Identification Of Cloud Clusters And Rain Cells And Their Features From Multi-Satellite Observations

The development of satellite remote sensing technique helps to reveal the characteristics of clouds and precipitation.Based on the entirety of the cloud and precipitation systems,this study identified and parameterized the pixel-level products observed by the spacebome instruments,and obtained several cloud-cluster/rain-cell datasets.Then,the temporal/spatial variation,scale,shape,and intensity of cloud and precipitation systems were revealed,and the role of some special cloud systems on the Earth's radiation budget were recognized.On this basis,some uncertainties in the use of satellite remote sensing data were clarified.The main work of this paper is as follows:1)Rain-cell identification and its featuresIn order to reveal the characteristics of rain cells in the tropics,rain cells were identified by grouping adjacent precipitation pixels utilizing Tropical Rainfall Measuring Mission(TRMM)Precipitation Radar(PR)data.A minimum bounding rectangle method was used to fit rain cells,and a variety of geometrical parameters(including scale,morphology,etc.)and physical parameters(including precipitation intensity,convective ratio,etc.)were defined.Results indicate about 50%of rain cells occurring at dimensions of 20 km(length)and 15 km(width),and the mean length and width of rain cells keep at a log-linear relationship.Land rain cells are "taller and thinner",whereas ocean rain cells are "shorter and fatter".Some rain-cell physical parameters have certain correlations with geometric parameters such as rain-cell area.The characteristics of the rain cells varies greatly in different regions,and the relationships between parameters are also different.These results help to understand the nature of the rain cell and provide observations for further estimation of the physical parameters using the geometrical morphology.2)Cross-Section cloud-cluster identification and its featuresA cloud-cluster dataset was presented using CloudSat Cloud Profile Radar observations from June to August,between 2006 and 2010.The cloud clusters were defined by grouping contiguous cloud pixels using a defined criteria,and then the characteristics of identified cloud clusters were analyzed.Based on the complex topography of the southern Himalayas,four adjacent regions were selected.The characteristics of the cloud clusters in the four regions are different,and the distribution of the cloud-top height gradually changes from bimodal(3 km and 15 km)over the plain,to unimodal(7-9 km)over the Plateau.From the plain to the Plateau,the average cloud-top height and cloud-base height increase after an initial decrease,whereas the maximum reflectivity decrease after an initial increase,which suggests that the cloud clusters may produce precipitation on the slope.3)Horizontal cloud-cluster identification and its featuresBy using satellite observations,the latest reanalysis data and the Hybrid Single-Particle Lagrangian Integrated Trajectory model,we reviewed an extreme event occurred in the middle and lower reaches of the Yangtze River in 2016,and analyzed the meteorological conditions and cloud-cluster movement that led to it.Satellite observations reveal that the storm system formed on the Tibetan Plateau.After continuous development,disappearance,and redevelopment,it gradually moved out of the Tibetan Plateau and affected the Yangtze River Basin in the downstream direction.The reanalysis data show that the South Asian High and the Subtropical High in this process were far stronger than the climate mean state.The Tibetan Plateau Vortex and the Southwest Vortex superimposed.The generation and movement of such a cloud system and vortex is an important disturbance for the generation of heavy precipitation in the downstream direction.4)Identification of cirrus cloud lateral boundary and its featuresDue to the thinness and small scale of cirrus cloud,its lateral boundary may be missed by conventional passive remote-sensing techniques and climate models.Here,using CALIPSO observations in June-August from 2006 to 2011,a global dataset for the cirrus cloud lateral boundary(CCLB)was established.The results indicate that the optical properties,such as the lidar backscatter and the optical depth,sharply decrease from cloudy regions to clear-sky regions.There are significant regional differences in the features of the CCLB.Based on a quantitative estimation,the strongest longwave warming effects(>0.3 W m-2)are found near the Equator and over tropical continents.The global average longwave warming effect of the CCLB is at least 0.07 W m-2,which is much larger than some of the radiative forcings considered in the Intergovernmental Panel on Climate Change reports.Specifically,the CCLB in traditional "clear-sky"region may be totally missed by current models,which contributes 28.25%(0.02 W m"2)of the whole CCLB radiative effect,twice greater than contrail effect.It is recommended that the CCLB should be taken account in future climate models.5)Uncertainties in satellite remote sensingThere are some uncertainties in the use and calculation of satellite remote sensing data.Here,we clarified the uncertainties from the aspects of partial-filling effect,radar sensitivity and gridding process.Based on the merged dataset of TRMM Visible and Infrared Scanner and PR observations,we found that clear-sky pixels and“cold”pixels probably exist in some apparent warm-rain cases(60.5%and 11.2%of the time,respectively).Statistical analysis shows that the existence of clear-sky pixels has a huge influence on the characteristics of the warm-rain pixels.The implications of this study are that many of the conventional warm-rain cases are in fact not warm rain.When studying warm rain,the situation whereby the edges of pixels are clear sky should be fully considered.The echo-top height observed by the TRMM PR has been used by some studies as an approximate calculation of the precipitating-cloud-top height to simulate radiative forcing or to identify overshooting convection.However,due to the low sensitivity(?17 dBZ)of PR,the PR-echo-top height is lower than the actual precipitating-cloud-top height.Here,the echo-top heights of the tropical precipitating cloud detected by three spaceborne radars were investigated to evaluate the underestimation of the PR-echo-top height to the actual precipitating-cloud-top height.The results show that there are significant spatial variations in the underestimates.The model simulation shows that these underestimates led to an underestimation of the radiative forcing of the Earth system,the relative error of which is?10%with 1-km underestimation.Therefore,the underestimates of precipitating-cloud-top height by PR should be taken into consideration when using PR-echo-top height.The gridding process applied to satellite-retrieved cloud properties results in the loss of certain information.Here,we analyzed the error associated with using gridded cloud optical depth(T)in calculating radiative forcing from the perspective of the distribution pattern of t.The greatest relative error occurs in the cases when ? fits a two-point or uniform distribution,reaching 10-20%,while this error is below 5%when ?follows a binomial distribution.Real cases show that ? within one grid point(1°×1°)could not be simply described by a normal distribution.Although using the logarithmic mean of ? controls the error effectively,the error can still be up to 4%.It suggests that using gridded data(especially the arithmetic mean)to calculate radiative forcing may result in uncertainty to a certain extent,which depends strongly on the distribution pattern.The probability distribution functions should be comprehensively considered in the gridding process.