Atmospheric gravity waves and turbulence generate small-scale fluctuations of wind, pressure, density, and temperature in the atmosphere. These fluctuations represent a real hazard for commercial aircraft and are known by the generic name of clear-air turbulence (CAT). Numerical weather prediction models do not resolve CAT and therefore provide only a probability of occurrence. A ground-based Rayleigh lidar was designed and implemented to remotely detect and characterize the atmospheric variability induced by turbulence in vertical scales between 40 m and a few hundred meters. Field measurements were performed at Observatoire de Haute-Provence (OHP, France) on 8 December 2008 and 23 June 2009. The estimate of the mean squared amplitude of bidimensional fluctuations of lidar signal showed excess compared to the estimated contribution of the instrumental noise. This excess can be attributed to atmospheric turbulence with a 95% confidence level. During the first night, data from collocated stratosphere-troposphere (ST) radar were available. Altitudes of the turbulent layers detected by the lidar were roughly consistent with those of layers with enhanced radar echo. The derived values of turbulence parameters Cn2 or CT2 were in the range of those published in the literature using ST radar data. However, the detection was at the limit of the instrumental noise and additional measurement campaigns are highly desirable to confirm these initial results. This is to our knowledge the first successful attempt to detect CAT in the free troposphere using an incoherent Rayleigh lidar system. The built lidar device may serve as a test bed for the definition of embarked CAT detection lidar systems aboard airliners.

We present the results of a study on the meteorological phenomena of oceanic and Mediterranean origins that could lead to exceptional floods in the Seine catchments upstream Paris. This analysis has been made feasible thanks to the exploitation of continuous time series during the 20(th) century of rain fall observations (MF and AESN rain gauges on more than 20 locations). In addition re-analysed meteorological fields have been used from NCEP as well as rebuilt ground pressure from WetterZentrale in order to extract the daily weather typologies from 1900 according to a simple extraction technique to sort out Mediterranean and oceanic origins of precipitations. Three studies have been conducted (meteorological fields, rain precipitation fields and hydrologic response) by using statistical analysis in both spatial and temporal dimension for the three Seine catchments upstream Paris (Yonne, Marne, Seine). Attempts have been performed to make appear causality chains between main phenomena. One interesting result is the rate flow of Seine in Paris for return period of 100 to 1000 years. In addition, perspectives given through similar use of longer time series (1850-2000) of ground pressure are presented to be applied to characterise the meteorological critical situation for the "point triple" that is the intersection location between Loire, Saone and Yonne catchments for which historical concomitant floods are known (e.g. 1866).

Atmospheric gravity waves and turbulence generate small-scale fluctuations of wind, pressure, density and temperature in the atmosphere. These fluctuations represent a real danger for commercial aircrafts and are known under the generic name of Clear Air Turbulence (CAT). They are not resolved in weather forecast models and are therefore unpredictable. A ground-based Rayleigh lidar was designed and implemented to remotely detect and characterize the atmospheric variability induced by gravity waves and turbulence in vertical scales between 10m and 1000m. Field measurements at Observatoire de Haute-Provence (France) have shown that the built lidar device was actually able to detect episodes of turbulence. This is to our knowledge the first Rayleigh lidar system able to detect clear air turbulence. The built lidar device may serve as a test bed for the definition of embarked CAT detection lidar systems on board airliners.

Spectroscopy for the Investigation of the Characteristics of the Atmosphere of Mars (SPICAM) is the first instrument orbiting a planet other than Earth that is dedicated to the technique of stellar occultation. During the first year of operation on board Mars Express, SPICAM observed more than 500 star occultations, yielding vertical profiles of CO2, ozone, and dust/clouds/aerosols. We review the principles of a star occultation in the absorptive regime, emphasizing two advantages of this method: an absolute value is obtained from a relative measurement without the need for an absolute calibration of the instrument, and the altitude of the measurement is accurately known because it depends only on the position of the spacecraft and not on the pointing of the instrument. We describe a general algorithm used for all occultations. First, we derive from the raw data the transmission of the atmosphere as a function of wavelength, T(λ, z), taking account of instrument‐specific factors. Then a spectral inversion retrieves the slant densities (local densities integrated along the line of sight) of all absorbing species for each measurement of the transmission T(λ, z) during the occultation. Finally, a vertical inversion retrieves the vertical distribution of the local densities from the series of the slant density measurements. This vertical inversion includes a new scheme of Tikhonov regularization. This paper will serve as a reference for the SPICAM Mars Express data which will be systematically made available to the public in the PDS‐like archive managed by ESA.

GOMOS (Global Ozone Monitoring by Occultation of Stars) on Envisat measures ozone, NO/sub 2/, NO/sub 3/, H/sub 2/O, aerosols, neutral air density, and temperature in the stratosphere and mesosphere by detecting the absorption of starlight in UV, visible and infrared wavelengths. During bright limb conditions, GOMOS also observes scattered solar radiation. GOMOS delivers ozone concentration profiles at altitudes 15-100 km with a vertical resolution of about 1.5 km and with a global coverage. As a self-calibrating method, stellar occultation measurements provide a basis for a long-term monitoring of ozone profiles. We present results achieved during the first year of the GOMOS validation program. The validation is based on comparisons with lidars, ozone sondes, balloon borne instruments, other satellites as well as with climatological and meteorological data.

RESUME This paper summarises the status of the GOMOS calibration assessed by the GOMOS CAL/VAL team after nine months of the instrument in-flight life. The first GOMOS occultation was successfully measured on March 22, 2002, starting the calibration and verification activity phase. Since this date, more than 50000 occultations have been performed until end of 2002. The GOMOS CAL team was in charge of i) the full in-flight calibration and recharacterisation of GOMOS, ii) all the activities required to achieve the complete verification of the level 1b processing models, including: algorithms, auxiliary products, instrument performance, measurements geolocation and validation of the atmosphere related parameters. Lastly, the GOMOS CAL team provided the relevant inputs to the GOMOS mission planning monitoring, and to the future routine calibration operations and validation activities. The GOMOS instrument in-flight performances are globally very good, in line with expected budget, except for CCD behaviour (hot pixels, RTS). All measured performances are described hereafter, together with assessment methodology and recommendations. 1 GOMOS INSTRUMENT PERFORMANCE 1.1 Brief reminder of the GOMOS instrument measurement concept The GOMOS operating principle relies on the stellar occultation method, which consists to observe a star outside the atmosphere (thus providing a reference stellar spectrum) and track it as it sets through the atmosphere (thus providing spectra with absorption features). When these occulted spectra are divided by the reference spectrum, nearly calibrationfree horizontal transmission spectra are obtained, assuming the instrument response function does not change during one occultation; lasting typically 30 to 40 seconds. These transmissions provide the basis for retrieval of atmospheric constituents density profiles, which benefits from the fact that observing the star light confines the measurement to a thin well-defined volume. 1.2 Method The verification of the instrument health and performance activities have started on March 22, 2002, just after the SODAP phase which had validated the correct behaviour of the Service Module, Payload Equipment Bay and Payload instruments, checked the GOMOS level 0 products format, and roughly verified that the instrument can be commanded as expected. Due to launch and in-orbit environments, several instrument characteristics may have been modified like the instrument response function, thermal distortion and settling of the optical bench, ... and this calls for calibration and monitoring of GOMOS parameters behaviour and performance: electronic gain chain, read-out noise and offset, non linearity, dark charge, pixel response non-uniformity, radiometric sensitivity, spectral line spread function, wavelength assignment, vignetting, and straylight. All measured performances are described hereafter. The method and tools used during the Calibration phase are briefly described. Then, the measurement results are analysed and compared to the expectations and the specifications. Finally, for each calibration parameter, recommendations are given for the routine mission. __________________________________________________________________________________________________________ Proc. of Envisat Validation Workshop, Frascati, Italy, 9 – 13 December 2002 (ESA SP-531, August 2003) 1.3 Acquisition, detection and pointing performance The objective is to verify if the GOMOS instrument is able to acquire and track stars outside and through the atmosphere. This task will allow to assess tracking limits (altitude range versus star magnitude) and to verify the dynamic component of the pointing errors that contribute to the dynamic spatial and spectral Line Spread Functions. The detection and acquisition performance have been analysed: • during SODAP, for the acquisition and tracking ranges • by analysing the successful acquisition of faint stars The mispointing observed during SODAP is linked with the acquisition cone. It has been analysed using the Most Illuminated Pixel position in detection given in the auxiliary data. Tracking capability has been analysed using the SATU noise monitoring (auxiliary data) and looking at the altitude corresponding to the tracking loss. The tools used for these analyses are mainly the Calibration software CALEX and the GOMOS processing prototype GOPR. The acquisition range and tracking Field Of View are consistent with the delivery status. Margins have been taken in order to be sure not to hit the limits. Table 1: Ranges and duration performance 10 s 2°/s > 2°/s Rallying speed 7°; 5° > 7.1°; 5.3° 7.4°; 6.5° Tracking FOV -10 – 90° 62 – 68° > -10.8 – 90.8° > 62 – 68.8° -11 – 91° 61.7 – 69° SFM range spec measured expected Acquisition The duration of rallying and centring phases are also confirmed. For the performance in detection, a limited number of stars have been tested and the detection threshold has been kept to a safe value. Indeed, a too small value can lead to a class C anomaly which put instrument in Refuse mode. Even if there is no danger for the instrument, we decided not to test the limits of detection. Indeed, the actual number of tested stars is already much larger than the scientific need (and than the specification). Table 2: Detection probability results 3.2 > 4 4.4 Cold star bright limb 1.6 > 3.2 2.2 Hot star bright limb 4 > 4.5 6.8 Cold star dark limb 2.4 > 3.7 4.6 Hot star dark limb spec measured expected Detection capability Acquisition mispointing performance is the difference between expected and observed position of the star in the SATU FOV after rallying. The expected value is less than 0.16°. The observed value during SODAP is 0.3° in elevation 'El) and 0.1° in azimuth (Az). The analysis of the unexpected mispointing in Elevation has shown: • a contribution due to a forgetting of GOMOS optical frame / GOMOS reference cube alignment in the configuration matrix (0.1° in elevation and 0.06° in azimuth). • a possible impact of the gravity compensation of the mechanism. A shift of a few tens of microns could easily induce a bias of 0.1° in elevation. • a possible contribution due to the calibration curves of SFA angles. This difference has been corrected and the stars are now detected at the centre of the SATU. No trend has been observed since then (Fig. 1). Fig. 1: Position of the detection of stars in SATU before and after the correction. Tracking noise and robustness: the expected value of the SATU noise equivalent angle in dark limb (at 100 Hz) is 13 μrad (3s) for a specification of 18 μrad. The measured value is better than 3 μrad in dark limb (DL) and 10 μrad in bright limb (BL) (35 km altitude). Noise is less than 1 μrad (3σ) for Sirius. Fig. 2 presents the mispointing error reported by the SATU data. Movement in elevation around 100 km is due to the presence of the O2 emission layer. The refraction effect can be seen in elevation below 40 km. -50 -40 -30 -20 -10 0 10 20 30 40 50 0 20 40 60 80 100 120 140 Elevation Azimuth μrad SATU noise at 100 Hz for Sirius Dark limb km -50 -40 -30 -20 -10 0 10 20 30 40 50 0 20 40 60 80 100 120 140 Elevation Azimuth μrad SATU noise at 100 Hz for ID 154, Bright limb

We present spectrally analyzed high‐resolution balloon measurements of vertical temperature and horizontal wind in the troposphere and lower stratosphere in French Guyana (≈5°N, 52°W) made in 1996/1997. The paper is principally concerned with spectra at vertical scales from a few kilometers down to 100 m. A special balloon has been developed to solve scales below 500 m. For the first time, the equatorial region has been studied extensively at this resolution. We have found wind and temperature spectra following the canonical shape in the logarithmic representation with a slope of ≈−3 for wavenumbers ≧10−3 cycle/m with day‐to‐day significant amplitude variations but a mean spectrum constant with the seasons. We have found isotropic wave horizontal direction of propagation for the troposphere and anticorrelated with the quasi‐biennial oscillation in the stratosphere. The vertical propagation of wave has been found symmetric in the troposphere but highly asymmetric in the stratosphere with dominant upward propagation. We were not able to find any correlation between deep moist convection and wave activity. The ratio kinetic to potential energy spectra, which is the ratio of velocity spectra to temperature spectra multiplied by N2T¯2b/g2 for the troposphere and stratosphere, is constant versus the vertical wavenumber and in good agreement with a model intrinsic frequency spectrum varying as ω−p with 5/3 ≤ p ≤ 2 and without necessity of energy enhancement near the inertial frequency which is close to zero at the latitude measurements.

The results of a program to estimate the limiting accuracy of atmospheric wind profiles obtained by tracking a new balloon, the Rubsonde, with the Global Positioning System (GPS) are presented. In the first part, the authors detail the data processing and find the total precision of wind measurements in the case of a Rubsonde tracked by a very high resolution centimetric radar, the Aquitaine, is as low as 0.3 m s(-1) for wind measurements over 25-m layers. In the second part, the Rubsonde-GPS wind measurements are compared with those of the Jimsphere tracked by the Aquitaine, and it is found that the instrumental accuracies are the same. The difference between these two 25-m atmospheric wind profiles, separated by a 10-min time interval, is found to be less than 0.6 m s(-1). The validity of the Rubsonde-GPS system for measuring wind variability is then demonstrated.

In this paper we present the first rotary spectra of the horizontal wind shear measured in the stratosphere with a resolution of 25m. A low frequency gravity wave propagating upward with a vertical wavelength of ≈ 1000 m is clearly identified. This result is quite consistent with many previous observations and may explain why the horizontal velocity spectra may be biased and indicates a power variation law versus vertical wave number steeper than those of both temperature and vertical velocity. Nevertheless, around the small vertical wavelengths (100 m), the asymptotic slope of temperature and horizontal velocity spectra are both close to ‐3 and so in better agreement with the saturated model predictions.

The first simultaneous power spectral densities of high resolution lower stratospheric temperature and horizontal/vertical velocity fluctuations are presented. Amplitudes are within a factor of 2 or 3 of other observations reported in literature. However, the measured amplitude ratios, which are absolute characteristics of the fluctuating wave field, disagree significantly (multiplicative factor up to 10) with predictions of the saturation theory. Our results disagree with the concept of a universal spectrum controlled by saturated waves.

We present in this paper the first high resolution analysis of wind shears and temperature gradient measured over 25 m in the low stratosphere. Their power spectral densities deduced by two different methods show that for vertical wavelengths greater than 500 m the behaviors of the temperature and vertical velocity fluctuating field are significantly different from the saturated wave model predictions.

We present in this paper one of the first high resolution analysis of temperature profiles measured in the lower stratosphere. The temperature fluctuations level is shown to depend only on the static stability. Moreover, in the 12 ‐ 1,500 m vertical wavelength range, the temperature spectrum has a mean slope around ‐ 2.3, a value significantly above the ‐ 3 saturated model value.