B. Guillemet

Neural network retrieval of cloud parameters of inhomogeneous and fractional clouds: Feasibility study

One of the major issues of cloud parameter retrieval is how to optimize the improved observational capability of new radiometers to retrieve additional information about cloud characteristics. To investigate this problem, we developed a neural network approach for simultaneous retrieval of cloud parameters of inhomogeneous clouds with fractional cloud cover. We defined a simple inverse inhomogeneous cloud model with four parameters: mean optical depth, effective radius, relative cloud inhomogeneity, and fractional cloud cover. The retrieval algorithm, based on the use of a mapping neural network (MNN) with two hidden layers, was implemented and tested with synthetic multispectral reflectance data prepared for 2D clouds generated with a modified bounded cascade cloud model. We found that these cloud parameters could be retrieved from the moderate-resolution multispectral reflectance data with reasonable accuracy: for example with our data base, optical depth has a root mean square error of 1.7 for an ensemble of 1000 samples with optical depth up to 30. However, this accuracy depends on measurements errors and noises. A comparison with plan-parallel hypothesis shows the expected improvement of such an inhomogeneous cloud model. We show that the relevance of these cloud parameters is a function of the horizontal scale of averaging due to the net horizontal photon transport to and from adjacent cloud pixels. We tested the inclusion of ancillary data (reflectance of the neighbouring pixels) into the retrieval algorithm, and showed that the use of these ancillary data could partially correct the modeling error and significantly improve the performance of cloud parameter retrieval.

Effective radiative properties of bounded cascade nonabsorbing clouds: Definition of the equivalent homogeneous cloud approximation

In the present study we investigated the radiative properties of inhomogeneous nonabsorbing clouds under the Equivalent plane-parallel Homogeneous Cloud Approximation (EHCA), by using the one-dimensional (1-D) bounded cascade inhomogeneous clouds. The effective optical depth was defined under the EHCA by requiring the identity of the radiant flux components of the radiation budget between the inhomogeneous clouds and their equivalent homogeneous counterparts. Such requirement provides a rational framework to define the effective optical depth of the inhomogeneous nonabsorbing clouds. We analyzed the dependency of the effective optical depth on the horizontal scale of averaging and solar incidence angle and specified the conditions under which an inhomogeneous cloud segment could be treated as a plane-parallel homogeneous cloud. A parameterization of the effective optical depth was proposed as a function of the mean optical depth and a relative cloud inhomogeneity parameter. Finally, we compared the EHCA with the effective thickness approximation, both based on the definition of the effective optical depth, and discussed the difference between their respective effective optical depths.

Pre-EUCREX Intercomparison of Airborne Humidity Measuring Instruments

Abstract During the pre-EUCREX (European Cloud and Radiation Experiment) intercomparison of airborne instrumentation in January 1992, nine hygrometers mounted on three different aircraft were compared. Although the different instruments are based on completely different principles and the three aircraft have very different flying characteristics, humidity data from both vertical profiles as well as horizontal flight legs showed good agreement. Despite the different aircraft limitations the intercomparison was done with the aircraft in close formation. In terms of relative difference in mixing ratio, most instruments agreed to within ±5% for values down to about 0.1 g kg−1. For mixing ratios between 0.03 and 0.1 g kg−1 most instruments agreed to within ±15%. Systematic differences between the instruments suggest that in joint experiments where data will be shared, the same algorithms for evaluating and converting humidity parameters should be used whenever possible.

Effective radiative properties of bounded cascade absorbing clouds: Definition of an effective single-scattering albedo

We applied the equivalent homogeneous cloud approximation (EHCA) to the bounded cascade inhomogeneous absorbing clouds and defined their effective radiative properties. It is found that we have to introduce an effective single-scattering albedo in addition to an effective optical depth to treat the inhomogeneous absorbing clouds under the plane-parallel homogeneous cloud assumption. For an inhomogeneous absorbing cloud, a pair of the effective parameters can be estimated from each one of three possible pairs taken from the area-averaged reflectance, transmittance and absorptance. We found that the behavior of these effective properties was quite similar to those observed for the inhomogeneous non absorbing clouds except that two effective parameters were to be examined instead of only one effective parameter for the nonabsorbing clouds. Empirical relations for both the effective optical depth and the single-scattering albedo were given as a function of the local mean optical depth and relative local cloud inhomogeneity. We showed that the effective single-scattering albedo could not be properly introduced under the effective thickness approximation (ETA), which indicates an important conceptual difference between the EHCA and the ETA. Finally, we discussed possible consequences of the effective single-scattering albedo, defined in this study, with respect to the anomalous absorption phenomenon.

Molecular dissipation of turbulent fluctuations in the convective mixed layer part I: Height variations of dissipation rates

During the Limagne and Beauce experiments, the INAG-IGN Aerocommander FL 280 aircraft made extensive ‘in situ’ measurements of turbulent fluctuations in diurnally evolving convective boundary layers. In this paper, these measurements were used to investigate characteristics of the molecular dissipation of turbulent fluctuations through the mixed layer and well into the overlying stable layer. The dimensionless dissipation rates of turbulent kinetic energy, temperature and humidity variances, and temperature-humidity covariance (ψ, ψθ, ψqand ψθq) were computed and their height variations analysed.The behaviour of the dissipation rate ψ was found to differ significantly from those observed for the other rates. In the lowest region of the mixed layer, ψ does not obey the local free convection prediction. Instead, it follows practically a relationship similar to the one established in the surface layer by Wyngaard et al. (1971). The dissipation rate ψ remains fairly constant in the bulk of the mixed layer (0.3 ≤ z/Zi≤ 0.8) and shows a very rapid decrease above the inversion. These results confirm those reported previously from the Minnesota and Ashchurch data by Kaimal et al. (1976), Caughey and Palmer (1979), etc.The height variations for the other dissipation rates were found to obey, as expected, the (z/Zi)-4/3 decrease predicted under the local free convection similarity hypothesis in the lowest region of the mixed layer. This region extends to the height z/Zi- 0.4, 0.1, and 0.3, respectively, for ψθ, ψq, and ψθq. Above these levels, the dissipation rates ψθ and ψq show, on average, a slight increase to reach peak-values near the mixed-layer top, while the ‘dissipation’ rate ψθqchanges sign from positive to negative around the height z/Zi, - 0.7. These characteristics confirm the fact that the structures of temperature and humidity fluctuations are considerably affected by their entrainment-induced fluctuations. Therefore, an attempt has been made to non-dimensionalize the dissipation rates near the mixed-layer top with the interfacial scaling factors.

Inhomogeneity effects of 1D and 2D bounded cascade model clouds on their effective radiative properties

Abstract Inhomogeneous clouds generated with the bounded cascade process were used in a number of recent studies on the interaction between the radiative transfer process and inhomogeneous clouds. In this study, we investigate how the EHCA could be applied to 1D and 2D bounded cascade inhomogeneous absorbing clouds generated with different pairs of fractal parameters H and f . Firstly, we found that the empirical formulas we previously established were still applicable to other 1D and 2D inhomogeneous clouds with different single scattering albedo and with various types of fluctuations of the optical depth, except the case of very intermittent inhomogeneous clouds generated with H = 0 whatever f and H = 0.25 and f ≥ 0.3 (i.e. ϱ τ > 1.5 ∼ 2, where ϱ τ is a relative cloud inhomogeneity parameter defined as the standard deviation of the optical depth normalized by the mean optical depth). Secondly, we also found that the bias in the bidirectional reflectance factor between 1D and 2D inhomogeneous clouds radiance and their EHCA counterpart remained rather small (≤ 5%), again except the case of very intermittent inhomogeneous clouds with ϱ τ > 1.5 ∼ 2, for which this bias reached 20% to 60%.

Neural network analysis of the radiative interaction between neighboring pixels in inhomogeneous clouds

In this study, we analyzed the effect of the radiative interaction between neighboring pixels on the high-resolution radiant flux of bounded cascade inhomogeneous clouds by using a one-layer mapping neural network as generalized regression analysis. The analysis was done for reflectance, transmittance, and absorptance at different wavelengths under different conditions of illumination. The sign and magnitude of output coefficients indicate how neighboring pixels contribute to the radiant flux of a target pixel. We found that the variation of output coefficients with the distance from the target pixel changes significantly in its shape and horizontal extent not only with the type of radiant flux we consider but also with the wavelength and solar zenith angle. The mapping neural network clearly reveals the asymmetric feature of radiative interaction between neighboring pixels under oblique illumination, which illustrates the shadowing and enhancing effects of local cloud inhomogeneity. The present analysis shows that the mapping neural network is a flexible method of analysis when used as a generalized regression analysis.

Neural network retrieval of cloud parameters from high-resolution multispectral radiometric data: A feasibility study

Abstract One of the important issues of the cloud parameter retrieval is how to optimize the improved observational capability of new radiometers. In this study, we examined a neural network approach to retrieve simultaneously optical depth and effective radius of overcast bounded cascade clouds from high-resolution multiwavelength radiometric data. The high-resolution retrieval allows the assumption of uniform cloud parameters within the target pixel but also requires the integration of radiometric data of neighboring pixels as ancillary data because of the net horizontal transport of photons. The performance of the mapping neural network (MNN) high-resolution retrieval was evaluated under conditions of vertical and oblique illumination using six pairs of wavelengths with 0.64, 1.6, 2.2, and 3.7 μm. Two types of clouds are used: inhomogeneous clouds with horizontal uniform effective radius and inhomogeneous clouds with horizontally variable optical depth and effective radius. The results show that we can retrieve these cloud parameters with a reasonable accuracy, which varies with the spectral channels used for the retrieval. The application of a “one-neuron” MNN for the cloud parameter retrieval shows that the effective radius estimation depends on visible wavelength when used with another having only a small absorption as 1.6 or 2.2 μm.

Analyse des hétérogénéités spatiales des stratocumulus et cumulus

Abstract Spatial inhomogeneities in clouds have a lot of consequences especially in both radiative properties and internal mechanisms in relation with effects of mixing. We are concerned by the representation of these inhomogeneities for spatial scale in the range of 10 m to several ten kilometers in cases of stratocumulus and cumulus clouds. These clouds are characterized by means of threshold of the droplet concentration measured with an aircraft during thwo experiments: NEPHOS-1982 and FRONTS-1984. In the cases of stratocumulus (NEPHOS-1982) we use only droplets with diameters less than 30 μm, while in the case of cumulus clouds (FRONTS-1984) we use both little droplets (diameters less than 30 μm) and hydrometeors (probably ice crystals) with diameters greater than 60 μm. The probability distribution to obtain a cloudy segment or a clear-air segment with a fixed length is described by a power-law which is characteristics for a geometrical scaling range. A possibility for the existence of scaling exponents is discussed with a very simple model of cloud mixing in relation to turbulence. We discuss consequences of scaling law for a parameterization of the cloud surface and the possibility for the non-unicity of this scaling.

Mesoscale cloud pattern classification over ocean with a neural network using a new index of cloud variability

The purpose of this study is to determine the feasibility of a mesoscale (<300 km) cloud classification using infrared radiance data of satellite‐borne instruments. A new method is presented involving an index called the diversity index (DI), derived from a parameter commonly used to describe ecosystem variability. In this respect, we consider several classes of value ranges of standard deviation of the brightness temperature at 11 µm (σBT). In order to calculate DI for 128×128 km2 grids, subframes of 8 km×8 km are superimposed to the satellite image, and then σBT is calculated for all 256 subframes and assigned to one of the classes. Each observed cloud pattern is associated with an index characterized by the frequency of σBT classes within the scene, representative of a cloud type. Classification of different clouds is obtained from Advanced Very High Resolution Radiometer (AVHRR)‐NOAA 16 data at 1 km resolution. Stratus, stratocumulus and cumulus are specifically recognized by this window analysis using a DI threshold. Then, a six‐class scheme is presented, with the standard deviation of the infrared brightness temperature of the entire cloud scene (σc) and DI as inputs of a neural network algorithm. This neural network classifier achieves an overall accuracy of 77.5% for a six‐class scheme, and 79.4% for a three‐class scheme, as verified against the analyses of nephanalists as verified against a cloud classification from Météo France. As an application of the proposed methodology, regional cloud variability over Pacific is examined using cloud patterns derived from the Moderate Resolution Imaging Spectroradiometer (MODIS) carried aboard Earth Observing System (EOS) Terra polar orbiter platform, for February 2003 and 2004. The comparison shows regional change in monthly mean cloud types, associated with 2003 El Niño and 2004 neutral events. A significant increase in the occurrence of convective clouds (+15%) and a decrease in stratiform clouds (−10%) are observed between the two months.