Michel Desbois

International Global Precipitation Measurement (GPM) Program and Mission: An Overview

Eric A. Smith , Ghassem Asrar , Yoji Furuhama , Amnon Ginati , Christian Kummerow , Vincenzo Levizzani , Alberto Mugnai , Kenji Nakamura , Robert Adler , Vincent Casse , Mary Cleave , Michele Debois , John Durning , Jared Entin , Paul Houser , Toshio Iguchi , Ramesh Kakar , Jack Kaye , Masahiro Kojima , Dennis Lettenmaier , Michael Luther , Amita Mehta , Pierre Morel , Tetsuo Nakazawa , Steven Neeck , Ken’ichi Okamoto , Riko Oki , Garudachar Raju , Marshall Shepherd , Erich Stocker , Jacques Testud , and Eric Wood 19

A methodology study of the validation of clouds in GCMs using ISCCP satellite observations

The cloudiness fields simulated by a general circulation model and a validation using the International Satellite Cloud Climatology Project (ISCCP) satellite observations are presented. An adapted methodology is developed, in which the issue of the sub-grid scale variability of the cloud fields, and how it may affect the comparison exercise, is considered carefully. In particular different assumptions about the vertical overlap of cloud layers are made, allowing us to reconstruct the cloud distribution inside a model grid column. Carrying out an analysis directly comparable to that of ISCCP then becomes possible. The relevance of this method is demonstrated by its application to the evaluation of the cloud schemes used in Laboratoire de Météoroligie Dynamique (LMD) general circulation model. We compare cloud properties, such as cloud-top height and cloud optical thickness, analysed by ISCCP and simulated by the LMD GCM. The results show that a direct comparison of simulated low cloudiness and that shown from satellites is not possible. They also reveal some model deficiencies concerning the cloud vertical distribution. Some of these features depend little on the cloud overlap assumption and may reveal inadequate parameterisation of the boundary layer mixing or the cloud water precipitation rate. High convective clouds also appear to be too thick.

Global precipitations and climate change

In regard to global change, emphasis is generally placed on the increase in global temperature, but large changes in the distribution of precipitation are also likely to occur. Such changes have been redorded in the past by paleoclimatological studies or in the field of climatology. Different approaches to monitoring and forecasting the evolution of climate-scale precipitation are reviewed by paleoclimatologists, hydrologists, satellite meteorologists, and climate modellers.

Satellite estimation of the tropical precipitation using the METEOSTAT and SSM/I data

Abstract The purpose of this study is to test a satellite method for estimating precipitation using infrared and microwave data. The method is validated by comparing the rain estimates to the ground precipitation measurements, over continental tropical regions (West Africa). The rain estimation method used is based on an automatic classification algorithm combining infrared and microwave satellite data. The method takes advantage of both, the good time and space resolution of infrared satellite images and the rain related information retrieved from microwave images. This method of rain and cloud classification (RACC) gives homogeneous classes characterising the different types of clouds or rain-rates. A set of microwave images from the 86 GHz channel of the passive microwave radiometer SSM/I and of the coincident infrared images from Meteosat is used in the learning phase of the classification process, while the full set of half-hourly infrared images is needed for the application phase. The areal rain estimates are computed with the microwave-infrared combined RACC method for time periods ranging from one to twelve hours and for different areas up to 120 km × 120 km. The correlation with the ground rainfall data given by the raingauge network of a validation site in Niger are estimated and compared to the correlations obtained for the rain estimated derived from a method based on a single infrared threshold and using the Meteosat images only. The improvement brought by the RACC method is discussed.

Construction of cloud trajectories to study the cloud life cycle

Abstract Cloud motion winds are usually computed on a regulared grid from geostationary satellite images. An extension of this type of methods enables the calculation of cloud trajectories and the study of their evolution. It has been applied to a series of Meteosat window images covering one day (18 Oct. 1989) of the International Cirrus Experiment (ICE). A test based on the reverse calculation of the cloud motion vector can give indications about its quality.

Comparison of radiance fields observed by satellite and simulated by the LMD general circulation model

Abstract A time series of International Satellite Cloud Climatology Project (ISCCP) B2 data has been used to study the spatio-temporal variability of cloud radiance fields and to compare it with similar diagnostics obtained from a numerical simulation with the Laboratoire de Meteorologie Dynamique general circulation model (GCM). We first compare zonal means of the observed and simulated fluxes: the largest differences appear mainly above land, probably due to a diurnal cycle effect. A computation of the cloud radiative forcing is made to distinguish model errors in the clear sky or in cloudy areas: it shows that the radiative effect of the simulated clouds is generally smaller than that of observed ones. The influence of the spatial resolution on the variability of the visible and infrared radiance fields is examined. The results show that the variability due to the better spatial resolution of the satellite data cannot be simulated by the GCM; however, the variability of the simulated radiances is comparable to that of the satellite data when they are spatially averaged on the GCM mesh. The temporal variability of the spatially averaged observed and simulated radiances has a similar spatial distribution but the model results are slightly smaller. The memory of the precedent state shown by the autocorrelation function is longer for the GCM than for the observation. However, the periods obtained by a spectral analysis in the mid-latitude regions are approximately the same in the observation and the simulation.

A note on: Estimation of rainfall due to squall lines over West Africa using meteosat imagery

SummaryThe results of the first step of a project to develop a method to estimate precipitation over the Soudano-Sahelian belt of West Africa are reported.The study has been performed over the period from 10 June to 9 July 1986 using hourly METEOSAT infrared images. 122 individual cloud clusters associated with squall lines or tropical storms have been tracked. For each event, the time variations of a convection index giving the volume of cloud cooler than −40°C has been determined every hour. The convection index exhibits a strong diurnal cycle. From daily rainfall amounts obtained at about 300 stations, and assuming a time apportion of rainfall within a cloud cluster, the time variations of the hourly total rainfall produced by the cluster can be determined and represents the precipitation index. Because of insufficient rainfall, the precipitation index has been determined for only 17 events. For 2/3 of the 17 cases, there is a significant correlation between the two indices. For each of the 17 events, precipitation has been regressed on the associated convective index and relative time variations. In that case, the results indicate that a convective index representing the life history of the cloud cluster can be calibrated with corresponding raingage measurements provided raingage data are available. Then, estimation of rainfall due to that event over data void regions can be obtained. However, this study shows that no universal relationship exists between precipitation: no rainfall can be estimated if there is not enough raingage measurements to construct a precipitation index for a tracked cloud cluster. This represents a limitation to the method.

An attempt to retrieve low cloud motion winds over land in the African Monsoon flow on Meteosat pictures

It has been possible to extract cloud motion winds corresponding to the African monsoon flow during the rainy season over land from the set of all cloud wind vectors provided by a crude cloud tracking method on VIS Meteosat pictures. The selection cannot be based only on IR radiances, due to the presence of semi-transparent clouds and partial coverages of the pixels. Several tests have been proposed with an additional selection, based on the analysis of the histogram of cloud motion directions. Cloud motion winds obtained on VIS pictures are clearly separated in this diagram and justify the selection of a direction range where only clouds embedded in the monsoon flow are tracked. These winds are also validated by comparisons with NCEP analysed fields.

DYNAMIC CLASSIFICATION OF MESOSCALE CLOUD PATTERNS

Abstract Cumulative histograms taken from eight consecutive infrared pictures of GOES 1 are used to study the spatio-temporal evolution of mesoscale cloud patterns. A model is proposed to approximate the volume of cloud lying in a particular area. Then, the cloud volume variation between successive pictures in the same area is deduced. This volume variation is used to define a cloud development index. Development or dissipation of cloud formations can be distinguished using this index.

Monitoring the life cycle of cirrus clouds using Meteosat‐4 data during ICE‐1989

During the International Cirrus Experiment (ICE), cirrus cloud properties over the North Sea were observed from local surface-based and aircraft instruments. In order to place these measurements in the context of the general space-time evolution of the cirrus systems and smaller cloud elements, Meteosat images were used to study the evolution of the clouds along their trajectories. The trajectories were established with techniques derived from those used to compute cloud motion winds from geostationary satellite images. Thermal infrared (IR) and water vapour (WV) channel tracking was performed to take advantage of the properties of each channel: better contrast of the clouds over the surface in the IR, and better discrimination of upper layers in the WV channel. Although long tracks, up to 60 hours, could be observed, they did not correspond to the same cirrus elements, or even the same system along the track. The evolution of the brightness temperatures along the tracked elements showed periodic fluctuations. The representativeness of these fluctuations in terms of cloud life cycle is discussed, as it may be related to the quality of the tracking. Copyright