atmosphere - MDPI

4 downloads 0 Views 5MB Size Report
May 2, 2018 - droplets, or cold ice clouds with low concentrations of small ice crystals. ...... Wang, Z.; Vane, D.; Stephens, G.; Reinke, D. Level 2 Combined ...
atmosphere Article

Clouds over East Asia Observed with Collocated CloudSat and CALIPSO Measurements: Occurrence and Macrophysical Properties Xuebin Li 1, *, Xianming Zheng 1 , Damao Zhang 2 , Wenzhong Zhang 1 , Feifei Wang 1 , Ye Deng 3 and Wenyue Zhu 1 1

2 3

*

Key Laboratory of Atmospheric Optics, Chinese Academy of Sciences, Hefei 230031, China; [email protected] (X.Z.); [email protected] (W.Z.); [email protected] (F.W.); [email protected] (W.Z.) Brookhaven National Laboratory, Upton, NY 11973, USA; [email protected] China Mobile Group Design Institute Co., Ltd., Beijing 100038, China; [email protected] Correspondence: [email protected]  

Received: 5 January 2018; Accepted: 21 April 2018; Published: 2 May 2018

Abstract: Cloud occurrences, vertical structures, and along-track horizontal scales over East Asia are studied using four years (2007–2010) of CloudSat 2B-CLDCLASS-LIDAR data. The CloudSat 2B-CLDCLASS-LIDAR data employs combined CloudSat radar and Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO) measurements to provide by far the most accurate detections of cloud boundaries and their vertical structures. The mean cloud occurrence frequency over East Asia is 66.3%, which is 13.8% and 21.6% higher than that from the Cloud–Aerosol LIdar with Orthogonal Polarization (CALIOP) level 2 5-km cloud layer product and the CloudSat 2B-GEOPROF product, respectively. Cloud-top heights over East Asia show three local peaks at approximately 1.5 km, 10 km, and 15 km above ground level (AGL), indicating different mid-altitude cloud formation mechanisms from those over the tropics. Significant fractions of low-level cloud, mid-level cloud, and high-level cloud have thicknesses smaller than 0.5 km, indicating that models with vertical resolutions lower than 0.5 km have difficulties resolving those clouds. The average cloud along-track horizontal scale over East Asia is 82.0 km. Probability distribution functions (PDFs) of cloud along-track horizontal scales suggest that approximately 81.2% of the clouds over East Asia cannot be resolved by climate models with a grid resolution of 1◦ . The results from this study can be used to improve cloud parameterizations in climate models and validate model simulations of clouds over East Asia. Keywords: cloud occurrence; cloud macrophysics; CALIPSO; CloudSat

1. Introduction Clouds cover more than 65% of Earth’s surface and play a critical role in Earth’s radiation budget [1,2]. They exert a cooling effect on Earth’s surface by reflecting solar radiation back to space and a warming effect by trapping infrared radiation that is emitted by Earth’s surface and low troposphere. The radiative properties of clouds are determined by their macrophysical and microphysical properties [3]. So far, parameterizations of cloud processes and cloud feedbacks still represent the largest uncertainty in general circulation models (GCMs) for projecting future climate change [4]. Accurate observations of clouds and their vertical distributions, as well as other macrophysical and microphysical properties under different dynamic and thermodynamic environments, are required to improve cloud parameterizations in GCMs and validate model outputs [5]. East Asia (15–55◦ N, 70–140◦ E) is a unique region that is characterized with high cloud occurrences at low, middle, and high altitudes [2,6,7]. Clouds have significant impacts on climate change and Atmosphere 2018, 9, 168; doi:10.3390/atmos9050168

www.mdpi.com/journal/atmosphere

Atmosphere 2018, 9, 168

2 of 16

hydrometeor cycling over this region [8–10]. The topography in East Asia is high in the west and flat in the east. The Tibetan Plateau on the northwest with an averaged elevation of over 4000 m above sea level exerts dramatic influences on the climate system and cloud distributions over East Asia [11–13]. In turn, cloud-radiation feedbacks may contribute substantially to the recently accelerated warming trend and snow melting over the Tibetan Plateau [14,15]. The eastern East Asia climate system and precipitation are greatly influenced by summer monsoon activities [16] Clouds and their vertical structures are critical components in a monsoon system [17,18]. However, cloud distributions and vertical structures during monsoon seasons are not well understood [9,19]. Motivated by the important role of clouds in the climate system and hydrological cycle, several studies have been dedicated to analyzing clouds and their vertical structures over East Asia using various satellite and ground-based remote sensing measurements. For example, Li et al. [20] examined cloud spatial distributions and their seasonal variations over East Asia using the International Satellite Cloud Climatology Project (ISCCP) data and surface observations. However, passive remote sensing measurements cannot provide cloud vertical structure information, and have difficulties in detecting multilayer clouds. Yin et al. [7,21] studied cloud vertical profiles and three-dimensional (3D) cloud structures over East Asia using CloudSat radar measurements and the GCM-Oriented Cloud-Aerosol light detection and ranging (Lidar) and Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO) Cloud Product (CALIPSO-GOCCP) data [22], respectively. However, CloudSat radar with a sensitivity of approximately −30 dBZ cannot detect thin cirrus clouds and thin liquid clouds [6,23]. CALIPSO lidar signals can be quickly attenuated by liquid layers, failed to detect cloud layers and their vertical structures below [24]. Therefore, using radar-only or lidar-only measurements as presented in Yin et al. [7,21] might miss a remarkable amount of clouds and their vertical structure information. To improve weather forecast and projection of future climate change over East Asia, reliable observations of cloud macrophysical and microphysical properties, as well as cloud vertical and horizontal extensions over the region, are still needed [7,19]. Up until now, the models still have difficulties in representing the vertical structures of clouds [25–27]. Recent multi-model evaluations with CALIPSO-GOCCP data [26] showed that CMIP5/CFMIP2 models have a ~10–20% underestimation of high cloud fractions and a ~15–25% underestimation of middle cloud fractions over East Asia region, while the multi-model mean bias in low cloud fraction is quite small (within ±5%) in this region. However, CALIPSO-GOCCP is a lidar-only product. The lidar-only product tends to underestimate the middle and especially low cloud fractions due to the strong attenuation by the optically thick clouds above (which will be shown in Section 3.1). Therefore, the true mode biases in the middle and low cloud fractions over East Asia region should be considerably larger than the numbers presented in Cesana and Waliser [26]. Previous studies have shown that combined radar and lidar measurements have the advantages in resolving the vertical structure of cloud layers, ranging from optically thin cirrus and boundary layer clouds to deep optically thick precipitating systems [28,29]. The combined radar and lidar method thus provides a powerful tool to derive a more complete view of cloud macrophysical and microphysical properties than radar-only or lidar-only measurements. In this study, we document cloud occurrences, cloud vertical structures, and along-track horizontal scales, and analyze their characteristics and seasonal variations over East Asia using four years (2007–2010) of collocated CloudSat radar and CALIPSO lidar measurements. The results from this study can be used to improve cloud parameterizations in climate models and validate model simulations of clouds over East Asia. The paper is organized as follows. A brief description of the CloudSat 2B-CLDCLASS-LIDAR product and ancillary data for providing cloud boundaries is given in Section 2. In Section 3, distributions and seasonal variations of cloud occurrences and their vertical and horizontal structures in terms of cloud thickness and along-track horizontal cloud scale are presented. Conclusions and a summary of the study are given in Section 4.

Atmosphere 2018, 9, 168

3 of 16

2. Datasets and Methodology The latest version (R04) of the CloudSat 2B-CLDCLASS-LIDAR product [30] (Wang et al., 2012) between 2007–2010 is employed to study the cloud distributions and other macrophysical properties over East Asia. The sample number over the study region is shown in Figure 1a. The 2B-CLDCLASS-LIDAR product combined measurements of the Cloud–Aerosol LIdar with Orthogonal Polarization (CALIOP) on CALIPSO and 94.05 GHz Cloud Profiling Radar (CPR) on CloudSat for cloud scenario classification. CALIOP is a near nadir-viewing two-wavelength lidar at 532 nm and 1064 nm, and linear polarization measurements at 532 nm [31]. Its vertical and horizontal resolutions are 30 m and 333 m below 8.2 km, and 60 m and 1000 m between 8.2–20.2 km, respectively. Due to its high vertical resolution, CALIOP is sensitive to most of the clouds and aerosols, and can detect their fine structure [24]. However, CALIOP is unable to penetrate optically thick clouds to detect anything beyond them due to strong attenuation at its short wavelength. CPR has a detection sensitivity of approximately −30 dBZ, with a vertical resolution of about 240 m between the surface and 30 km above ground level (AGL) and a horizontal footprint of 1.3 km cross-track and 1.7 km along-track. CPR is sensitive to large-sized particles such as ice and drizzle drops with relatively coarser resolution [32,33], and can penetrate optically thick clouds to detect multilayer cloud systems. However, its long wavelength limits its capability to detect water clouds with relatively small water droplets, or cold ice clouds with low concentrations of small ice crystals. Therefore, CALIOP and CPR measurements provide unique complementary capabilities for each other [32]. The 2B-CLDCLASS-LIDAR product firstly uses the radar cloud mask from CloudSat 2B-GEOPROF [33] and the CALOP cloud mask identified from the attenuated backscattering coefficient profile [34,35] as inputs to search for cloud clusters and generate cloud feature information, including cloud boundary height, temperature, thickness, fraction, precipitation, etc. CALIOP attenuated backscattering coefficients are averaged to the CPR horizontal resolution, but keep its native vertical resolutions for better cloud boundary and phase identifications. Temperature profiles from the European Center for Medium-Range Weather Forecast are interpolated to each CPR vertical bin, and are provided in the ECMWF-AUX product to determine cloud temperatures and assist in cloud classification [36]. Then, the CloudSat 2B-CLDCLASS-LIDAR product uses these cloud features, together with a fuzzy logical-based algorithm, to further classify clouds into different cloud types and determine cloud thermodynamic phases. Eight cloud types were classified, including cumulus (Cu), stratocumulus (Sc), stratus (St), altocumulus (Ac), altostratus (As), nimbostratus (Ns), cirrus (ci), and deep convective clouds (Dc). Furthermore, three cloud thermodynamic phases were determined: water, ice, and mixed phase. The 2B-CLDCLASS-LIDAR product can provide more accurate detections of cloud boundaries and their vertical structures over the global scale than lidar-only or radar-only measurements. Figure 1 presents an example of A-Train satellite measurements and cloud detections from different products on 31 May 2008. Various cloud types and multilayer clouds can be observed from CALIOP total attenuated backscattering (TAB) at 532 nm (Figure 1b) and CPR radar reflectivity (Ze , Figure 1c). The lidar-only cloud mask (red and green patches) from the CALIOP level 2 5-km cloud layer product [37] and radar-only cloud mask (blue and green patches) from the CloudSat 2B-GEOPROF product [38] are shown in Figure 1d. It can be seen that CPR is not able to detect most of thin Ci and non-precipitating liquid clouds, while CALIOP is not able to penetrate liquid-topped clouds that occur within latitudes between 20.5–30.0 N to provide cloud information below the layer. With combined CPR and CALIOP measurements, a complete view of the cloud can be derived, which is shown as the cloud mask from 2B-CLDCLASS-LIDAR in Figure 1e. Note that although the 2B-CLDCLASS-LIDAR is by far the most accurate satellite-based cloud product, it still has some limitations. For example, this product cannot detect some non-precipitation water clouds if there is optically thick cloud above, due to the detection limitations of radar and lidar measurements. Furthermore, the 2B-CLDCLASS-LIDAR keeps the CALIOP high vertical resolution measurements to derive the fine structure of the clouds, especially for optically thin clouds, which are

Atmosphere 2018, 9, 168 Atmosphere 2018, 9, x FOR PEER REVIEW

4 of 16 4 of 16

transparent for identification Lidar. However, thick clouds such as liquid, the which identification of the cloud as liquid, the of for theoptically cloud bottom mainly relies on CRP, could result in an bottom mainly relies on CRP, which could result in an underestimation of cloud thicknesses due to its underestimation of cloud thicknesses due to its low sensitivity for small-size particles. Furthermore, low sensitivity for small-size particles. Furthermore, the CALIOP sparse temporal and space sampling of CPR the sparse temporal and space sampling of CPR and measurements may also result in and CALIOP measurements may also result in biases our results. non-convective cloudsstrong have biases in our results. The non-convective clouds haveinsmall diurnalThe variations [39–41], while small diurnal variations [39–41], while strongclouds. diurnalDue variations are seen for convective clouds. Due to diurnal variations are seen for convective to its sun-synchronous orbit with only two its sun-synchronous only local two overpasses at about 0130 cannot and 1330 local time, A-Train satellites overpasses at about orbit 0130 with and 1330 time, A-Train satellites capture the strongest or most cannot capture the strongest or most frequent afternoon convective period [39]. However, the results frequent afternoon convective period [39]. However, the results from averaging both the day and from averaging both the daymeasurements and the night A-train satellite close to the mean Rainfall derived the night A-train satellite are close to themeasurements mean derivedare from The Tropical from The Tropical Rainfall Measuring full-dayresults samples [39]. Therefore, with Measuring Mission (TRMM) full-day Mission samples(TRMM) [39]. Therefore, with both day andresults night-time both day and night-time measurements were presented in this study. Although reliable observations measurements were presented in this study. Although reliable observations are currently lacking, are currently lacking, further investigations andproduct evaluations of this product arewill stillbe needed, and will further investigations and evaluations of this are still needed, and done in future be done in future study. study.

Figure 1. 1. An and cloud detections on on 31 May 2008. (a) Figure Anexample exampleofofA-Train A-Trainsatellite satellitemeasurements measurements and cloud detections 31 May 2008. Sample number of the four-year A-Train satellite measurements over the study region, overlaid with (a) Sample number of the four-year A-Train satellite measurements over the study region, overlaid the magenta line line indicating the the location of of thethe case; Orthogonal with the magenta indicating location case;(b) (b)Cloud–Aerosol Cloud–AerosolLIdar LIdar with with Orthogonal Radar Polarization (CALIOP) total attenuated backscattering (TAB) at 532 nm; (c) Cloud Profiling Radar (CPR) reflectivity reflectivity (Z (Zee); (d) clouds detected with CloudSat 2B-GEOPROF (blue) and CALIOP level 2 (CPR) 5-km cloud cloudlayer layerproduct product (red), green regions represent clouds both products; (e) 5-km (red), green regions represent clouds seen seen from from both products; and (e)and clouds clouds detected with CloudSat 2B-CLDCLASS-LIDAR detected with CloudSat 2B-CLDCLASS-LIDAR product.product.

3. Results Results and and Discussions Discussions 3. 3.1. Cloud Occurrences

profile amount amount over the total measurement Cloud occurrence is defined as the ratio of a cloudy profile profile amount in boxbox in this study. In a statistical point point of view, profile in aa1° 1◦(latitude) (latitude)× 1° × (longitude) 1◦ (longitude) in this study. In a statistical of cloud view, occurrence is equivalent to thetocommonly used used cloudcloud fraction in models. Cloud occurrence is a cloud occurrence is equivalent the commonly fraction in models. Cloud occurrence fundamental parameter through which clouds exertexert impacts on atmosphere radiation budgets. On is a fundamental parameter through which clouds impacts on atmosphere radiation budgets. average, cloud occurrence over East Asia is is66.3% On average, cloud occurrence over East Asia 66.3%with withaastandard standarddeviation deviation of of 12.6%, compared with mean mean cloud cloud occurrences occurrences of of 52.5% from the CALIOP level 2 5-km cloud layer product and 44.7% with from the the CloudSat CloudSat 2B-GEOPROF 2B-GEOPROF product. product. from the CloudSat CloudSat The cloud occurrence distributions over East Asia from four years of the 2B-CLDCLASS-LIDAR product productare aregiven giveninin Figure From Figure 2a, cloud occurrences 2B-CLDCLASS-LIDAR Figure 2a.2a. From Figure 2a, cloud occurrences largerlarger than thanoccur 80% downstream occur downstream of the Plateau Tibetan and Plateau overChina the East Sea. Large cloud 80% of the Tibetan over and the East Sea. China Large cloud occurrences occurrences downstream ofPlateau the Tibetan Plateau are consistent with Li and Gu which [12]’s study, which downstream of the Tibetan are consistent with Li and Gu [12]’s study, showed that showed that the dynamic and thermodynamic effects of the Tibetan Plateau provide favorable conditions for stratiform clouds. Figure 2b presents the differences of cloud occurrence frequencies

Atmosphere 2018, 9, 168

5 of 16

Atmosphere 2018, 9, x FOR PEER REVIEW

5 of 16

the dynamic and thermodynamic effects of the Tibetan Plateau provide favorable conditions for stratiform clouds. Figure 2b presents the differences of cloud occurrence frequencies between between the CloudSat 2B-CLDCLASS-LIDAR product and the CALIOP Level 2 5-km cloud layer the CloudSat 2B-CLDCLASS-LIDAR product and the CALIOP Level 2 5-km cloud layer product. product. In general, cloud occurrences from the CloudSat 2B-CLDCLASS-LIDAR product are In general, cloud occurrences from the CloudSat 2B-CLDCLASS-LIDAR product are approximately approximately 20% larger those that from the CALIOP Level 2 5-km cloud layer product along the 20% larger those that from the CALIOP Level 2 5-km cloud layer product along the south of Tibetan south of Tibetan Plateau and over the South and East China Sea area. This is caused by the Plateau and over the South and East China Sea area. This is caused by the occurrence of large, deep, occurrence of large, deep, convective mid-level and low-level clouds over this region. Deep convective mid-level and low-level clouds over this region. Deep convective and mid-level clouds convective and mid-level clouds attenuate the CALIOP signal dramatically, which prevents attenuate the CALIOP signal dramatically, which prevents CALIOP from detecting low-level clouds. CALIOP from detecting low-level clouds. Note that the performance of the CALIOP Level 2 5-km Note that the performance of the CALIOP Level 2 5-km cloud layer product is similar to that of cloud layer product is similar to that of the CALIPSO-GOCCP reported in Yin et al. [7], which the CALIPSO-GOCCP reported in Yin et al. [7], which determines the cloud with a CALIOP scattering determines the cloud with a CALIOP scattering ratio at the resolution of the GCM [22]. Figure 2c ratio at the resolution of the GCM [22]. Figure 2c shows the cloud occurrence differences between shows the cloud occurrence differences between the CloudSat 2B-CLDCLASS-LIDAR product and the CloudSat 2B-CLDCLASS-LIDAR product and the 2B-GEOPROF product. Differences larger than the 2B-GEOPROF product. Differences larger than 30% occur over most of the southern area of East 30% occur over most of the southern area of East Asia, which is caused by large cirrus cloud occurrences Asia, which is caused by large cirrus cloud occurrences that are out of the detection limit of the CPR that are out of the detection limit of the CPR over the region. over the region.

Figure 2. (a) 2. Cloud (a) occurrence Cloud occurrence frequencies from four-year CloudSat Figure frequencies from four-year (2007–2010) CloudSat(2007–2010) 2B-CLDCLASS-LIDAR 2B-CLDCLASS-LIDAR product; (b) differences of cloud occurrence frequencies between the product; (b) differences of cloud occurrence frequencies between the CloudSat 2B-CLDCLASS-LIDAR CloudSatand 2B-CLDCLASS-LIDAR and layer the CALIOP 2 5-km cloud layerthe product; (c) product the CALIOP Level 2product 5-km cloud product;Level (c) differences between CloudSat differences between the CloudSat 2B-CLDCLASS-LIDAR product and the 2B-GEOPROF product. 2B-CLDCLASS-LIDAR product and the 2B-GEOPROF product. CC in the figure represents combined CC inand theCALIOP figure represents combined CPR in and measurements employedproduct. in the CloudSat CPR measurements employed theCALIOP CloudSat 2B-CLDCLASS-LIDAR 2B-CLDCLASS-LIDAR product.

Figure 3 shows the zonally averaged vertical distribution of cloud occurrence over East Asia from Figure 3 shows the zonally averaged vertical distribution of cloud occurrence over East Asia the CloudSat 2B-CLDCLASS-LIDAR product and its differences from that of the CALIOP level 2 from the CloudSat 2B-CLDCLASS-LIDAR product and its differences from that of the CALIOP level 5-km cloud layer product and that of the CloudSat 2B-GEOPROF product. The height bin uses 2 5-km cloud layer product and that of the CloudSat 2B-GEOPROF product. The height bin uses a a CPR vertical resolution of 0.24 km, and the latitude bin size is 1◦ . From Figure 3a, we can see large CPR vertical resolution of 0.24 km, and the latitude bin size is 1°. From Figure 3a, we can see large cloud occurrences in altitudes between 10–18 km AGL at latitudes lower than 25 N, with maximum cloud occurrences in altitudes between 10–18 km AGL at latitudes lower than 25 N, with maximum values greater than 30% at approximately 15 km AGL. There are also significant amounts of cloud values greater than 30% at approximately 15 km AGL. There are also significant amounts of cloud below 2 km AGL at latitudes lower than 25 N. Within 25 N to 35 N, cloud occurrences decrease below 2 km AGL at latitudes lower than 25 N. Within 25 N to 35 N, cloud occurrences decrease gradually from 30% near the surface to lower than 10% at about 11 km AGL. Within 45 N and 50 N, gradually from 30% near the surface to lower than 10% at about 11 km AGL. Within 45 N and 50 N, clouds mainly occur at altitudes between 4–10 km AGL, with an average occurrence of approximately clouds mainly occur at altitudes between 4–10 km AGL, with an average occurrence of 15%. At latitudes higher than 50 N, cloud occurrences decrease slightly, with heights below 11 km approximately 15%. At latitudes higher than 50 N, cloud occurrences decrease slightly, with heights AGL. From Figure 3b, the CALIOP level 2 5-km cloud layer product significantly underestimates below 11 km AGL. From Figure 3b, the CALIOP level 2 5-km cloud layer product significantly cloud occurrences below 11 km AGL at latitudes lower than 25 N, because deep convective clouds underestimates cloud occurrences below 11 km AGL at latitudes lower than 25 N, because deep strongly attenuate CALIOP signals. At latitudes higher than 25 N, significant underestimations of convective clouds strongly attenuate CALIOP signals. At latitudes higher than 25 N, significant cloud occurrences from the CALIOP level 2 5-km cloud layer product mainly occur at altitudes underestimations of cloud occurrences from the CALIOP level 2 5-km cloud layer product mainly lower than 5 km AGL. Figure 3c shows that compared with the CloudSat 2B-CLDCLASS-LIDAR occur at altitudes lower than 5 km AGL. Figure 3c shows that compared with the CloudSat product, the 2B-GEOPROF product greatly underestimates high-level clouds at latitudes lower than 2B-CLDCLASS-LIDAR product, the 2B-GEOPROF product greatly underestimates high-level clouds 25 N and mid-altitude clouds at latitudes higher than 40 N. The 2B-GEOPROF product also slightly at latitudes lower than 25 N and mid-altitude clouds at latitudes higher than 40 N. The underestimates low-level non-precipitating clouds, as shown in Figure 3c. 2B-GEOPROF product also slightly underestimates low-level non-precipitating clouds, as shown in Figure 3c.

Atmosphere 2018, 9, 168 Atmosphere2018, 2018,9,9,xxFOR FORPEER PEERREVIEW REVIEW Atmosphere

6 of 16 66ofof16 16

.

Figure Zonally averaged averagedvertical verticaldistributions distributions cloud occurrences Figure3.3. (a) (a) Zonally of of cloud occurrences overover East East Asia Asia from from the the CloudSat 2B-CLDCLASS-LIDAR product; differencesofofcloud cloudoccurrence occurrencefrequencies frequenciesbetween between CloudSat 2B-CLDCLASS-LIDAR product; (b)(b) differences theCloudSat CloudSat2B-CLDCLASS-LIDAR 2B-CLDCLASS-LIDAR product Level 2 5-km cloud layerlayer product; (c). the productand andthe theCALIOP CALIOP Level 2 5-km cloud product; the CloudSat 2B-CLDCLASS-LIDAR andand thethe 2B-GEOPROF product. (c)differences differences between the CloudSat 2B-CLDCLASS-LIDAR product 2B-GEOPROF product. Figure 3. between (a) Zonally averaged vertical distributions of product cloud occurrences over East Asia from the CloudSat 2B-CLDCLASS-LIDAR product; (b) differences of cloud occurrence frequencies between CPR and CALIOP measurements also provide high-resolution vertically resolved detections of CPR and CALIOP measurements also also provide high-resolution vertically resolved detections ofof CPR and CALIOP measurements provide high-resolution resolved detections the CloudSat 2B-CLDCLASS-LIDAR product and the Level 2vertically 5-km cloud layer and product; (c) cloud structures and their seasonal variations. Figure 4CALIOP shows the seasonal variations anomalies cloudstructures structures and theirthe seasonal variations. Figure 44 shows shows the seasonal variations and differences between CloudSat 2B-CLDCLASS-LIDAR product and the 2B-GEOPROF cloud and their seasonal variations. theover seasonal variations andanomalies anomalies of zonally averaged vertical distributions of Figure cloud occurrence East Asia fromproduct. the CloudSat

of zonally averaged vertical distributions of cloud occurrence over East Asia from the CloudSat of zonally averaged vertical distributions cloudwe occurrence over seasonal East Asiavariations from theofCloudSat 2B-CLDCLASS-LIDAR product. From theoffigure, can see large cloud 2B-CLDCLASS-LIDAR product. From the figure, we can see large seasonal variations of cloud 2B-CLDCLASS-LIDAR From thelatitudes figure, lower we can see25large seasonal variations of cloud occurrence at mid andproduct. high altitudes over than N, with maximum values during occurrence at mid and high altitudes over latitudes lower than 25 N, with maximum values during June, July, (JJA) and minimum values during December, January, and February occurrence atand midAugust and high altitudes over latitudes lower than 25 N, with maximum values(DJF). during June, July, and August (JJA) and minimum values during December, January, and February (DJF). This feature is directly caused more and stronger convections during summer, and less and June, July, and August (JJA) and by minimum values during December, January, and February (DJF). This feature is directly caused by more and stronger convections during summer, and less and weaker convections during winter. For latitudes higher than 25 N, seasonal variations of cloud This feature is directly caused by more and stronger convections during summer, and less and weaker weaker convections during winter. For latitudes higher than 25 N, seasonal variations of cloud occurrenceduring mainlywinter. occur atFor midlatitudes and low higher altitudes, with values duringof March, and convections than 25 maximum N, seasonal variations cloudApril, occurrence occurrence mainly occur at mid and low altitudes, with maximum values during March, April, and May occur (MAM) and and minimum values with during September, October, November (SON). mainly at mid low altitudes, maximum values duringand March, April, and May These (MAM) May (MAM) and minimum values during September, October, and November (SON). These be more clearly seen in October, the seasonal figures (asThese shownsignatures in the lower andsignatures minimumcan values during September, andanomaly November (SON). canpanel). be more signatures can be more clearly seen altitudes in the seasonal anomaly figures (as shown in the lower panel). Cloud occurrences at mid and high at latitudes lower than 25 N have seasonal variations clearly seen in the seasonal anomaly figures (as shown in the lower panel). Cloud occurrences at as mid Cloud at mid40%, and while high altitudes at latitudesatlower thanlow 25 Naltitudes have seasonal variations as large occurrences as approximately cloud occurrences mid and at latitudes higher and high altitudes at latitudes lower than 25 N have seasonal variations as large as approximately 40%, large 40%, while cloudofoccurrences at Interestingly, mid and low clear altitudes at latitudes higher than as 25 approximately N only have seasonal variations less than 10%. seasonal variations of a while cloud occurrences at mid and low of altitudes at10%. latitudes higher than 25 N onlyvariations have seasonal than 25 N only have seasonal variations less than Interestingly, clear seasonal of a narrow mid-altitude cloud band at altitudes between 8–12 km AGL at latitudes higher than 35 N can variations of less thancloud 10%. band Interestingly, clear seasonal variations oflatitudes a narrow mid-altitude cloud narrow mid-altitude at altitudes between 8–12 km AGL at higher than 35 N can be seen in the seasonal anomaly figures. This narrow mid-altitude cloud band has maximum band at altitudes between anomaly 8–12 km figures. AGL at latitudes higher than 35 Ncloud can beband seen has in the seasonal beoccurrences seen in the seasonal This narrow mid-altitude maximum during MAM and JJA, and minimum occurrences during DJF. The seasonal variability anomaly figures. This narrow cloudoccurrences band has maximum occurrences during MAM occurrences during MAM and mid-altitude JJA, minimum during DJF. The seasonal variability of this mid-altitude cloud band mayand relate to the seasonal variation of the sub-tropic westerly, which and JJA, and minimum occurrences during DJF. The seasonal variability of this mid-altitude cloud ofplays this mid-altitude may relate to thevapor seasonal of China the sub-tropic which a dominant cloud role inband transporting water intovariation northwest [42–45]. westerly, The formation band may relate the variation of the sub-tropic westerly, which plays aThe dominant role plays a dominant roleseasonal in transporting vapor into northwest China formation mechanism andtoseason variation of thiswater narrow mid-altitude cloud band will[42–45]. be further explored in inmechanism transporting water vapor into northwest China [42–45]. The formation mechanism and season and season variation of this narrow mid-altitude cloud band will be further explored in future study. variation of this narrow mid-altitude cloud band will be further explored in future study. future study.

Figure 4. Seasonal variations (upper panel) and anomalies (lower panel) of zonally averaged vertical distributions of cloud occurrences over Eastand Asiaanomalies from the CloudSat 2B-CLDCLASS-LIDAR product. Figure4.4. Seasonal (upper panel) (lower(lower panel) of zonally vertical Figure Seasonalvariations variations (upper panel) and anomalies panel) of averaged zonally averaged MAM represents March–April–May, JJA represents June–July–August, SON represents September– distributions of cloudofoccurrences over East AsiaEast from the from CloudSat 2B-CLDCLASS-LIDAR product. vertical distributions cloud occurrences over Asia the CloudSat 2B-CLDCLASS-LIDAR October–November, and DJF represents MAM represents March–April–May, JJA December–January–February. represents SON represents September– product. MAM represents March–April–May, JJA June–July–August, represents June–July–August, SON represents October–November, and DJF represents December–January–February. September–October–November, and DJF represents December–January–February.

Atmosphere 2018, 9, 168

7 of 16

Figure 5 2018, shows the PEER probability Atmosphere 9, x FOR REVIEW distribution functions (PDF) of cloud top and base heights, 7 of 16 top temperature, and their seasonal variations over East Asia. Johnson et al. [46] showed trimodal cloud Figure shows the probability (PDF)the of cloud top and baseFigure heights, distributions at 5approximately 2 km,distribution 5 km, andfunctions 15 km over tropics. From 5a,topcloud temperature, and their seasonal variations over East Asia. Johnson et al. [46] showed trimodal cloud occurrence over East Asia has similar low-altitude (1.5 km) and high-altitude (15 km) peaks all of distributions at approximately 2 km, 5 km, and 15 km over the tropics. From Figure 5a, cloud the time, and a very small peak around 5–6 km in JJA and MAM. Unlike over the tropics, there is occurrence over East Asia has similar low-altitude (1.5 km) and high-altitude (15 km) peaks all of the another high-altitude peak at 10 km over East Asia all of the time. The two high-altitude peaks at 15 km time, and a very small peak around 5–6 km in JJA and MAM. Unlike over the tropics, there is and 10 km from the high-altitude (mainly cirrus), consistent with cloud anotherresult high-altitude peak at 10 km overclouds East Asia all of the time. which The twoishigh-altitude peaksthe at 15 vertical shown in the Figure 4. The cirrus has(mainly quite different behaviors in cloud-top height kmdistributions and 10 km result from high-altitude clouds cirrus), which is consistent with the over the sub-tropic and mid-latitude (~10–11 km), and tropics (~15 km) The peak cloud vertical distributions shownregions in Figure 4. The cirrus has over quite the different behaviors in [43]. cloud-top height the sub-tropic and seasonal mid-latitude regionsthan (~10–11 andpeak, over the tropics (~15 at 10 km hasover a relatively smaller variation thekm), 15-km which may bekm) the [43]. result of The peak at 10 km has a relatively smaller seasonal variation than the 15-km peak, which may be thefrom different formations of Ci over different regions. Over mid-latitudes, the cirrus tends to form result of different formations of Ci over different regions. Over mid-latitudes, the cirrus tends to cloud top ice nucleating zones, where ice supersaturations are relatively high, from the homogeneous form from cloud top ice nucleating zones, where ice supersaturations are relatively high, from the nucleation of haze particles [47]. However, over the tropics, cirrus is mainly formed by strong updrafts homogeneous nucleation of haze particles [47]. However, over the tropics, cirrus is mainly formed in deep convective clouds, bringing water vapor and cloud nuclei from the near surface upward until by strong updrafts in deep convective clouds, bringing water vapor and cloud nuclei from the near it is blocked by the top of the tropopause (~15 km). Therefore, the 15-km peak is largest in JJA and surface upward until it is blocked by the top of the tropopause (~15 km). Therefore, the 15-km peak smallest in DJF,incorresponding to in theDJF, deep convection activates in these two seasons. cloud is largest JJA and smallest corresponding to the deep convection activatesAs in for these two base height, PDFs decrease quickly until 2 km AGL, and then start to decrease slowly with altitude. Similar seasons. As for cloud base height, PDFs decrease quickly until 2 km AGL, and then start to decrease to Figure 5a, JJA has the smallest PDFs at low altitudes and largest values at high altitudes, and slowly with altitude. Similar to Figure 5a, JJA has the smallest PDFs at low altitudes and largest vice at high vicetop versa for DJF. The have PDFs of cloud top temperature large values versavalues for DJF. Thealtitudes, PDFs ofand cloud temperature large values and smallhave seasonal variations seasonal range variations within the◦ temperature between −30 °C and 10 are °C. two At low withinand thesmall temperature between −30 C and 10 ◦ C.range At low temperatures, there distinct ◦ ◦ temperatures, there are two distinct local PDF peaks at approximately −60 ° C and −85 ° local PDF peaks at approximately −60 C and −85 C, corresponding to the 10-km and 15-kmC,peaks corresponding to the 10-km and 15-km peaks in cloud-top height, respectively. The different cloud in cloud-top height, respectively. The different cloud top temperature indicates different ice growth top temperature indicates different ice growth habits and rates of Ci between the tropics and habits and rates of Ci between the tropics and mid-latitudes. mid-latitudes.

Figure 5. Probability distribution functions (PDF) of (a) cloud-top height, (b) base height, and (c) top

Figure 5. Probability distribution of (a)and cloud-top height, base height, and (c) top temperature and their seasonal functions variations.(PDF) Cloud-top base heights have(b) a bin size of 1 km, and temperature and their seasonal variations. Cloud-top and base heights have a bin size of 1 km, and their their top temperature has a bin size of 6 °C. top temperature has a bin size of 6 ◦ C. 3.2. Cloud Distributions at Different Altitude Levels

3.2. CloudTo Distributions at Different Altitude Levels better understand the dynamic and thermodynamic environments associated with cloud distributions, the cloudsthe are dynamic classified into (cloud base ≥ 6 km AGL),associated mid-level (6 km cloud > To better understand andhigh-level thermodynamic environments with cloud base ≥ 2 km AGL), and low-level (cloud base < 2 km AGL and cloud thickness < 6 km) clouds distributions, the clouds are classified into high-level (cloud base ≥ 6 km AGL), mid-level (6 km and vertically extended deep cloud systems (cloud base < 2 km AGL and cloud thickness ≥ 6 km), > cloud base ≥ 2 km AGL), and low-level (cloud base < 2 km AGL and cloud thickness < 6 km) following Adhikari et al. [48]’s criteria. The adopted thresholds of cloud base height are commonly clouds and extended deep cloud systems (cloud < 2 Office km AGL and cloud thickness used in vertically the literature and meteorological departments (such asbase the Met and National Weather ≥ 6 km), following Adhikari et al. [48]’s criteria. The adopted thresholds of cloud base height Service). The threshold of 6-km cloud thickness to distinguish low clouds from deep cloud systems are commonly in the literature meteorological departments as thestratiform Met Office and National is alsoused chosen based on the and consideration that the low clouds (such are mainly clouds and Weather Service). The threshold of 6-km cloud thickness to distinguish low clouds from deep cloud systems is also chosen based on the consideration that the low clouds are mainly stratiform clouds and

Atmosphere 2018, 9, 168 Atmosphere 2018, 9, x FOR PEER REVIEW

8 of 16 8 of 16

stratocumulus clouds, whose cloud thicknesses are mostly smaller than 6. Figure 6 shows distributions stratocumulus whose cloud thicknesses are mostly smaller than 6. East Figure 6 shows of cloud occurrenceclouds, for low-level, mid-level, high-level, and deep clouds over Asia from four distributions of cloud occurrence for low-level, mid-level, high-level, and deep clouds over East years of the CloudSat 2B-CLDCLASS-LIDAR product. From the figure, low-level cloudsAsia mainly from four years of the CloudSat 2B-CLDCLASS-LIDAR product. From the figure, low-level clouds occur in the southwest of China, with occurrences larger than 60%, and over the East and South mainly occur in the southwest of China, with occurrences larger than 60%, and over the East and China Sea, with occurrences of approximately 50%. Mid-level clouds are widely distributed over East South China Sea, with occurrences of approximately 50%. Mid-level clouds are widely distributed Asia, over and have of approximately 30%, with larger occurrences of approximately East occurrences Asia, and have occurrences of approximately 30%, with larger occurrences 40% of in the middle of China. High-level clouds are concentrated within southern East Asia, with occurrences approximately 40% in the middle of China. High-level clouds are concentrated within southern East generally than 70%. Deep clouds are mainly located the Bay Bengallocated and the China Asia,larger with occurrences generally larger than 70%. Deepatclouds areofmainly at East the Bay of Sea, Bengal and theofEast China Sea, with occurrences approximately 10%.bar Please note that the bar6d is with occurrences approximately 10%. Please of note that the color magnitude in color Figure magnitude Figure is smaller smaller than thatinfor the 6d other three. than that for the other three.

6. Distributions of cloud occurrence for (a) low-level, (b) mid-level, (c) high-level, and (d) FigureFigure 6. Distributions of cloud occurrence for (a) low-level, (b) mid-level, (c) high-level, and (d) deep deep clouds. Please note the color bar magnitude in (d) is smaller than that for the other three. clouds. Please note the color bar magnitude in (d) is smaller than that for the other three.

Seasonal variations and anomalies of cloud occurrence distributions for low-level, mid-level,

Seasonal and anomalies of in cloud occurrence distributions for low-level, mid-level, high-level,variations and deep clouds are presented Figures 7 and 8, respectively. From Figure 7 (first panel), high-level, and deep clouds are presented in Figures 7 and 8, respectively. From Figure 7 (first panel), the largest occurrences of low-level clouds are in the southwest of China during JJA, with occurrences larger than 70%, andclouds over the andsouthwest South China during DJF,JJA, withwith occurrences the largest occurrences of low-level areEast in the of Sea China during occurrences than 60%. Figure panel),China these two seasonal variations of 60%. largerlarger than 70%, and From over the East8 (first and South Sea regions duringalso DJF,have withlarge occurrences larger than low-level cloud occurrences, of up to 40%. As for mid-level clouds, the largest cloud occurrences From Figure 8 (first panel), these two regions also have large seasonal variations of low-levelof cloud 50% occur over the northwest of China during MAM. From Figure 8 (second row), the northwest of occurrences, of up to 40%. As for mid-level clouds, the largest cloud occurrences of 50% occur over China also has strong seasonal variations of mid-level cloud occurrences, with positive anomalies the northwest of China during MAM. From Figure 8 (second row), the northwest of China also has during MAM and negative anomalies during SON. In addition, southwestern East Asia has strong strongseasonal seasonal variations of mid-level occurrences, with positive anomalies during MAM and variations of mid-level cloudcloud occurrences, with positive anomalies during JJA and negative negative anomalies during SON. In addition, southwestern East Asia has strong seasonal variations anomalies during DJF. For high-level clouds, as shown in Figure 7 (third row), cloud occurrences of mid-level cloud occurrences, with positive anomalies during JJA and negative anomalies during during JJA not only have higher magnitudes of 90%, this also extends to higher latitudes. The DJF. For high-level clouds, shown in Figure (third with row),large cloud occurrences during signatures can be clearlyas seen in Figure 8 (third7panel), positive anomalies duringJJA JJAnot andonly negativemagnitudes anomalies during DJF. From 8 (third panel), there is another band of can high-level have higher of 90%, this alsoFigure extends to higher latitudes. The signatures be clearly clouds over northwestern East large Asia positive that has anomalies strong seasonal variations positive anomalies seen in Figure 8 (third panel), with during JJA andwith negative anomalies during during MAM and negative anomalies during JJA and SON. From Figure 7 (fourth row), similar to East DJF. From Figure 8 (third panel), there is another band of high-level clouds over northwestern high-level cloud distributions, deep clouds have their largest occurrences over southern East Asia Asia that has strong seasonal variations with positive anomalies during MAM and negative anomalies during JJA, and their smallest occurrences during DJF. This region also has the strongest seasonal during JJA and SON. From Figure 7 (fourth row), similar to high-level cloud distributions, deep clouds variations, with positive anomalies during JJA and negative anomalies during DJF. have their largest occurrences over southern East Asia during JJA, and their smallest occurrences during DJF. This region also has the strongest seasonal variations, with positive anomalies during JJA and negative anomalies during DJF.

Atmosphere 2018, 9, 168

9 of 16

Atmosphere 2018, 9, x FOR PEER REVIEW Atmosphere 2018, 9, x FOR PEER REVIEW

9 of 16 9 of 16

Figure 7.7. Seasonal Seasonal variations variations of of cloud cloud occurrence occurrence distributions distributions for for low-level, low-level, mid-level, mid-level, high-level, high-level, Figure Figure 7. Seasonal variations of cloud occurrence distributions for low-level, mid-level, high-level, and deep clouds. and deep clouds. and deep clouds.

Figure 8. Similar to Figure 7, but for the seasonal anomalies. Figure 8. Similar Similar to to Figure Figure 7, 7, but but for for the the seasonal seasonal anomalies. anomalies. Figure 8.

3.3. Discussions about Cloud Vertical and Horizontal Scales 3.3. Discussions Discussions about about Cloud Cloud Vertical Vertical and and Horizontal Horizontal Scales Scales 3.3. The radiative impacts of clouds are directly affected by their horizontal coverage and vertical The radiative radiative impacts of clouds are directly directly affectedprovide by their their horizontal coverage and vertical vertical extensions. Collocated CPRof and CALIOP measurements reliable detections of cloud systems The impacts clouds are affected by horizontal coverage and extensions. Collocated CPR and CALIOP measurements provide reliable detections of cloud systems and their along-track scales. measurements provide reliable detections of cloud systems extensions. Collocatedhorizontal CPR and CALIOP and their along-track horizontal scales. Figure 9 shows the time series of monthly averaged cloud-top height and cloud thickness for and their along-track horizontal scales. Figuremid-level, shows the the time series series of monthly averaged cloud-top height and and mid-level cloud thickness thickness for low-level, high-level, andof deep clouds. From Figure 9, low-level cloud-top Figure 99 shows time monthly averaged cloud-top height cloud for low-level, mid-level, high-level, and monthly deep clouds. clouds. From Figure Figure 9, low-level low-level and mid-level mid-level cloud-top heights and thicknesses have little variations. For high-level clouds, cloud-topcloud-top height is low-level, mid-level, high-level, and deep From 9, and heights and thicknesses have little monthly variations. For high-level clouds, cloud-top height is about 2and km thicknesses larger during JJA, butmonthly cloud thicknesses little monthly variations. For deep heights have little variations.have For high-level clouds, cloud-top height isclouds, about about 2 km larger during JJA, but cloud thicknesses have little monthly variations. For deep clouds, height are about have 2 km little larger during JJA. 2both km cloud-top larger during JJA,and butthickness cloud thicknesses monthly variations. For deep clouds, both both cloud-top height and thickness are about 2 km larger during JJA. cloud-top height and thickness are about 2 km larger during JJA.

Atmosphere 2018, 9, 168 Atmosphere 2018, 9, x FOR PEER REVIEW Atmosphere 2018, 9, x FOR PEER REVIEW

10 of 16 10 of 16 10 of 16

Figure 9. Time series of monthly averaged cloud-top height and cloud thickness for low-level, Figure 9. Time series of monthly averaged cloud-top heightheight and cloud for low-level, mid-level, Figure 9. Time series of monthly averaged cloud-top and thickness cloud thickness for low-level, mid-level, high-level, and deep clouds. The vertical bars represent the standard deviations of the high-level, deep clouds. Theclouds. vertical represent the standard deviations of the monthly mid-level,and high-level, and deep Thebars vertical bars represent the standard deviations of the monthly mean values. mean values. monthly mean values.

To take a closer look at cloud vertical distributions, Figure 10 shows the PDFs of cloud thickness To take a closer lookatatcloud cloudvertical verticaldistributions, distributions, Figure 10 the cloud thickness To take a closer look Figure 10shows shows thePDFs PDFsofof cloud thickness for low-level, mid-level, high-level, and deep clouds. Their mean cloud-top heights (km AGL) and for low-level, mid-level, high-level, and deep clouds. Their mean cloud-top heights (km AGL) and forcloud low-level, mid-level, high-level, and deep clouds. Their mean cloud-top heights (km AGL) and thicknesses are given in Table 1. From the figure, the PDFs of low-level, mid-level, and cloud thicknesses are given in Table 1. From the figure, the PDFs of low-level, mid-level, and cloud thicknesses given in Table figure, of low-level, mid-level, and decrease high-level high-level cloudarethickness have 1. a From peak the value that the is PDFs smaller than 0.5 km, and then high-level cloud thickness have a peak value that is smaller than 0.5 km, and then decrease cloud thickness have a peak value that smaller 0.5 km, and then than decrease gradually with cloud gradually with cloud thickness, withisthe meanthan thicknesses smaller ~2 km in Table 1. This gradually with cloud thickness, with the mean thicknesses smaller than ~2 km in Table 1. This thickness, with the mean thicknesses smaller than ~2 km in Table 1. This indicates that models with indicates that models with vertical resolutions lower than 0.5 km have difficulties resolving those indicates that models with vertical resolutions lower than 0.5 km have difficulties resolving those vertical resolutions thancloud 0.5 km have difficulties The PDFsItofshould deep cloud clouds. The PDFslower of deep thickness generallyresolving decreasethose with clouds. cloud thickness. be clouds. The PDFs of deep cloud thickness generally decrease with cloud thickness. It should be noted that due to decrease the strongwith attenuation of the lidar signal by layers and sensitivity of thickness generally cloud thickness. It should be liquid noted that due to the the low strong attenuation noted that due to the strong attenuation of the lidar signal by liquid layers and the low sensitivity of CPR signal and attenuation of the radar signal by strong convective clouds, the ofCloudSat of the the lidar by liquid layers and the low sensitivity of the CPR and attenuation the radar the CPR and attenuation of the radar signal by strong convective clouds, the CloudSat 2B-CLDCLASS-LIDAR product may underestimate cloud thicknesses for non-precipitating liquid signal by strong convective clouds, the underestimate CloudSat 2B-CLDCLASS-LIDAR product may underestimate 2B-CLDCLASS-LIDAR product may cloud thicknesses for non-precipitating liquid clouds and deepfor convective clouds. liquid clouds and deep convective clouds. cloud thicknesses non-precipitating clouds and deep convective clouds.

Figure for low-level, low-level, mid-level, mid-level, high-level, high-level, and anddeep deepclouds. clouds. Figure 10. 10. PDFs PDFs of of cloud cloud thickness thickness for Figure 10. PDFs of cloud thickness for low-level, mid-level, high-level, and deep clouds.

Atmosphere 2018, 9, 168

11 of 16

Table 1. Mean cloud-top height (km above ground level, or AGL) and cloud thickness for low-level, 11 of 16 mid-level, high-level, and deep clouds. The standard deviation of the mean is given in parenthesis.

Atmosphere 2018, 9, x FOR PEER REVIEW

Table 1. Mean cloud-top height (km above ground level, or AGL) and cloud thickness for low-level, Cloud Class Cloud-Top Height deviation (km) Cloud Thickness mid-level, high-level, and deep clouds. The standard of the mean is given (km) in parenthesis.

Low-level Cloud Class Mid-level Low-level High-level Mid-level Deep High-level Deep

2.4Height (1.6) (km) Cloud-Top 5.8 (2.4) 2.4 (1.6) 11.8 (3.0) 5.8 (2.4) 10.0 (2.4) 11.8 (3.0) 10.0 (2.4)

1.4 (1.5) (km) Cloud Thickness 1.9 (2.0) 1.4 (1.5) 1.5 (1.2) 1.9 (2.0) 9.0 (2.4) 1.5 (1.2) 9.0 (2.4)

For cloud horizontal scales, a cloud system is first identified using the CloudSat For cloud horizontal a cloud is first identified using thescale CloudSat 2B-CLDCLASS-LIDAR product scales, following Zhangsystem et al. [49]’s method. The horizontal of a cloud 2B-CLDCLASS-LIDAR product following Zhang et al. [49]’s method. The horizontal scale of a cloud system along the CALIPSO/CloudSat track is then determined by calculating the continuous profiles system along the CALIPSO/CloudSat track is then determined by calculating the continuous profiles (N) of a detected cloud system. Considering the 1.7-km along-track resolution of CPR measurements, (N) of a detected the 1.7-km along-track resolution of CPR the along-track scalecloud (L in system. km) of Considering a cloud system is L = N × 1.7. The distribution of measurements, cloud along-track the along-track scale (L in km) of a cloud system is L = N × 1.7. The distribution of cloud along-track horizontal scales over East Asia is shown in Figure 11. The averaged cloud along-track horizontal horizontal scales over East Asia is shown in Figure 11. The averaged cloud along-track horizontal scale over East Asia is 82.0 km, with a standard deviation of 142.0 km. From the figure, the largest scale over East Asia is 82.0 km, with a standard deviation of 142.0 km. From the figure, the largest cloud along-track horizontal scales are located in southwestern East Asia with values larger than cloud along-track horizontal scales are located in southwestern East Asia with values larger than 120 120 km, cloudalong-track along-trackhorizontal horizontal scales at the middle of western km, while while the the smallest smallest cloud scales areare at the middle of western East East Asia,Asia, with with values smaller than 50 km. values smaller than 50 km.

Figure 11. Distributions of cloud along-track horizontal scales over East Asia.

Figure 11. Distributions of cloud along-track horizontal scales over East Asia.

Figure 12 shows the PDFs and cumulative PDFs of along-track horizontal scales for total, low-level, deep clouds. a typical grid resolution of 1° in climate Figure 12 mid-level, shows thehigh-level, PDFs and and cumulative PDFsAssuming of along-track horizontal scales for total, low-level, models,high-level, which is about km (blue dashed line in approximately the clouds mid-level, and 110 deep clouds. Assuming a Figure typical12), grid resolution of81.2% 1◦ inof climate models, over East Asia have horizontal scales smaller than the model grid resolution on a global average, and which is about 110 km (blue dashed line in Figure 12), approximately 81.2% of the clouds over East be resolvedscales by climate models. From Figure the PDFs on decrease dramatically as the Asia cannot have horizontal smaller than the model grid12,resolution a global average, and cannot along-track horizontal scale increases for all of the cloud types. The mean along-track horizontal be resolved by climate models. From Figure 12, the PDFs decrease dramatically as the along-track scales for low-level, mid-level, high-level, and deep clouds are 56 km, 88 km, 110 km, and 99 km, horizontal scale increases for all of the cloud types. The mean along-track horizontal scales for low-level, respectively. Comparing the along-track horizontal scale for different cloud types, low-level clouds mid-level, high-level, and deep clouds are scales, 56 km,while 88 km, 110 km,clouds and 99have km,the respectively. Comparing have the smallest along-track horizontal high-level largest along-track the along-track horizontal types,oflow-level clouds have the smallest along-track horizontal scales. Notescale that for thedifferent statisticalcloud behaviors along-track horizontal scale for different horizontal scales, while high-level clouds have the largest along-track horizontal scales. Note that cloud types over the global are similar to those presented here.

the statistical behaviors of along-track horizontal scale for different cloud types over the global are similar to those presented here.

Atmosphere 2018, 9, 168

12 of 16

Atmosphere 2018, 9, x FOR PEER REVIEW

12 of 16

Figure 12. PDFs and cumulative PDFs of along-track horizontal scales for total, low-level, mid-level, Figure 12. PDFs and cumulative PDFs of along-track horizontal scales for total, low-level, mid-level, high-level, high-level, and and deep deep clouds. clouds.

4. Summary and Conclusions 4. Summary and Conclusions In this study, four years (2007–2010) of CloudSat 2B-CLDCLASS-LIDAR product are analyzed In this study, four years (2007–2010) of CloudSat 2B-CLDCLASS-LIDAR product are analyzed to study cloud occurrences, vertical distributions, and along-track horizontal scales over East Asia. to study cloud occurrences, vertical distributions, and along-track horizontal scales over East Asia. The CloudSat 2B-CLDCLASS-LIDAR product uses combined CPR and CALIOP measurements for The CloudSat 2B-CLDCLASS-LIDAR product uses combined CPR and CALIOP measurements for cloud scenario classification, and provides by far the most accurate detections of cloud boundaries cloud scenario classification, and provides by far the most accurate detections of cloud boundaries and and their vertical structures on a global scale. their vertical structures on a global scale. The mean occurrence frequency over East Asia derived from the CloudSat The mean occurrence frequency over East Asia derived from the CloudSat 2B-CLDCLASS-LIDAR 2B-CLDCLASS-LIDAR product is 66.3% with a standard deviation of 12.6%, which is approximately product is 66.3% with a standard deviation of 12.6%, which is approximately 13.8% and 21.6% higher 13.8% and 21.6% higher than that from the CALIOP level 2 5-km cloud layer product and the than that from the CALIOP level 2 5-km cloud layer product and the CloudSat 2B-GEOPROF product, CloudSat 2B-GEOPROF product, respectively. These differences are even larger than the respectively. These differences are even larger than the model-observation differences, as is pointed model-observation differences, as is pointed out in the introduction section, suggesting the out in the introduction section, suggesting the importance of using more reliable cloud datasets importance of using more reliable cloud datasets for model validations. Significant differences in for model validations. Significant differences in cloud occurrence between that of the CloudSat cloud occurrence between that of the CloudSat 2B-CLDCLASS-LIDAR product and that of the 2B-CLDCLASS-LIDAR product and that of the CALIOP Level 2 5-km cloud layer product occur CALIOP Level 2 5-km cloud layer product occur along the south of the Tibetan Plateau and over the along the south of the Tibetan Plateau and over the South and East China Sea area, while significant South and East China Sea area, while significant differences between that of the CloudSat differences between that of the CloudSat 2B-CLDCLASS-LIDAR product and that of the 2B-GEOPROF 2B-CLDCLASS-LIDAR product and that of the 2B-GEOPROF product occur over most of the product occur over most of the southern East Asia area. For vertical distributions, clouds occur southern East Asia area. For vertical distributions, clouds occur frequently at high and low altitudes, frequently at high and low altitudes, with significant seasonal variations at latitudes lower than 25 N, with significant seasonal variations at latitudes lower than 25 N, and at mid and low altitudes, with and at mid and low altitudes, with less seasonal variations at latitudes higher than 25 N. Cloud-top less seasonal variations at latitudes higher than 25 N. Cloud-top heights over East Asia have three heights over East Asia have three local peaks at approximately 1.5 km AGL corresponding to the low local peaks at approximately 1.5 km AGL corresponding to the low clouds, and 10 km and 15 km clouds, and 10 km and 15 km AGL corresponding to the high clouds over mid-latitude regions and AGL corresponding to the high clouds over mid-latitude regions and the near-tropic regions, the near-tropic regions, respectively. respectively. To better understand cloud dynamic and thermodynamic environments, clouds are also classified To better understand cloud dynamic and thermodynamic environments, clouds are also into low-level clouds, mid-level clouds, high-level clouds, and vertically extended deep cloud systems. classified into low-level clouds, mid-level clouds, high-level clouds, and vertically extended deep cloud systems. Except for over the East and South China Sea, low-level clouds mainly occur in the southwest of China around the Tibetan Plateau, as a result of the topography effect of the Tibetan

Atmosphere 2018, 9, 168

13 of 16

Except for over the East and South China Sea, low-level clouds mainly occur in the southwest of China around the Tibetan Plateau, as a result of the topography effect of the Tibetan Plateau on the circulation and thus the transport of water vapor. Mid-level clouds are widely distributed over East Asia, with larger occurrences in the middle of China, forming a mid-altitude cloud band related to the water vapor that is transported by the sub-tropic westerly. High-level clouds are concentrated within southern East Asia, and mostly occur in JJA, due to the deep convection and strong water vapor transport by the updraft. Deep clouds are mainly located over the Bay of Bengal and East China Sea. Significant seasonal variations for low-level clouds, mid-level clouds, high-level clouds, and deep clouds are found in different regions, which are related to different cloud formation mechanisms. Low-level and mid-level cloud-top heights and thicknesses have little monthly variations. For high-level clouds, cloud-top heights are about 2 km larger during JJA, but cloud thicknesses have little monthly variations. For deep clouds, both cloud-top height and thickness are about 2 km larger during JJA. The vertical and horizontal scales of the clouds were analyzed in detail. Significant fractions of low-level cloud, mid-level cloud, and high-level cloud have thicknesses smaller than 1 km, which indicates that models with vertical resolutions lower than 0.5 km have difficulties reliably resolving those clouds. The averaged cloud along-track horizontal scale over East Asia is 82.0 km, with a standard deviation of 142.0 km. The PDFs of cloud along-track horizontal scales suggest that approximately 81.2% of the clouds over East Asia cannot be resolved by climate models with a grid resolution of 1◦ . Mean along-track horizontal scales for low-level clouds, mid-level clouds, high-level clouds, and deep clouds are 56 km, 88 km, 110 km, and 99 km, respectively. Low-level clouds have the smallest along-track horizontal scales, while high-level clouds have the largest along-track horizontal scales. This study provides a unique and reliable dataset of cloud occurrences and morphology characteristics over East Asia. The dataset can be used to improve cloud parameterizations in climate models and validate model simulations of clouds over East Asia. Further analyses of cloud microphysical properties such as droplet effective radius, particle size, and water content retrieved from combined CloudSat radar and CALIPSO lidar measurements over East Asia are currently being conducted in our group. Author Contributions: X.L. and W.Z. conceived and designed the study; X.Z. and W.Z. analyzed the data; F.W. downloaded the data; Y.D. contributed analysis tools; X.L. and X.Z. drafted the manuscript. D.Z. edited and critically revised the manuscript. Funding: This research was funded by National Key Basic Research Program of China (2017YFC0209801) and Nature Science Research Project of Anhui province (1708085MD95). Acknowledgments: CloudSat data are available at the CloudSat Data Processing Center (http://www.cloudsat. cira.colostate.edu). CALIOP data are available online at the NASA Langley Atmospheric Sciences Data Center website (https://eosweb.larc.nasa.gov/order-data). Many thanks go to the CALIPSO/CloudSat data group. The authors acknowledge the editors’ and referees’ efforts in improving the manuscript. Conflicts of Interest: The authors declare no conflict of interest.

References 1.

2. 3. 4.

Wielicki, B.A.; Barkstrom, B.R.; Harrison, E.F.; Lee, R.B., III; Louis Smith, G.; Cooper, J.E. Clouds and the Earth’s Radiant Energy System (CERES): An earth observing system experiment. Bull. Am. Meteorol. Soc. 1996, 77, 853–868. [CrossRef] Sassen, K.; Wang, Z. Classifying clouds around the globe with the CloudSat radar: 1-Year of results. Geophys. Res. Lett. 2008, 35. [CrossRef] Chen, T.; Rossow, W.; Zhang, Y. Radiative effects of cloud-type variations. J. Clim. 2000, 13, 264–286. [CrossRef] Stocker, T.F.; Qin, D.; Plattner, G.-K.; Tignor, M.; Allen, S.K.; Boschung, J.; Nauels, A.; Xia, Y.; Bex, V.; Midgley, P.M. (Eds.) Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change; Cambridge University Press: Cambridge, UK; New York, NY, USA, 2013; p. 1535.

Atmosphere 2018, 9, 168

5.

6. 7. 8. 9.

10. 11. 12. 13. 14.

15. 16. 17. 18. 19. 20. 21. 22. 23. 24.

25.

26. 27. 28.

14 of 16

Zhang, M.H.; Lin, W.Y.; Klein, S.A.; Bacmeister, J.T.; Bony, S.; Cederwall, R.T.; Genio, A.D.; Del Hack, J.J.; Loeb, N.G.; Lohmann, U.; et al. Comparing clouds and their seasonal variations in 10 atmospheric general circulation models with satellite measurements. J. Geophys. Res. 2005, 110. [CrossRef] Zhang, D.; Wang, Z.; Liu, D. A global view of midlevel liquid-layer topped stratiform cloud distribution and phase partition from CALIPSO and CloudSat measurements. J. Geophys. Res. 2010, 115. [CrossRef] Yin, J.; Wang, D.; Xu, H.; Zhai, G. An investigation into the three-dimensional cloud structure over East Asia from the CALIPSO-GOCCP Data. Sci. China Earth Sci. 2015, 58, 2236–2248. [CrossRef] Fan, J.; Leung, L.R.; Li, Z.; Morrison, H.; Chen, H.; Zhou, Y.; Qian, Y.; Wang, Y. Aerosol impacts on clouds and precipitation in eastern China: Results from bin and bulk microphysics. J. Geophys. Res. 2012, 117. [CrossRef] Liu, J.; Li, Z.; Zheng, Y.; Chiu, J.C.; Zhao, F.; Cadeddu, M.; Weng, F.; Cribb, M. Cloud optical and microphysical properties derived from ground-based and satellite sensors over a site in the Yangtze Delta region. J. Geophys. Res. Atmos. 2013, 118, 9141–9152. [CrossRef] Yan, Y.; Liu, Y.; Lu, J. Cloud vertical structure, precipitation, and cloud radiative effects over Tibetan Plateau and its neighboring regions. J. Geophys. Res. Atmos. 2016, 121, 5864–5877. [CrossRef] Yu, R.; Wang, B.; Zhou, T. Climate Effects of the Deep Continental Stratus Clouds Generated by the Tibetan Plateau. J. Clim. 2004, 17, 2702–2713. [CrossRef] Li, Y.; Gu, H. Relationship between middle stratiform clouds and large scale circulation over eastern China. Geophys. Res. Lett. 2006, 33. [CrossRef] Kang, S.; Xu, Y.; You, Q.; Flügel, W.A.; Pepin, N.; Yao, T. Review of climate and cryospheric change in the Tibetan Plateau. Environ. Res. Lett. 2010, 5, 015101. [CrossRef] Lau, W.K.-M.; Kim, M.-K.; Kim, K.-M.; Lee, W.-S. Enhanced surface warming and accelerated snow melt in the Himalayas and Tibetan Plateau induced by absorbing aerosols. Environ. Res. Lett. 2010, 5, 025204. [CrossRef] Duan, A.; Xiao, Z. Does the climate warming hiatus exist over the Tibetan Plateau? Sci. Rep. 2015, 5, 13711. [CrossRef] [PubMed] Ding, Y.; Chan, J.C. The East Asian summer monsoon: An overview. Meteorol. Atmos. Phys. 2005, 89, 117–142. Johnson, R.H.; Houze, R.A. Precipitating Cloud Systems of the Asian Monsoon, Monsoon Meteorology; Chang, C.-P., Krishnamurti, T.N., Eds.; Oxford University Press: New York, NY, USA, 1987. Berry, E.; Mace, G.G. Cloud properties and radiative effects of the Asian summer monsoon derived from A-Train data. J. Geophys. Res. Atmos. 2014, 119, 9492–9508. [CrossRef] Li, J.; Waliser, D.; Stephens, G.; Lee, S. Characterizing and Understanding Cloud Ice and Radiation Budget Biases in Global Climate Models and Reanalysis. Meteorol. Monogr. 2016, 56, 13.1–13.20. [CrossRef] Li, Y.; Yu, R.C.; Xu, Y.P.; Zhang, X.H. Spatial distribution and seasonal variation of cloud over China based on ISCCP data and surface observations. J. Meteorol. Soc. Jpn. 2004, 82, 761–773. [CrossRef] Yin, J.; Wang, D.; Zhai, G.; Wang, Z. Observational characteristics of cloud vertical profiles over the continent of East Asia from the CloudSat data. Acta Meteorol. Sin. 2013, 27, 26–39. [CrossRef] Chepfer, H.; Bony, S.; Winker, D.; Cesana, G.; Dufresne, J.L.; Minnis, P.; Stubenrauch, C.J.; Zeng, S. The GCM-Oriented CALIPSO Cloud Product (CALIPSO-GOCCP). J. Geophys. Res. 2010, 115. [CrossRef] Bühl, J.; Seifert, P.; Myagkov, A.; Ansmann, A. Measuring ice- and liquid-water properties in mixed-phase cloud layers at the Leipzig Cloudnet station. Atmos. Chem. Phys. 2016, 16, 10609–10620. [CrossRef] Winker, D.M.; Pelon, J.; Coakley, J.A., Jr.; Ackerman, S.A.; Charlson, R.J.; Colarco, P.R.; Flamant, P.; Fu, Q.; Hoff, R.M.; Kittaka, C. The CALIPSO mission: A global 3D view of aerosols and clouds. Bull. Am. Meteorol. Soc. 2010, 91, 1211–1230. [CrossRef] Su, H.; Jiang, J.H.; Zhai, C.; Perun, V.S.; Shen, J.T.; Genio, A.D.; Nazarenko, L.S.; Donner, L.J.; Horowitz, L.; Seman, C.; et al. Diagnosis of regime-dependent cloud simulation errors in CMIP5 models using “A-Train” satellite observations and reanalysis data. J. Geophys. Res. Atmos. 2013, 118, 2762–2780. [CrossRef] Cesana, G.; Waliser, D.E. Characterizing and understanding sys-tematic biases in the vertical structure of clouds in CMIP5/CFMIP2 models. Geophys. Res. Lett. 2016, 43, 10538–10546. [CrossRef] Myers, T.A.; Norris, J.R. On the Relationships between Subtropical Clouds and Meteorology in Observations and CMIP3 and CMIP5 Models. J. Clim. 2015, 28, 2945–2967. [CrossRef] Luo, Y.; Zhang, R.; Wang, H. Comparing Occurrences and Vertical Structures of Hydrometeors between Eastern China and the Indian Monsoon Region Using CloudSat/CALIPSO Data. J. Clim. 2009, 22, 1052–1064. [CrossRef]

Atmosphere 2018, 9, 168

29.

30.

31.

32.

33.

34. 35. 36.

37.

38. 39. 40.

41. 42. 43.

44. 45. 46. 47.

15 of 16

Mace, G.G.; Zhang, Q.; Vaughan, M.; Marchand, R.; Stephens, G.; Trepte, C.; Winker, D. A description of hydrometeor layer occurrence statistics derived from the first year of merged Cloudsat and CALIPSO data. J. Geophys. Res. 2009, 114. [CrossRef] Wang, Z.; Vane, D.; Stephens, G.; Reinke, D. Level 2 Combined Radar and Lidar Cloud Scenario Classification Product Process Description and Interface Control Document; Jet Propulsion Laboratory; California Institute of Technology: Pasadena, CA, USA, 2012; 22p, Available online: http://www.cloudsat.cira.colostate.edu/sites/default/files/ products/files/2B-CLDCLASS-LIDAR_PDICD.P_R04.20120522.pdf (accessed on 22 May 2012). Winker, D.M.; Pelon, J.R.; McCormick, M.P. The CALIPSO mission: Spaceborne Lidar for observation of aerosols and clouds. In Lidar Remote Sensing for Industry and Environment Monitoring III, Proceedings of the 3rd International Asia-Pacific Environmental Remote Sensing Remote Sensing of the Atmosphere, Ocean, Environment, and Space, Hangzhou, China, 23–27 October 2002; Singh, U.N., Itabe, T., Lui, Z., Eds.; International Society for Optical Engineering: Washington, DC, USA, 2003; Volume 4893, pp. 1–11. Stephens, G.L.; Boain, R.J.; Mace, G.G.; Sassen, K.; Wang, Z.; Illingworth, A.J.; O’Connor, E.J.; Rossow, W.J.; Durden, S.L.; Miller, S.D.; et al. The CloudSat mission and the A-Train. Bull. Am. Meteorol. Soc. 2002, 83, 1771–1790. [CrossRef] Stephens, G.L.; Tanelli, S.; Im, E.; Durden, S.; Rokey, M.; Reinke, D.; Partain, P.; Mace, G.G.; Austin, R.; L’Ecuyer, T.; et al. CloudSat mission: Performance and early science after the first year of operation. J. Geophys. Res. 2008, 113. [CrossRef] Mace, G.G.; Marchand, R.; Zhang, Q.; Stephens, G. Global hydrometeor occurrence as observed by CloudSat: Initial observations from summer 2006. Geophys. Res. Lett. 2007, 34. [CrossRef] Wang, Z.; Sassen, K. Cloud type and macrophysical property retrieval using multiple remote sensors. J. Appl. Meteorol. 2001, 40, 1665–1682. [CrossRef] Partain, P. Cloudsat ECMWF-AUX Auxiliary Data Process Description and Interface Control Document; Cooperative Institute for Research in the Atmosphere, Colorado State University: Fort Collins, CO, USA, 2007; p. 80523. Liu, Z.; Vaughan, M.; Winker, D.; Kittaka, C.; Getzewich, B.; Kuehn, R.; Omar, A.; Powell, K.; Trepte, C.; Hostetler, C. The CALIPSO Lidar Cloud and Aerosol Discrimination: Version 2 Algorithm and Initial Assessment of Performance. J. Atmos. Ocean. Technol. 2009, 26, 1198–1213. [CrossRef] Marchand, R.; Mace, G.G.; Ackerman, T.; Stephens, G. Hydrometeor Detection Using CloudSat—An Earth-Orbiting 94-GHz Cloud Radar. J. Atmos. Ocean. Technol. 2008, 25, 519–533. [CrossRef] Liu, C.; Zipser, E.J.; Mace, G.; Benson, S. Implications of the differences between daytime and nighttime CloudSat observations over the tropics. J. Geophys. Res. 2008, 113. [CrossRef] Bouniol, D.; Couvreux, F.; Kamsu-Tamo, P.-H.; Leplay, M.; Guichard, F.; Favot, F.; O’Connor, E.J. Diurnal and Seasonal Cycles of Cloud Occurrences, Types, and Radiative Impact over West Africa. J. Appl. Meteorol. Clim. 2012, 51, 534–553. [CrossRef] Zuidema, P.; Mapes, B. Cloud Vertical Structure Observed from Space and Ship over the Bay of Bengal and the Eastern Tropical Pacific. J. Meteorol. Soc. Jpn. 2008, 86, 205–218. [CrossRef] Li, W.L.; Wang, K.L.; Fu, S.M.; Jiang, H. The interrelationship between regional westerly index and the water vapor budget in northwest China. J. Glaciol. Geocryol. 2008, 30, 34–38. (In Chinese) Chen, T.; Guo, J.; Li, Z.; Zhao, C.; Liu, H.; Cribb, M.; Wang, F.; He, J. A CloudSat Perspective on the Cloud Climatology and Its Association with Aerosol Perturbations in the Vertical over Eastern China. J. Atmos. Sci. 2016, 73, 3599–3616. [CrossRef] Ren, G.Y.; Yuan, Y.J.; Liu, Y.J.; Ren, Y.Y.; Wang, T.; Ren, X.Y. Changes in precipitation over northwest China. Arid Zone Res. 2016, 33, 1–19. (In Chinese) [CrossRef] Peng, D.; Zhou, T. Why was the arid and semiarid northwest China getting wetter in the recent decades? J. Geophys. Res. Atmos. 2017, 122, 9060–9075. [CrossRef] Johnson, R.H.; Rickenbach, T.M.; Rutledge, S.A.; Ciesielski, P.E.; Schubert, W.H. Trimodal Characteristics of Tropical Convection. J. Clim. 1999, 12, 2397–2418. [CrossRef] Sassen, K.; Wang, Z.; Liu, D. Global distribution of cirrus clouds from CloudSat/Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO) measurements. J. Geophys. Res. 2008, 113. [CrossRef]

Atmosphere 2018, 9, 168

48. 49.

16 of 16

Adhikari, L.; Wang, Z.; Deng, M. Seasonal variations of Antarctic clouds observed by CloudSat and CALIPSO satellites. J. Geophys. Res. 2012, 117. [CrossRef] Zhang, D.; Luo, T.; Liu, D.; Wang, Z. Spatial scales of altocumulus clouds observed with collocated CALIPSO and CloudSat measurements. Atmos. Res. 2014, 149, 58–69. [CrossRef] © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).