| Clouds are the important part of the atmosphere,and they can act on the earth-atmosphere system by balancing the radiation budget,regulating the water cycle,and affecting the atmospheric circulation.Uncertainty of macroscopic cloud properties on global climate system models is a major obstacle in studying cloud effects on the Earth-atmosphere system,cloud fraction in the horizontal direction and cloud vertical structure in the vertical direction are important aspects to improve cloud parameterization schemes in climate models and to enhance their simulation capabilities.For cloud fraction,the cloud overlap properties such as combined cloud fraction are crucial for calculating radiative transfer;for cloud vertical structure,the cloud layer properties such as cloud top height are the key to affect the surface radiation effect.Currently,cloud observation can be obtained from ground-and satellite-based active or passive sensors.However,data products from various sources will inevitably vary,considering differences in their observation platforms,instrument performance,and retrieval methods.To reduce the uncertainty of cloud detection and improve the reliability of the research results,it is necessary to combine the observations of different instruments and make a validation comparison.Therefore,this paper will consider the following aspects: Utilizing the vertical cloud profile data provided by ground-based cloud radar to analyze the effects of seasonal variations and cloud type changes on cloud overlap properties;Combining satellite-based active sensors to quantify the main differences in cloud overlap properties and cloud vertical structure from different products;Considering the influence of cloud top thickness and cloud layer structure to compare the difference of cloud top height observed by ground-based cloud radar and satellite-based active/passive sensors.The research in this paper shows that:(1)Using the continuous vertical cloud data from KAZR,based on limitations of model and observation on vertical resolution as well as the effect of temporal resolution on cloud fraction and sample size,the resolution for cloud overlap studies is determined as 1hour and 360 m.On this basis,the total cloud fraction and cloud overlap properties varying with seasons are analyzed and found to have significant periodicity.Generally,the maximum and the minimum total cloud fraction appear in winter-spring and summer-autumn respectively.The seasonal variation of cloud overlap factor is basically synchronized with the total cloud fraction,and the seasonal variation of decorrelation distance lags behind by one quarter.By analyzing the cloud overlap properties of different cloud types,it is found that the combined cloud fraction bias of random overlap is the smallest for LMH,MH,and LM clouds;the combined cloud fraction bias of maximum overlap is the smallest for L and M clouds;for H clouds,both random and maximum overlap are minimized in combined cloud fraction error.(2)Together with observations from satellite-based active sensors(including Cloud Sat,CALIPSO,and combined Cloud Sat-CALIPSO products),the specific cases involved in the study are determined by counting the total cloud fraction and sample size from different data.On this basis,the cloud overlap properties of different data are comparatively analyzed,and the research shows that the random overlap enhances and then weakens with increasing cloud layer separation.Furthermore,for cloud layer separation below 1km and above 1km,the true overlapping cloud fraction is close to maximum and random overlap assumption,respectively.When the cloud layer separation corresponding to cloud overlap factor is negative,the combined cloud fraction bias of random overlap from different data is minimized.When considering the influence of different cloud types,random overlap is more effective in simulating cloud fraction for LMH and MH clouds with larger sample sizes.(3)Assessment of cloud top height reliability retrieved by different satellite instruments(i.e.,MODIS/Terra,MODIS/Aqua,MISR,HW8,Cloud Sat,CALIPSO)using ground-based observation data,shows that the total cloud fraction and total cloud occurrence from both ground and satellite observations are minimum differences optimally consistent over a space area of 0.5° × 0.5° and a time interval of40 minutes.When considering the effects of cloud top thickness,satellite instruments are more reliable in retrieving cloud top height for thick clouds,and vice versa for thin clouds.For CALIPSO,the retrieved cloud top height is slightly higher than that of KAZR due to the influence of low-level thin clouds;for MODIS/Terra,MODIS/Aqua,MISR,HW8,and Cloud Sat,the retrieved cloud top height is lower than that of KAZR,and the bias is mainly from high-level thin clouds.(4)Based on KAZR-detected single-layer profile ratio to distinguish non-overlapping and overlapping clouds,and analyze cloud top height difference between satellites and KAZR,respectively.The study shows that the cloud top height underestimation of satellite-based passive sensors is more significant in overlapping clouds,and the mean difference and correlation between their retrieved cloud top height and KAZR are more susceptible to cloud inhomogeneity,especially for MISR and HW8,which can be appropriately improved by controlling the increase of total cloud fraction.For satellite-based active sensors,the cloud top heights retrieved by Cloud Sat and CALIPSO are relatively consistent with KAZR,and are less affected by seasonal changes and cloud inhomogeneity.Furthermore,the cloud top height underestimation for Cloud Sat is greater in overlapping clouds,and the cloud top heights overestimation for CALIPSO is greater in non-overlapping clouds. |
PDF Full Text Request