term on the right hand side is denoted as smoothing error which accounts for the fact that ne structures are not resolved in the retrieved prole. For a proper calculation of the smoothing error the statistics of the true state must be known. In our case, S a is just a very rough estimate of the covariance of the true state and hence the smoothing error that is given below might be only a poor approximation of this error. However, if the retrieved prole is regarded as an estimate of the true state smoothed by the averaging kernel, then the smoothing error has not to be considered. In other words: If we compare our retrieved water vapor proles to a highly resolved reference prole that has been smoothed with the averaging kernel of the retrieval [equation (4) in Haefele et al. in press], the smoothing error has not to be considered. The error of the retrieval is then given by the terms 2 to 6 in equation (3.35). The covariance matrices in equation (3.35) are diagonal matrices with S i,i = σ 2, where σ is an estimate of the standard deviation of the corresponding parameter (the same for all frequencies and all pressures in the case of measurement noise and temperature). The standard deviations for the ve error sources under consideration are given in Table 3.4.

37 CHAPTER 3. MIDDLE ATMOSPHERIC WATER VAPOR RADIOMETER MIAWARA 33 p [hpa] Onsala, Summer Onsala,Winter p [hpa] Bern, Summer Bern,Winter p [hpa] Seoul, Summer Seoul,Winter Mauna Loa Lauder z [km] Meas. Resp. / AVK*5 z [km] Meas. Resp. / AVK*5 Figure 3.11: Averaging kernels of the ve instruments that have been considered in the validation study [Haefele et al. in press]. At Onsala, Bern and Seoul the opacity shows a strong seasonal cycle and the averaging kernels are given for summer and winter conditions. The red curve indicates the response of the retrieved prole to the measurement.

38 34 Table 3.4: Estimates of the errors of some forward model parameters. In case of the line intensity and the air broadening the standard deviations are derived from the values given in [Haefele et al. in press, Table 1]. The error in the troposphere correction factor is further determined by a numerical simulation assuming the following uncertainties: τ = 5 %, φ = 0.5, T hot/ref = 2 K, T eff = 3.5 K (see Section 3.3). The resulting uncertainties of the retrieved prole are given in Figure 3.12 Parameter Standard Dev. Unit Measurement noise 0.01 K Temperature 5 K Troposphere correction 4 % Line intensity S 6.81e-21 m 2 Hz Air broadening γ air 1014 Hz/Pa The error of the troposphere correction factor is derived from a numerical perturbation calculation of equation (3.10) assuming the uncertainties given in Table 3.1. The standard deviations of the line strength, S, and of the air broadening parameter, γ, are calculated from the values given in Haefele et al. [in press, Table 4]. Figure 3.12 shows the vertically resolved errors of the retrieved proles. In the following discussion the smoothing error will be neglected as this error is of no importance if we compare our retrieval to the true state smoothed by the averaging kernel. The measurement noise, the troposphere correction factor and the air broadening parameter are the dominating error sources. Note that the troposphere correction factor includes the uncertainties of the pointing. The fact that the line intensity contributes much less to the error than the air broadening is because the line intensity is assumed to be better known (σ S = 0.5 %) than the air broadening parameter (σ γ = 3.7 %). The total systematic error is 7 % at 0.1 hpa and consideres all error sources except the measurement noise and the smoothing error. But note, that the error of the troposphere correction factor and the temperature may be random to some extent and that the total systematic error is hence an upper limit while the random error due to measurement noise is a lower limit.

39 CHAPTER 3. MIDDLE ATMOSPHERIC WATER VAPOR RADIOMETER MIAWARA MIAWARA v17 Pressure [hpa] 10 1 Measurement noise 10 0 Temperature Tropospheric correction Line intensity Air broadening Total systematic error Smoothing Relative Error [%] Figure 3.12: Errors in the retrieved prole resulting from the measurement noise and uncertainties in the temperature prole, the troposphere correction factor, the line intensity and the air broadening parameter (see Table 3.4). The total systematic error includes all sources except for the measurement noise and the smoothing error. See text.

40 Chapter 4 Seoul Water Vapor Radiometer SWARA 4.1 Introduction The Seoul Water Vapor Radiometer, SWARA, is a sister instrument of the Middle Atmospheric Water Vapor Radiometer, MIAWARA. It is a joint project of the University of Bern and the Sookmyung's Women University in Seoul, South Korea. The instrument was built at the Institute of Applied Physics in Bern and was shipped to Seoul in September Within this thesis a major part of the measurement software has been written from scratch or adopted from MIAWARA. The SWARA operating manual [Haefele 2006] is a rough guide to the software. Since the rst measurements strong artefacts were present in the spectra that did not show the structure and behavior one would expect from spectral features originating from standing waves. Instead, these artefacts could be attributed to resonances in the corrugated horn antenna that alter the antenna pattern as a function of frequency [De Wachter et al. 2009]. In the attempt to overcome the diculties with the antenna the instrument has been run in dierent congurations with dierent antennas and reference absorbers as described in detail in De Wachter et al. [2009]. As a consequence of these artefacts the practical bandwidth is very much limited and water vapor retrievals are thus conned to the mesosphere. 4.2 Instrument Description SWARA is a heterodyne receiver running in a single sideband mode. The radiation from the calibration targets and the sky is directed into a corrugated horn antenna by a rotating mirror at an inclination of 45. The incoming signal at a radio frequency (RF) of GHz is amplied and downconverted in two steps to an intermediate frequency (IF) of 0.5 GHz where it is spectrally analyzed by a digital FFT. The diagram of Figure 3.3 is valid for SWARA as well. Key specications of the instrument are given in Table

41 CHAPTER 4. SEOUL WATER VAPOR RADIOMETER SWARA 37 Figure 4.1: The Seoul Water Vapor Radiometer, SWARA, in Seoul, South Korea. Table 4.1: Key specications of SWARA. Calibration Technique Balancing calibration Operational mode Single sideband mode RF amplication 2 uncooled HEMT ampliers (35 db and 30 db) Mirror Plane mirror Antenna Corrugated horn antenna (Θ HP BW = 6 ) Receiver temperature 140/190 K (depending on conguration) Radio frequency range ± 0.5 GHz Spectral analysis Acqiris digital FFT: f = 0 1 GHz, f = 61 khz

42 Results Figure 4.2 shows calibrated and corrected spectra (see T b in Section 3.3) measured in January 2007 and 2009 in Seoul. The spectrum from January is heavily contaminated by the resonant eects in the horn antenna. The horn has been replaced by a spare horn which performs better, as can be seen in the bottom panel of Figure 4.2. However, all water vapor data presented in this work are based on a bandwidth of only 15 MHz centered around line center. This central part is marginally aected by the resonant features and does show negligible dependence on the conguration of the instrument. With this retrieval setup a continuous time series since October 2006 could be generated. The retrieval is described in Haefele et al. [in press] and the mesospheric part of the error analysis presented in Section 3.4 applies for SWARA retrievals as well. SWARA averaging kernels are shown in Figure 3.11 for summer and winter conditions. A time series of mesospheric water vapor as measured by SWARA since the start of operation in October 2006 until the time of writing is shown in Figure 4.3 in comparison with observations made by the Microwave Limb Sounder (MLS) onboard the Aura satellite. MLS data are convolved with the averaging kernles of SWARA. A more detailed validation is given in Haefele et al. [in press]. SWARA is taken care of in the frame of an other dissertation and for more details about the instrument and the evolution of the measurements it is referred to separate publications.

43 CHAPTER 4. SEOUL WATER VAPOR RADIOMETER SWARA SWARA, Seoul (37N, 126E) Tb [K] Tb [K] f [GHz] Figure 4.2: Calibrated, balanced and corrected spectra measured by SWARA on December 1, 2007 (top panel) and on January 1, 2009 (bottom panel). The spectrum in the upper panel is contaminated by strong features originating from resonances in the horn antenna. For the retrieval only the very center of the line (±7.5 MHz) is used in this data version. The spectrum of the bottom panel is measured with the same instrument but with a dierent antenna. Data courtesy of E. De Wachter. 0.2 H 2 O VMR [ppmv] SWARA, Seoul, p=0.03 hpa ( 72km) MLS conv SWARA Jan07 Jul07 Jan08 Jul08 Jan09 Jul09 Figure 4.3: Time series of water vapor on the p = 0.03 hpa pressure level as observed by SWARA (red) and by MLS (grey). MLS data have been convolved with the averaging kernels of SWARA.

44 Bibliography Bevilacqua, R. M., W. J. Wilson, W. B. Ricketts, P. R. Schwartz, and R. J. Howard. Possible seasonal variability of mesospheric water vapor. Geophys. Res. Lett., 12(6): , Brasseur, G. and S. Solomon. Aeronomy of the Middle Atmosphere. Springer, Brewer, A. W. Evidence for a world circulation provided by measurements of helium and water vapour distribution in the stratosphere. Q. J. Roy. Meteorol. Society, Buehler, S. A., P. Eriksson, T. Kuhn, A. von Engeln, and C. Verdes. Arts, the atmospheric radiative transfer simulator. J. of Quantitative spectroscopy and Radiative Transfer, 91: 6593, doi: /j.jqsrt De Wachter, E., A. Murk, C. Straub, A. Haefele, S. Ka, J. J. Oh, and N. Kämpfer. Eects of Resonances in Corrugated Horn Antennas for a 22 GHz Balancing Radiometer. IEEE Geoscience and Remote Sensing Letter, 6(1):37, doi: /LGRS Deuber, B. Microwave Remote Sensing of Water Vapour in the Middle Atmosphere - Calibration, Validation and Applications in Atmospheric Studies. PhD thesis, Philosophisch- Naturwissenschaftliche Fakultät, Universität Bern, Deuber, B., N. Kämpfer, and D. G. Feist. A new 22-GHz Radiometer for Middle Atmospheric Water Vapour Prole Measurements. IEEE Transactions on Geoscience and Remote Sensing, 42(5): , Deuber, B., A. Haefele, D. G. Feist, L. Martin, N. Kämpfer, G. E. Nedoluha, V. Yushkov, S. Khaykin, R. Kivi, and H. Vömel. Middle Atmospheric Water Vapour Radiometer - MIAWARA: Validation and rst results of the LAUTLOS / WAVVAP campaign. J. Geophys. Res., 110(D13306), doi: /2004JD Eriksson, P., C. Jiménez, and S. A. Buehler. Qpack, a general tool for instrument simulation and retrieval work. J.of Quantitative Spectroscopy and Radiative Transfer, 91:4764, doi: /j.jqsrt

45 BIBLIOGRAPHY 41 Evans, S. J., R. Toumi, J. E. Harries, M. R. Chippereld, and J. M. R. III. Trends in stratospheric humidity and the sensitivity of ozone to these trends. J. Geophys. Res., 103(D8): , Fiedler, J., G. Baumgarten, and F.-J. Lübken. NLC observations during one solar cycle above ALOMAR. Journal of Atmospheric and Solar-Terrestrial Physics, 71:424433, Flury, T., K. Hocke, A. Haefele, N. Kämpfer, and R. Lehmann. Ozone Depletion, Water Vapor Increase and PSC 1 Generation at Mid-Latitudes by the 2008 Major Stratospheric Warming. J. Geophys. Res., in press. doi: /2009JD Forkman, P., P. Eriksson, and A. Winnberg. The 22 GHz radio-aeronomy reciver at Onsala Space Observatory. J. Quant. Spectrosc. Radiat. Transfer, 77(1):2342, doi: /S (02) Forster, P., V. Ramaswamy, P. Artaxo, T. Berntsen, R. Betts, D. Fahey, J. Haywood, J. Lean, D. Lowe, G. Myhre, J. Nganga, G. R. R. Prinn, M. Schulz, and R. V. Dorland. Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, chapter 2. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, Forster, P. M. d. F. and K. P. Shine. Assessing the climate impact of trends in stratospheric water vapor. Geophys. Res. Lett., 29(6), doi: /2001GL Fueglistaler, S., M. Bonazzola, P. Haynes, and T. Peter. Stratospheric water vapor predicted from the Lagrangian temperature history of air entering the stratosphere in the tropics. J. Geophys. Res., 110(D8), doi: /2004JD Haefele, A. Miawara operating manual: 2nd edition. Technical report, University of Bern, Institute of Applied Physics, Haefele, A. Swara operating manual. Technical report, University of Bern, Institute of Applied Physics, Haefele, A. and N. Kämpfer. Tropospheric water vapor proles retrieved from pressure broadened emission spectra at 22 GHz. J. of Atm. and Oc. Tech., subm. Haefele, A., K. Hocke, N. Kämpfer, P. Keckhut, M. Marchand, S. Bekki, B. Morel, T. Egorova, and E. Rozanov. Diurnal changes in middle atmospheric H2O and O3: Observations in the Alpine region and climate models. J. Geophys. Res., 113(D17303), doi: /2008JD Haefele, A., E. D. Wachter, K. Hocke, N. Kämpfer, G. E. Nedoluha, R. M. Gomez, P. Eriksson, P. Forkman, A. Lambert, and M. J. Schwartz. Validation of ground based microwave

46 42 radiometers at 22 GHz for stratospheric and mesospheric water vapor. J. of Geophys. Res., in press. doi: /2009JD Han, Y. and E. R. Westwater. Analysis and improvement of tipping calibration for groundbased microwave radiometers. IEEE Transactions on Geoscience and Remote Sensing, 38(3): , Held, I. M. and B. Soden. Water vapor feedback and global warming. Annu. Rev. Energy Environ., 25:441475, Janssen, M. A. Atmospheric Remote Sensing by Microwave Radiometry. John Wiley and Sons, Kley, D., J. M. R. III, and C. P. (editors). SPARC Assessment of Upper Tropospheric and Stratospheric Water Vapour Klose, B. Meteorologie. Springer, Lahoz, W. A., E. S. Carr, L. Froidevaux, R. S. Harwood, J. B. Kumer, J. L. Mergenthaler, G. E. Peckham, W. G. Read, P. D. Ricaud, A. E. Roche, and J. W. Waters. Hemisphere Mid-Stratosphere Vortex Processes Diagnosed from H2O, N2O and Potential Vorticity. Geophys. Res. Letters, 20(23): , Lambert et al., A. Validation of the aura microwave limb sounder middle atmosphere water vapor and nitrous oxide measurements. J. Geophys. Res., 112(D24S36), doi: /2007JD Marsh, D., A. Smith, and E. Noble. Mesospheric ozone response to changes in water vapor. J. Geophys. Res., 108(D3), doi: /2002JD Mote, P. W., K. H. Rosenlof, J. R. Holton, R. S. Harwood, and J. W. Waters. Seasonal variations of water vapor in the tropical lower stratosphere. Geophysical Research Letters, Müller, S. C., A. Murk, C. Monstein, and N. Kämpfer. Intercomparison of Digital Fast Fourier Transform and Accousto Optical spectrometers for microwave radiometry of the atmosphere. IEEE Trans. on Geosc. and Remote Sensing, 47: , Nedoluha, G. E., R. M. Bevilacqua, R. M. Gomez, W. B. Waltman, B. C. Hicks, D. L. Thacker,, and W. A. Matthews. Measurements of water vapor in the middle atmosphere and implications for mesospheric transport. J. Geophys. Res., 101(D16):21,18321,193, Nedoluha, G. E., R. M. Bevilacqua, R. M. Gomez, D. E. Siskind, B. C. Hicks, J. M. R. III, and B. J. Connor. Increases in middle atmospheric water vapor as observed by the Halogen Occultation Experiment and the ground-based Water Vapor Millimeter-wave Spectrometer from Geophys. Res. Let., 103(D3): , 1998.

47 BIBLIOGRAPHY 43 Oltmans, S. J. and D. J. Hofmann. The increase in stratospheric water vapor from balloonborne, frostpoint hygrometer measurements at washington, d.c. and boulder, colorado. Geophys. Res. Lett., Palm, M., N. Wieters, M. Sinnhuber, S. H. W. Golchert, C. G. Homann, J. Notholt, and G. Hochschild. Stratospheric and Mesospheric O 3 above Spitsbergen modeled by a 3D chemical transport model and measured by groundbased millimeter wave radiometry. subm. J. of Geophys. Res., Pan, L. L., W. J. Randel, H. Nakajima, S. T. Massie, H. Kanzawa, Y. Sasano, T. Yokota, T. Sugita, S. Hayashida, and S. Oshchepkov. Satellite observation of dehydration in the arctic polar stratosphere. Geophys. Res. Lett., 29(8):1184, doi: /2001GL Parrish, A., R. L. dezafra, P. M. Solomon, and J. W. Barrett. A groundbased technique for millimeter wave spectroscopic observations of stratospheric trace constituents. Radio Science, 23:106118, Randel, W. J., F. Wu, and D. J. Gaen. Interannual variability of the tropical tropopause derived from radiosonde data and NCEP reanalyses. J. Geophys. Res., 105(D12):15,509 15,523, Randel, W. J., F. Wu, H. Vömel, G. E. Nedoluha, and P. Forster. Decreases in stratospheric water vapor after 2001: Links to changes in the tropical tropopause and the Brewer- Dobson circulation. J. Geophys. Res., 111(D12312), doi: /2005JD Randel et al., W. J. An update of observed stratospheric temperature trends. J. Geophys. Res., 114(D02107), doi: /2008JD Remsberg, E. E., J. M. R. III, L. L. Gordley, J. C. Gille, and P. L. Bailey. Implications of the Stratospheric Water Vapor Distribution as Determined from the Nimbus 7 LIMS Experiment. J. of the Atm. Sci., 41: , doi: / (1984)041<2934:IOTSWV>2.0.CO;2. Rodgers, C. D. Inverse Methods for Atmospheric Soundings. World Scientic Publishing Co. Pte. Ltd., Rogers, A., M. Lekberg, and P. Pratap. Seasonal and Diurnal Variations of Ozone near the Mesopause from Observations of the GHz Line. J. of Atm. and Oc. Tech., doi: /2009JTECHA Rosenlof, K. H. and G. C. Reid. Trends in the temperature and water vapor content of the tropical lower stratosphere: Sea surface connection. J. Geophys. Res., 113(D06107), doi: /2007JD Rosenlof et al., K. H. Stratospheric water vapor increases over the past halfcentury. Geophys. Res. Lett., 28(7): , 2001.

48 44 Schneider, N., F. Selsis, J. Urban, L. Olivier, J. Noë, and P. Ricaud. Seasonal and Diurnal Ozone Variations: Observations and Modeling. Journal of Atmospheric Chemistry, 50 (1):2547, Seele, C. and P. Hartogh. Water vapor of the polar middle atmosphere: Annual variation and summer mesosphere Conditions as observed by groundbased microwave spectroscopy. Geophys. Res. Lett., 26(11): , Seele, C. and P. Hartogh. A case study on middle atmospheric water vapor transport during the February 1998 stratospheric warming. Geophys. Res. Lett., 27(20): , Simmons, A. J., A. Untch, C. Jakob, P. Kåallberg, and P. Undén. Stratospheric water vapour and tropical tropopause temperatures in ECMWF analyses and multi-year simulations. Q. J. R. Meteorol. Sor., 125:353386, Siskind, D. E. and M. E. Summers. Implications of enhanced mesospheric water vapor observed by HALOE. Geophys. Res. Lett., 25(12): , Soden, B. J., D. L. Jackson, V. Ramaswamy, M. D. Schwarzkopf, and X. Huang. Rhe Radiative Signature of Uper Tropopheric Moistening. Science, 310:841844, Sonnemann, G. R., M. Grygalashvyly, and U. Berger. Autocatalytic water vapor production as a source of large mixing ratios within the middle to upper mesosphere. J. Geophys. Res., 110(D15303), doi: /2004JD Sonnemann, G. R., P. Hartogh, M. Grygalashvyly, S. Li, and U. Berger. The quasi 5- day signal in the mesospheric water vapor concentration at high latitudes in 2003-a comparison between observations at ALOMAR and calculations. J. of Geophys. Res., 113(D04101), doi: /2007JD Summers, M. E. and R. R. Conway. Insights into Middle Atmospheric Hydrogen Chemistry from Analysis of MAHRSI OH Observations. AGU, Summers, M. E. and D. E. Siskind. Surface recombination of O and H2 on meteoric dust as a source of mesospheric water vapor. Geophys. Res. Lett., 26(13): , Ulaby, F. T., R. K. Moore, and A. K. Fung. Microwave Remote Sensing. Reading, MA: AddisonWesley, von Zahn, U. and U. Berger. Persistent ice cloud in the midsummer upper mesosphere at high latitudes: Three-dimensional modeling and cloud interactions with ambient water vapor. J. Geophys. Res., 108(D8):8451, doi: /2002JD Zommerfelds, W. C., K. F. Kunzi, M. E. Summers, R. M. Bevilacqua, D. F. Strobel, M. Allen, and W. J. Sawchuck. Diurnal variations of mesospheric ozone obtained by groundbased microwave radiometry. J. Geophys. Res., 94(D10):12,81912,832, 1989.

49 Chapter 5 Diurnal Changes in Middle Atmospheric H 2 O and O 3 : Observations in the Alpine Region and Climate Models Published in the Journal of Geophysical Research 45

50 Click Here for Full Article JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 113, D17303, doi: /2008jd009892, 2008 Diurnal changes in middle atmospheric H 2 OandO 3 : Observations in the Alpine region and climate models A. Haefele, 1 K. Hocke, 1 N. Kämpfer, 1 P. Keckhut, 2 M. Marchand, 3 S. Bekki, 3 B. Morel, 4 T. Egorova, 5 and E. Rozanov 5,6 Received 31 January 2008; revised 26 June 2008; accepted 2 July 2008; published 5 September [1] In this paper we investigate daily variations in middle atmospheric water vapor and ozone based on data from two ground-based microwave radiometers located in the Alpine region of Europe. Temperature data are obtained from a lidar located near the two stations and from the SABER experiment on the TIMED satellite. This unique set of observations is complemented by three different three-dimensional (3-D) chemistry-climate models (Monitoring of Stratospheric Depletion of the Ozone Layer (MSDOL), Laboratoire de Météorologie Dynamique Reactive Processes Ruling the Ozone Budget in the Stratosphere (LMDz-REPROBUS), and Solar Climate Ozone Links (SOCOL)) and the 2-D atmospheric global-scale wave model (GSWM). The first part of the paper is focused on the first Climate and Weather of the Sun-Earth System (CAWSES) tidal campaign that consisted of a period of intensive measurements during September Variations in stratospheric water vapor are found to be in the order of 1% depending on altitude. Meridional advection of tidal nature is likely to be the dominant driving factor throughout the whole stratosphere, while vertical advection becomes more important in the mesosphere. Observed ozone variations in the upper stratosphere and lower mesosphere show amplitudes of several percent in accordance with photochemical models. Variations in lower stratospheric ozone are not solely governed by photochemistry but also by dynamics, with the temperature dependence of the photochemistry becoming more important. The second part presents an investigation of the seasonal dependence of daily variations. Models tend to underestimate the H 2 O diurnal amplitudes, especially during summer in the upper stratosphere. Good agreement between models and observations is found for ozone in the upper stratosphere, which reflects the fact that the O 3 daily variations are driven by the photochemistry that is well modeled. Citation: Haefele, A., K. Hocke, N. Kämpfer, P. Keckhut, M. Marchand, S. Bekki, B. Morel, T. Egorova, and E. Rozanov (2008), Diurnal changes in middle atmospheric H 2 O and O 3 : Observations in the Alpine region and climate models, J. Geophys. Res., 113, D17303, doi: /2008jd Institute of Applied Physics, University of Bern, Bern, Switzerland. 2 Service d Aéronomie du CNRS, University of Versailles, Verrières-le Buisson, France. 3 Service d Aéronomie, University Pierre et Marie Curie, Paris, France. 4 Université delaréunion, La Réunion, France. 5 Physical-Meteorological Observatory, World Radiation Center, Davos, Switzerland. 6 Institute for Atmospheric and Climate Science, ETH Zurich, Zurich, Switzerland. Copyright 2008 by the American Geophysical Union /08/2008JD009892$ Introduction [2] One important driver of daily variations in middle atmospheric trace constituents is photodissociation. Depending on altitude and constituent these changes can show large amplitudes and abrupt transitions at sunrise or sunset quite similar to step functions as in the case of ozone in the upper stratosphere and mesosphere. In addition the periodic heating of the atmosphere by absorption of solar radiation by stratospheric ozone and tropospheric water vapor as well as by latent heat release excites planetary-scale waves that propagate in the whole atmosphere. The sun-synchronous waves of wavenumbers 1 and 2 are denoted as the diurnal and semidiurnal migrating tides. These tides are spatiotemporal disturbances in temperature, density, pressure and wind which in turn produce variations in the distribution of constituents through modulation of temperature-dependent reaction rates, compression, and advection given a nonvanishing gradient in concentration. [3] The diurnal variation of upper stratospheric and mesospheric ozone has been intensively discussed [Prather, 1981; Pallister and Tuck, 1983; Vaughan, 1984]. Odd oxygen (O x =O+O 3 ) is produced during the day through photolysis of molecular oxygen. In the middle stratosphere (up to km) all odd oxygen is in the form of ozone. As a consequence, ozone volume mixing ratios are enhanced during the day and reach a maximum in the late afternoon. At higher levels the [O]/[O 3 ] ratio increases owing to its inverse dependence on density. Therefore more and more of 46 D of13

51 D17303 HAEFELE ET AL.: DIURNAL CHANGES IN H 2 O AND O 3 D17303 the odd oxygen resides as atomic oxygen and ozone shows a strong depletion during the day. After dusk all the O recombines with O 2 to O 3. In addition, odd oxygen is catalytically destroyed during the day by NO x,clo x and HO x. These processes lead to a characteristic daytime variation in ozone and to a dependence of the night/day ratio of ozone on water vapor, which is the main source of HO x compounds [Vaughan, 1984; Marsh et al., 2003]. [4] The temperature dependence of the ozone concentration near the stratopause was first noted by Barnett and Pyle [1975]. Rood and Douglass [1985] and Douglass et al. [1985] investigated the relationship between ozone and temperature perturbations taking photochemistry and dynamics into account and found that a wide range of phase relationships can be observed depending on the importance of the dynamical terms. Froidevaux et al. [1989] analyzed the coupling of ozone and temperature taking into account only photochemistry. In that case the phase relationship is governed by the ratio of the ozone chemical relaxation time, t c,to the period of the temperature perturbation, t p. For t p /t c 1 photochemical equilibrium is always maintained and the perturbation in ozone mixing ratio, f 0 O 3, can be estimated from the temperature perturbation, T 0,as f 0 O 3 ¼ Q Ef O3 T 2 T 0 ; ð1þ where f O3 is the ozone mixing ratio, T is temperature, Q E is the equilibrium sensitivity coefficient and its value is in the range of 1400 K [Barnett and Pyle, 1975] depending on altitude and for pure oxygen photochemistry. In case of photochemical equilibrium, ozone and temperature are out of phase by 12 h. For temperature perturbations on a timescale of less than about 4t c, equilibrium cannot be maintained and as a consequence the sensitivity decreases and ozone leads temperature by 6 h [Froidevaux et al., 1989]. In case of very fast temperature perturbations, t p /t c < 0.1, ozone no longer shows any response. Some initial results on the nature of short-term (1 3 h) fluctuations of the ozone volume mixing ratio around the stratopause have been reported by Hocke et al. [2006]. [5] Adiabatic vertical transport also constitutes an important link between temperature and ozone revealing a positive correlation for positive vertical gradients (ozone increases with height) and negative correlation for negative vertical gradients. The quantitative connection between temperature perturbation and vertical displacement is given in section 2.3. [6] In contrast to ozone little is known about the diurnal behavior of middle atmospheric water vapor because the variations are small and difficult to measure [Randel, 1990]. H 2 O enters the middle atmosphere through the tropical tropopause or is formed by methane oxidation in the stratosphere [see Randel et al., 1998, and references therein]. The long chemical lifetime of water vapor makes it a good tracer for atmospheric transport processes. Daily variations of H 2 O in the middle atmosphere are therefore expected to be dominated by advection. Large vertical gradients are found in the lower mesosphere and meridional gradients are maximal during autumn in the stratosphere. However, diurnal amplitudes in vertical and meridional winds are small at these altitudes and hence amplitudes in water vapor are expected in the order of 1% or less, only. To our knowledge this effect has so far not been observed successfully. The small amplitudes require a large amount of observations at a fixed location in order to reduce statistical noise. These requirements are met by the ground based radiometers reported here. [7] In order to improve our knowledge of the composition change under the combined effect of photochemistry and tidal advection processes, observations are required. Up to now there exists a variety of numerical models dedicated to climate investigations, data assimilation, ozone-climate processes, planetary waves or other issues. Many of such models have been investigated with respect to diurnal variations [see, e.g., McLandress, 1997; Hecht et al., 1998; Morel et al., 2004]. Since the daily variations result from dynamical-radiative interaction processes, their investigation and validation with observations provide helpful information for the evolution and improvement of the numerical models. [8] In this work we investigate daily variations in middle atmospheric water vapor, ozone and temperature observed at Alpine NDACC stations (Network for the Detection of Atmospheric Composition Change). The observed daily variations are compared to the 3-D climate models MSDOL [Bertaux et al., 1999], LMDz [Eyring et al., 2006] (both are expected to implicitly include tides), SOCOL [Egorova et al., 2005] and to the 2-D linearized global scale wave model GSWM [Hagan et al., 1999], specifically developed to reproduce tides of the dynamical parameters. The use of both observational and model data should increase the confidence in the derived diurnal cycles and allow for a better characterization. [9] The paper is organized as follows. In section 2 the data sets obtained from measurements and models are presented, and the derivation of daily variations is described. Section 3 is dedicated to the first CAWSES campaign with a detailed description and qualitative comparison of the daily variations derived from observational and model data. In section 4 the analysis is extended to a longer time period, and seasonal variations in the daily cycles are investigated with special focus on lower mesospheric water vapor. A summary and conclusions are given in section Data Sets and Analysis 2.1. Observations MIAWARA [10] The composition of the middle atmosphere is investigated with two ground based microwave radiometers operated in Switzerland. The Middle Atmospheric Water Vapor Radiometer, MIAWARA [Deuber et al., 2004], is run by the University of Bern. The measurement site is close to Bern (47 N,7 E) and is one of the Alpine NDACC stations (Network for the Detection of Atmospheric Composition Change). MIAWARA measures the pressure broadened emission line of H 2 O at 22 GHz which allows to retrieve vertical water vapor profiles between 25 and 65 km at a sampling rate of one profile per hour. The back end consists of an acousto-optical spectrometer with a frequency resolution of 1.2 MHz and for the inversion a bandwidth of 300 MHz centered around the line center frequency has been chosen. The ARTS [Buehler et al., 2of13 47

52 D17303 HAEFELE ET AL.: DIURNAL CHANGES IN H 2 O AND O 3 D ] and QPack [Eriksson et al., 2005] software packages are used for the retrieval. The width of the averaging kernels lies between 12 and 18 km at altitudes of 30 and 50 km, respectively. The measurement is equally accurate for daytime and nighttime. Rain, however, does not allow to take measurements. The total error (sum of systematic error, noise and retrieval smoothing) is less than 30% between 35 and 55 km and less than 40% between 30 and 65 km. Precision, however, is more important in the derivation of daily variations and is estimated from the standard deviation of the retrieved water vapor values and is less than 6% over the whole altitude range. In order to detect variations in H 2 O of less than 1% a time period of at least one month is required. [11] Within the investigation of such small variations also instrumental effects and calibration issues have to be considered. The spectra measured at the ground have to be corrected for the troposphere before information from the middle atmosphere can be retrieved. This correction is a pure scaling of the spectrum and it depends mainly on the tropospheric opacity. Any time-dependent error in this scaling factor could produce diurnal variations in the retrieved water vapor values. We thus compared carefully the diurnal cycles in opacity with those in water vapor and could not find any correlation that is persistent throughout the whole year. Also the seasonal evolution of the diurnal amplitude in opacity differs significantly from the one in water vapor. The same analysis has been done with receiver temperature and water vapor revealing the same results. There is no evidence that the diurnal variations in water vapor reported in this paper are driven by the tropospheric correction or by a diurnally evolving error in the calibration SOMORA [12] The Stratospheric Ozone Monitoring Radiometer, SOMORA, is run by MeteoSwiss at Payerne (47 N,7 E), another Alpine NDACC station [Hocke et al., 2007]. It measures the pressure broadened emission line of O 3 at 142 GHz allowing to retrieve vertical profiles of volume mixing ratio between 25 and 55 km with a vertical resolution of 8 10 km and a time resolution of 30 min by using the optimal estimation method of Rodgers [1976]. The back end contains two acousto-optical spectrometers and the total of 3072 channels is reduced to 57 channels with a channel width of MHz. The total error (sum of systematic error, noise, and retrieval smoothing error) is less than 15% at altitudes from 20 to 40 km and around 30% in the lower mesosphere [Calisesi, 2003]. The precision is derived from the standard deviation of the nighttime ozone values between 20 and 4 h local solar time and lies between 2 and 3% of the mean nighttime value below 40 km and increases to 7% at 55 km. The instrument is in operation since the year Both radiometers are run continuously within NDACC Retrieval Issues [13] In the retrieval process information from the measurement, for example, the intensity spectrum of a transition line of the molecule of interest, is added to a first guess of the vertical distribution of the constituent under consideration, namely to the a priori profile. The retrieved profile is thus a mix of an a priori guess and information from the measurement and it is characterized by the averaging kernel, AVK, which is the response of the retrieved profile to a change in the true atmosphere. The FWHM and the area of the AVK functions are measures of the vertical resolution and the sensitivity of the retrieval to the measurement. To compare the retrieved profile to an other measured or modeled profile with higher vertical resolution, the AVK should be taken into account. However, we decided not to take the AVK into account in the analysis for the following reasons: In case of water vapor only little is known about daily variations in the middle atmosphere. As it is the aim to characterize these variations it does not make sense to impose the deficiencies of the observations to the model data even though this would improve the comparison. Interesting features would get lost in the model data that are likely to happen in the real atmosphere as well but that do not show up in the observations owing to their limitations in resolution and sensitivity. As a control most of the analysis has been done also under consideration of the AVK and whenever large discrepancies between observations and models can be attributed to AVK effects it is discussed in the text. [14] The altitude range where the contribution of the measurement is high is mainly given by the spectral bandwidth and the resolution of the measurement. The thermal noise on the measured intensity spectrum is mapped to the ozone and water vapor profiles by the retrieval. Hence the retrieved mixing ratio values are most noisy at altitudes where the contribution of the measurement is largest and decrease toward the upper and lower boundary with increasing contribution of the a priori profile. This explains why the time series of water vapor as shown in Figure 3 in section are more noisy at 3.14 hpa than at 0.10 hpa, for instance. The decreasing contribution of the measurement to the retrieved profile leads to an underestimation of the daily variations at the upper and lower limit of the altitude range as the daily variations are not accounted for in the a priori profile. The investigated altitude range is restricted to the levels where the contribution of the a priori value is less than 40%. A priori information on H 2 O is based on the US standard atmosphere and is kept constant for all seasons. The a priori profiles of O 3 are two climatological mean profiles for summer and winter. Pressure and temperature profiles are also required to retrieve the volume mixing ratio profiles and ECMWF analyses at a 6-hourly basis are used OHP Lidar [15] Since 1979 routine lidar measurements are conducted at the Observatory of Haute-Provence (OHP) in France (42 N,7 E), another Alpine NDACC station. This site is located 3 south from Bern and 360 km apart which is quite small compared to tidal scales. During clear nights, the lidar can be operated continuously and provides hourly temperature profiles between 30 and 80 km. The statistical noise is better than 1 K below 70 km and improves at lower altitudes up to a few tenth of degree at 30 km [Hauchecorne and Chanin, 1980]. Because this lidar uses the 532 nm wavelength, it cannot operate during daytime with the same accuracy and only part of the diurnal cycle can be retrieved. However, the diurnal evolution can be observed during nighttime and the tidal signature can be extracted from a composite time series (see section 2.3) Models MSDOL [16] The Monitoring of Stratospheric Depletion of the Ozone Layer system, MSDOL [Bertaux et al., 1999], is a 48 3of13

53 D17303 HAEFELE ET AL.: DIURNAL CHANGES IN H 2 O AND O 3 D D dynamics-chemistry-transport mechanistic model running with an interactive ozone chemistry using the REPROBUS (Reactive Processes Ruling the Ozone Budget in the Stratosphere) code [Lefèvre et al., 1994, 1998]. The ozone field can be fully interactive or adjusted by data assimilation [Bertaux et al., 1999]. The model covers the altitude range from 10 to 80 km with a resolution of about 3 km in the vertical, 5 in latitude and 11.6 in longitude. For this study MSDOL has been running in an interactive mode without data assimilation. However, to produce realistic values horizontal wind, temperature and geopotential fields below 10 hpa have been nudged to European Center for Medium range Weather Forcast (ECMWF) daily reanalysis data that are available every 6 h. The ability of MSDOL to reproduce tides has been investigated by Morel et al. [2004]. Radiative forcing is included in the dynamical equations based on solar heating due to O 3 and O 2 absorption, and CO 2 infrared cooling. The solar heating due to the absorption of UV radiation by O 3 and O 2 is computed using the parameterization of Strobel [1978], with coefficients updated from Zhu [1994]. CO 2 cooling rates for the 15 and 4.3 mm bands are calculated using the algorithm supplied by Fels and Schwarzkopf [1981] and Schwarzkopf and Fels [1985]. Gravity wave drag is simulated using a simple Rayleigh friction. The chemical model includes 33 chemical species and 5 chemical families. The species are grouped into two categories: short-lived species are assumed to be in photochemical equilibrium in their family; long-lived species are affected by both photochemistry and transport. The advection scheme implemented in MSDOL is a van Leer Eulerian scheme [van Leer, 1977] LMDz [17] LMDz-REPRO (LMD: Laboratoire de Météorologie Dynamique, REPRO: see section 2.2.1) is a fully coupled 3-D Chemistry-Climate Model (CCM) [Eyring et al., 2006] that is one of the components of the IPSL (Institut Pierre-Simon Laplace) earth system model. The dynamical model used is the stratospheric extension of the LMDz fourth-generation atmospheric GCM described by Lott et al. [2005]. It is a gridpoint model. In the horizontal direction, the equations are discretized on a staggered latitude-longitude Arakawa-C grid. It currently uses a uniform resolution of 2.5 in latitude and 3.75 in longitude. The vertical coordinate is a hybrid sigma pressure. It currently uses 50 levels with the upper boundary near 65 km. The resolution in the stratosphere varies slowly from 1 km at 12 km to 3 km at 50 km and reaches 6 km at the model top. The salient features of the physical parameterizations used in the model are a radiation scheme based on the ECMWF scheme [Morcrette, 1989], a convection scheme based on work by Tiedke [1989], a Subgrid Scale Orography (SSO, which forces orographic gravity waves) scheme based on work by Lott and Miller [1997] and Lott [1999], and a Doppler-spread nonorographic gravity waves scheme based on work by Hines [1997] and adapted from Manzini and McFarlane [1998]. The transport of tracers is calculated using the Van Leer scheme I (a first-order volume finite scheme with slope limitation) [van Leer, 1977; Hourdin and Armengaud, 1999]. The model is interactively coupled to the module of atmospheric chemistry from the REPROBUS chemistry transport model. The module contains a comprehensive stratospheric chemistry scheme. It describes a large range of chemical reactions associated with species from O x, HO x,clo x,bro x and NO x families and source gases. The model includes both gas-phase chemistry and heterogeneous chemistry on aerosols and polar stratospheric clouds. The model-calculated fields of radiatively active species such as ozone or CH 4 are used as input variables of the radiative routine. As a result, chemistry impacts dynamics. In return, the model-calculated winds are used to transport the chemical tracers and the model-calculated temperatures are used in the chemical module. Therefore, dynamics impacts chemistry. Because of this configuration, most CCMs are fully interactive and are probably the most suitable tools to make predictions of the atmospheric composition in a changing climate SOCOL [18] The chemistry-climate model SOCOL (Solar Climate Ozone Links) is a combination of the middle atmosphere MA-ECHAM4 General Circulation Model and a modified version of the atmospheric chemistry-transport model MEZON (Model for the Evaluation of Ozone Trends). The MA-ECHAM4 model is the middle atmosphere version of ECHAM4 (European Center/Hamburg 4), which has been developed at the Max Planck Institute for Meteorology in Hamburg, Germany. It is a spectral model with T30 horizontal truncation resulting in a grid spacing of about 3.75 ; in the vertical direction the model has 39 levels in a hybrid sigma-pressure coordinate system spanning the model atmosphere from the surface to 0.01 hpa; a semiimplicit time stepping scheme with weak filter is used with a time step of 15 min for dynamical processes and physical process parameterizations; full radiative transfer calculations are performed every 2 h, but heating and cooling rates are calculated every 15 min. The model includes the parameterizations for the orographic gravity wave and momentum flux deposition due to a continuous spectrum of vertically propagating gravity waves. A more detailed description of MA-ECHAM4 is given by Manzini and McFarlane [1998, and references therein]. The chemicaltransport part of the model [Rozanov et al., 1999; Egorova et al., 2003] simulates 41 chemical species from the oxygen, hydrogen, nitrogen, carbon, chlorine and bromine groups, which are determined by 202 gas-phase and photolysis reactions. The model also takes into account 16 heterogeneous reactions. The chemical solver is based on the pure implicit iterative Newton-Raphson scheme. The reaction coefficients are taken from Sander et al. [2000]. The photolysis rates are calculated at every 2-hour-long chemical-transport step using a look-up-table approach. This parameterization takes into account the photodissociation in the spectral region between 120 and 170 nm, which is significant for the chemistry of the mesosphere. The transport of all considered species is calculated using the hybrid numerical advection scheme proposed by Zubov et al. [1999]. MA-ECHAM4 and MEZON are interactively coupled by the radiative forcing induced by ozone, water vapor, methane, nitrous oxide, and chlorofluorocarbons. The CCM SOCOL has been comprehensively evaluated by Egorova et al. [2005] and Rozanov et al. [2005] as well as in the framework of the SPARC CCMval campaign [Eyring et al., 2006] and reviled some problems. Here in this study we use 4of13 49

54 D17303 HAEFELE ET AL.: DIURNAL CHANGES IN H 2 O AND O 3 D17303 version 2.0 of CCM SOCOL in which found problems have been corrected and described in detail by Schraner et al. [2008] GSWM [19] The GSWM is a 2-D, linearized, steady state numerical tidal and planetary wave model which extends from the ground to the thermosphere. Background temperatures and densities are specified by MSISE90 [Hedin, 1991]. Below 20 km the background winds are taken from the semiempirical model of Groves [1995, 1997], but the stratomesospheric jets and mesopause region winds are based upon Upper Atmosphere Research Satellite (UARS) High Resolution Doppler Interferometer (HRDI) climatologies [Hagan et al., 1999]. Above 125 km mean zonal winds are from HWM93 [Hedin, 1991, 1996]. GSWM employs the tropospheric tidal heating formulae of Groves [1982] which are based on 3-month averaged global models of specific humidity centered on January, April, July, and October. These heating rates were linearly interpolated for the GSWM-00 calculations. In the stratosphere, throughout the mesosphere, and into the lower thermosphere GSWM tidal heating is based upon a parameterization reported by Strobel [1978]. [20] For this study we used GSWM-00 which produces monthly migrating tidal climatologies and which is a extension to GSWM-98 [Hagan et al., 1999]. MSDOL and GSWM have been compared in the work of Morel et al. [2004] with respect to tides. The diurnal amplitude in temperature shows a maximum at midlatitudes at around 45 km, corresponding to the region of maximum heating by ozone. This maximum is described well by both models but is generally smaller in GSWM. Above 60 km however, the diurnal amplitudes predicted by GSWM are larger than in MSDOL by a factor of Derivation of Daily Variations [21] A statistical approach has to be chosen to derive daily variations from climate model and radiometer data: Each obtained value of volume mixing ratio, f, is transformed to the relative deviation from the daily mean: f *= (f f )/f, where f is the daily mean. These deviations are binned into i intervals of 24/i h based on local solar time (i = 24 for ozone and i = 12 for water vapor). The mean values of each of these intervals give finally the mean deviation from the daily mean, the daily variation. p The ffiffiffiffi errors of these mean values are estimated by s i / n i where s i is the standard deviation, and n i is the number of data points within the interval i. As noise on the retrieved values by far exceeds the amplitudes of daily variations, the analysis requires at least one month of data. This makes it impossible to study day to day variations but still allows to study the seasonal behavior of the daily cycle. Additionally, sine and cosine waves with periods of 12 and 24 hour periods are fitted to the daily cycles and the amplitude and phase of the diurnal and semidiurnal component are derived from the coefficients. The phase is defined as the local solar time of the maximum. Even though harmonic decomposition may not always be a good representation of the physics behind the daily variations of atmospheric composition and step or tent functions might be more appropriate it is still useful for direct comparison purposes. [22] In case of lidar a lot of resources are required for the operation. Full night measurements are only performed on a campaign basis. In addition to the partial coverage of the diurnal cycle, tides are also masked by the propagation of gravity waves. Several nights are usually averaged to get a better signal to noise ratio. However the mean diurnal cycle due to tides can also interact with longer periods and the analysis is not trivial [Morel et al., 2002]. With some assumption about the phase, or during winter time when longer dark periods allow to cover 14 h of continuous measurements, tidal amplitudes of temperature can be retrieved by fitting a 12- and a 24-hour cosine wave to the time series [Gille et al., 1991; Keckhut et al., 1996]. [23] In case of GSWM the daily constituent variations are calculated analytically as follows from GSWM winds and monthly mean meridional gradients derived from MLS/ Aura satellite data [Froidevaux et al., 2006] and monthly mean vertical gradients derived from radiometer data. To a first approximation the contributions from zonal, meridional and vertical advection can be calculated separately and added together in order to get the total perturbation. The ith component of the displacement, d i (t, z), at time t and altitude z is obtained by integration of the according wind component. Multiplication of d i (t, z) with the gradient in volume mixing ratio along the dimension i, r i f, leads to the constituent perturbation f 0 ðt; zþ ¼ d i ðt; zþr i f : ð2þ If adiabatic vertical motion is assumed the vertical displacement, d z (t, z), can also be estimated from the temperature perturbation as follows [Ehhalt et al., 1983]: d z ¼ T 0 g T N 2 ; where T 0, T and N 2 are the temperature perturbation, the mean temperature and the Brunt-Väisälä frequency squared, respectively. 3. Observations During the First CAWSES Campaign [24] CAWSES (Climate and Weather of the Sun-Earth System) is an international program sponsored by SCO- STEP (Scientific Committee on Solar-Terrestrial Physics) focused on the space environment and its impacts on life and society. Within this project coordinated campaigns on tides are organized to study their influence from the troposphere to the thermosphere. The first CAWSES campaign, herein after referred to as C1, took place in September and October [25] During C1 both radiometers were running continuously collecting 600 water vapor and 1800 ozone profiles. The lidar was in operation whenever the conditions allowed it covering the times specified in Figure 1. As discussed in section 2.2 the MSDOL model has been nudged to ECMWF reanalysis data to produce realistic atmospheric conditions while LMDz was run in fully interactive mode for the year ð3þ 50 5of13

55 D17303 HAEFELE ET AL.: DIURNAL CHANGES IN H 2 O AND O 3 D17303 mesosphere, above peak height) and results are shown in the top two panels of Figure 3. At 3.14 hpa a diurnal cycle is apparent in the observations with a double bumped maximum of 1% of the daily mean between 6 and 12 h and a clear minimum at 18 h. The MSDOL and LMDz models are in good agreement with the observations showing enhanced water vapor values between midnight and noon followed by a minimum short after 18 h. The observed diurnal evolution at 0.10 hpa is similar to the one observed at the lower level with a pronounced maximum at 4 h but an amplitude less than 0.5%. MSDOL shows a strong semidiurnal component at this level. It has to be noted that the measurement does not have the full sensitivity at this level and the amplitude is Figure 1. Time coverage of lidar measurements at OHP from 1700 local time (LT) to 0500 LT during the first CAWSES campaign in September/October The SOCOL run does not cover the year 2005 and is not taken into account in this Section Time Series [26] The vertical distribution of water vapor and ozone are characterized by a pronounced maximum at km and km, respectively, as shown in Figure 2. These maxima are herein after called peak heights. Large variability is found among the different data sets in case of both ozone and water vapor. The peak heights range from 50 km (MIAWARA) to 60 km (SOCOL) in case of water vapor with maximum values between 5 ppmv (LMDz) and 6.5 ppmv (SOCOL). The peak heights in ozone are found between 30 km (MSDOL) and 36 km (SOMORA) and the maximum values are between 7 ppmv (SOCOL) and 8.25 ppmv (MSDOL) Water Vapor [27] Water vapor has been analyzed at 3.14 hpa (upper stratosphere, below peak height) and at 0.10 hpa (lower Figure 2. Mean water vapor (solid line) and ozone (dashed line) profiles of the first CAWSES campaign in September/ October MIAWARA-radiometer (black), SOMORAradiometer (black), MSDOL (red), LMDz (blue), and SOCOL (cyan). Figure 3. Mean relative daily variations of H 2 O and O 3 volume mixing ratios at fixed pressure levels as a function of local solar time and averaged for the first CAWSES campaign in September/October 2005 derived from observations (black line) and from models: MSDOL (red line) and LMDz (blue line). The error bars represent the errors of the mean values as described in section of13 51

56 D17303 HAEFELE ET AL.: DIURNAL CHANGES IN H 2 O AND O 3 D17303 Figure 4. Nighttime evolution of the temperature anomaly with respect to the nighttime mean between 1700 LT and 0500 LT at 45 km. UARS refers to the statistical model described by Keckhut et al. [1996]. likely to be underestimated. The LMDz model is close to its top and is not shown at this level Ozone [28] Ozone cycles are derived for the levels at hpa (lower stratosphere, below peak height), 3.14 hpa (stratosphere, above peak height) and 0.55 hpa (stratopause, above peak height). Observations show a pronounced diurnal cycle at hpa with a maximum of 1% short before midnight. Both MSDOL and LMDz are out of phase with the observations by 6 h and show smaller amplitudes. In the observations a phase change of 12 h takes place between the and the 3.14 hpa level. At 3.14 hpa observed ozone shows a maximum around 16 h as a consequence of the production of odd oxygen through photolysis and the low [O]/[O 3 ] ratio. The peak to peak amplitude is around 5% which is in agreement with photochemical models [Pallister and Tuck, 1983; Ricaud et al., 1994]. The SOMORA measurements confirm nicely the predictions of photochemical models about midstratospheric ozone. Observational evidence of this feature has been reported by Connor et al. [1994], Huang et al. [1997] and very recently by Huang et al. [2008] who s results derived from TIMED/SABER are in good agreement with our observations. MSDOL is almost in perfect agreement with the observations at this level. The flat shaped daytime enhancement is an indication of catalytic destruction of ozone. LMDz does not show any indication of daytime depletion at this level and overestimates the amplitude in ozone. At 0.55 hpa ozone is dramatically depleted in the daytime as most O x resides in the form of atomic oxygen because the [O]/[O 3 ] ratio increases with decreasing air density. The peak to peak amplitude is around 25% (Figure 3) again in agreement with photochemical models [Vaughan, 1984; Ricaud et al., 1994; Schneider et al., 1999]. The night-day differences are well reproduced by MSDOL and LMDz but a time lag of 2 h with respect to the observations is noted Temperature [29] The diurnal evolution of temperature can be observed in the stratosphere by lidar. The comparison with the different models reveal a quite good agreement as shown in Figure 4. It has to be noted that the SOCOL run does not cover the year 2005 and the data shown in Figure 4 is for the year A statistical tidal model was developed using UARS data [Keckhut et al., 1996] for the OHP latitude and is also in good agreement with OHP observations with an amplitude slightly larger and a small phase shift of a few hours. The recombined time series from monthly mean tidal analysis from SABER data [Zhang et al., 2006] also show larger amplitudes but agree well in terms of phase. No accurate hourly temperature data could be retrieved in the mesosphere from lidar measurements as the variability is too high and no systematic diurnal evolution can be extracted from the noise Vertical Structure [30] Diurnal anomalies around the altitudes where direct forcing applies (photodissociation or direct radiative heating) are not expected to follow sinusoid functions. However, to investigate the diurnal evolution with altitude the mean daily variations of water vapor and ozone derived from observations have been decomposed into the diurnal and semidiurnal components at all pressure levels. The vertical structure of the amplitude and phase of the diurnal component is shown in Figure 5. The phase is defined as the local solar time of the maximum. Figure 5. (top) Vertical profiles of the amplitude and phase of the diurnal component of water vapor variations derived from observations (black line), MSDOL (red line), and LMDz (blue line) averaged for the first CAWSES campaign in September/October The phase is defined as the local solar time of the maximum. The contributions of meridional (red dashed line) and vertical advection (red dash-dotted line), respectively, derived from MSDOL winds and volume mixing ratio gradients according to equation (2) are shown as well. (bottom) Vertical profile of amplitude and phase of the daily variation in ozone from observations (black line), MSDOL (red line), and LMDz (blue line). The temperature phases derived from MSDOL (red dashed line) and SABER data (green line) are also shown. 52 7of13

57 D17303 HAEFELE ET AL.: DIURNAL CHANGES IN H 2 O AND O 3 D17303 Table 1. Coverage of the Data Sets Data Set Time Period MIAWARA Jan 2003 to Sept 2006 SOMORA Jan 2003 to Dec 2006 SABER MSDOL Sept 2004 to March 2006 LMDz 2005 SOCOL Jan 1998 to Dec Water Vapor [31] The observed diurnal amplitude in H 2 O shows two distinct maxima at 10 hpa and 0.6 hpa with values around 1%. Above 0.6 hpa the amplitude does not increase as one would expect because of the decreasing contribution of the measurement to the retrieved value (see section 2.1). The phase change from 9 h to 3 h at 3 hpa coincides with the pressure level of a local minimum in amplitude. [32] The diurnal amplitudes derived from the model data show similar vertical structures with local maxima in the stratosphere but they are smaller than the observed amplitudes. The maxima are found at 0.7 hpa in LMDz and at 2 hpa in MSDOL data, respectively. In addition to the full MSDOL model also the contributions from meridional and vertical advection according to equation (2) are presented in Figure 5. According to MSDOL meridional advection is the main driver of the diurnal variation in stratospheric water vapor. Semidiurnal variations are caused by vertical and meridional advection to the same extent but reach only half of the amplitude of the diurnal component (not shown here). The interaction of vertically decreasing meridional gradients with increasing wind amplitudes above 3 hpa leads to a local maximum in the diurnal amplitude at 2 hpa. Above 0.10 hpa vertical advection starts to govern daily variations and the diurnal amplitude shows a fast increase with height because of the amplification of the tidal waves with decreasing density. LMDz does not show this feature because the upper boundary of this model is at 0.10 hpa. In the stratosphere simulated amplitudes are generally smaller than observed. The phases as simulated by MSDOL and LMDz are very constant throughout the stratosphere and show only little variation below 0.2 hpa. Compared to the observations, MSDOL is 3 h in advance below 3 hpa and 3-4 h behind above 3 hpa. Above 0.2 hpa MSDOL reveals very different amplitudes and phases suggesting different mechanisms for daily variations Ozone [33] The vertical structure of diurnal amplitude and phase in ozone is presented in the bottom panels of Figure 5. Observations show a local maximum in amplitude of 3% at 10 hpa coincident with the level of the maximum ozone volume mixing ratio (Figure 2). Above 2 hpa amplitudes increase to high values as predicted by photochemistry. Between 20 and 3 hpa the diurnal phase is constant at 14 h and an abrupt change occurs at 2 hpa where the phase changes to midnight. This can be understood from photochemistry as summarized in section 1. Below 20 hpa, however, the phase is found short after midnight close to the phase of observed temperature variations (Figure 5). This is a strong indication that vertical advection is the dominating process behind the ozone variations at this level because it is below the peak height. Both models are in good agreement with each other and reproduce all features found in the observations. However, the local maximum in amplitude is found at higher levels at around 3 hpa and the amplitudes are substantially smaller than observed below 4 hpa. The phases as derived from the observations and models agree well above 20 hpa and major discrepancies between observations and models are found below this level. But we note that ozone and temperature modeledbymsdolareinphasewhichconfirmsthat vertical transport may be the dominating process in the lower stratosphere. 4. Seasonal Investigations [34] In order to characterize the daily variations in ozone and water vapor for different seasons, data of months 3, 6 and 9 have been averaged over 2 3 years. The coverage of each data set is given in Table 1. In order to sufficiently Figure 6. Relative daily variations in water vapor and ozone at constant pressure levels as function of local solar time derived from observations (black line), MSDOL (red line), LMDz (blue line), and SOCOL (cyan line). Data have been averaged over several years (see Table 1) and over months (left) 3, (middle) 6, and (right) 9. The error bars represent the errors of the mean values as described in section of13 53

58 D17303 HAEFELE ET AL.: DIURNAL CHANGES IN H 2 O AND O 3 D17303 Figure 7. Vertical profiles of the amplitude and phase of the diurnal component of water vapor derived from observations (black line), MSDOL (red line), LMDz (blue line), and SOCOL (cyan line). The plots represent data averaged over several years (see Table 1) and over months (left) 3, (middle) 6, and (right) 9. reduce noise MIAWARA data have been averaged over the months 3/4, 6/7 and 9/ Time Series Water Vapor [35] In case of dynamically driven changes it has to be noted that the seasonality is a superposition of the seasonality in the winds and the gradients of composition. Therefore equinox conditions had to be split in spring and fall equinox as the gradient in water vapor is of different peculiarity for these seasons. The meridional gradient in water vapor, as derived from Aura/MLS data [Froidevaux et al., 2006], is maximal for the months 9 and 10 at 2 hpa reaching ppm/km, twice as large as in spring. [36] Observations and models reveal no significant cycles during spring equinox (Figure 6). During summer and fall a diurnal component with a minimum at around 18 h is dominating in the observations and the MSDOL and LMDz models. Analysis of the MSDOL data reveals that during the months 6 and 9 the meridional gradients are large and lead to amplitudes of more than 0.3% while in spring amplitudes of the full model do not exceed 0.1% and both meridional and vertical advection are important (not shown here). Only weak cycles are generally observed at 0.10 hpa and maximum amplitudes of 0.1% are found for fall equinox (Figure 6). In contrast to the observations the MSDOL and SOCOL models show amplitudes up to 1% at this level. Again, observed amplitudes are likely to be underestimated because of the lower sensitivity at this level Ozone [37] At hpa, no systematic cycle can be found in ozone during spring neither in observational nor in model data. However in summer a clear minimum is found at 6 h followed by a maximum at 16 h which is also simulated by MSDOL with somewhat lower amplitude. Observations reveal a maximum of 1% at 2 h during fall equinox when the models are in good agreement with each other showing similar amplitudes as observed but are 6 h behind the observations. At 3.14 hpa the enhancement during daytime is largest in summer when the solar zenith angle is smallest and hence the production of odd oxygen is largest. This is clearly reproduced by LMDz and SOCOL while MSDOL shows strong daytime depletion during all seasons. The smallest amplitude is observed during spring equinox, where again a flat-shaped daytime enhancement could be an indication of catalytic destruction of ozone. However, there is no evidence for daytime ozone depletion in LMDz and SOCOL data Vertical Structure [38] The vertical profiles of the diurnal amplitude and phase derived from daily variations in ozone and water vapor are shown in Figures 7 and 8 (see also section 3.2) Water Vapor [39] The largest amplitudes in water vapor are observed during summer and fall at 1 2 hpa. The double-peak structure observed during C1 (Figure 5) is not characteristic for fall equinox but we note that the amplitude at 10 hpa is clearly largest during this season. MSDOL shows a very similar behavior but the local maximum is predicted somewhat lower at 2 3 hpa. Below 1 hpa SOCOL and LMDz show the small amplitudes during all periods which is a consequence of the small meridional gradients in the H 2 O field of SOCOL and LMDz (see Figure 9). The strong increase above 0.2 hpa found in MSDOL and SOCOL data is expected to be present in the real atmosphere as well because of the strong vertical gradient in H 2 O and the amplification of the tidal winds with increasing altitude. However, the measurement fails to detect this increase because its sensitivity gradually decreases above 0.10 hpa. 54 9of13

59 D17303 HAEFELE ET AL.: DIURNAL CHANGES IN H 2 O AND O 3 D17303 Figure 8. Same as Figure 7 but for ozone. The temperature phases derived from SABER temperature data (green line) are also shown. LMDz is not reliable at the upper levels because it is close to its upper boundary at 0.10 hpa Ozone [40] Observations and models reveal generally a consistent picture of daily ozone variations in the middle atmosphere. Large amplitudes of 10 to 20% are found above 2 hpa where ozone is dramatically depleted during daytime. Daytime maxima occur below 2 hpa and are largest in summer but do not exceed 4% in amplitude. Large discrepancy in amplitude is found among all data sets during spring equinox between 3 and 6 hpa and between observations and models below 20 hpa in summer and fall. The phases are very consistent above 20 hpa and disagree substantially below 20 hpa in spring. [41] The diurnal phase of temperature derived from SABER data is shown as well in Figure 8. Observations reveal the following phase relationship between ozone and temperature: The phase lag of 12 h in spring at 6 hpa shrinks to 6 h at 40 hpa where ozone leads temperature. The phase profiles in summer reveal no systematics. During fall equinox temperature and ozone are out of phase at 20 hpa and change to in phase at 40 hpa. MSDOL and SOCOL behave very similar in terms of the phase relationship between O 3 and T. However, in MSDOL the phase lag does not exceed 6 h indicating that photochemical equilibrium is never maintained (see section 1). SOCOL indicates a strong temperature dependence of ozone below 20 hpa in summer where ozone leads temperature by 6 h Comparison With GSWM [42] The seasonal evolution of amplitude and phase in H 2 O at 65 km has been investigated and compared to the GSWM (Figure 10). For this purpose observed daily variations have been decomposed in a diurnal and a semidiurnal Figure 9. Vertical profiles of the meridional gradient derived from MSDOL (red line), LMDz (blue line), SOCOL (cyan line), and Aura/MLS measurements (black line) for months 3 (spring equinox), 6 (summer solstice), and 9 (fall equinox). 10 of 13 55

60 D17303 HAEFELE ET AL.: DIURNAL CHANGES IN H 2 O AND O 3 D17303 Figure 10. Monthly mean (left) diurnal and (right) semidiurnal (top) amplitudes and (bottom) phases at 65 km derived from MIAWARA (circles) and from GWSM winds and H 2 O volume mixing ratio gradients (Aura/MLS data) according to equation (2) (crosses). For comparison, harmonic functions with periods of 12 months are fitted to the observational data (solid line) and to the model data (dashed line). component. For comparison amplitudes and phases of windinduced composition changes have been estimated according to equation (2) using winds from GSWM and the vertical gradient in volume mixing ratio derived from radiometer data. For comparison a harmonic function with a 12 month period has been fitted to the amplitudes and phases. The results derived on altitude surfaces differ from those derived on pressure surfaces because the pressure surface itself is moving up and down as part of the tidal wave. As a consequence the amplitudes derived on pressure surfaces are generally smaller than those derived on altitude surfaces as shown in Figure 10. [43] Enhanced diurnal amplitudes are observed during summer while semidiurnal amplitudes are maximal in winter (Figure 10). Diurnal phases are very constant throughout the whole year and semidiurnal phases show a strong seasonality. The agreement with GSWM is generally good. Maximal diurnal amplitudes are predicted somewhat later in the year and are generally larger by 50%. Good agreement is found for the semidiurnal amplitude in terms of amplitude and seasonality. A constant offset between observations and GSWM is found for the phases but the seasonal dependence is in good agreement. 5. Summary and Conclusions [44] The correct treatment of tides in atmospheric general circulation models is necessary for an improved description of interactions between radiative, chemical and dynamical processes on timescales from hours to days. Also, a correct parameterization of processes operating on short timescales in the atmosphere are in turn important for the simulation of long-term changes. In this study, data from ground based radiometers and lidar have been analyzed in order to characterize daily variations in middle atmospheric water vapor, ozone and temperature over the Alps and detailed comparisons with the chemistry-climate models MSDOL, LMDz and SOCOL have been made. [45] Water vapor is a long-lived trace gas in the stratosphere. Consequently, changes on a timescale of a day are governed by advection. By means of a simplified advection scheme and on the basis of MSDOL winds and water vapor fields, we have found that the daily variations of H 2 Oare mainly induced by meridional advection in the stratosphere and by vertical advection in the mesosphere. We have also noted that both LMDz and MSDOL underestimate absolute values of water vapor over the whole vertical range by 20% compared to observations while SOCOL overestimates water vapor by more than 10% compared to the observations. This reflects the difficulty to properly model middle atmospheric water vapor. [46] Photochemistry is the dominant process for daily variations in ozone in the middle-upper stratosphere and mesosphere. Two regimes can be identified: Above 2 hpa, ozone is dramatically depleted during daytime while, below 2 hpa, a daytime enhancement in ozone is observed. This behavior is attributed to the [O]/[O 3 ] ratio which depends strongly on air density and UV levels; the ratio increases with height. The transition layer below which dynamical and temperature-dependent processes become more important than photochemistry lies at 10 hpa and varies with season (being located at lower levels in summer). [47] One of the main focuses of this work was to analyze the first CAWSES campaign of September/October In the stratosphere, at 3.14 hpa, the daily variations of water vapor show amplitudes between 0.5 and 1% with a distinct minimum at around 18 h. The observed vertical structure of the diurnal amplitude showed an enhancement in the stratosphere and a decrease toward the mesopause. This behavior is likely to be due to the interaction of daily oscillations in the meridional wind and the meridional of 13

61 D17303 HAEFELE ET AL.: DIURNAL CHANGES IN H 2 O AND O 3 D17303 gradient in water vapor. All models under consideration could qualitatively simulate these features. However, in terms of absolute magnitude, significant differences were found between the observations and the model simulations and between the different model simulations. In the mesosphere the observed variations are generally smaller than what is predicted by the models even under consideration of the low sensitivity of the measurement in the mesosphere. [48] The main feature in O 3 daily variations, at 3.14 hpa, is an enhancement during the day reaching a maximum of about 2% shortly before 18 h. All models were able to reproduce these features and also to resolve some smallscale features. However, there are still uncertainties pertaining to the contributions of the different catalytic cycles of ozone destruction. [49] We also performed an analysis on seasonal timescales. The results indicate that uncertainties exist with respect to the amplitude of daily variations in stratospheric water vapor which arise partly from the differences in the meridional gradient that in turn is mainly governed by the large-scale meridional circulation. In case of the H 2 O observations, an underestimation of the amplitudes in the lower mesosphere arises from the influence of a priori information to the retrieval. [50] Generally, in the case of ozone, the seasonal dependencies of the diurnal variations derived from the observations and in the model simulations are found to be rather consistent. Diurnal variations in the stratosphere can be as large as 4% in summer and large variability is noted in terms of catalytic destruction during daytime depending on season and model. The absolute values of water vapor, the source of HO x, could not explain the differences. Large discrepancies are also found in the day-night differences of mesospheric ozone and in the amplitude and phase in the lower stratosphere where dynamical and temperaturedependent processes play an important role. [51] Monthly mean diurnal and semidiurnal amplitudes and phases in water vapor at 65 km were derived at different seasons. The diurnal amplitudes are found to be maximal in summer, which is consistent with the GSWM simulations. However, the maximum occurs 2 months later compared to the observations. Model-calculated diurnal amplitudes are twice as large as observed; this is probably related to the limited sensitivity of the measurement at this altitude. Diurnal phases do not exhibit a distinct seasonal cycle in contrast to the semidiurnal phase that changes by 4 h throughout the whole year. There is a systematic offset in diurnal and semidiurnal phase of 2 to 4 h between GSWM simulations and MIAWARA observations. [52] Acknowledgments. This work has been supported by the Swiss National Science foundation under grant /1 as well as through the project SHOMING financed by MeteoSwiss within GAW. We acknowledge the support of the European Commission through the GEOMON Integrated Project under the 6th Framework Program (contract FOP Global ). The development and maintenance of the CCM SOCOL is supported by ETH Zurich grant PP-1/04-1. We would like to thank J.M. Forbes for providing SABER data and M.A. Hagan for providing GSWM data. References Barnett, J. J., and J. A. Pyle (1975), The temperature dependence of the ozone concentration near the stratopause, Q. J. R. Meteorol. Soc., 101, Bertaux, J. L., A. Hauchecorne, A. Mangin, C. Cot, O. Talagrand, P. Simon, E. Kyrölä, H. Roscoe, O. Hembise, and B. P. Brasseur (1999), The MSDOL project: Assimilation of GOMOS ozone data in a 3-D chemistry-transport model, Phys. Chem. Earth, 24, Buehler, S. A., P. Eriksson, T. Kuhn, A. von Engeln, and C. Verdes (2005), ARTS, the atmospheric radiative transfer simulator, J. Quant. Spectrosc. Radiat. Transfer, 91, Calisesi, Y. (2003), The stratospheric ozone monitoring radiometer SOMORA: NDSC application document, Tech. Rep , Inst. of Appl. Phys., Univ. of Bern, Bern. Connor, B. J., D. E. Siskind, J. J. Tsou, A. Parrish, and E. E. Rempsberg (1994), Ground-based microwave observations of ozone in the upper stratosphere and mesosphere, J. Geophys. Res., 99, 16,757 16,770. Deuber, B., N. Kämpfer, and D. G. Feist (2004), A new 22-GHz radiometer for middle atmospheric water vapour profile measurements, IEEE Trans. Geosci. Remote Sens., 42, Douglass, A. R., R. B. Rood, and R. S. Stolarski (1985), Interpretation of ozone temperature correlations: 2. Analysis of SBUV ozone data, J. Geophys. Res., 90, 10,693 10,708. Egorova, T., E. Rozanov, V. Zubov, and I. Karol (2003), Model for investigating ozone trends (MEZON), Phys. Atmos. Ocean, 39, Egorova, T., E. Rozanov, V. Zubov, E. Manzini, W. Schmutz, and T. Peter (2005), Chemistry-Climate Model SOCOL: A validation of the presentday climatology, Atmos. Chem. Phys., 5, Ehhalt, D. H., E. P. Röth, and U. Schmidt (1983), On the temporal variance of stratospheric trace gas concentrations, J. Atmos. Chem., 1, Eriksson, P., C. Jiménez, and S. A. Buehler (2005), Qpack, a general tool for instrument simulation and retrieval work, J. Quant. Spectrosc. Radiat. Transfer, 91, Eyring, V., et al. (2006), Assessment of temperature, trace species, and ozone in chemistry-climate model simulations of the recent past, J. Geophys. Res., 111, D22308, doi: /2006jd Fels, S. B., and M. D. Schwarzkopf (1981), An efficient, accurate algorithm for calculating CO 2 15 mm band cooling rates, J. Geophys. Res., 86, Froidevaux, L., M. Allen, S. Berman, and A. Daughton (1989), The mean ozone profile and its temperature sensitivity in the upper stratosphere and lower mesosphere: An analysis of lims observations, J. Geophys. Res., 94, Froidevaux, L., et al. (2006), Early validation analyses of atmospheric profiles from EOS MLS on the aura satellite, IEEE Trans. Geosci. Remote Sens., 44, Gille, S. T., A. Hauchecorne, and M.-L. Chanin (1991), Semidiurnal and diurnal tidal effects in the middle atmosphere as seen by Rayleigh lidar, J. Geophys. Res., 96, Groves, G. V. (1982), Hough components of water vapor heating, J. Atmos. Terr. Phys., 44, Groves, G. V. (1995), A global reference atmosphere from 18 to 80 km, Tech. Rep. TR , Air Force Geophys. Lab., Bedford, Mass. Groves, G. V. (1997), Final scientific report, Tech. Rep , Air Force Off. of Sci. Res., Bolling Air Force Base, Washington, D. C. Hagan, M. E., M. D. Burrage, J. M. Forbes, J. Hachney, W. J. Randel, and X. Jhang (1999), GSWM-98: Results for migrating solar tides, J. Geophys. Res., 104, Hauchecorne, A., and M.-L. Chanin (1980), Density and temperature profiles obtained by lidar between 35 and 70 km, Geophys. Res. Lett., 7, Hecht, J. H., et al. (1998), A comparison of atmospheric tides inferred from observations at the mesopause during ALOHA-93 with the model predictions of the TIME-GCM, J. Geophys. Res., 103, Hedin, A. E. (1991), Extension of the MSIS thermosphere model into the middle and lower atmosphere, J. Geophys. Res., 96, Hedin, A. E. (1996), Empirical wind model for the upper, middle, and lower atmosphere, J. Atmos Terr. Phys., 58, Hines, C. O. (1997), Doppler spread parameterization of gravity wave momentum deposition in the middle atmosphere. Part 1: Basic formulation, J. Atmos Terr. Phys., 59, Hocke, K., N. Kämpfer, D. G. Feist, Y. Calisesi, J. H. Jiang, and S. Chabrillat (2006), Temporal variance of lower mesospheric ozone over Switzerland during winter 2000/2001, Geophys. Res. Lett., 33, L09801, doi: / 2005GL Hocke, K., et al. (2007), Comparison and synergy of stratospheric ozone measurements by satellite limb sounders and the ground-based microwave radiometer SOMORA, Atmos. Chem. Phys., 7, Hourdin, F., and A. Armengaud (1999), The use of finite-volume methods for atmospheric advection trace species: 1. Tests of various formulations in a general circulation model, Mon. Weather Rev., 127, Huang, F. T., C. A. Reber, and J. Austin (1997), Ozone diurnal variations observed by UARS and their model simulation, J. Geophys. Res., 102, 12,971 12, of 13 57

62 D17303 HAEFELE ET AL.: DIURNAL CHANGES IN H 2 O AND O 3 D17303 Huang, F. T., H. G. Mayr, J. M. Russel III, M. G. Mlynczak, and C. A. Reber (2008), Ozone diurnal variations and mean profiles in the mesosphere, lower thermosphere, and stratosphere, based on measurements from SABER on TIMED, J. Geophys. Res., 113, A04307, doi: / 2007JA Keckhut, P., et al. (1996), Semi-diurnal and diurnal temperature tides (30 55 km): Climatology and effect on uars-lidar data comparisons, J. Geophys. Res., 101, 10,299 10,310. Lefèvre, F., G. P. Brasseur, I. Folkins, A. K. Smith, and P. Simon (1994), Chemistry of the 1991/1992 stratospheric winter: Three-dimensional model simulations, J. Geophys. Res., 99, Lefèvre, F., F. Figarol, K. S. Carslaw, and T. Peter (1998), The 1997 Arctic ozone depletion quantified from three-dimensional model simulations, Geophys. Res. Lett., 25, Lott, F. (1999), Alleviation of stationary biases in a GCM through a mountain drag parameterization and a simple representation of mountain lift forces, Mon. Weather Rev., 127, Lott, F., and M. Miller (1997), A new subgrid scale orographic drag parameterization: Its testing in the ECMWF model, Q. J. R. Meteorol. Soc., 123, Lott, F., L. Fairhead, F. Hourdin, and P. Levan (2005), The stratospheric version of LMDz: Dynamical climatologies, arctic oscillation, and impact on the surface climate, Clim. Dyn., 25, Manzini, E., and N. A. McFarlane (1998), The effect of varying the source spectrum of a gravity wave parameterization in a middle atmosphere general circulation model, J. Geophys. Res., 103, 31,523 31,539. Marsh, D., A. Smith, and E. Noble (2003), Mesospheric ozone response to changes in water vapor, J. Geophys. Res., 108(D3), 4109, doi: / 2002JD McLandress, C. (1997), Seasonal variability of the diurnal tide: Results from the Canadian middle atmosphere general circulation model, J. Geophys. Res., 102, 29,747 29,764. Morcrette, J.-J. (1989), Description of the radiative scheme in the ECMWF model, Tech. Rep. N260165, 26 pp., Eur. Cent. for Medium-Range Weather Forecasts, Reading, U. K. Morel, B., H. Bencherif, P. Keckhut, S. Baldy, and A. Hauchecorne (2002), Evidence of tidal perturbations in the middle atmosphere over southern tropics as observed by Rayleigh lidar, J. Atmos. Sol. Terr. Phys., 64, Morel, B., P. Keckhut, H. Bencherif, A. Hauchecorne, G. Megie, and S. Baldy (2004), Investigation of the tidal variations in a 3-Df dynamicschemistry-transport model of the middle atmosphere, J. Atmos. Sol. Terr. Phys., 66, Pallister, R. C., and A. F. Tuck (1983), The diurnal variation of ozone in the upper stratosphere as a test of photochemical theory, Q. J. R. Meteorol. Soc., 109, Prather, M. J. (1981), Ozone in the upper stratosphere and mesosphere, J. Geophys. Res., 86, Randel, W. J. (1990), Kelvin wave-induced trace constituent oscillations in the equatorial stratosphere, J. Geophys. Res., 95, 18,641 18,652. Randel, W. J., J. M. Russel, A. Roche, and J. W. Waters (1998), Seasonal cycles and QBO variations in stratospheric CH 4 and H 2 O observed in UARS HALOE data, J. Atmos. Sci., 55, Ricaud, P., G. Brasseur, J. Brillet, J. de La Noë, J.-P. Parisot, and M. Pirre (1994), Theoretical validation of ground-based microwave ozone observations, Ann. Geophys., 12, Rodgers, C. D. (1976), Retrieval of atmospheric temperature and composition from remote measurements of thermal radiation, Rev. Geophys., 14, Rood, R. B., and A. R. Douglass (1985), Interpretation of ozone temperature correlations: 1. Theory, J. Geophys. Res., 90, Rozanov, E., V. Zubov, E. M. Schlesinger, F. Yang, and N. Andronova (1999), The UIUC 3-D Stratospheric Chemical Transport Model: Description and evaluation of the simulated source gases and ozone, J. Geophys. Res., 104, 11,755 11,781. Rozanov, E., M. Schraner, C. Schnadt, T. Egorova, M. Wild, A. Ohmura, V. Zubov, W. Schmutz, and T. Peter (2005), Assessment of the ozone and temperature variability during with the chemistry-climate model SOCOL, Adv. Space. Res., 35, Sander, S. P., et al. (2000), Chemical kinetics and photochemical data for use in stratospheric data, supplemented to evaluation 12: Update of key reactions, JPL Publ., 00-3, 73 pp. Schneider, N., F. Selsis, J. Urban, O. Lezeaux, J. de la Noë, and P. Ricaud (1999), Seasonal and diurnal ozone variations: Observations and modeling, J. Atmos. Chem., 50, Schraner, M., et al. (2008), Chemistry-climate model SOCOL: Version 2.0 with improved transport and chemistry/microphysics schemes, Atmos. Chem. Phys. Discuss., 8, 11,103 11,147. Schwarzkopf, M. D., and S. B. Fels (1985), Improvements to the algorithm for computing CO 2 transmissivities and cooling rates, J. Geophys. Res., 90, 10,541 10,550. Strobel, D. F. (1978), Parameterization of the atmospheric heating rate from 15 to 120 km due to O 2 and O 3 absorption of solar radiation, J. Geophys. Res., 83, Tiedke, M. (1989), A comprehensive mass flux scheme for cumulus parameterization in large scale models, Mon. Weather Rev., 117, van Leer, B. (1977), Towards the ultimate conservative difference scheme: IV. A new approach to numerical convection, J. Comput. Phys., 23, Vaughan, G. (1984), Mesospheric ozone Theory and observation, Q. J. R. Meteorol. Soc., 110, Zhang, X., J. M. Forbes, M. E. Hagen, J. M. Russel III, S. E. Palo, C. J. Mertens, and M. G. Mlynczak (2006), Monthly tidal temperatures km from TIMED/SABER, J. Geophys. Res., 111, A10S08, doi: /2005ja Zhu, X. (1994), An accurate and efficient radiation algorithm for middle atmosphere models, J. Atmos. Sci., 51, Zubov, V. A., E. V. Rozanov, and M. E. Schlesinger (1999), Hybrid scheme for three-dimensional advective transport, Mon. Weather Rev., 127, S. Bekki and M. Marchand, Service d Aéronomie, University Pierre et Marie Curie, 4, Place Jussieu, F Paris, France. T. Egorova, Physical-Meteorological Observatory, World Radiation Center, Dorfstrasse 33, CH-7260 Davos, Switzerland. (t.egorova@pmodwrc.ch) A. Haefele, K. Hocke, and N. Kämpfer, Department of Microwave Physics, Institute of Applied Physics, University of Bern, Sidlerstrasse 5, CH-3012 Bern, Switzerland. (alexander.haefele@mw.iap.unibe.ch) P. Keckhut, Service d Aéronomie, University of Versailles, Saint Quentin, B.P. 3, F Verrières-le Buisson, France. B. Morel, Université de la Réunion, 15 avenue René Cassin, BP 7151, F La Réunion, France. (beatrice.morel@univ-reunion.fr) E. Rozanov, Institute for Atmospheric and Climate Science, ETH Zurich, CH-8092 Zürich, Switzerland. (e.rozanov@pmodwrc.ch) of 13

63 Chapter 6 Validation of Ground Based Microwave Radiometers at 22 GHz for Stratospheric and Mesospheric Water Vapor Accepted for publication in the Journal of Geophysical Research 59

64 JOURNAL OF GEOPHYSICAL RESEARCH, VOL.???, XXXX, DOI: /2009JD Validation of ground based microwave radiometers at 22 GHz for stratospheric and mesospheric water vapor A. Haefele 1, E. De Wachter 1, K. Hocke 1,2, N. Kämpfer 1,2, G. E. Nedoluha 3, R. M. Gomez 3, P. Eriksson 4, P. Forkman 4, A. Lambert 5, and M. J. Schwartz 5 Abstract. We present a detailed intercomparison of five ground based 22 GHz microwave radiometers for stratospheric and mesospheric water vapor. Four of these instruments are members of the Network for the Detection of Atmospheric Composition Change, NDACC. The global measurements of middle atmospheric water vapor of the Microwave Limb Sounder (MLS) onboard the Aura satellite serve as reference and allow intercomparison of the ground based systems that are located between 45 S and 57 N. The retrievals of water vapor profiles from the ground based radiation measurements have been made consistent to a large extent: For the required temperature profiles we used the global temperature measurements of MLS and we agreed on one common set of spectroscopic parameters. The agreement with the reference measurements is better than ±8 % in the altitude range from 0.01 to 3 hpa. Strong correlation is found between the ground based and the reference data in the mesosphere with respect to seasonal cycle and planetary waves. In the stratosphere the measurements are generally more noisy and become sensitive to instrumental instabilities towards lower levels (pressures greater than 3 hpa). We further present a compilation of a NDACC data set based on the retrieval parameters described herein but using a temperature climatology derived from the MLS record. This makes the ground based measurements independent of additional information and allows to extend the data set for years in a homogeneous manner. 1. Introduction The main source of water vapor in the atmosphere is the evaporation from the Earth s surface. Water vapor enters the stratosphere through the tropical tropopause, which acts as a cold trap and renders the stratosphere and mesosphere typically a thousand times dryer than the lower troposphere. Methane oxidation is the second important source of water vapor in the stratosphere and provides a link between stratospheric humidity and human activities [Forster et al., 2007]. As the dominant greenhouse gas, water vapor has a strong impact on the radiative budget of the atmosphere and hence on the Earth s surface temperature. The long lifetime of water vapor in the stratosphere makes it a good tracer and gives valuable information about the atmospheric circulation and waves. Water vapor in the upper stratosphere and mesosphere is mainly observed by space borne and ground based remote sensing instruments in passive modes. While single ground based instruments can not provide any information on the horizontal distribution of water vapor, they are characterized by long operational life times. They are hence of major importance for water vapor monitoring and for the intercomparison of consecutive satellite missions. The merging 1 Institute of Applied Physics, University of Bern, Switzerland 2 Oeschger Center for Climate Change Research, University of Bern, Switzerland 3 Naval Research Laboratory, Washington, D.C., USA 4 Department of Radio and Space Science, Chalmers University of Technology, Gothenburg, Sweden 5 Jet Propulsion Laboratory, California Institute of Technology, Pasadena, USA Copyright 2009 by the American Geophysical Union /09/$ of data sets of consecutive satellite missions is an important task to generate global and homogeneous long term data sets that are essential for the climate community. A network of ground based instruments allows to detect biases between satellite experiments and the geographical dependence of these biases and plays a key role in the attempt to merge satellite data sets. However, this requires that the network is consistent, hence, that its members are in agreement with each other under consideration of the errors. This has been formulated by Harris [1976] as follows: It therefore seems imperative that workers [...] should combine to compare their instrumental sensitivities and errors so that some standardization of results may be achieved. Without such standardization, much information about the real nature and behavior of stratospheric humidity is being lost. At the time of writing several new ground based 22 GHz radiometers for middle atmospheric water vapor are being developed [Motte et al., 2007; Straub et al., 2008] and the network of such instruments will increase in the near future. There is thus a need for a standardization of the data reduction to increase the significance of data intercomparisons and to retrieve the highest amount of information from the network. In this paper we present an intercomparison of five ground based microwave radiometers for middle atmospheric water vapor that are members of the Network for the Detection of Atmospheric Composition Change (NDACC). The global data set of middle atmospheric water vapor of the Earth Observing System (EOS) Microwave Limb Sounder (MLS) serves as reference and allows to perform an intercomparison of the ground based instruments that are located at different sites. The retrievals of the radiometer systems have been standardized to a large extent: Temperature information, that is required in the water vapor retrieval, is taken from the global temperature record of MLS. Also, a common set of spectroscopic parameters has been used. The paper is organized as follows: The intercomparison strategy is elucidated in Section 2. In Sections 3 and 4 the

65 X - 2 HAEFELE ET AL.: VALIDATION OF GROUND BASED RADIOMETERS ground based instruments and the space born MLS are introduced, respectively. Section 1 is dedicated to the retrieval of water vapor profiles from measured radiation spectra. In Section 4 and 7 we present and discuss the results and implications of the intercomparison. 2. Intercomparison Strategy In order to compare measurements of ground based instruments that are not located at the same place a reference measurement is needed that is available at each location. Under the assumption that the reference measurement does not have a location dependent bias it is then possible to estimate differences between the instruments under consideration from the differences between the instruments and the reference according to the approach of the double differences by Hocke et al. [2007]. As a travelling standard for ground based microwave radiometry is not yet available we have chosen the Microwave Limb Sounder (MLS) on board of the Aura satellite (see Section 4) as the reference for this study because its water vapor product has extensively been validated [Lambert et al., 2007; Nedoluha et al., 2007; Vömel et al., 2007] and has proven to be of good quality. Furthermore the data set shows excellent continuity and covers the latitudes up to 82 north and south. The retrieval of a water vapor profile from a measured spectrum requires a first guess of the H 2O profile and its error, i.e. the a priori profile and its covariance, auxiliary information like pressure and temperature profiles and a forward model that does the radiative transfer calculations and establishes a relation between an atmospheric state and the measured intensity spectrum. These retrieval parameters are generally chosen in an attempt to optimize the retrievals from a particular type of instrument, and the specific choice can affect the retrieved water vapor values significantly. All of the retrievals of this study use a common set of retrieval parameters to eliminate biases between the instruments that simply originate from differences in the spectroscopic parameters or in the temperature information, for instance. A detailed description of the applied inversion parameters and their values is given in Section Instruments The five microwave radiometers under consideration are operated at the four NDACC sites Onsala (57 N, 12 E, 50 m amsl), Bern (47 N, 7 E, 900 m amsl), Mauna Loa H 2 O Radiometers of the NDACC (20 N, 156 W, 3500 m amsl) and Lauder (45 S, 170 E, 200 m amsl), and at Seoul (37 N, 127 E, 50 m amsl), which is not a NDACC site yet. The geographical distribution of the radiometers is shown in Figure 1. All systems measure the rotational transition line of H 2O at GHz in a balancing mode including a line measurement at a low elevation angle (20 40 ) and a reference measurement in zenith direction according to the method first introduced by Parrish et al. [1987]. Small modifications, however, ared made and for details we refer to the instrument papers referenced in Table 1. The receiver front ends consist of a horizontally aligned horn antenna and a rotating mirror at 45 inclination allowing the beam to be pointed at different elevation angles. The optical components ared followed by a heterodyne receiver which amplifies the incoming signal and converts it to a lower frequency range to be analyzed by a spectrometer. Important specifications and references of the five systems are given in Table 1. In the hot-cold calibration mode, that is applied by three of the five instruments (Onsala, Bern and Seoul), the atmospheric signal is compared to the signals from two reference targets at known temperatures. The hot load, which is a microwave absorber at ambient temperature, and the cold load, which is the sky itself. The temperature of the sky and the opacity of the atmosphere are derived on regular intervals of 30 min from a tipping curve measurement [Han and Westwater, 2000]. The tipping curve is a set of atmospheric measurements taken at different elevation angles to which a model of the atmosphere is fitted with its opacity as a free parameter revealing both τ and the temperature of the sky. The noise diode calibration mode, applied at Lauder and Mauna Loa, uses a noise diode as reference and its temperature is determined on a weekly basis by means of a hot-cold calibration using an ambient temperature and a liquid nitrogen calibration load. The atmospheric opacity is derived from tipping curve measurements as well. The Water Vapor Millimeter wave Spectrometers (WVMS) at Mauna Loa and Lauder have been in operation since 1996 and 1992, respectively, and have undergone extensive validation (see references in Table 1). The Onsala and Bern systems started operation in The instrument from Bern showed drifts in the properties of the previously used acousto optical spectrometer before March 2007 and we will present only data after the implementation of a new digital FFT spectrometer in March The Stratospheric Water vapor RAdiometer (SWARA) began with routine measurements in October Features in the spectra that originated from a frequency dependence in the antenna pattern strongly limited the practical bandwidth and hence do not allow to retrieve water vapor below 1 hpa [De Wachter et al., 2008]. This problem has been solved but at the time of this work there was no new data version available and we will thus not show any data from Seoul for pressures greater than 1 hpa N 45 S Mauna Loa Bern Onsala 90 W 0 90 E member candidate Seoul Lauder Figure 1. Location of the five 22 GHz radiometers for middle atmospheric water vapor. 4. MLS Because the temperature as well as the water vapor product of the Earth Observing System Microwave Limb Sounder (EOS MLS), herein after referred to as MLS, are extensively used in this study we give here a short overview of this instrument. MLS is operated on board of the Aura satellite that was launched July 15, 2004, and is part of NASA s A train group, which is a formation of six satellites flying in close proximity. The Aura satellite is on a near polar orbit covering 82 S to 82 N latitudes. The measurements are taken at fixed local solar times. A detailed description of MLS is given from Waters et al. [2006]. MLS observes thermal microwave emission by the atmosphere in five spectral regions from 115 GHz to 2.5 THz. 61

66 HAEFELE ET AL.: VALIDATION OF GROUND BASED RADIOMETERS X - 3 Table 1. Key specifications of the five microwave radiometers for stratospheric and mesospheric water vapor. The spectral resolution refers to the best resolution at line center. References: 1 [Forkman et al., 2003]; 2 [Deuber et al., 2004, 2005]; 3 [Thacker et al., 1995; Nedoluha et al., 1995, 1997, 2007]. Onsala 1 Bern 2 Seoul Mauna Loa 3 Lauder 3 Project name MIAWARA SWARA WVMS-3 WVMS-1 Spectrometer Autocorrelator digital FFT digital FFT Filter bank Filter bank Spectral resolution 25 khz 61 khz 61 khz 50 khz 200 khz Bandwidth 20 MHz 100 MHz 15 MHz 60 MHz 40 MHz Receiver Temperature 170 K 135 K 140 K 170 K 100 K Preamplifier (HEMT) uncooled uncooled uncooled cooled cooled Calibration hot-cold hot-cold hot-cold noise diode noise diode The Earth s limb is scanned vertically from the ground to 96 km. These scans are synchronized to the Aura orbit such that vertical scans are made at essentially the same latitudes each orbit. Temperature is derived from observations near the 118 GHz O 2 spectral line and the 243 GHz O 18 O spectral line. The vertical resolution of the MLS temperature measurement, taken to be the full width at half maximum of the averaging kernels, is 4 km at 10 hpa, 8 km at 1 hpa, 9 km at 0.1 hpa, 14 km at 0.01 hpa and 15 km at hpa. In the horizontal along-track direction, the temperature data have single profile resolution of 165 km through most of the profile, degrading to 185 km at 0.01 hpa and to 220 km at hpa [Schwartz et al., 2008]. Schwartz et al. [2008] present a detailed validation of the MLS temperature data version 2.2 and report a cold bias of 1-3 K compared to SABER, ACE and HALOE in the upper stratosphere and mesosphere. Regarding the error in the water vapor retrieval of the ground based instruments with respect to an error in the temperature profile of -2 %/5 K (see Section 5.5), the cold bias in MLS temperatures leads to an oversestimation of water vapor of < 1.5 %. The seasonal and latitudinal dependence of the bias in MLS temperatures with respect to SABER temperatures is < 3 K up to 0.1 hpa between 50 S and 50 N [Schwartz et al., 2008]. Water vapor profiles are retrieved from the limb emission measurements at GHz. The vertical resolution is better than 4 km below the stratopause and increases to > 10 km in the mesosphere and the along-track horizontal resolution is in the order of 400 km [Lambert et al., 2007]. Lambert et al. [2007] present a detailed validation of the MLS H 2O product and reports a bias compared to ACE- FTS of ±5% for pressures hpa and a bias compared to HALOE of +2% to +10% for pressures hpa. The precision on individual profiles is 0.4 ppmv at 0.1 hpa, 0.3 ppmv at 1 hpa and 0.2 ppmv at 10 hpa. No latitudinal dependence of the biases is reported. This is of particular importance for this study as MLS serves as reference for the ground based instruments that are spread between 45 S and 58 N. 5. Retrieval 5.1. The Optimal Estimation Algorithm The forward model implements a radiative transfer calculation and provides the relation between an atmospheric state and a measured spectrum, accounting also for instrumental properties like antenna pattern, side band suppression or spectrometer resolution. The inverse problem is the derivation of the atmospheric state from a measured radiation spectrum. It is ill posed, which means that an infinite number of solutions exists. A statistical approach is used to find the most likely atmospheric state given a measured radiation spectrum. For the ground based instruments we used the optimal estimation method by Rodgers [1976], which minimizes the following cost function derived from Bayes theorem: c = [y F (ˆx)] T S 1 y [y F (ˆx)] + [ˆx x a] T S 1 [ˆx x a](1) a The variables are described in Table 2. Costs are generated by deviations from the measured spectrum in the observation space, y F (ˆx), weighted with the inverse covariance of the measurement, S y, and by deviations from the a priori estimate in the state space, ˆx x a, accordingly weighted by the inverse a priori covariance, S a. The second term constrains the solution to physically meaningful states. The low concentration of water vapor in the stratosphere and mesosphere allows to use the linearization, F (ˆx) F (x a) + K(ˆx x a) (2) where K is the partial derivative of F with respect to x. Inserting (2) in (1) and solving dc/dˆx = 0 for ˆx yields ˆx = x a + (S 1 a + K T S 1 y K) 1 K T S 1 y (y F (x a)) (3) The retrieved profile thus consists of an a priori estimate to which we add a fraction of the H 2O profile corresponding to (y F (x a)). The amount of information that is added to x a depends on the error covariances of x a and y, namely S a and S y, and on the kernel K which describes the physical sensitivity of the spectrum to changes in x. K approaches zero at low pressures (high altitudes) where the absolute amount of H 2O becomes very small and where the line width becomes smaller than the spectral resolution of the instrument. For the five instruments the level where the contribution of the measurement becomes smaller than 60% lies typically between 70 and 80 km. Also, K decreases at high pressures (low altitudes), where the line width becomes larger than the bandwidth of the instrument due to the pressure broadening. In practice, the lower boundary for valid water vapor retrievals lies much higher than the theoretical value due to spectral artefacts from internal reflections. These artefacts are often referred to as baselines and are sine wave like structures that are superimposed on the spectrum. Baselines are accounted for in the forward model with an empirically determined set of sine waves of one or more known periods, or a polynomial fit of low order to the measured spectrum. Variations over time in the baseline cause unnatural variability in the retrieved Table 2. Description of the variables used in the optimal estimation retrieval of H 2 O profiles. y F (x) K x x a ˆx S y S a measured spectrum at frequencies depending on spectrometer calculated spectrum based on an atmospheric state x derivative of F with respect to x the true atmospheric state a priori assumption of the atmospheric state, i.e. of the H 2 O distribution the retrieved atmospheric state error covariance matrix of the measured spectrum error covariance matrix of the a priori assumption 62

67 X - 4 HAEFELE ET AL.: VALIDATION OF GROUND BASED RADIOMETERS water vapor values mainly at the lower levels. For the five instruments the lower boundary where the measurements are reliable lies between 1 and 3 hpa. For some instruments special retrieval setups allow retrievals to reach 10 hpa, but this is not further discussed here. The instruments from Onsala, Bern and Seoul use the retrieval software package QPack that is a user friendly implementation of the optimal estimation retrieval [Eriksson et al., 2005]. In Sections 5.3 to 5.6 the most important retrieval parameters are discussed Averaging Kernels The averaging kernel, A, characterizes the response of the retrieved profile to a perturbation in the true profile: A = ˆx/ x. It accounts for the limited vertical resolution and, at least as important, for the sensitivity of the retrieval that decreases towards higher and lower altitudes. A depends upon the measurement covariance matrix, S y. To account for possible variations in the signal to noise ratio that is given by S y the averaging kernels are calculated for each retrieved profile. To derive the profile as it would be measured by the radiometer system, ˆx ref, from a colocated reference profile, x ref, the averaging kernels are considered as follows: p [hpa] Onsala 57N/12E Bern 47N/7E p [hpa] ˆx ref = x a + A(x ref x a) (4) If not mentioned otherwise the MLS water vapor data are convolved with the averaging kernels of the microwave systems according to Equation (4) where x a is the a priori profile of the ground based instrument. The vertical resolution of MLS at 0.10 hpa is better by a factor of 2 than that of the ground based instruments and is being neglected. p [hpa] p [hpa] vmr [ppmv] Seoul 37N/127E vmr [ppmv] Lauder 45S/170E vmr [ppmv] vmr [ppmv] Mauna Loa 19N/156W vmr [ppmv] Mean value Standard deviation Min/max value Figure 2. Mean MLS profiles of the time period from August 2004 to September These profiles were used as a priori profiles in the retrievals of the ground based instruments. Table 3. Parameters of the a priori covariance matrices. The standard deviation, σ, is given in [ppmv] and the correlation length, l c, in [km]. The correlation is assumed to follow a Gaussian curve (see text Section 5.3). Bern Mauna Loa Onsala Seoul Lauder p [hpa] σ l c σ l c σ l c p [hpa] 5.3. A Priori Information on H 2 O Given the strong latitudinal dependence of the vertical distribution and the variability of middle atmospheric water vapor, it is not appropriate to use one single a priori profile in the retrievals of all instruments in the attempt to establish similar conditions for the retrievals of all instruments. Instead, individual a priori profiles were constructed for each instrument site by taking the mean of all MLS H 2O profiles within 200 km in latitude and 400 km in longitude from the period A smoothing has subsequently been applied to get rid of oscillations in the mean MLS H 2O profiles. The a priori profiles of each site are presented in Figure 2. The fact that mean MLS water vapor profiles are used as a priori profiles in the retrievals, which are subsequently compared to MLS measurements, is not an issue, since the comparison is restricted to the altitude ranges where the contribution of the a priori profile is very low. By using a priori profiles that are constant in time we assure that all the seasonal variations in the H 2O retrievals come from the measurements. The a priori covariance matrix, S a, defines the error of the a priori profile and controls the strength of the constraint of the retrieved profile to the a priori profile. The covariance matrix is defined by the standard deviation, σ, on each pressure level and the correlation length, l c, giving the distance over which the correlation between two levels decreases below 30 %. The correlation is assumed to follow a Gaussian Opacity Onsala Bern Seoul Mauna Loa Lauder Month Figure 3. Monthly means of the tropospheric opacity at GHz at the different sites. 63

68 HAEFELE ET AL.: VALIDATION OF GROUND BASED RADIOMETERS X - 5 Table 4. Spectroscopic parameters of the transition of H 2 O for T = 296 K. ν 0 : resonant frequency, S: line intensity, γ air (γ self ): air (self) broadening parameter, n air (n self ): temperature dependence of γ air (γ self ). The framed values have been used for all retrievals in this study. References: 1 [Rothman et al., 2004]; 2 [Cohen et al., 2001]; 3 [Poynter and Pickett, 1985] ν 0 S E γ air n air γ self n self [GHz] [m 2 Hz] [J] [Hz/Pa] [ ] [Hz/Pa] [ ] HITRAN e e JPL e e-21 JPL e e-21 Cazzoli et al. [2007] Payne et al. [2008] Liebe [1989] curve. The choice for the values for σ and l c is only partially motivated by the numbers derived from observational data sets but also by the requirement for a stable and sensitive retrieval. The covariance matrices are thus specific for each instrument but this can be accepted as S a has only a minor influence on the bias of the retrieved water vapor data. The values for σ and l c that were used in the retrievals are given in Table Spectroscopy In order to compute the atmospheric emission due to a given transition of a given molecule the absorption coefficient has to be calculated as a function of frequency. The four essential line parameters are the resonant frequency, ν 0, the line intensity, S(T ), the line width, ν(p, T ), and the energy of the lower quantum state, E. The line width accounts for the natural, Doppler and pressure broadening of the spectral line and the latter requires a further set of parameters consisting of the self and air broadening parameters, γ air and γ self, accounting for water-air and waterwater collisions, respectively, and of the exponents for the temperature dependence, n air and n self, respectively. The pressure broadened line halfwidth ν(p, T ) for a gas at pressure p, temperature T and partial pressure p s is given by: ( Tref ) nair ( Tref ) nself ν(p, T ) = γair(p p s) + γself p s T T (5) Values for all of these parameters base on measurements or calculations or both and are provided by spectral catalogues like JPL or HITRAN and by a wealth of publications. Table 4 shows a selection of values that can be found in the literature. For our study we used the broadening parameters from Liebe [1989], the line intensity, lower state energy and the line center frequency from the JPL 1985 catalogue [Poynter and Pickett, 1985]. These values are used in the WVMS retrievals since 1992 and reveal good validation results and are hence well suited for this study as well. In the context of a network, the consistency resulting from the use of a common set of values is more important than the absolute values. A Voigt line shape accounts for pressure and Doppler broadening. For the Mauna Loa and Lauder systems an adapted version of the radiative transfer model by Liebe [1989] is used while the Atmospheric Radiative Transfer Simulator, ARTS [Buehler et al., 2005], is used for the other systems Temperature Information The emission of the atmosphere depends on the actual temperature profile. Underestimating the temperature at a particular level will cause the retrieval to overestimate the water vapor amount required to emit the observed signal. The relative error in H 2O mixing ratio depends on altitude and is in the order of -2 %/5 K. For the retrievals of the ground based microwave systems the temperature profiles are routinely taken from different analyses as provided by NCEP, ECMWF or from models or climatologies like WACCM or MSISE90. But upper stratospheric and mesospheric temperature data provided by analyses or models are purely modeled and afflicted with considerable uncertainties. Observations as provided by SABER or MLS for an extended altitude range are assumed to be a better data source. Nedoluha et al. [2007] investigated the performance of MLS temperature observations in the WVMS retrievals, and reported an improvement in the reanalyzed water vapor retrievals mainly with respect to interannual variations. In this study the MLS temperature observations at the specific locations were used in the retrievals for all microwave systems and hence the validation is free of effects that are related to the use of different temperature data sources. As a consequence of the sun synchronous orbit, the MLS measurements are taken at constant local solar times, with daytime measurements on the ascending branch. Because of the coincidence criterion (see Section 4) and data availability there are not always both day and nighttime measurements available at one site for a 24 hour interval. This can introduce biases in daily mean profiles as differences between day and nighttime measurements are in the order of 5 K at 0.10 hpa. To create data sets of daily temperature profiles that are more representative for 24 hours we thus calculated 3 day running means where day and nighttime measurements were equally weighted, e.g. the mean of the mean daytime and the mean nighttime profile of a 72 hour interval Measurement Integration Time As the tropospheric opacity attenuates the signal from the middle atmosphere it has a major influence on the signal to Pressure [hpa] X=(X X )/X [%] MW MLS,conv MLS,conv STD( X) [%] Pressure [hpa] Bern (126) Onsala (578) 10 0 Lauder (222) Mauna Loa (181) Seoul (100) Figure 4. Mean value (left panel) and standard deviation (right panel) of the relative differences between the ground based measurements and the reference data (MLS). The numbers in brackets in the legend indicate the number of availabel profile pairs at each site. 64

69 X - 6 HAEFELE ET AL.: VALIDATION OF GROUND BASED RADIOMETERS VMR [ppmv] Onsala 0.03 hpa VMR [ppmv] Bern 0.03 hpa VMR [ppmv] hpa VMR [ppmv] 0.10 hpa VMR [ppmv] Seoul 0.03 hpa VMR [ppmv] Mauna Loa 0.03 hpa hpa 0.10 hpa VMR [ppmv] VMR [ppmv] VMR [ppmv] Lauder 0.03 hpa VMR [ppmv] hpa Figure 5. Time series of mesospheric H 2O at 0.10 and 0.03 hpa as observed by the ground based instruments (red) and MLS (blue). The MLS profiles have been convolved with the averaging kernels of the ground based instruments to account for differences in vertical resolution and sensitivity. noise ratio of the ground based measurements. The instruments are located between 45 S and 57 N and at altitudes from 50 to 3400 amsl and thus encounter very different conditions in terms of opacity. Figure 3 shows the mean annual cycle of the tropospheric opacity at the different sites. The signal to noise ratio governs the contribution of the a priori profile to the retrieval and must be kept constant. This in turn requires longer integration times for higher opacities. In Onsala, Bern and Seoul the number of spectra to be averaged before a retrieval depends on the actual value of the opacity to conserve the signal to noise ratio. Thus, the time between two retrievals changes dramatically with season; At Seoul the integration time ranges from a couple of hours in winter up to 4 weeks in summer. On the other hand, at Mauna Loa and Lauder a constant number of 500 spectra ( 1 week) is averaged before being fed into the inversion routine. Such a long integration time and the rather low opacities encountered at Mauna Loa and Lauder lead to a reasonably constant signal to noise ratio throughout the whole year. Due to the non linearity of the radiative transfer, the inversion of a mean spectrum is not the same as the mean of the retrievals from the individual spectra. The systematic bias due to the combined effect of non linearity of the radiative transfer and long integration times was found to be < 0.25 % up to the 0.01 hpa level (based on simulations) and is further neglected in this study. Furthermore, an integration time of one day and more reduces the effect of daily variations of water vapor on the comparison. However, diurnal variations are expected to have an amplitude of less 65

70 HAEFELE ET AL.: VALIDATION OF GROUND BASED RADIOMETERS X - 7 than 1 % up to 65 km [Haefele et al., 2008] and should not significantly affect the comparison. Also, any bias arising from the MLS sampling at fixed local solar times is hence expected to be of minor importance. While statistical errors like noise can be reduced by longer integration times, this does no apply for systematic errors. But systematic spectral errors are much smaller than the standard deviation of the target noise for all instruments and are not further discussed here. One exception, however, are baselines, which are features in the measured spectrum that originate from internal reflections. The structure of baselines is usually well known and they are thus not treated as spectral errors but included in the forward model (see Section 5.1). However, if the baseline is not well known or if it changes over time, baselines may affect the retrievals and can lead to unrealistic fluctuations (see Section 4). 6. Intercomparison of the Ground Based Measurements For each ground based retrieval a correlative MLS profile is generated by taking the mean profile of all MLS measurements that are available within ±200 km in latitude and ±400 km in longitude and within the integration time necessary for the ground based observation (hours to days). With this approach we get the following totals of coincidences: Onsala: 578, Bern: 128, Seoul: 109, Mauna Loa: 186 and Lauder: 224. We then calculated the differences between the ground based retrievals and the correlative MLS profiles. The mean values and the standard deviations of these differences are presented in Figure 4 and the corresponding time series are presented in Figure 5 and 7. The agreement between the ground based instruments and the reference is better than 8 % in the whole altitude range. The biases of the Mauna Loa and Lauder instruments with respect to MLS are on average between -3 and 3 % and between -3 and 5 %, respectively, and the standard deviations are lower than 7 % for both instruments, indicating very good correlation with the reference data. However, the decrease of the standard deviation of the Lauder data above 0.04 hpa is related to the increase of the a priori contribution which in turn is a consequence of the degradation of the averaging kernels due to the limited spectral resolution (see Section 5.1). These values are in agreement with those reported by Nedoluha et al. [2007]. The Bern instrument agrees with MLS within -3 to 3 %. Unlike the other instruments under consideration, we found a wet bias of 2 % between 0.20 and 0.03 hpa for this instrument. This wet bias, however, is only apparent during the summer season, particularly during summer 2007 (see Figure 5). If only the winter months (10-5) are considered the same analysis reveals a mean difference of -2 % between 0.10 and 0.03 hpa in excellent agreement with the WVMS instruments and the Onsala system. The Seoul instrument shows a wet bias of 5 % at 0.30 hpa and a dry bias of -7 % at 0.03 hpa. The mean difference between the Onsala instrument and MLS is between -7 % at 0.7 hpa and -3 % at 0.07 hpa. The large differences in the standard deviations in Figure 4 are mainly due to differences in the performance of the instruments and to some extent due to the increase in the natural variability with latitude. The standard deviation of variations on time scales of less than 90 days at 0.10 hpa derived from the MLS H 2O data are as follows: Onsala: 11 % (0.7 ppm); Bern: 9 % (0.6 ppm); Seoul: 7 % (0.4 ppm); Mauna Loa: 4 % (0.3 ppm); Lauder: 8 % (0.5 ppm). The noisy nature of the Bern and Seoul data becomes evident in relation to these numbers and compared to the performance of the Lauder instrument, which shows lower standard deviations and is located at a comparable latitude. At altitudes above the 0.10 hpa level the standard deviations of the Bern and Seoul data are comparable to the natural variability of H 2O, while for the other instruments the scatter is less than the natural variability revealing better correlation with the reference data on time scales below 90 days (not shown). For the Bern instrument the standard deviation is reduced by 20 % when the data of the summer time periods are not considered. The standard deviations of the differences between the ground based instruments and MLS are a good estimate of their statistical uncertainties. From Figure 4 one can see that the mean differences between the ground based instruments are smaller than their combined statistical uncertainties. In other words, the ground based instruments do not differ significantly from each other and, in this sense, the five ground based instruments build a consistent network. Figure 5 shows time series of H 2O at 0.10 and 0.03 hpa as observed by the ground based instruments and MLS. The instruments represent well the seasonal cycle in the mesosphere revealing an increase in amplitude towards higher latitudes and also show a lot of small scale features like planetary waves that are particularly evident at mid and high latitudes during the spring season (Lauder, Bern and Onsala). A secondary wintertime maximum is characteristic for the seasonal cycle in the mesosphere especially at mid latitudes and is well represented in the Bern and Lauder data. The systems from Bern, Seoul and Lauder generally overestimate the water vapor content during the summer period at 0.10 hpa and thus overestimate the seasonal variations. This is also reflected in Figure 6 that shows the monthly mean differences between the ground based instruments and the reference since Please note that in the case of the Seoul H 2 O GBMW H 2 O MLS,conv [ppmv] Monthly mean differences 0.03 hpa hpa hpa hpa Bern Onsala Mauna Loa Lauder Seoul Figure 6. Monthly mean differences between the ground based instruments and MLS since

71 X - 8 HAEFELE ET AL.: VALIDATION OF GROUND BASED RADIOMETERS VMR [ppmv] Onsala 1.00 hpa VMR [ppmv] 8 7 Bern 1.00 hpa hpa VMR [ppmv] VMR [ppmv] Lauder 1.00 hpa VMR [ppmv] Mauna Loa 1.00 hpa VMR [ppmv] hpa VMR [ppmv] hpa Figure 7. Time series of stratospheric H 2O at 1 and 3 hpa as observed by the ground based instruments (red) and MLS (blue). The MLS profiles have been convolved with the averaging kernels of the ground based instruments to account for differences in vertical resolution and sensitivity. system the monthly mean differences for the months May to October are based on single measurements only. Furthermore, Figure 6 reveals that the seasonal amplitude in the monthly mean differences of the Bern instrument could be significantly reduced between 2007 and Figure 7 shows time series of stratospheric measurements. Generally the time series of the ground based instruments become more unstable at these levels which is mainly because baseline artefacts in the spectra start to interfere (see Section 5.1). The instruments from Mauna Loa and Lauder perform well at 1 hpa representing nicely the weak seasonal cycle. However, on time scales of weeks not much of the variations can be expected to be real. We would like to emphasize the abrupt drop in the H 2O time series of WVMS and MLS at 1 hpa over Mauna Loa at the beginning of 2006 (see Figure 7). This feature is not present at 3 hpa which demonstrates that these two layers are independent to a great extent. In early 2007 the 3 hpa measurements from Mauna Loa developed a positive bias relative to the MLS measurements which persisted until the end of the timeseries. If this bias jump is corrected for then the seasonal variations from early 2007 onwards at 3 hpa appear reasonable. The Onsala instrument shows a dry bias and does not catch small scale features at 1 hpa but the seasonal cycle is apparent. The limited bandwidth of the Onsala system does not allow retrieval of water vapor below 2 hpa. The stratospheric data of the Bern instrument are of good quality during the winter season but show unrealistic fluctuations in summer. This effect is to the largest extent related to the fact that the observation geometry is slightly different in summer to account for the higher opacity. This in turn causes an unfortunate change of the baseline in the calibrated spectra that effectively destroys the retrievals for the stratosphere. This has been recognized recently and in future the change in geometry will be minimized thus allowing reliable retrievals also for the mid stratosphere in summer. Due to the limited practical bandwidth we do not show stratospheric measurements from the Seoul system. 7. NDACC data set We have standardized the retrievals as much as possible given differences in receiver and measurement locations. These retrievals will be available for each instrument and the entire measurement time period at in the section Microwave Group, and are processed as described in this paper with one exception. As some of the instrumental records reach further back than 2004, when MLS started its operation, and as the records of all systems are expected to last for several years no single and continuous global temperature data set exists to be used in the retrievals. We thus created a temperature climatology from the MLS record by building daily averages from the 4 years of observation that are smoothed with a 11 day rectangle filter. The use of this climatology in the retrievals makes the H 2O data of the ground based instruments independent of any additional information about the state of the atmosphere. While the mean differences between the NDACC retrievals and MLS are not affected, the standard deviation is slightly degraded for all instruments as H 2O variations on timescales of days and weeks that are accompanied by large variations in temperature can not be represented correctly using the temperature climatology. Also, any trend in temperature will not 67

72 HAEFELE ET AL.: VALIDATION OF GROUND BASED RADIOMETERS X - 9 be accounted for and could show up as trend in water vapor according to the 2 %/5 K error in H 2O. Remsberg and Deaver [2005] analyzed HALOE temperature data between 1991 and 2004 and report a linear trend of K in the tropical upper stratosphere and the subtropical mesosphere. The solar cycle effect is found to be < 1.7 K in the upper stratosphere and mesosphere. Beig et al. [2003] present a review of mesospheric temperature trends and summarize that linear trends in the midlatitudinal lower mesosphere are around -2 K/decade. Observations in the polar mesosphere are sparse and the results contradictory. It is thus necessary to assess the temperature climatology in regular intervals and to evaluate, based on recent literature, whether a linear trend should be included. 8. Conclusions The intercomparison of the five radiometer systems of NDACC with the global measurement from MLS reveals good agreement among the instruments within 10 % given that the retrievals of all ground based systems are made with the same spectroscopic parameters and with the same temperature data set. Consistency can be attributed to the network in the sense that the biases between the individual instruments are smaller than their combined statistical uncertainties. This consistency is essential for the network and it allows us to use the measurements of the individual instruments as one single data set. All instruments show a strong seasonal cycle and planetary scale wave features in the mesosphere in good agreement with the reference data set from MLS. While the WVMS systems from Lauder and Mauna Loa perform very well in the upper stratosphere, the performance of the other instruments is slightly degraded at 1 hpa. The main reason for this are instrumental instabilities that lead to variations in the baseline (see Section 5.1) which in turn disturb the water vapor retrieval at and below the 1 hpa level. This degradation is particularly a problem for the instruments with a low bandwidth (Onsala and Seoul) for which it is difficult to characterize the baseline. Baseline problems become even more evident in the mid stratosphere (3 hpa) and below, where the detection of the seasonal cycle and planetary waves is possible, but where all of the instruments have difficulties to provide stable long term measurements. Major efforts are being made to improve and understand instrumental stability in order to get stable retrievals in the mid and lower stratosphere. The compilation of a NDACC data set using a temperature climatology derived from the MLS temperature record shows good consistency within ±10 %. It is independent of additional observations or analyses and will be extended into the future and back to 1992 when the WVMS system in Lauder started its operation still allowing to keep the data sets homogeneous. The overall good performance of this network for middle atmospheric water vapor has been demonstrated and it should become a standard reference for any validation study dealing with water vapor in the stratosphere or mesosphere. Acknowledgments. This work has been supported by the Swiss National Science foundation under grant /1 as well as through the project SHOMING financed by MeteoSwiss within GAW. We acknowledge the support of the European Commission through the GEOMON Integrated Project under the 6th Framework Program (contract number FOP Global ) and the support by NASA under the Upper Atmosphere Research Program and by the Naval Research Laboratory. We also would like to thank the teams at Lauder, Mauna Loa and Seoul for their technical support to run the instruments. Work at the Jet Propulsion Laboratory, California Institute of Technology, was carried out under a contract with the National Aeronautics and Space Administration. References Beig, G., P. Keckhut, R. P. Lowe, R. G. Roble, M. G. Mlynczak, J. Scheer, V. I. Fomichev, D. Offermann, W. J. R. French, M. G. Shepherd, A. I. Semenov, E. E. Remsberg, C. Y. She, F. J. Lbken, J. Bremer, B. R. Clemesha, J. Stegman, F. Sigernes, and S. Fadnavis. Review of mesospheric temperature trends. Rev. Geophys., 41(4), 1015, doi: /2002rg000121, Buehler, S. A., P. Eriksson, T. Kuhn, A. von Engeln, and C. Verdes. Arts, the atmospheric radiative transfer simulator. J. Quant. Spectrosc. Radiat. Transfer, 91:65 93, Cazzoli, G., C. Puzzarini, G. Buffa, and O. Tarrini. Experimental and theoretical investigation on pressure-broadening and pressure-shifting of the 22.2GHz line of water. J. Quant. Spectrosc. Radiat. Transfer, 105: , Cohen, E. A., M. L. Delitsky J. C. Pearson H. S. P. Mller H. M. Pickett, R. L. Poynter. Submillimeter, millimeter, and microwave spectral line catalog. J. Quant. Spectrosc. Radiat. Transfer, 60: , Deuber, B., N. Kämpfer, and D.G. Feist. A new 22-GHz Radiometer for Middle Atmospheric Water Vapour Profile Measurements. IEEE Transactions on Geoscience and Remote Sensing, 42(5): , Deuber, B., A. Haefele, D. G. Feist, L. Martin, N. Kämpfer, G. E. Nedoluha, V. Yushkov, S. Khaykin, R. Kivi, and H. Vömel. Middle Atmospheric Water Vapour Radiometer - MIAWARA: Validation and first results of the LAUT- LOS / WAVVAP campaign. J. Geophys. Res., 110, D13306, doi: /2004jd005543, 2005 De Wachter, E., A. Murk, C. Straub, A. Haefele, S. Ka, J. J. Oh, and N. Kämpfer. Effects of Resonances in Corrugated Horn Antennas for a 22 GHz Balancing Radiometer. IEEE Geoscience and Remote Sensing Letter, 6, doi: /lgrs , Eriksson, P., C. Jiménez, S. A. Buehler. Qpack, a general tool for instrument simulation and retrieval work. J. Quant. Spectrosc. Radiat. Transfer, 91:47 64, Forkman, P., P. Eriksson, and A. Winnberg. The 22 GHz radioaeronomy reciver at Onsala Space Observatory. J. Quant. Spectrosc. Radiat. Transfer, 77(1):23 42, Forster, P., V. Ramaswamy, P. Artaxo, T. Berntsen, D.W. Fahey R. Betts, J. Haywood, J. Lean, D.C. Lowe, G. Myhre, J. Nganga, G. Raga R. Prinn, M. Schulz, and R. Van Dorland. Changes in Atmospheric Constituents and in Radiative Forcing. In: Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, Haefele, A., K. Hocke, N. Kämpfer, P. Keckhut, M. Marchand, S. Bekki, B. Morel, T. Egorova, and E. Rozanov Diurnal changes in middle atmospheric H2O and O3: Observations in the Alpine region and climate models. J. Geophys. Res., 112, doi: /2008jd009892, Han, Y., and E.R. Westwater. Analysis and improvement of tipping calibration for ground-based microwave radiometers. IEEE Transactions on Geoscience and Remote Sensing, 38(3): , Harris, J.E. The Distribution of Water Vapor in the Stratosphere. Rev. of Geophysics and Space Physics, 14: , Hocke, K., N. Kämpfer, D. Ruffieux, L. Froidevaux, A. Parrish, I. Boyd, T. von Clarmann, T. Steck, Y.M. Timofeyev, A.V. Polyakov, and E. Kyrölä. Comparison and synergy of stratospheric ozone measurements by satellite limb sounders and the ground-based microwave radiometer SOMORA. Atmospheric Chemistry and Physics, 7: , Lambert, A. et al. Validation of the Aura Microwave Limb Sounder middle atmosphere water vapor and nitrous oxide measurements. J. Geophys. Res., 112, D24S36, doi: /2007jd008724, Liebe, H.J. MPM - a atmospheric millimeter-wave propagation model. Int. J. of Infrared and Millimeter Waves, 10: ,

73 X - 10 HAEFELE ET AL.: VALIDATION OF GROUND BASED RADIOMETERS Motte, E., P. Ricaud, M. Niclas, B. Gabard, and F. Gangneron. A 22 GHz mobile microwave radiometer for the study of stratospheric water vapor IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2007, Nedoluha, G. E., R. Bevilacqua, R. Gomez, W. Waltman, B. Hicks, D. Thacker, J. Russell III, M. Abrams, H. Pumphrey, and B. Connor. A comparative study of mesospheric water vapor measurements from the ground-based water vapor millimeter-wave spectrometer and space-based instruments. J. Geophys. Res., 102(D14): , Nedoluha, G. E., R. M. Bevilacqua, R. M. Gomez, D. L. Thacker, W. B. Waltman, and T. A. Pauls. Ground-based measurements of water vapor in the middle atmosphere. J. Geophys. Res., 100(D2): , Nedoluha, G. E., R. M. Gomez, B. C. Hicks, R. M. Bevilacqua, J. M. Russell III, B. J. Connor, and A. Lambert. A comparison of middle atmospheric water vapor as measured by WVMS, EOS-MLS, and HALOE. J. Geophys. Res., 112, D24S39, doi: /2007jd008757, Parrish, A., R. L. dezafra, P.M. Solomon, and J. W. Barret. A ground based technique for millimeter wave spectroscopic observations of stratospheric trace constituents. Radio Science, 23: , Payne, V.H., J.S. Delamere, K.E. Cady-Pereira, R.R. Gamache, J-L. Moncet, S. Mlawer, and A. Clough. Air Broadened Half Widths of the 22 and 183 GHz Water Vapor Lines. IEEE Transactions on Geoscience and Remote Sensing, 46: , Poynter, R.L., and Pickett H.M. Submillimeter, millimeter, and microwave spectral-line catalog. Applied Optics, 24: , Remsberg, E. E., and L. E. Deaver Interannual, solar cycle, and trend terms in middle atmospheric temperature time series from HALOE J. Geophys. Res., 110, D06106, doi: /2004jd Rodgers, C. D. Retrieval of atmospheric temperature and composition from remote measurements of thermal radiation. Rev. Geophys., 14: , Rothman, L.S., D. Jacquemart, A. Barbe, D. Chris Benner, M. Birk, L.R. Brown, M.R. Carleer, C. Chackerian Jr., K. Chance, L.H. Coudert, V. Dana, V.M. Devi, J.-M. Flaud, R.R. Gamache, A. Goldman, J.-M. Hartmann, K.W. Jucks, A.G. Maki, J.-Y. Mandin, S.T. Massie, J. Orphal, A. Perrin, C.P. Rinsland, M.A.H. Smith, J. Tennyson, R.N. Tolchenov, R.A. Toth, J. Vander Auwera, P. Varanasi, and G. Wagner. The HITRAN 2004 molecular spectroscopic database. J. Quant. Spectrosc. Radiat. Transfer, 96: , Rothman, L.S., C.P. Rinsland, A. Goldman, S.T. Massie, D.P. Edwards, J.M. Flaud, A. Perrin, C. Camy-Peyret, V. Dana, J.Y. Mandin, J. Schroeder, A. McCann, R.R. Gamache, R.B. Wattson, K. Yoshino, K.V. Chance, K.W. Jucks, L.R. Brown, V. Nemtchinov, and P. Varanasi. The HITRAN molecular spectroscopic database and HAWKS (HITRAN Atmospheric Workstation): 1996 edition. J. Quant. Spectrosc. Radiat. Transfer, 60(5): , Schwartz, M. J. et al. Validation of the aura microwave limb sounder temperature and geopotential height measurements. J. Geophys. Res., 113, D15S11, doi: /2007jd008783, Straub, C., A. Murk, N. Kämpfer, D. Zardet, and B. Stuber. Development of a 22 GHz correlating Radiometer for the observation of Stratospheric Water Vapor Microwave Radiometry and Remote Sensing of the Environment (MICRORAD 2008),2008. Thacker, D. L., R. M. Bevilacqua, W. B. Waltman, T. A. Pauls, R. M. Gomez, G. E. Nedoluha, and P. R. Schwartz. Groundbased sensing of water-vapor in the stratosphere and mesosphere. IEEE Trans. on Instr. and Meas., 44(2): , Vömel, H., J.E. Barnes, R.N. Forno, M. Fujiwara, F. Hasebe, S. Iwasaki, R. Kivi, N. Komala, E. Kyro, T. Leblanc, B. Morel, S.Y. Ogino, W. Read, S.C. Ryan, S. Saraspriya, H. Selkirk, M. Shiotani, J. Valverde-Canossa, and D.N. Whiteman. Validation of Aura MLS Water Vapor by Balloon Borne Cryogenic Frostpoint Hygrometer Measurements. J. Geophys. Res., 112, D24S37, doi: /2007jd008698, Waters, Joe W. et al. The Eartch Observing System Microwave Limb Sounder (EOS MLS)on the Aura Satellite. IEEE Transactions on Geoscience and Remote Sensing, 44: , A. Haefele, Department of Microwave Physics, University of Bern, Switzerland. (haefele@iap.unibe.ch) 69

74 Chapter 7 Comparison of Insitu and Remote Sensing Measurements of Water Vapor made within the frame of the SHOMING Project over Switzerland Research report of the Institute of Applied Physics, University of Bern 70

75 Comparison of in situ and remote sensing measurements of water vapor made within the frame of the SHOMING project over Switzerland. A. Haefele Research Report No MW August 7, 2009 Institute of Applied Physics Microwave Physics Division Sidlerstr. 5 Tel. : CH-3012 Bern Fax : Switzerland iap @iap.unibe.ch 71

76 1 Introduction Water vapor is chemically and radiatively very active in the atmosphere and has thus a huge impact on its composition and radiative balance. Water evaporates from the surface and the combined eect of the temperature lapse rate and the strong temperature dependence of the saturation vapor pressure leads to an exponential decrease of water vapor with altitude with a scale height of approximately 2 km. The upper troposphere and the stratosphere is typically a thousand times dryer than the lower troposphere. Furthermore, the spatial and temporal variability of water vapor is very large. These circumstances make the measurement of water vapor a dicult task and there exists no single technique that can cover altitudes from the surface up to 100 km. The key to measure water vapor in the whole atmosphere is thus the combination of several measurement techniques. This, however, requires that the dierent measurements are consistent or are made consistent, for example with the approach proposed by Haefele [2005] or with the method suggested by Löhnert et al. [2007] (integrated proling technique). SHOMING (Stratospheric Humidity Observations and Monitoring) is a project within the Global Atmospheric Watch (GAW) program with the following aims: Improvement of existing measurements by optimization of the calibration concept Comparison with other sensors in campaigns and complementing data from dierent techniques in an optimal way Long term validation of satellites and combining the ground based measurements with satellite data Investigation of temporal variability in stratospheric water vapor and its relation to ozone Strengthen the NDACC alpine station Jungfraujoch/Bern Within the SHOMING project a variety of sensors were available as described below. Deployed from dierent platforms these sensors cover dierent altitude ranges and provide the possibility to characterize the water vapor prole from the ground up to 80 km. Within this initiative ve balloon soundings have been performed so far. These balloon soundings were coordinated with the activities of a microwave system for stratospheric and mesospheric water vapor and with the activities of a lidar system for tropospheric water vapor. In this work we present an intercomparison of the data of the dierent measurement techniques in their overlapping region. Knowing the performance or the information content of the individual measurements is an important precondition for the next step which is the combination of the various measurements. After a short description of the dierent water vapor sensors and measurement techniques in Section 2 the results are presented in Section 3. In Section 4 a summary is given and conclusions drawn. 72 1

77 Table 1: Available instruments for the FLASH campaign. Sensor Platform Technique Approx. Altitude Range RS92 Balloon Thinlm capacitor 0-10 Snow White Balloon Frost point 0-10 FLASH-B Balloon Lyman-α 8-35 Ralmo Ground based Lidar 0-10 Miawara Ground based Microwave spectrometer Data 2.1 Balloon Soundings The balloon soundings are performed by MeteoSwiss from the aerological site Payerne (46.82 N, 6.95 E, 491 m amsl), Switzerland. For each sounding the balloon payload was equipped with the sensors given in Table 1. Until the time of writing ve soundings could be realized at the following days: February 8, 2008; April 29, 2008; June 24, 2008; October 7, 2008; May 6, The Soundings are exclusively performed during night and 4 of them were successfully recovered and can be refurbished to some extent. RS92 The Vaisala RS92 is a thinlm capacitance sensor that directly measures relative humidity. The accuracy of the RS92 measurements is assessed in detail in Miloshevich et al. [2009] and is reported as ±4 % for nighttime soundings. For this study no correction was applied to the RS92 data. Snow White The Snow White hygrometer manufactured by Meteo Labor AG, Switzerland, is a low cost chilledmirror humidity sensor. The performance of Snow White is assessed in Fujiwara et al. [2003]; Vömel et al. [2003] and is found to be good up to the tropopause. The uncertainty is as good as 2 % RH. The Snow White version MRSSRSC34/04night was used in this project. FLASH-B The Fluorescent Advanced Stratospheric Hygrometer for balloon born application is a high end Lymanα humidity sensor for stratospheric conditions [Yushkov et al.]. This instrument took part in the LAUTLOS campaign and has been compared to the CMDL frost point hygrometer [Vömel et al. 2007], which is considered as a reference instrument for water vapor in the troposphere and stratosphere. FLASHB typically provides good data between the mid troposphere up to the burst altitude of the balloon (typically 35 km) with an uncertainty of 8 % in volume mixing ratio. More reliable measurements are obtained during the descent after the balloon burst. 2 73

78 2.2 Remote Sensing Instruments Ralmo Ralmo is a ground based raman lidar system providing water vapor proles at very high temporal and vertical resolution. It is deployed at the aerological station in Payerne (46.82 N, 6.95 E, 491 m asl). Under good conditions and with longer integration times the lidar system reach the lower stratosphere but it is most reliable in the lower and mid troposphere. Miawara The Middle Atmospheric Water Vapor Radiometer, Miawara, is a ground based microwave spectroradiometer dedicated to stratospheric and mesospheric water vapor (approx km) [Deuber et al. 2004]. This instrument has been validated in Deuber et al. [2005]; Haefele et al. [in press]. The uncertainty of single proles is 10 % below the 1 hpa level and increases to 30 % at 75 km. Miawara is deployed at Zimmerwald (46.88 N / 7.46 E, 907 m amsl) which is 40 km East of Payerne. 3 Results 3.1 Troposphere The measurements obtained in the troposphere are shown in Figure 1. The balloon borne measurements are taken during the ascent in the case of RS92 and Snow White and during the descent in the case of Flash. Note that RS92/Snow White and Flash are thus not perfectly collocated. No RS92 was onboard of the rst balloon sonde in February 2008 and no lidar data are available for the third ight in June The upper limit of the lidar measurements lies between 6 and 12 km for dry and wet conditions, respectively. It has to be noted also, that the proles of the ground based lidar system are not perfectly collocated with the balloon borne measurements. The Flash measurements in June 2008 and May 2009 were not successful because of a failure of the photo multiplier. The proles of the relative dierence in relative humidity between Snow White and RS92 are shown in Figure 1. Snow White shows a lot of variations that are not present in RS92 data and might be related to diculties with the temperature control of the mirror. However, on average Snow White shows no bias compared to RS92 in the lower and mid troposphere (< 6 km). Between 6 and 12 km the dierence proles reveal a large spread and indicate on average a wet bias of % compared to RS92. Despite the larger spread of the dierences at these altitudes the agreement is good enough for the data to be considered as reliable. Above 12 km the dierences between Snow White and RS92 get very large and no systematic bias can be determined. The measurements of Snow White, RS92 and Flash did not reveal considerable agreement above the tropopause and stratospheric measurements of Snow White and RS92 are not presented. 74 3

79 Temperature [K] SnowWhite Ralmo Flash Payerne, :04 UTC Relative Humidity [%] Temperature [K] RS92 SnowWhite Ralmo Flash Payerne, :34 UTC Relative Humidity [%] Height [km] Height [km] Temperature [K] RS92 SnowWhite Payerne, :13 UTC Relative Humidity [%] Temperature [K] RS92 SnowWhite Ralmo Flash Payerne, :18 UTC Relative Humidity [%] Height [km] Height [km] Temperature [K] RS92 SnowWhite Ralmo Payerne, :15 UTC Relative Humidity [%] (RH RH )/RH *100 [%] SW RS92 RS92 Height [km] Height [km] Figure 1: Water vapor proles measured by the dierent sensors between 0 and 12 km. No corrections have been made to the data. Temperature is also shown (black). Note, that RS92/Snow White measurements are made during ascend and Flash measurements are made during descent of the balloon ight. These measurements are hence not perfectly collocated. The bottom right plot shows the relative dierence in relative humidity between Snow White and RS

80 :34 MLS +/ 24 h Miawara +/ 12 h Flash :34 MLS +/ 24 h Miawara +/ 12 h Flash :18 MLS +/ 24 h Miawara +/ 48 h Flash Pressure [hpa] 10 1 Pressure [hpa] 10 1 Pressure [hpa] VMR [ppmv] VMR [ppmv] VMR [ppmv] Figure 2: Water vapor proles measured by the balloon borne Flash hygrometer (red), by the ground based microwave system Miawara (blue) and by the space borne Microwave Limb Sounder (black). Averaging kernels have not been considered and the contribution of the a priori prole to the Miawara retrievals is less than 35 %. 3.2 Stratosphere The stratospheric measurements taken by the balloon borne Flash hygrometer and by the ground based microwave radiometer Miawara are complemented by observations made by the Microwave Limb Sounder (MLS) onboard the Aura satellite. Figure 2 shows the data of the three successful Flash soundings together with collocated MLS and Miawara proles. As single proles of Miawara have a considerable uncertainty of 30 % Figure 2 shows the mean and the standard deviation derived from individual proles within ±12 h around the balloon sounding (23 proles). MLS proles are taken within ±200 km in latitude and ±400 km in longitude and within ±24 h in time. Neither the Flash nor the MLS data have been convolved with the averaging kernels of Miawara. The contribution of the a priori prole is less than 35 % above the 30 hpa level ( 24 km). Figure 3 shows the temporal evolution of water vapor at the 10 hpa level as measured by Miawara and in comparison with MLS and Flash measurements. The data gap during the months June to August is due to baseline problems. Nevertheless, the stability of the Miawara measurements at 10 hpa is remarkable as the retrieval of mid and lower stratospheric water vapor is extremely sensitive to instrumental instabilities like baselines, that arise from internal reections. 4 Summary and Conclusions In the frame of the SHOMING project coordinated eorts have been made to measure atmospheric water vapor from the surface up to the mesosphere. In this report an intercomparison of the data collected by the various in situ and remote sensing techniques is presented. The RH measurements of RS92, Snow White and the lidar system are in good agreement. In particular, no systematic dierence could be found between Snow White and RS92 in the lower and mid troposphere. In the upper troposphere the uncertainty of the Snow 76 5

81 H 2 O VMR [ppmv] p=10 hpa MLS MIAWARA FLASH Figure 3: Timeserie of water vapor at 10 hpa as measured by Miawara (blue), MLS (green) and Flash (red circles). Averaging kernels have not been considered and the contribution of the a priori prole to the Miawara retrievals is less than 35 % at this level. White and RS92 measurements increases but the agreement is considerably good up to an altitude of 12 km. Based on the data taken within this project, 12 km is found to be the upper limit for the Snow White and RS92 hygrometers. Three successful Flash measurements are available for intercomparisons in the stratosphere. The burst altitude of the balloon was between 35 adn 37 km. On the other hand, an arbitrary threshold of 35 % a priori contribution has been chosen for the microwave system which allows the retrieved prole to be extended downward to an altitude of 30 km. The agreement between Flash and Miawara in this overlap region is good. Stable long term measurements in the mid stratosphere down to km by ground based microwave radiometry is a major challenge and requires a very high level of instrumental stability. Any change in the baseline structure, that emerges from internal reections, aects the retrieval mostly below 40 km. The performance of Miawara in the mid stratosphere is good as revealed by this intercomparison but so far the timeseries at 10 hpa is not continuous and further eorts are required to achieve stable measurements at this altitude level throughout the year. Generally, the data collected within the SHOMING project are in good agreement and are a good basis to study the synergistic use of these data to retrieve water vapor pro- les of high quality from the ground up to the mesosphere using the best combination of information sources at every altitude. Acknowledgements We acknowledge the big eort made by the team of MeteoSwiss at Payerne to carry out such challanging soundings with the dierent kind of sensors. We also would like to thank the MeteoSwiss team at Payerne for providing RS92, SnowWhite and Ralmo data. 6 77

82 References Deuber, B., N. Kämpfer, and D. G. Feist. A new 22-GHz Radiometer for Middle Atmospheric Water Vapour Prole Measurements. IEEE Transactions on Geoscience and Remote Sensing, 42(5): , Deuber, B., A. Haefele, D. G. Feist, L. Martin, N. Kämpfer, G. E. Nedoluha, V. Yushkov, S. Khaykin, R. Kivi, and H. Vömel. Middle Atmospheric Water Vapour Radiometer - MIAWARA: Validation and rst results of the LAUTLOS / WAVVAP campaign. J. Geophys. Res., 110(D13306), doi: /2004JD Fujiwara, M., M. Shiotani, F. Hasebe, H. Vömel, S. J. Oltmans, P. W. Ruppert, T. Horinouchi, and T. Tsuda. Performance of the meteolabor snow white chilledmirror hygrometer in the tropical troposphere: Comparisons with the vaisala RS80 A/Hhumicap sensors. J. of Atm. and Oc. Tech., 20: , Haefele, A. Atmosphärische Wasserdampfprole von 060 km aus optimierter Kombination von Mikrowellendaten und Ballonsondierungen. Master's thesis, Universität Bern, Haefele, A., E. D. Wachter, K. Hocke, N. Kämpfer, G. E. Nedoluha, R. M. Gomez, P. Eriksson, P. Forkman, A. Lambert, and M. J. Schwartz. Validation of ground based microwave radiometers at 22 GHz for stratospheric and mesospheric water vapor. J. of Geophys. Res., in press. doi: /2009JD Löhnert, U., E. van Meijgaard, H. K. Baltink, S. Gross, and R. Boers. Accuracy assessment of an integrated proling technique for oparationally deriving proles of temperature, humidity, and cloud liquid water. J. of Geophys. Res., 112(D04205), doi: /2006JD Miloshevich, L. M., H. Vömel, D. H. Whiteman, and T. Leblanc. Accuracy Assessment and Correction of Vaisala RS92 Radiosonde Water Vapor Measurements. J. of Geophys. Res., 114(D11305), doi: /2008JD Vömel, H., M. Fujiwara, M. Shiotani, F. Hasebe, S. J. Oltmans, and J. E. Barnes. The behavior of the snow white chilledmirror hygrometer in extremely dry conditions. J. of Atm. and Oc. Tech., 20: , Vömel, H., D. David, and K. Smith. Accuracy of tropospheric and stratospheric water vapor measurements by the cryogenic frost point hygrometer (CFH): Instrumental details and observations. J. of Geophys. Res., 112(D08305), Yushkov, V., V. Astakhov, and S. Merkulov. Optical balloon hygrometer for uppertroposphere and stratosphere water vapor measurements. In Optical Remote Sensing of the Atmosphere and Clouds, SPIE Proceedings. 78 7

83 Chapter 8 Tropospheric Water Vapor Proles Retrieved from Pressure Broadened Emission Spectra at 22 GHz Submitted to the Journal of Atmospheric and Oceanic Technology 79

84 Generated using V3.0 of the official AMS LATEX template journal page layout FOR AUTHOR USE ONLY, NOT FOR SUBMISSION! Tropospheric water vapor profiles retrieved from pressure broadened emission spectra at 22 GHz Alexander Haefele Institute of Applied Physics, University of Bern, Bern, Switzerland Niklaus Kämpfer Institute of Applied Physics, University of Bern, Bern, Switzerland Oeschger Center for Climate Change Research, University of Bern, Switzerland ABSTRACT We present the analysis and the evaluation of the retrieval of tropospheric water vapor profiles from pressure broadened emission spectra at 22 GHz measured with a ground based microwave spectro radiometer. The spectra have a bandwidth of 1 GHz with a resolution of 20 MHz and are centered at GHz. Due to the small bandwidth the retrieval is insensitive to clouds and measurements are possible under almost all non precipitating weather conditions. The retrieved profiles are evaluated with a set of 200 coincident balloon soundings with RS92 sensors. The correlation coefficient between the microwave retrievals and the RS92 measurements lies above 0.7 up to 8 km and the retrievals show a wet bias compared to RS92 of 10 % at 2 km increasing to 30 % at 6 km. 1. Introduction Water vapor is the most important natural greenhouse gas of the atmosphere and has a large impact on its radiative properties and hence on its thermodynamic balance. Despite the importance of water vapor in the climate system and weather forecasting there is no technique available, neither ground based nor space borne, that can provide continuous measurements of the humidity profile under all weather conditions. Ground based microwave radiometers measuring the pressure broadened emission line of water vapor at 22 GHz are good candidates to fill this gap. Classical microwave profilers usually cover a frequency range from 20 to 30 GHz at a few individual channels for humidity measurements (Solheim et al. (1998); Crewell et al. (2001); Ware et al. (2003); Martin et al. (2006)). The instrument presented here, however, was originally designed to measure water vapor profiles in the stratosphere and mesosphere as part of the Network for the Detection of Atmospheric Composition Change, NDACC ( It covers a bandwidth of 1 GHz centered at GHz with very high spectral resolution of 61 khz as required for observations in the middle atmosphere. Thus, only the peak of the H 2 O line is measured, as shown in Figure 1. The spectral signature of clouds within the measured frequency interval of 1 GHz is in good approximation linear in frequency. The forward calculation in the retrieval is therefore based on a clear sky atmosphere and clouds are accounted for by adding an offset and a slope to the cal- α [1/km] Absorption Coefficient: p=850 hpa, T=280 K H2O (6 g/m 3 ) O2 CLW (2 g/m 3 ) Freq. [GHz] Fig. 1. Absorption coefficient between 10 and 40 GHz at 850 hpa for water vapor, cloud liquid water and oxygen. The shaded area marks the bandwidth of the radiometer. culated spectrum. This approach allows to retrieve water vapor profiles under almost all non precipitating weather conditions without additional information on clouds. The paper is organized as follows: A short description of the instrument and the calibration method is presented in Section 2. In Section 3 the retrieval is discussed and the key characteristics are derived. A validation of the retrievals is presented in Section

85 2. Instrument and Calibration The emission spectra are measured by the MIddle Atmospheric WAter vapor RAdiometer, MIAWARA (Deuber et al. (2004)), that is optimized for measuring the narrowband emission line from stratospheric and mesospheric water vapor. MIAWARA was developed and is operated by the University of Bern and is deployed close to Bern, Switzerland, at 47 N, 7 E at 900 m above sea level. It is a single sideband heterodyne receiver converting the incoming signal at GHz to an intermediate frequency of 0.5 GHz in two mixing steps. The signal is analyzed by a digital fast Fourier transform spectrometer with a spectral range from 0 to 1 GHz and a resolution of 61 khz. Hence, the radiometer system is able to measure the emission from water vapor at frequencies between and GHz at very high resolution, as required for observations in the middle atmosphere. This high resolution, however, is of no advantage for the retrieval of tropospheric profiles and for this study the high resolution raw spectra are binned into 20 MHz bins before being calibrated. The instrument is calibrated with a tipping curve measurement (Han and Westwater (2000)) that is performed every half an hour. The tipping curve includes six antenna positions per measurement cycle looking at (1) a microwave absorber at ambient temperature, herein after referred to as hot load, and (2-6) at the sky under zenith angles φ = [30 ; 36 ; 42 ; 48 ; 54 ]. Assuming an isothermal and stratified atmosphere the brightness temperature under a zenith angle φ is given by T b sky (φ) = T bg e τ/ cos(φ) + T eff (1 e τ/ cos(φ) ) (1) where T bg is the cosmic background radiation, T eff is the effective temperature of the atmosphere, derived from the ambient surface temperature T amb, T eff = 0.69(T amb 273) according to Han and Westwater (2000), and τ is the atmospheric opacity in zenith direction. On the other hand we can write the total power calibration assuming to know the brightness temperature of the sky at a reference angle φ ref : T b sky (φ) = (S sky (φ) S sky (φ ref )) T hot T b sky (φ ref ) S hot S sky (φ ref ) +T b sky (φ ref ) (2) Where S sky (φ) is the signal from the sky measured at a zenith angle φ, S hot is the signal from the hot load. Iterating equations (1) and (2), until they are fulfilled for all zenith angles of the tipping curve, yields τ. From numerical simulations the systematic error of τ is estimated to be 5 %. Equation (1) is used to calculate T b(φ ref ) and the error in τ converts to an error in T b(φ ref ) of 3 K. The random error depends strongly on the atmospheric conditions as the atmosphere is assumed to be homogeneously layered. If strong horizontal gradients in the distribution of water vapor are present, the random error is in the order of a few degrees Kelvin while it is below 1 K for a well stratified atmosphere. The receiver temperature can also be derived from tipping curve measurements and lies at 140±5 K. This is in agreement with the value derived from calibrations with a liquid nitrogen load. Tipping curves that reveal a receiver temperature of T rec < 120 K or T rec > 160 K are rejected. The thermal noise on a binned, single spectrum derived from a tipping curve is 0.2 K. For this study bunches of 16 spectra have been summed up before the retrieval is performed in order to reduce the noise to 0.05 K. As tipping curves are performed only every half an hour the retrieved profiles are averaged over 8 h. 3. Retrieval of Water Vapor Profiles An optimal estimation algorithm according to Rodgers (2000) is applied to retrieve water vapor profiles from the pressure broadened emission spectra. Given a state vector x and an additional set of parameters b, the measured radiation spectrum y is calculated with the forward model F : y = F (x, b). (3) The state vector x is the quantity to be retrieved and contains the water vapor profile. b contains parameters describing the sensor, spectroscopic parameters, pressure and temperature profiles as well as profiles of O 2, N 2 and CO 2. For the radiative transfer calculations and for the sensor modeling the software packages Arts (Buehler et al. (2005)) and QPack (Eriksson et al. (2005)) have been used. In fact, within Arts the absorption coefficients are calculated with the model PWR98 (Rosenkranz (1998)). Pressure and temperature profiles are taken from ECMWF reanalyses. Estimating the state vector ˆx, given the measured radiation spectrum y, is the inverse problem. As there exists an infinite number of states that solve the inverse problem the solution has to be constrained to realistic profiles. From the Bayes theorem the following cost function can be derived: c = [y F (ˆx)] T S 1 y [y F (ˆx)] + [ˆx x a ] T S 1 a [ˆx x a ]. (4) x a is an a priori assumption of the water vapor profile, the a priori profile, and S a is the error covariance of x a controlling the strength of the constraint of the solution ˆx to the a priori profile. The solution of the inverse problem, ˆx, minimizes the cost function c. As F is non linear in the troposphere with respect to x, an iterative search for the solution ˆx is required and a Marquardt Levenberg approach is used. Within the spectral range of 1 GHz the contribution from clouds to the measured brightness temperature can be reasonably well approximated with a linear term in the forward model. Thus, a polynomial of degree one is included 2 81

86 Fig. 2. Analysis of the simulated retrievals. Left panel: Matrix of correlation coefficients between the true and the retrieved profiles. As the true profiles are based on RS92 measurements they are only taken into account up to 11.5 km. Right panels: Scatter plots of the true and the retrieved data. r is the correlation coefficient. in the forward model to account for clouds and its coefficients (offset and slope) are part of the state vector x. Our a priori assumptions are as follows: The a priori profile, x a, decreases linearly from the surface volume mixing ratio, obtained from relative humidity and temperature measurements of the nearby weather station, to the statistical mean value at 500 hpa. Above this altitude, the a priori profile is equal to the statistical mean profile, that has been derived from one year of balloon soundings close to the measurement site. The a priori covariance matrix, S a, is calculated from the standard deviations, σ(i), and the correlation lengths, l c (i), at the altitudes z(i) assuming Gaussian statistics: ( ( ) ) 2 z(i) z(j) S a (i, j) = σ(i)σ(j) exp 4 (5) l c (i) + l c (j) The a priori standard deviation is expressed as fraction of the a priori profile and is set to 10 % at the surface increasing to 80 % at 500 hpa and staying constant at 80 % above. The correlation length is set to 2 km (one scale height of water vapor). The a priori values of the offset and slope were set to zero while the standard deviations are set to 9 K and 0.2 K/Hz, respectively. For the characterization of the retrieval a simulation has been performed. From a set of 241 balloon soundings we calculated synthetic spectra that were subsequently inverted using the approach and a priori information as described above. Gaussian noise with a standard deviation of 0.01 K was added to the spectra before inversion. It has to be noted that this low noise is not achieved by the real measurement within reasonable integration times because tipping curves are performed only every half an hour. The simulation should be regarded as representative for the best performance that could be achieved by this approach. Liquid water clouds have been included in the radiative transfer for the calculation of the synthetic spectra. The profile of the liquid water content, ρ, was built according to the following rule: { 0.2 g/m 3 for RH > 95% ρ = 0 g/m 3 (6) for RH 95% In the retrieval, however, the clouds were accounted for with a first order polynomial as described above. The comparison of the true profile with the retrieved profile gives insight in the characteristics of the retrieval. The left panel of Figure 2 shows the correlation coefficient between the true and the retrieved profiles. The high correlation in the first kilometer comes mainly from the influence of the a priori profile that is set to the measured surface value at the ground, as described above. Experiments with a constant a priori profile reveal significantly lower correlation between 1 and 2 km (not shown). The correlation length, here defined as the distance over which the correlation coefficient decreases below 0.5, gives an indication of the vertical resolution. Up to an altitude of 6 km the correlation length is 5 km. For this analysis the data have been deseasonalyzed by subtracting a second order polynomial fit from the one year data record. The seasonal cycle would introduce substantial correlations throughout the whole troposphere. It has to be noted also that the correlation length of water vapor measured by radiosondes is 2 km. The right panels in Figure 2 show the scatter plots of the retrieved and the true atmosphere at four pressure 3 82