| The cloud covers about 2/3 of the earth's surface.Cloud is one of the key regulators of Earth's radiation budget.Cloud characteristics and their temporal and spatial variations have great impact on global weather and climate change.The accurate determination of cloud phase not only contributes to the inversion of other cloud physical parameters,but also provides a basis for the research of weather forecast,climate model,Earth's radiation budget,atmospheric cycle and precipitation cycle.The accurate determination of cloud phase is conducive to the detection of supercooled water droplets for improving the safety of aircraft flight.In the field of weather forecast,cloud phase can improve the initial field of the model,and the results of the test and correction.It is helpful to determine the distribution and influence of cloud,precipitation,snow,typhoon and disastrous weather,so as to improve the understanding of weather and climate change and increase the accuracy of forecast.The existing cloud phase inversion of spaceborne single sensor has some limitations.The passive sensor has low resolution and the detecting depth is insufficient.The active sensor has the problem of insufficient profile detection.The combination of multi-sensors mainly focuses on horizontal or vertical single directional research.There is relatively little research on the synergy of cloud phase from the spatial perspective.The combination of satellite polarization,laser and microwave sensors can be used to overcome the above problems.The cloud phase spatial synergetic inversion method proposed in this thesis can break through the limitation of traditional single identification of the cloud phase by using the active and passive technologies.It provides new methods for atmospheric synergetic observation and inversion,and also provides new technology for weather forecast,climate change,artificial weather,extreme weather disaster prevention and flight safety flight.The main work of this thesis is as follows:1.Cloud detection and cloud phase identification is proposed for observation data of satellite polarization sensor.Using the theory of atmosphere-ocean radiation transmission,and combining the Fresnel reflection and the rough sea surface polarization radiation,the calculation formula of polarization radiation in the flare area is constructed.The dynamic detection model of the ocean glint based on the polarizing radiation data of the near infrared band is put forward.The model is used to dynamically obtain the boundary value of the glint angle,so as to realize the preprocessing of the ocean glint.For non-glint pixels,a fast detection of cloud pixels is realized by the improved cloud detecting method.The method uses the reflectance difference in the near infrared band of cloud and clear-sky pixel and the threshold of polarization reflectivity to identify the cloudy pixel,and further marks the clear-sky pixel by using the ratio of the reflectance of near infrared to visible light band.Cloud detection results are spatially fused to generate cloud,clear-sky and undetermined pixel products.The multi-angle polarization cloud phase recognition algorithm is constructed to distinguish ice and water cloud.The algorithm mainly uses the cloud particle polarization radiation to identify the ice/water cloud with the variation trend of the scattering angle in the main rainbow and the non-rainbow region.Compared with the POLDER3 operational product,the result of the method is better and the resolution is better.2.A vertical cloud phase fusion algorithm is constructed to achieve the phase synergy of laser and microwave sensor cloud profiles.The cloud phase data of the CALIOP lidar is stored in the VFM of level-2 product.The CloudSat profile cloud phase product(2B-CLDCLASS-LIDAR)is stored as the cloud layer stucture,and the vertical resolution of the two products is different.The vertical directional cloud phase coordination is realized by constructing the dynamic multi-target optimal rule.Taking the typhoon "Lupit" as the experimental object to verify the algorithm,the results of the cloud phase classification of two sensor's weighting coefficients from zero to one are calculated.It is found that when the weighting coefficient is 1/2,the statistics of the vertical cloudy pixel reach the maximum,and the cloud phase fusion structure reaches the optimal proportion.The results show that the vertical synergistic algorithm of laser and microwave cloud phase can effectively solve the defects of the cloud profile information detection of single sensor,enrich the particle radiation information in the cloud layer,and greatly improve the accuracy of the inversion of the vertical cloud phase.3.A multi-source data phase spatial synergistic algorithm is proposed based on satellite polarization,laser and microwave sensors.The algorithm aims at the difference of the storage structure and dimension of the three sensors,and unifies the differentiation structure and dimension by reducing dimension and ascending dimension transformation.We construct the multi-target fuzzy optimal rule to realize the horizontal fusion of three cloud phase products.Taking the typhoon "Lupit" as an example,the POLDER3 level-2 products obtained by two different methods are fused with the two cloud phase products of laser and microwave.The two different fusion results are obtained,and the analysis and discussion are carried out.The results of the two kinds of cloud top phase fusion are similar to those of CALIPSO and CloudSat,but the proportion of POLDER3 and MODIS is larger,and the standard deviations of ice clouds and water clouds are more than 10%.The reason is that POLDER3 is misjudged by the deviation of the rainbow characteristics of the cloud particles in the multilayer cloud and cloud boundary because of the multi-angle observation.The infrared radiation characteristics of MODIS in the thin cloud and the broken cloud have no significant difference in the light temperature discrimination of the cloud phase.Compared with the mean change trend of the two fusion cloud phase,the cloud top phase is close to the CALIPSO's cloud top phase,and close to the CloudSat in the center and bottom of cloud.The two fusion results combine the advantages of the laser in the cloud top detection and the microwave detection in the center and bottom,making it more accurate in the ice cloud and the water cloud classification.The division of the cloud top and bottom phase is more reasonable.The experimental results show that the fusion algorithm is very effective in improving the cloud phase classification,and can not be interfered by the error of a certain cloud phase result.It can effectively improve the efficiency and accuracy of the cloud phase classification.Spatial synergistic inversion is one of the main direction of future satellite remote sensing detection.It can not only overcome the shortcomings of single sensor inversion,but also overcome the single directional problem of multi-sensor's collaboration at present.It has opened up new research directions for detecting atmospheric components such as clouds and aerosols.This paper respectively coprocesses the horizontal and vertical directional data of satellite polarization,laser and microwave,and constructs the spatial synergistc inversion process and algorithm,which provides the method and technical support for the future satellite spatial detection and synergistc inversion in our country. |
PDF Full Text Request