[1] A method for estimating effective ice particle radius R_e at the tops of tropical deep convective clouds (DCC) is developed on the basis of precomputed look-up tables (LUTs) of brightness temperature differences (BTDs) between the 3.7 and 11.0 μm bands. A combination of discrete ordinates radiative transfer and correlated k distribution programs, which account for the multiple scattering and monochromatic molecular absorption in the atmosphere, is utilized to compute the LUTs as functions of solar zenith angle, satellite zenith angle, relative azimuth angle, R_e, cloud top temperature (CTT), and cloud visible optical thickness τ. The LUT-estimated DCC R_e agrees well with the cloud retrievals of the Moderate Resolution Imaging Spectroradiometer (MODIS) for the NASA Clouds and Earth's Radiant Energy System with a correlation coefficient of 0.988 and differences of less than 10%. The LUTs are applied to 1 year of measurements taken from MODIS aboard Aqua in 2007 to estimate DCC R_e and are compared to a similar quantity from CloudSat over the region bounded by 140°E, 180°E, 0°N, and 20°N in the Western Pacific Warm Pool. The estimated DCC R_e values are mainly concentrated in the range of 25–45 μm and decrease with CTT. Matching the LUT-estimated R_e with ice cloud R_e retrieved by CloudSat, it is found that the ice cloud τ values from DCC top to the vertical location where LUT-estimated R_e is located at the CloudSat-retrieved R_e profile are mostly less than 2.5 with a mean value of about 1.3. Changes in the DCC τ can result in differences of less than 10% for R_e estimated from LUTs. The LUTs of 0.65 μm bidirectional reflectance distribution function (BRDF) are built as functions of viewing geometry and column amount of ozone above upper troposphere. The 0.65 μm BRDF can eliminate some noncore portions of the DCCs detected using only 11 μm brightness temperature thresholds, which result in a mean difference of only 0.6 μm for DCC R_e estimated from BTD LUTs.

Key Points

A new method for estimating effective radius of tropical deep convection
Radius is mainly 25–45 micrometers and decreases with cloud top temperature
Mean value of cloud top optical depth corresponding to estimated radius is about 1.3

1. Introduction

[2] Knowledge of ice particle effective radius R_e of tropical deep convective clouds (DCCs) is important for understanding the important roles of DCCs in the physical and chemical processes occurring in the tropical tropopause layer (TTL) and the stratosphere [Fueglistaler et al., 2009, and references therein]. For example, R_e plays an important role in cloud radiative forcing, particularly shortwave radiative forcing of ice clouds including DCCs and thereby has been used as a key factor for the parameterization of ice cloud radiative properties implemented in climate models [Hong et al., 2009, and references therein]. Smaller ice particles are more likely to remain suspended in the air entering the stratosphere and therefore deposit moisture there [Sherwood, 2002; Iwasaki et al., 2010]. Jensen et al. [2007] found that the potential for irreversible dehydration near the tropical tropopause caused by DCCs depends on DCC R_e. Sherwood et al. [2006] reported that climatological maxima in lightning activity are associated with small ice particles near DCC tops and suggested this relationship strengthens the importance of aerosols as well as dynamics in electrification. Effective ice particle size is also critical for better understanding of cirrus clouds since the ice microphysics of DCCs has a strong influence on the characteristics and development of cirrus clouds [Yuan and Li, 2010, and references therein].

[3] Fan et al. [2010] suggested that more accurate retrievals or measurements of DCC R_e can provide better understanding of the heterogeneous and homogeneous freezing parameterizations applied in the numerical cloud models, which are important for characterizing effect of aerosols on DCC R_e [e.g., Fan et al., 2007; Jiang et al., 2011]. However, reliable characterization of DCC R_e from observations remains a major challenge [Iwasaki et al., 2010; Chiriaco et al., 2007] owing to the sensitivities of sensors to different viewing geometry and spatial and diurnal coverage in retrieving R_e [Comstock et al., 2007; Waliser et al., 2009]. It was emphasized by Minnis et al. [2011] that cloud retrieval algorithms utilizing channels common to most meteorological satellite imagers have the advantage of producing long records of consistent cloud properties including R_e, which is important for understanding climate variations and predictions [Waliser et al., 2009].

[4] The shortwave infrared (SIR) band centered at 3.7 μm is primarily sensitive to cloud R_e and has been extensively used to retrieve R_e and other cloud properties with combinations of additional solar bands and/or infrared bands [Minnis et al., 2011, and references therein]. Owing to the large visible optical depths τ of DCCs, the daytime reflected SIR solar radiance is essentially independent of the DCC τ [e.g., Minnis et al., 1998]. On the basis of this feature, 3.7 μm reflectance has been used to estimate optically thick ice cloud R_e by numerous researchers [e.g., Sherwood, 2002; Rosenfeld and Lensky, 1998; Lensky and Rosenfeld, 2006; Lindsey and Grasso, 2008]. Brightness temperature differences BTD between the 3.7 and the infrared (IR) 11.0 μm channels are sensitive to cloud τ and R_e during daytime [Stone et al., 1990]. When τ becomes large, the IR radiance effectively becomes a constant value, so the BTD over DCCs are essentially sensitive only to R_e. Dong et al. [2008] suggested that the BTDs over optically thick ice clouds are less influenced by atmospheric absorption than the 3.7 μm band alone.

[5] The objective of this study is to develop offline look-up tables (LUTs) on the basis of simulated BTDs from a rigorous radiative transfer model for estimating tropical DCC R_e, which remains a challenge. The motivation for developing the new LUT method is to quantify tropical radiative forcing over DCC more effectively than using currently available methodology and to specify more homogenous DCC calibration targets [Minnis et al., 2008a; Doelling et al., 2010]. Different from previous studies on estimating ice cloud R_e using passive satellite imagers using the 3.7 μm band, solar reflectance and thermal emission at the 3.7 μm band are not separated and combined together to build the BTD LUTs in section 2. In section 3, the LUT-estimated DCC R_e is first compared with the results of ice cloud R_e retrieved by the Moderate Resolution Imaging Spectroradiometer (MODIS) for the NASA Clouds and Earth's Radiant Energy System (CERES) Project (CERES-MODIS) [Minnis et al., 2011]. The LUTs are then applied to 1 year of measurements in 2007 taken from the MODIS aboard Aqua to estimate DCC R_e. The results are compared to the values of a similar quantity from CloudSat over the region bounded by 140°E, 180°E, 0°N, and 20°N in the Western Pacific Warm Pool. Conclusions are given in section 4. The advantage of this approach is that it is simple and requires minimal auxiliary information to implement (i.e., only measurements of brightness temperatures at 3.7 and 11.0 μm bands), making it straightforward for estimating DCC R_e.

2. Look-Up Table and Methodology

[6] The discrete ordinates radiative transfer model (DISORT) [Stamnes et al., 1988] is conjoined with correlated k distribution routines developed by Kratz [1995] for the MODIS bands to compute radiances at 3.7 and 11.0 μm during daytime. Visible (VIS; 0.65 μm) channel bidirectional reflectance distribution functions (BRDFs) are built to assist detection of DCCs on the basis of 11 μm brightness temperatures (BT₁₁) for an additional option.

[7] For optically thick clouds, surface reflectance and emission have little impact on BTD [Dong et al., 2008]. Particularly, because of the large τ of tropical DCCs, the surface albedo and emissivity variations can be neglected. Surface albedos of 0.05 and 0.02 are set for computing solar radiances at 0.65 and 3.7 μm, respectively. A surface emissivity of 0.98 is used for computing the SIR and IR radiances. Tropical DCCs have high vertical extensions approaching or penetrating the TTL. Thus, the atmospheric effects on the radiances are essentially from atmospheric water vapor above the upper troposphere. The precipitable water above 11 km altitude is less than 0.3% from the total atmospheric precipitable water [Badescu, 1988]. Furthermore, both 3.7 and 11.0 μm are located in atmospheric windows. So, the standard tropical atmospheric profile is used for computing atmospheric absorption at 3.7 and 11.0 μm. Only changes in column ozone above the upper troposphere are considered for the VIS BRDFs. The 11 ice cloud models from Minnis et al. [1998] are used for all computations. The R_e used in this study, which is defined as the ratio of total volume of cloud particles to the total projected area and multiplied by 0.75, can be derived from formula given by Minnis et al. [2011]. Six-dimensional BTD LUTs are built at 18 solar zenith and 18 satellite zenith angles, μ₀ and μ, respectively, and covering the range, 0°–85°; 37 relative azimuth angles Δϕ between 0°–180°; 11 R_e values from about 3 to 80 μm; 6 cloud top temperatures (CTT) from 190 to 215 K (cloud base height is set at 9 km and about −30°C); and 4 τ values from 25 to 100. Six-dimensional VIS BRDF LUTs are built for these same μ₀, μ, Δϕ, R_e, and τ and 7 stratospheric column ozone values of 100–400 DU. Owing to large values of DCC τ, the effect of variations in DCC cloud base heights on radiances over DCCs is neglectable [Hong et al., 2007; Garrett et al., 2009], particularly for BTDs and VIS BRDFs. The cloud base height is set at 9 km with a temperature of about −30°C for computing the LUTs.

[8] BT₁₁ thresholds have been extensively used to identify DCCs. The DCCs identified by BT₁₁ thresholds alone can be contaminated by optically thick ice clouds outside the core area of the cloud [e.g., Tian et al., 2004; Hong et al., 2005; Liu et al., 2007]. Some high, thick ice clouds that are identified as DCCs by BT₁₁ can be eliminated by setting a VIS reflectance threshold [Fu et al., 1990]. The VIS BRDF LUTs were developed to use together with BT₁₁ thresholds to eliminate such contamination of high, thick ice clouds. In the present study, two methods are used to identify tropical DCCs, one uses BT₁₁ thresholds alone and the other combines BT₁₁ thresholds with VIS BRDF LUTs.

[9] An example of the computed BTD LUTs is shown in Figure 1 as a function of R_e for DCCs having τ between 25 and 100 and CTT of 215 and 200 K at the specified viewing geometry (μ₀ = 30°, μ = 45°, and Δϕ = 60°). For a given DCC with a specified CTT, BTD is essentially only sensitive to R_e. This feature is particularly true for small R_e owing to stronger SIR solar reflectance for smaller particles [e.g., Minnis et al., 1998, 2011; Han et al., 1994; Platnick and Twomey, 1994; Lindsey and Grasso, 2008]. Optical depth has negligible influence on the BTD, particularly for larger τ. Figure 1 also shows that an accurate CTT is important for using the LUTs to estimate R_e. As a blackbody, the DCC CTT can be directly estimated using the satellite-observed BT₁₁ [Smith and Platt, 1978] or an adjusted value, since BT₁₁ was found to be slightly warmer than CTT [Minnis et al., 2008b]. The 6 CTTs are then converted to 6 BT₁₁ values at nadir over DCCs on the basis of CTT and the simulated BT₁₁ at nadir. Before using the LUTs, the initial observed BT₁₁ at off nadir is adjusted to nadir using a correction factor that is built as functions of μ and BT₁₁ at the μ using the simulated BT₁₁. The adjustments of BT₁₁ from off nadir to nadir have root-mean-square errors (RMSEs) below 0.3 K for all DCCs.

Details are in the caption following the image — **Figure 1**
Open in figure viewer PowerPoint

BTD as a function of R_e for tropical DCC with τ of 25, 50, 75, and 100 and CTT of 215 and 210 K at μ₀ = 30°, μ = 45°, and Δϕ = 60°.

[10] Figure 2a shows the VIS BRDF as a function of R_e at the specified viewing geometry for DCCs having τ between 25 and 100 and column ozone of 250 DU in the stratosphere. As expected, the VIS BRDF is primarily sensitive to τ and only slightly affected by R_e. Essentially the VIS BRDF is small when τ is small and R_e is large. So, the LUTs having τ = 25 with the specified viewing geometry and column ozone are re-constructed using the minimum BRDF among the computed BRDFs at all R_e. In this way, the LUTs are downsized from six dimensions to four dimensions of μ₀, μ, Δϕ, and column ozone. Figure 2b shows the remarkable angular dependence of the VIS BRDF from the four-dimensional LUTs. The patterns are shown at μ₀ = 30° and μ₀ = 60° with μ varying from 0° to 85° along the radius and Δϕ between 0° and 180° along the counterclockwise direction. The effect of changing the column ozone above the tropopause on the VIS BRDF is investigated by showing the BRDF differences resulting from increasing the above-cloud ozone loading from 200 DU to 300 DU. The ozone effects (Figure 2c) are distinct at large μ₀ and μ and can be over 0.1 with a change of 100 DU.

[11] The process for retrieving the DCC R_e is straightforward. First, BT₁₁ measurements at μ are corrected to nadir using the regressions developed above. Then, a pixel is identified as a DCC if the corrected BT₁₁ at nadir is less than the specified BT₁₁ threshold (range of 190–215 K in the present study) for detecting DCCs. When measurements of the VIS reflectance are employed to constrain DCC detection, an additional criterion is that the reflectance measurement should exceed those from the four-dimensional VIS BRDFs in the LUTs. After DCCs are identified, R_e can be estimated using the six-dimensional LUTs of BTD on the basis of known μ₀, μ, Δϕ, above-cloud column ozone, BT₁₁ at nadir, and a specified τ at 25, 50, 75, or 100.

3. Application and Discussions

3.1. Comparison With CERES-MODIS-Retrieved R_e

[12] To simplify complex, highly variable mixtures of habits and distributions of ice crystal sizes, most passive satellite retrieval algorithms assume a vertically homogeneous ice cloud with a prescribed set of habits and size distributions to perform retrievals of ice cloud microphysical properties. Typically, each algorithm makes different assumptions about the particle habits used in the retrievals. Also, different retrieval algorithms can use different passive satellite imagers, so that the retrieved results may be dependent on the retrieval algorithms and sensors used [Chiriaco et al., 2007; Comstock et al., 2007; Lindsey and Grasso, 2008; Waliser et al., 2009]. On the basis of these facts and the danger of measuring DCC properties in situ, it is difficult to validate retrieved ice cloud microphysical properties. To avoid the differences caused by different ice cloud models used for retrieving the DCC R_e, the LUT-estimated DCC R_e is compared with the results from CERES-MODIS analyses [Minnis et al., 2011] that utilize consistent ice cloud models.

[13] Figure 3 shows the comparison results for an Aqua MODIS granule over the Western Pacific Warm Pool at 02:35 UTC on 1 February 2007. Only BT₁₁ threshold of 215 K is used for DCC detection and the DCC fraction is about 12% in this granule. It is clear that the geographical distribution of LUT-estimated DCC R_e is in good agreement with that for the CERES-MODIS-retrieved R_e (Figures 3b and 3c). The 1-D and 2-D histogram distributions of the comparison of R_e are shown in Figure 4. Statistically, the LUT-estimated DCC R_e (33.05 ± 5.57 μm, mean value and standard deviation) values are essentially the same as those from CERES-MODIS (32.93 ± 6.21 μm). The estimated R_e is highly correlated with a correlation coefficient of 0.988. The RMSE is about 1.13 μm. In 67.8% and 99.8% of the DCCs, the differences in R_e are less than the RMSE and three times the RMSE (3.39 μm), respectively. From these results, it is concluded that the differences between LUT-estimated DCC R_e and CERES-MODIS R_e are less than 10%.

3.2. Comparison With CloudSat-Retrieved R_e

[14] Ice cloud R_e values are estimated on the basis of 3.7 μm band is near the cloud top [e.g., Sherwood, 2002] since the radiation at 3.7 μm at the top of atmosphere comes mostly from cloud top [Rosenfeld et al., 2004]. The vertical distribution of DCC R_e can be estimated by radar measurements, thus, providing an opportunity to investigate the depth from cloud top to the location where the 3.7 μm estimated DCC R_e is located.

[15] The constructed LUTs are applied to BTDs taken from the MODIS/Aqua Calibrated Radiances 1 km 5 Min 1B Narrow Swath Subset along CloudSat (MAC021S0) to estimate the DCC R_e. The VIS reflectances used for assisting DCC detection (optionally) are also taken from MAC021S0. The MAC021S0 cross-track width of 10 km consists of 11 MODIS pixels, which results in a horizontal resolution of ∼1 km (http://mirador.gsfc.nasa.gov/collections/MAC021S0__002.shtml). The LUT-estimated R_e is then compared to the radar-only ice R_e from the CloudSat product, 2B-CWC-RO (version 008), that consists of one hundred twenty-five 240 m thick vertical bins with a horizontal resolution of about 1.3 km × 1.7 km [Austin et al., 2009]. The 2B-CWC-RO ice cloud products are generated from CloudSat reflectivity measurements using the retrieval algorithms described by Austin et al. [2009]. The MODIS/Aqua Geolocation Fields 1 km 5 min 1A Narrow Swath Subset along CloudSat V2 (MAC03S0, http://mirador.gsfc.nasa.gov/collections/MAC03S0__002.shtml) is used to collocate the MAC021S0 with 2B-CWC-RO using the nearest-neighbor method within a maximum distance of 0.5 km. This study uses all of the 2007 collocated data over the region bounded by 140°E, 180°E, 0°N, and 20°N in the Western Pacific Warm Pool. In 2007, the collocation of near 2600 MODIS granules and 1200 CloudSat cross tracks over this region. Half were taken during daytime and generate over one million matched MODIS and CloudSat measurements. Using a BT₁₁ threshold of 215 K, the detected DCC fraction over the region is about 2.7% which is close to the result from Fu et al. [1990]. With additional VIS BRDF LUTs, the DCC fraction decreases to about 1.8% by removing some contaminated high, thick ice clouds associated with DCCs.

[16] Figure 5 shows an example of comparing LUT-estimated DCC R_e from the MODIS data with R_e derived from CloudSat. DCC systems are shown along the CloudSat track at about 02:36–02:39 UTC (167.35°E, 2.78°S to 165.40°E, 6.36°N) on 1 February 2007 by CloudSat 1B-CPR reflectivities (version 008). The 1B-CPR has the same vertical and horizontal resolutions as 2B-CWC-RO and a minimum detectable reflectivity of approximately −26 dBZ [Stephens et al., 2002]. Cloud tops from 2B-GEOPROF-LIDAR [Mace et al., 2007] are also shown in Figure 5a since the CALISPO lidar [Winker et al., 2007] is more sensitive to small ice particles and thin cloud layers with respect to the CloudSat radar [McGill et al., 2004]. Figure 5a shows that some of the DCC cloud tops penetrate into the TTL.

[17] The DCCs, identified using a BT₁₁ threshold of 215 K, are shown as the light blue and black dots in Figure 5b. The black dots correspond to only those DCCs identified using both the BT₁₁ threshold and the thresholds from the VIS BRDF LUTs. Using the VIS BRDF thresholds eliminates some DCCs identified by the BT₁₁ threshold beyond the convective cores. This feature is clearly shown around 02:36:40 and 02:38:20 UTC in Figure 5b. The above-cloud column ozone amount used for the VIS BRDF LUTs was taken from ozone profiles in ECMWF-AUX product (version 008). The BTD LUTs for τ = 25 are then applied to the identified DCCs to obtain a value of R_e. Vertical profiles of 2B-CWC-RO ice cloud R_e are shown along the CloudSat track. The LUT-estimated R_e values are compared with 2B-CWC-RO R_e profiles to locate the altitude where the two values are the same. This altitude, shown by the red dots in Figure 5b, indicates that that LUT-estimated R_e is quite consistent with the 2B-CWC-RO R_e near the DCC radar tops, but well below the CALIPSO tops.

[18] Figure 6a shows probability density functions (PDFs) of the LUT-estimated R_e using τ = 25 for DCCs identified using the BT₁₁ thresholds of 215 and 200 K alone or combined with the VIS thresholds over the chosen region for all of 2007. The LUT-estimated DCC R_e's are mainly concentrated in the range of 25–45 μm. The DCCs identified with the lower BT₁₁ threshold have smaller R_e values than those identified with the higher BT₁₁ threshold. This feature is in good agreement with earlier findings that demonstrate a decrease in ice cloud particle sizes with diminishing cloud temperatures [Yuan and Li, 2010, and references therein]. Although inclusion of the VIS BRDF thresholds can eliminate some noncore areas of DCCs detected using BT₁₁ thresholds alone, these additional constraints have negligible impact on the R_e values retrieved using the BTD LUTs. This conclusion is clearly supported by the similarity of PDFs for DCCs with and without the VIS thresholds.

[19] Total ice cloud τ from the DCCs cloud top (z = 0) to the location (z = Z) where the LUT-estimated R_e equals that from CloudSat are calculated by

$urn:x-wiley:01480227:media:jgrd17584:jgrd17584-math-0001$

where IWC(z) and r_e(z) are the vertical profiles of 2B-CWC-RO ice cloud water content and R_e, respectively, Q_e is visible effective efficiency at the corresponding r_e(z) and is derived from the 11 ice cloud models [Minnis et al., 1998] to be consistent with those used for the LUT computations, and ρ is the ice density of 0.917 × 10³ kg/m³. The cumulative distribution functions (CDFs) of calculated τ are shown in Figure 6b for DCCs identified by BT₁₁ thresholds of 215 and 200 K alone or with the additional VIS BRDF thresholds. The CDFs show that over 86% and 94% of the LUT-estimated R_e values match their CloudSat counterparts at depths corresponding to τ < 2.5 for DCCs identified with BT₁₁ < 215 K and BT₁₁ < 200 K, respectively. This feature corresponds well to theoretical results [Platnick, 2000; Kokhanovsky, 2004] and to other observations [Dong et al., 2008]. Moreover, the mean values of cloud top τ corresponding to the estimated R_e for all DCCs are around 1.2. Please note that it was found that thin cirrus clouds detected by CALIPSO, but missed by CloudSat have a τ ranging between 0.02 and 0.3 with a mean value of 0.1 [Haladay and Stephens, 2009]. Assuming that value holds for the tops of DCCs, the mean value of cloud top τ (about 1.3) is very close to the optical depth below cloud top where the altitude corresponding to BT₁₁ is located [Minnis et al., 2008b]. Including the VIS BRDF thresholds for DCC detection has negligible effect on these features and only slight differences are shown for DCCs with a high BT₁₁ threshold.

3.3. Effect of Variations in DCC τ on LUT-Estimated R_e

[20] In order to investigate effects of variations in DCC τ on LUT-estimated R_e, the BTD LUTs for τ = 25, 50, 75, and 100 are used to estimate the DCC R_e separately. Figure 7 shows the mean values of R_e for DCCs using BT₁₁ thresholds of 190–215 K with and without using the VIS BRDF thresholds. It is clearly seen again that DCC R_e values decrease with cloud temperature. The differences in DCC R_e estimated with different τ LUTs are less than 3.3 μm (about <10%) for all BT₁₁ thresholds and LUTs. The distinct differences are between LUTs with τ = 25 and τ = 50. This feature agrees with the impact of τ on the LUTs seen in Figure 1. The physics of this feature is that most of the contributions to the BTD are from the reflected SIR solar radiance, which reaches its saturated value when τ is above 32 [e.g., Minnis et al., 1998].

[21] Including the VIS BRDF thresholds for DCC detection essentially has no influence on R_e. Differences less than 0.6 μm in R_e are found between those without BRDF constraints and with BRDF constraints for DCC detection. Moreover, the differences decrease with BT₁₁ thresholds and tend to be zero. This characteristic is as expected since DCCs detected using lower BT₁₁ thresholds are more likely to be associated with convective cores than those using greater BT₁₁ thresholds. Considering the negligible effects of including the VIS BRDF on the LUT-estimated DCC R_e, it is concluded that R_e can be efficiently estimated by BTD with a BT₁₁ threshold, a method that provides the advantage of using only two bands.

4. Summary

[22] DCC R_e plays important role in better understanding physical and chemical processes in the TTL and stratosphere. Accurate knowledge of DCC R_e is a key factor of both ice cloud radiative and physical parameterizations implemented in numerical weather/climate models [Waliser et al., 2009, and references therein]. However, it is still a major challenge for observing reliable characterization of DCC R_e [Iwasaki et al., 2010; Chiriaco et al., 2007].

[23] On the basis of rigorous radiative transfer modeling, six-dimensional BTD LUTs for estimating DCC R_e are established as functions of μ₀, μ, Δϕ, R_e, CTT, and τ. Four-dimensional LUTs of VIS BRDF are also built as function of μ₀, μ, Δϕ, and column amount of ozone above upper troposphere to use optionally together with BT₁₁ thresholds for removing contamination by non-DCC, optically thick ice clouds that cannot be distinguished from DCCs identified by BT₁₁ thresholds alone. After identifying DCCs on the basis of BT₁₁ or with the additional, optional criterion of VIS BRDF, DCC R_e are estimated straightforwardly and efficiently using offline BTD LUTs by interpolating on the basis of the observed viewing geometry, BT₁₁, and τ that can be supposed to be a value between 25 and 100.

[24] Using the same ice cloud models as the CERES-MODIS cloud retrievals [Minnis et al., 2011], the LUT-estimated DCC R_e for an Aqua MODIS granule over the Western Pacific Warm Pool at 02:35 UTC on 1 February 2007 was compared to results from CERES-MODIS. The estimated DCC R_e values agree well with a correlation coefficient of 0.988 and an RMSE of about 1.13 μm. The differences between LUT-estimated DCC R_e and CERES-MODIS R_e are less than 10%.

[25] MAC021S0 data for all of 2007 over the chosen region (140°E–180°E, 0°N–20°N) in Western Pacific Warm Pool were analyzed with the BTD LUTs to estimate DCC R_e. The LUT-estimated DCC R_e values are mainly concentrated in the range of 25–45 μm and they decrease with CTTs. The LUT-estimated R_e were then compared to the radar-only ice R_e from collocated 2B-CWC-RO. The LUT-estimated R_e are in agreement good with 2B-CWC-RO R_e near DCC tops. It is found that ice cloud τ values from the DCC top to the vertical location of 2B-CWC-RO R_e corresponding to the LUT-estimated R_e values are mostly less than 2.5 with a mean value of about 1.3 after accounting for the highest layer of thin cirrus (mean of τ is about 0.1) missed by CloudSat [Haladay and Stephens, 2009]. Applying the BTD LUTs requires a DCC τ in the range of 25–100. The different choices of DCC τ can result in differences of less than 3.3 μm (<10%) for the LUT-estimated DCC R_e from LUTs for DCCs identified by BT₁₁ thresholds alone or with additional VIS BRDF thresholds. Although the use of additional VIS LUTs eliminates some thick non-DCC ice clouds detected by BT₁₁ thresholds used alone, the estimated DCC R_e distributions are essentially the same with or without the extra constraints. Therefore, the BTD LUTs constructed here can be used by themselves with BT₁₁ to estimate DCC R_e.

[26] Owing to our limited knowledge of the natural variability of ice cloud microphysical properties including ice particle size distributions, habits and their typical mixing percentages, the practical approach is to use some typical and homogenous values for computing radiative properties of general ice clouds. Different ice cloud models have been used extensively [Baran, 2009] and they can induce uncertainties in the retrieved ice cloud microphysical properties [e.g., Chiriaco et al., 2007; Yang et al., 2007, 2008a, 2008b]. Because vertical photon transport in ice clouds is affected by both vertical distributions of cloud optical thickness and particle size [Platnick, 2000; Kokhanovsky, 2004], solar reflectance and thermal emission can be affected by the vertical inhomogeneity of R_e [Wang et al., 2009; Zhang et al., 2010]. Using the homogeneous assumption for ice cloud microphysical properties, commonly used for ice cloud retrieval algorithms, the vertical inhomogeneity of ice cloud R_e can cause potential errors in the present LUT-estimated R_e, but no more than any of the other techniques employing the 3.7 μm channel for retrieving R_e. Investigating the impact is a challenge owing to variable vertical inhomogeneity in ice cloud R_e values. At this point, our statistical investigation of the ice cloud τ at DCC top above the vertical location of the CloudSat-retrieved R_e corresponding to the LUT-estimated R_e provides a step in assessment of this issue.

Acknowledgments

[27] We thank the NASA CloudSat project for providing the 2B-CWC-RO, 1B-CPR, ECMWF-AUX, and 2B-GEOPROF-LIDAR data used in this study, which are taken from the CloudSat Data Processing Center at Colorado State University. The MODIS data are archived at NASA's Goddard Earth Sciences Data and Information Services Center. This research was supported by NASA through the CERES, CALIPSO, and CloudSat programs.

References

Austin, R. T., A. J. Heymsfield, and G. L. Stephens (2009), Retrieval of ice cloud microphysical parameters using the CloudSat millimeter-wave radar and temperature, J. Geophys. Res., 114, D00A23, doi:10.1029/2008JD010049.
Badescu, V. (1988), New formulae to evaluate the atmospheric layers of precipitable water and gases, applicable in solar radiation computing models, Rev. Phys. Appl., 23, 1869–1883, doi:10.1051/rphysap:0198800230120186900.
Baran, A. J. (2009), A review of the light scattering properties of cirrus, J. Quant. Spectrosc. Radiat. Transf., 110, 1239–1260, doi:10.1016/j.jqsrt.2009.02.026.
Chiriaco, M., et al. (2007), Comparison of CALIPSO-like, LaRC, and MODIS retrievals of ice cloud properties over SIRTA in France and Florida during CRYSTAL-FACE, J. Appl. Meteorol. Climatol., 46, 249–272, doi:10.1175/JAM2435.1.
Comstock, J. M., et al. (2007), An intercomparison of microphysical retrieval algorithms for upper tropospheric ice clouds, Bull. Am. Meteorol. Soc., 88, 191–204, doi:10.1175/BAMS-88-2-191.
Doelling, D. R., G. Hong, D. Morstad, R. Bhatt, A. Gopalan, and X. Xiong (2010), The characterization of deep convective cloud albedo as a calibration target using MODIS reflectances, Proc. SPIE Int. Soc. Opt. Eng., 7862, 78620I, doi:10.1117/12.869577.
Dong, X., P. Minnis, B. Xi, S. Sun-Mack, and Y. Chen (2008), Comparison of CERES-MODIS stratus cloud properties with ground-based measurements at the DOE ARM Southern Great Plains site, J. Geophys. Res., 113, D03204, doi:10.1029/2007JD008438.
Fan, J., R. Zhang, G. Li, and W.-K. Tao (2007), Effects of aerosols and relative humidity on cumulus clouds, J. Geophys. Res., 112, D14204, doi:10.1029/2006JD008136.
Fan, J., J. M. Comstock, M. Ovchinnikov, S. A. McFarlane, G. McFarquhar, and G. Allen (2010), Tropical anvil characteristics and water vapor of the tropical tropopause layer: Impact of heterogeneous and homogeneous freezing parameterizations, J. Geophys. Res., 115, D12201, doi:10.1029/2009JD012696.
Fu, R., A. D. Delgenio, and W. B. Rossow (1990), Behavior of deep convective clouds in the tropical pacific deduced from ISCCP radiances, J. Clim., 3, 1129–1152, doi:10.1175/1520-0442(1990)003<1129:BODCCI>2.0.CO;2.
Fueglistaler, S., A. E. Dessler, T. J. Dunkerton, I. Folkins, Q. Fu, and P. W. Mote (2009), Tropical tropopause layer, Rev. Geophys., 47, RG1004, doi:10.1029/2008RG000267.
Garrett, K. J., P. Yang, S. L. Nasiri, C. R. Yost, and B. A. Baum (2009), Influence of cloud-top height and geometric thickness on a MODIS infrared-based ice cloud retrieval, J. Appl. Meteorol. Climatol., 48, 818–832, doi:10.1175/2008JAMC1915.1.
Haladay, T., and G. Stephens (2009), Characteristics of tropical thin cirrus clouds deduced from joint CloudSat and CALIPSO observations, J. Geophys. Res., 114, D00A25, doi:10.1029/2008JD010675.
Han, Q., W. B. Rossow, and A. A. Lacis (1994), Near-global survey of effective droplet radii in liquid water clouds using ISCCP data, J. Clim., 7, 465–497, doi:10.1175/1520-0442(1994)007<0465:NGSOED>2.0.CO;2.
Hong, G., G. Heygster, J. Miao, and K. Kunzi (2005), Detection of tropical deep convective clouds from AMSU-B water vapor channels measurements, J. Geophys. Res., 110, D05205, doi:10.1029/2004JD004949.
Hong, G., P. Yang, H.-L. Huang, B. A. Baum, Y.-X. Hu, and S. Platnick (2007), The sensitivity of ice optical and microphysical passive satellite retrievals to cloud geometric thickness, IEEE Trans. Geosci. Remote Sens., 45, 1315–1323, doi:10.1109/TGRS.2007.894549.
Hong, G., P. Yang, B. A. Baum, A. J. Heymsfield, and K.-M. Xu (2009), Parameterization of shortwave and longwave radiative properties of ice clouds for use in climate models, J. Clim., 22, 6287–6312, doi:10.1175/2009JCLI2844.1.
Iwasaki, S., T. Shibata, J. Nakamoto, H. Okamoto, H. Ishimoto, and H. Kubota (2010), Characteristics of deep convection measured by using the A-train constellation, J. Geophys. Res., 115, D06207, doi:10.1029/2009JD013000.
Jensen, E. J., A. S. Ackerman, and J. A. Smith (2007), Can overshooting convection dehydrate the tropical tropopause layer?, J. Geophys. Res., 112, D11209, doi:10.1029/2006JD007943.
Jiang, J. H., et al. (2011), Influence of convection and aerosol pollution on ice cloud particle effective radius, Atmos. Chem. Phys., 11, 457–463, doi:10.5194/acp-11-457-2011.
Kokhanovsky, A. A. (2004), The depth of sunlight penetration in cloud fields for remote sensing, IEEE Geosci. Remote Sens. Lett., 1, 242–245, doi:10.1109/LGRS.2004.832228.
Kratz, D. P. (1995), The correlated k-distribution technique as applied to the AVHRR channels, J. Quant. Spectrosc. Radiat. Transf., 53, 501–517.
Lensky, I. M., and D. Rosenfeld (2006), The time-space exchangeability of satellite retrieved relations between cloud top temperature and particle effective radius, Atmos. Chem. Phys., 6, 2887–2894, doi:10.5194/acp-6-2887-2006.
Lindsey, D. T., and L. Grasso (2008), An effective radius retrieval for thick ice clouds using GOES, J. Appl. Meteorol. Climatol., 47, 1222–1231, doi:10.1175/2007JAMC1612.1.
Liu, C., E. J. Zipser, and S. W. Nesbitt (2007), Global distribution of tropical deep convection: Different perspectives from TRMM infrared and radar data, J. Clim., 20, 489–503, doi:10.1175/JCLI4023.1.
Mace, G. G., R. Marchand, Q. Zhang, and G. Stephens (2007), Global hydrometeor occurrence as observed by CloudSat: Initial observations from summer 2006, Geophys. Res. Lett., 34, L09808, doi:10.1029/2006GL029017.
McGill, M. J., L. Li, W. D. Hart, G. M. Heymsfield, D. L. Hlavka, P. E. Racette, L. Tian, M. A. Vaughan, and D. M. Winker (2004), Combined lidar-radar remote sensing: Initial results from CRYSTAL-FACE, J. Geophys. Res., 109, D07203, doi:10.1029/2003JD004030.
Minnis, P., D. P. Garber, D. F. Young, R. F. Arduini, and Y. Takano (1998), Parameterization of reflectance and effective emittance for satellite remote sensing of cloud properties, J. Atmos. Sci., 55, 3313–3339, doi:10.1175/1520-0469(1998)055<3313:PORAEE>2.0.CO;2.
Minnis, P., D. R. Doelling, L. Nguyen, W. Miller, and V. Chakrapani (2008a), Assessment of the visible channel calibrations of the TRMM VIRS and MODIS on Aqua and Terra, J. Atmos. Oceanic Technol., 25, 385–400, doi:10.1175/2007JTECHA1021.1.
Minnis, P., C. R. Yost, S. Sun-Mack, and Y. Chen (2008b), Estimating the physical top altitude of optically thick ice clouds from thermal infrared satellite observations using CALIPSO data, Geophys. Res. Lett., 35, L12801, doi:10.1029/2008GL033947.
Minnis, P., et al. (2011), CERES Edition-2 cloud property retrievals using TRMM VIRS and Terra and Aqua MODIS data, Part I: Algorithms, IEEE Trans. Geosci. Remote Sens., 49, 4374–4400, doi:10.1109/TGRS.2011.2144601.
Platnick, S. (2000), Vertical photon transport in cloud remote sensing problems, J. Geophys. Res., 105, 22,919–22,935, doi:10.1029/2000JD900333.
Platnick, S., and S. Twomey (1994), Determining the susceptibility of cloud albedo to changes in droplet concentrations with the Advanced Very High Resolution Radiometer, J. Appl. Meteorol., 33, 334–347, doi:10.1175/1520-0450(1994)033<0334:DTSOCA>2.0.CO;2.
Rosenfeld, D., and I. M. Lensky (1998), Satellite based insights into precipitation formation processes in continental and maritime convective clouds, Bull. Am. Meteorol. Soc., 79, 2457–2476, doi:10.1175/1520-0477(1998)079<2457:SBIIPF>2.0.CO;2.
Rosenfeld, D., E. Cattani, S. Melani, and V. Levizzani (2004), Considerations on daylight operation of 1.6-versus 3.7-mm channel on NOAA and Metop satellites, Bull. Am. Meteorol. Soc., 85, 873–881, doi:10.1175/BAMS-85-6-873.
Sherwood, S. C. (2002), Aerosols and ice particle size in tropical cumulonimbus, J. Clim., 15, 1051–1063, doi:10.1175/1520-0442(2002)015<1051:AAIPSI>2.0.CO;2.
Sherwood, S. C., V. T. J. Phillips, and J. S. Wettlaufer (2006), Small ice crystals and the climatology of lightning, Geophys. Res. Lett., 33, L05804, doi:10.1029/2005GL025242.
Smith, W. L., and C. M. R. Platt (1978), Comparison of satellite-deduced cloud heights with indications from radiosonde and ground-based laser measurements, J. Appl. Meteorol., 17, 1796–1802, doi:10.1175/1520-0450(1978)017<1796:COSDCH>2.0.CO;2.
Stamnes, K., S.-C. Tsay, W. J. Wiscombe, and K. Jayaweera (1988), Numerically stable algorithm for discrete-ordinate-method radiative transfer in multiple scattering and emitting layered media, Appl. Opt., 27, 2502–2509, doi:10.1364/AO.27.002502.
Stephens, G. L., et al. (2002), The CloudSat mission and the A-TRAIN: A new dimension to space-based observations of clouds and precipitation, Bull. Am. Meteorol. Soc., 83, 1771–1790, doi:10.1175/BAMS-83-12-1771.
Stone, R. S., G. L. Stephens, C. M. R. Platt, and S. Banks (1990), The remote sensing of thin cirrus cloud using satellites, lidar and radiative transfer theory, J. Appl. Meteorol., 29, 353–366, doi:10.1175/1520-0450(1990)029<0353:TRSOTC>2.0.CO;2.
Tian, B., B. J. Soden, and X. Wu (2004), Diurnal cycle of convection, clouds, and water vapor in the tropical upper troposphere: Satellites versus a general circulation model, J. Geophys. Res., 109, D10101, doi:10.1029/2003JD004117.
Waliser, D. E., et al. (2009), Cloud ice: A climate model challenge with signs and expectations of progress, J. Geophys. Res., 114, D00A21, doi:10.1029/2008JD010015.
Wang, X., K. N. Liou, S. S. C. Ou, G. G. Mace, and M. Deng (2009), Remote sensing of cirrus cloud vertical size profile using MODIS data, J. Geophys. Res., 114, D09205, doi:10.1029/2008JD011327.
Winker, D. M., W. H. Hunt, and M. J. McGill (2007), Initial performance assessment of CALIOP, Geophys. Res. Lett., 34, L19803, doi:10.1029/2007GL030135.
Yang, P., L. Zhang, G. Hong, S. L. Nasiri, B. A. Baum, H.-L. Huang, M. D. King, and S. Platnick (2007), Differences between Collection 004 and 005 MODIS ice cloud optical/microphysical products and their impact on radiative forcing simulations, IEEE Trans. Geosci. Remote Sens., 45, 2886–2899, doi:10.1109/TGRS.2007.898276.
Yang, P., G. Hong, G. W. Kattawar, P. Minnis, and Y. X. Hu (2008a), Uncertainties associated with the surface texture of ice particles in satellite-based retrieval of cirrus clouds: Part II. Effect of particle surface roughness on retrieved cloud optical thickness and effective particle size, IEEE Trans. Geosci. Remote Sens., 46, 1948–1957, doi:10.1109/TGRS.2008.916472.
Yang, P., G. W. Kattawar, G. Hong, P. Minnis, and Y. X. Hu (2008b), Uncertainties associated with the surface texture of ice particles in satellite-based retrieval of cirrus clouds: Part I. Single-scattering properties of ice crystals with surface roughness, IEEE Trans. Geosci. Remote Sens., 46, 1940–1947, doi:10.1109/TGRS.2008.916471.
Yuan, T. L., and Z. Q. Li (2010), General macro- and micro-physical properties of deep convective clouds as observed by MODIS, J. Clim., 23, 3457–3473, doi:10.1175/2009JCLI3136.1.
Zhang, Z., S. Platnick, P. Yang, A. K. Heidinger, and J. M. Comstock (2010), Effects of ice particle size vertical inhomogeneity on the passive remote sensing of ice clouds, J. Geophys. Res., 115, D17203, doi:10.1029/2010JD013835.

Citing Literature

Volume117, IssueD6

27 March 2012

Estimating effective particle size of tropical deep convective clouds with a look-up table method using satellite measurements of brightness temperature differences

Abstract

Key Points

1. Introduction

2. Look-Up Table and Methodology

3. Application and Discussions

3.1. Comparison With CERES-MODIS-Retrieved R_e

3.2. Comparison With CloudSat-Retrieved R_e

3.3. Effect of Variations in DCC τ on LUT-Estimated R_e

4. Summary

Acknowledgments

References

Citing Literature

Figures

References

Information

RESOURCES

PUBLICATION INFO

About Wiley Online Library

Help & Support

Opportunities

Connect with Wiley

Estimating effective particle size of tropical deep convective clouds with a look-up table method using satellite measurements of brightness temperature differences

Abstract

Key Points

1. Introduction

2. Look-Up Table and Methodology

3. Application and Discussions

3.1. Comparison With CERES-MODIS-Retrieved Re

3.2. Comparison With CloudSat-Retrieved Re

3.3. Effect of Variations in DCC τ on LUT-Estimated Re

4. Summary

Acknowledgments

References

Citing Literature

Figures

References

Related

Information

RESOURCES

PUBLICATION INFO

3.1. Comparison With CERES-MODIS-Retrieved R_e

3.2. Comparison With CloudSat-Retrieved R_e

3.3. Effect of Variations in DCC τ on LUT-Estimated R_e