1. Introduction
Tropical storm anvils are very interactive with solar and infrared (IR) radiation. It is common knowledge that these anvils reflect a large fraction of the incident solar radiation back to space and radiate in the IR at a colder temperature than the surface, but, like other cloud types, their optical and microphysical properties and global amounts are not well known (e.g., Gates et al. 1999). The rate at which the ice crystals fall and eventually dissipate is also a very important but little understood factor in modeling the radiative and dynamic properties of storm anvils and their lifetimes, and in estimating a radiation balance.
The storm anvil observations described in this paper were taken as part of the Maritime Continent Thunderstorm Experiment (MCTEX), held at a tropical site on the Tiwi Islands north of Darwin, Northern Territory, Australia. The main aim of MCTEX was to study the life cycle of the intense storms (“Hectors”) that build up daily on the islands before the start of the monsoon season (e.g., Keenan et al. 1990, 1994, 2000). In this paper, we describe light detecting and ranging (lidar) and infrared radiometer observations on four storm anvils that were advected over the observational site. The experiment also included various radars (Sekelsky et al. 1999) and a microwave radiometer that measured water vapor path and liquid water path. In Part I of this article, Platt et al. (2002, hereafter Part I) describe the properties of synoptic-type cirrus and detection of small particles in high, cold cirrus during MCTEX.
Sekelsky et al. (1999) also studied the properties of MCTEX storms using millimeter and S-band radar to derive values of radar reflectivity, vertical velocity, and effective radius. The present observations complement these radar anvil observations by studying in detail the part of the anvil near cloud base that was evaporating strongly. Employing the lidar radiometer (LIRAD) technique, an apparent reduction in emittance, and thus temperature, is found above cloud base. This paper concentrates on this phenomenon. The backscatter-to-extinction ratio of the cloud particles was also found to vary as the anvil dissipated.
Previous work on evaporation and cooling at cloud base and formation of mammatus is sparse and little advanced from the early work of Ludlam and Scorer (1953) and Ludlam (1980). However, more recent work by Martner (1995) and Winstead et al. (2001) using Doppler radar has confirmed the early hypotheses of anvil evaporation and cooling causing destabilization of the atmosphere near cloud base (although other mechanisms such as gravity waves could also contribute). Thus, it appears that the new results obtained in MCTEX could shed some further light on the problem.
2. Instrumentation and observations
Observations were made at a site near the Pirlangimpi Golf Club (11.40°S, 130.41°E), Pularumpi, Melville Island, Northern Territory, Australia. The Commonwealth Scientific and Industrial Research Organisation (CSIRO) operated a three-wavelength, dual-polarization lidar and an IR fast filter radiometer. An Atmospheric Radiation Measurement (ARM) program microwave radiometer was used to measure water vapor path and liquid water path continuously. A two-frequency millimeter radar (Sekelsky and McIntosh 1996) was operated a few meters away from the lidar trailer by the University of Massachusetts. The CSIRO instruments were operational from 18 November to 8 December 1995. Details of the site, the above instruments, and the CSIRO instrument calibrations can be found in Part I.
The present paper presents observations on four storm anvils that occurred during the period 27 November to 7 December 1995. Typically, the anvils were advected over the observation site from island storm cells during the afternoon and evening periods and progressively covered an appreciable fraction of the sky before eventually dissipating. In total, 18 h of anvil observations were obtained.
Lidar backscatter profiles, each being the average of up to 256 individual profiles, were obtained at intervals of either 1 min or 30 s. The IR radiance was measured continuously by the CSIRO/ARM Mark I radiometer and recorded at 1-s intervals. The radiance was calibrated against blackbodies at liquid nitrogen and stirred ice–water temperatures.
3. Theory and analysis
Theory and analysis methods have been summarized in Part I and described in detail in Platt et al. (1998, hereafter P98).
The quantities of interest to this paper are the absorption emittance ϵa and the cloud lidar integrated attenuated backscatter γ′(π). These are calculated in a similar manner to other (nonanvil) cirrus clouds. Methods of obtaining these properties through the measured values of attenuated backscatter coefficient
The IR absorption coefficient σa is related to the visible extinction coefficient σc through the quantity α in (6) of P98. (The ratio α is known to have a value of about 2 for the large particles that occur at cloud base, as shown in Part I.) The multiple scattering factor
4. Results
The four storm anvils observed in this study are described below, and the analyzed values of ϵa and γ′(π) are presented and discussed.
a. Storm anvil on 27 November
This case represents the typical cycle of storm anvil fallout and dissipation during one afternoon and evening period. The lidar time–height backscatter image is shown in Fig. 1a. This illustrates the typical rapid decrease in height of the anvil base and the subsequent leveling off between about 7 and 9 km. There is evidence of mammata/stalactites between about 1700 and 2000 local time (LT) (LT = UTC + 9 h, 30 min), signifying intense cooling near cloud base (e.g., Ludlam and Scorer 1953). An equivalent 9.05-mm radar image (Austin et al. 1999; Sekelsky et al. 1999) is shown in Fig. 1b. The radar-measured cloud top remained at 14–15 km until about 1700 LT, when it descended to 12 km and remained there until about 1930 LT, after which there was a steady further descent. These heights are seen to be well above the lidar-measured cloud-top heights (the latter defined here as “effective cloud tops”), confirming the strong attenuation of the lidar pulse. The radar-measured cloud-base height, however, follows that of the lidar quite closely.
The observed and analyzed quantities are shown in Figs. 2a–e. Figure 2a shows the measured IR radiance Ls [(10) in P98] and total water vapor path. Figure 2b shows the calculated cloud radiance after removal of water vapor effects. This is the radiance Lc before scattering components are subtracted [(10) and (11) in P98]. The radiance increases as the cloud base falls and becomes warmer, and then levels out, as expected. Finally, the radiance fluctuates considerably as the cloud base breaks up. The midcloud temperature shown in Fig. 2c applies only to the cloud depth measured by the lidar. Consequently, it is seen to increase and to follow much the same fluctuations as the cloud radiance. The calculated cloud absorption emittance is shown in Fig. 2d. Before 1530 LT, the lowered emittance and integrated backscatter (Fig. 2e) imply that the cloud is semitransparent. After about 1600 LT, the emittance becomes unity, indicating that the LIRAD calculation is working correctly with the available temperature profile and that the cloud has become optically thick. However, although the lidar attenuation remains strong, as indicated by the radar cloud depth, the emittance after 1600 LT falls unexpectedly to a value of about 0.8, rising gradually to 0.9. The anvil is seen to finally break up after about 2000 LT. Figure 2e shows the values of γ′(π) over the same time interval. The variations in γ′(π) are not entirely random with time, but rather show some systematic, but fluctuating, decreases during the time that the cloud was fully attenuating.
As discussed in section 5, the apparent reduction in emittance is shown to be due to a cooling of the layers just above cloud base. We define the reduced emittance here as the “effective emittance,” ϵe.
b. Storm anvils on 28 November, 6 December, and 7 December
For the three other cases studied, thunderstorms and squall lines occurred in the vicinity of the site, leaving high anvil clouds in their wake. The time–height images of such anvils are shown in Figs. 3a, 4a, and 5a. Mammata/stalactites are particularly apparent in Fig. 5a. The calculated emittance is shown in Figs. 3b, 4b, and 5b, and integrated attenuated backscatter, γ′(π), in Figs. 3c, 4c, and 5c. The images illustrate the rather different patterns of dissipation and fallout, on different days, which were dependent on where and when the initial storm occurred. On each day, the millimeter radar time–height image of reflectivity was used to gauge when the anvil was fully attenuating. On 28 November, the cloud above cloud base was only fully attenuating for about the first 30 min, when the cloud base lifted and the cloud became semitransparent. The radar had indicated precipitation 2 h earlier, with the cloud base rising and the cloud top decreasing from about 1800 to 2000 LT, when the lidar–radiometer observations commenced. Values of effective emittance ϵe are seen to be about 0.8 for the first 30 min, when the cloud was still fully attenuating. This value is similar to the value after 1600 LT in Fig. 2d. Over the same period, the integrated backscatter showed a slow decrease. Figure 3a shows an upper layer that gradually descends with time, but the anvil base stays constant at about 10 km.
The anvil on 6 December, shown in Fig. 4a, was fully attenuating for about the first 40 min, and during this time ϵe fluctuated between about 0.65 and 0.75, after which it decreased steadily to quite low values. Values of γ′(π) were initially quite high, with a maximum of about 0.75. Such high values would indicate detection of some oriented platelike crystals. This anvil was obviously close to breaking up when observations commenced. High layers near the tropopause appeared from about 1500 LT onward as the cloud optical depth decreased and the cloud became semitransparent.
The final anvil on 7 December showed some typical regions of descending crystals combined with intense backscatter. The millimeter radar image showed a cloud-top height of 14 km until 1545 LT, after which the top gradually decreased, but with some fluctuations. The lidar pulse was, therefore, completely attenuated until at least 1600 LT. Once again, the effective emittance in Fig. 5b shows a value of 0.8, increasing at times to 0.95, despite fluctuations in the height of the emitting cloud layer during this time. Values appeared to be lowest near the lowest parts of the mammata, where maximum cooling might be expected. Values of γ′(π) fluctuated systematically, including a peak of high values at about 1520 LT, indicating again the presence of oriented crystals. Apart from this peak, γ′(π) is seen to decrease slowly with time.
c. Comparison of the measured effective emittance values with previous observations
Earlier observations of dissipating storm anvils at Darwin in 1981 by Platt et al. (1984) provided a useful comparison with the present results. Of the four known fully attenuating (for at least part of the time) anvil cloud episodes on different days, the (effective) emittance never rose above about 0.85, and was below 0.8 in three of the cases, showing very similar behavior to the MCTEX cases. This feature had not been pursued previously. A histogram of emittance observations for the four Darwin anvils is shown in Fig. 6. The values of emittance lower than about 0.6 indicate periods when the absorption emittance was indeed less than unity. The anvils were observed in the same fashion as in MCTEX with the LIRAD method, and the analysis procedure was similar, except that microwave observations of water vapor path were not available, and radiosonde data were used instead. There were also extensive periods when the anvils were semitransparent and, therefore, the “correct” emittance was 0.6 or less. Although millimeter radar data were not available, molecular backscatter from above cloud top could be detected. This would explain the values of ϵa below about 0.6 that occur in the histograms of Fig. 6. That this separate set of data showed evidence of depressed effective emittance at the base of tropical storm anvils, similar to that in MCTEX, would indicate that depressed effective emittance might at least be a common feature of tropical storm anvils.
5. Interpretation and analysis of effective emittance
The occurrence of the depressed effective emittance during episodes of strong lidar attenuation appeared to be so prevalent, both in the MCTEX and earlier Darwin 1981 data, that the phenomenon deserved further analysis. The nature and vertical extent of the radiating cloud layers (i.e., those layers whose radiation reaches cloud base) are considered first, to remove the possibility that the depressed emittance might be an artifact of the analysis. A simple model to explain at least some of the cooling is then discussed, together with cooling calculations from the millimeter radar data.
a. Emission from layers above cloud base
The measured absorption radiance at cloud base is the integral of emission from cloud layers that are sufficiently close to cloud base that their emitted IR radiation is not completely attenuated by the intervening cloud layers. In the case of the attenuating anvils, as shown in Fig. 1, the lidar pulse is attenuated completely far below the radar-measured cloud top. These regions of complete attenuation are the regions having an effective emittance that is less than unity. It is important, therefore, to determine whether the IR emission reaching cloud base is solely due to the effective lidar layer, or whether a significant proportion of IR radiation emitted from above this layer also reaches cloud base. With that information, the apparent cooling of the layers can then be estimated.
To examine the emitting layer in detail, we consider a typical case at 1730 LT on 27 November. A selection of retrieved profiles of cloud backscatter coefficient at various times indicates that the profile at 1730 LT (Fig. 7) is quite representative. It should be noted that the retrieved profile of Bc(π) is not accurate, except just above cloud base, because of the attenuating nature of the cloud and the necessity for a very accurate value of the backscatter-to-extinction ratio to retrieve an accurate profile. The value of Bc(π) is seen to decrease toward effective cloud top at 8.7 km, a typical effect in retrievals of attenuating clouds (Platt 1979). The value is also forced to zero at effective cloud top. Note that, because the emission can be shown to originate mainly from the lower levels of the cloud, it is not essential for our purposes that the backscatter profile is accurately retrieved over the whole depth of the cloud. We know that the cloud extends to about 13 km in reality, but we are interested in the backscatter coefficient Bc(π) only in that region where the IR emission reaches cloud base.
The effective lidar-layer infrared characteristics were calculated using (1) and the other required parameters shown in Table 1. As discussed in section 3, the ratio α is assumed to have a value of 2 for the large particles that occur at cloud base. The radiant emission from different layers in the cloud is shown in Fig. 7. This emission is forced to zero at effective cloud top. The calculations were made using the measured radiosonde temperature profile. Resultant values of absorption emittance ϵax and cloud radiance Latx using the α = 2 assumption are shown in Table 1. The value of ϵax is found to be 0.87, with a value of radiance Latx equal to 1.56 W m−2 sr−1. The effective emittance ϵae (using LIRAD-iterated α and k) is 0.84, giving a measured radiance of 1.506 W m−2 sr−1. Thus, although the calculated value of the emittance is less than unity, the measured value (which should be close to unity, as at 1600 LT) is even lower. It is also evident that emission is coming from the layers above the lidar layer, or at least from a lidar layer that is correctly retrieved where the backscatter coefficient does not fall artificially to zero at 8.7 km. To simulate the effect of such a layer, a layer of backscatter coefficient equal to 1.5 km−1 was added from 8.0 to 9.8 km, as shown in Fig. 8. A profile of millimeter radar reflectivity taken at 1730 LT (Fig. 9) indicates that this is a reasonable approximation. The calculated emittance ϵaxc of the combined layers is 0.992, or nearly black, indicating that very little emission from above 9.8 km is reaching cloud base. The calculated maximum radiance Latxc that can reach cloud base with this model is 1.75 W m−2 sr−1, which is much greater than the measured radiance La, as shown in Table 1. Thus, again the measured radiance shows that the temperatures above cloud base must be less than the radiosonde values. The effective emittance ϵaec is now 0.86, slightly higher than before because the emission is coming from colder layers above 8.7 km. Thus, in order to calculate effective values of cloud cooling above the base, the entire layer up to 9.8 km needs to be considered. Of course, adding a layer with a different backscatter coefficient will give a different answer. A layer of backscatter coefficient of 1 km−1, instead of 1.5 km−1, between 8.3 and 9.8 km yields an emittance of 0.98, which is not very different. The estimation of cooling of the layers above the cloud base due to the measured depressed effective emittance is given below.
b. Apparent reduction in temperature of layers above cloud base
In order to obtain quantitative values for the depression of the temperature in the layers above cloud base, three somewhat artificial cooling or cloud evaporation scenarios have been considered. These are shown in Table 2 in terms of cooling ΔT below the radiosonde temperature T(z). In the calculation of cooling, ΔT is varied until the known measured radiance of 1.506 W m−2 sr−1 is obtained. The parameters in Table 1 remain the same. Each scenario gives different calculated values of ΔT. For the linear reduction in cooling above cloud base, a very strong and unrealistic temperature change of 28.5°C is calculated at cloud base. In the other two cases, the cooling is much lower and decreases as the thickness of the evaporating layer increases. In the case of a cooling layer extending up to 9.8 km, a minimum value ΔTm2 is attained and would apply even if the entire anvil vertically had cooled to this value. This is because little emission reaches the anvil cloud base from above 9.8 km.
Values of ΔTm4 estimated using the above procedure are shown in Table 3 for the three other anvils studied. The periods listed in this table were selected such that the anvils were (approximately) horizontally level and fully attenuating within 1–2 km above cloud base. There is a consistent cooling at the various anvil bases of 6°–18°C.
It could be argued that the observed cooling was simply due to a change in the “clear air” temperature profile. However, three radiosonde ascents at 1100, 1400, and 2000 LT on 27 November showed no temporal variation of atmospheric temperature greater than about 0.6°C between 6 and 8 km.
6. Calculations of cooling from a simple model and from radar data
Wagner (1948), Ludlam and Scorer (1953), and Warner (1973) hypothesized that mammata at the base of anvil clouds are formed, at least partially, by the chilling of clear air just below cloud base from the evaporation of small but copious particles. Martner (1995) found that Doppler radar observations of mammata gave ample evidence that evaporation was a significant factor in the formation of such mammata. The present results thus confirm these findings. To examine this cooling in more detail, we consider a simple model of cooling by evaporation of the ice water content (IWC) at and above cloud base to compare with estimated values, ΔTm, of cooling obtained from observations. An investigation of evaporative cooling from radar-deduced IWC is also discussed.
a. Model of cloud layer cooling
A rigorous 2D model including cloud microphysics could be used to estimate the cooling rate, but such a study is beyond the scope of this article. The anvil base in the dissipative phase was situated typically between 6- and 8-km altitude. The structure of the bottom of the anvil from both lidar (Fig. 7) and radar (Fig. 9) profiles, with rapid falloff in reflectivity toward the cloud base, would indicate the presence of strong evaporation and, therefore, cloud cooling, as noted in section 4a. Cloud particles ejected near the top of the storm cloud and advected horizontally into the anvil will eventually start falling, evaporate into the drier air beneath, and, therefore, moisten the layers into which they fall. The degree of subsaturation, together with the temperature, determines the maximum cooling that can result before saturation of the layer occurs. In this model, the difference between the ice saturation vapor density and the ambient vapor density places an upper limit on the cooling that can occur in the parcel. Because the cloud-base altitude remains fairly constant, drier air must continuously replace the moistened air at and for some distance above cloud base, presumably by the action of wind shear and turbulence. The emitted radiance from which the emission temperature is measured must come from the evaporating ice particles themselves, but in an equilibrium situation we assume that cloud particle temperature is close to the temperature of the air in that layer.
Both the observed and theoretical cooling, and thus the amount of sublimated IWC, is seen to decrease with decreasing temperature. These lower values represent cooling when the anvil base was higher, as shown in Fig. 10. Calculations show that the value of ΔT in (5) becomes smaller rapidly with decreasing temperature, as shown in Fig. 11. The air then has only a small capacity to hold the evaporated ice. This explains the increase in the values of ΔTm4 in Fig. 10 after 1600 LT. Before 1600 LT, the cloud base was even higher, and little cooling at the base appeared to be occurring. The cooling ΔTm4 gradually decreased with time as the anvil evaporated. It should be noted that before about 1530 LT, the lidar pulse was penetrating to cloud top, and the decrease in ϵa before 1530 LT was a real decrease due to the anvil depth being thin as it first covered the zenith. There was some indication of increased cooling in the protruding mammata at 1700 LT and again just before 1830 LT.
b. Estimation of cooling from radar-derived anvil ice water content
Sekelsky et al. (1999) measured 33-GHz millimeter radar reflectivity, Ze, simultaneously with the lidar and radiometer data. A time–height image of radar reflectivity for the 27 November anvil is shown in Fig. 1b, and a profile taken at 1730 LT is shown in Fig. 9.
In Table 4, the cloud-base temperatures corresponding to the times in column 1 are shown in column 2. Values of maximum effective radar reflectivity Ze and experimental values of ΔTm are shown in columns 3 and 4, respectively. The resultant values of Do are given in column 5. The range of values corresponds to the range of values of A that is used in (6). Corresponding values of W calculated from (6) are shown in column 6.
Column 7 of Table 4 shows the values of Do that were estimated from the lidar–radar method. This method retrieves an effective particle diameter, De, by obtaining the ratio of the radar reflectivity to the lidar backscatter coefficient. The lidar/radar ratio has a dependence of
The values of Do from the lidar–radar method agree with those from the radar and infrared observations within about a factor of two. Uncertainties arise from several sources. Sekelsky et al. (1999) quote formulas relating IWC and radar reflectivity from several other authors that give varying values of IWC. Because the present data were for the same set of storm anvils as that studied by Sekelsky et al. (1999), we have used their expressions. As they pointed out, their expressions have the advantage of containing a mean volume diameter that can be retrieved by the lidar/radar method.
Again, in the lidar–radar method, Austin et al. (1999) used a two-component cloud particle size distribution, whereas Sekelsky et al. (1999) used a modified gamma distribution. Further, Austin et al. (1999) treated the cloud particles as spheres in the radar reflectivity, causing a small error. However, representative values were used for lidar backscatter from ice crystals. The total uncertainty in the comparisons in terms of the value of Do is thus estimated as a factor of about two. Of course, as discussed in section 5a, the amount of ice evaporated is limited by the state of saturation of the atmospheric layers. Thus, the high values of Do in the range shown in column 7 (Table 4) are probably unrealistic.
The real situation is obviously much more complex. Large-scale subsidence and strong downdrafts associated with mammata were not considered here but may be important. The formation of downdrafts in the anvil coupled with incomplete warming of the crystals as they fall through the cloud could also contribute (e.g., Ludlam and Scorer 1953). As stated before, a more complex and accurate model of cooling is beyond the scope of this article.
c. Behavior of the backscatter-to-extinction ratio
A decrease in k/2
7. Discussion and conclusions
The LIRAD method has been used to investigate the evolution of anvil cloud bases and associated cooling processes. Although the observed anvil clouds were quite black to IR radiation initially, dissipation of the anvils led to rapid thinning and decay of the cloud layers, with resultant reduction in cloud optical depth and penetration of the lidar beam through the cloud layers. The temperature at the anvil cloud base has been measured for the first time by the LIRAD method. A significant cooling of about 8°–10°C was found compared with the noncloudy atmosphere when the anvil was completely attenuating to the lidar pulse. The cooling was detected through a reduction in the apparent cloud emittance below unity. This cooling was a feature of all the anvils studied. A calculation of radiative transfer through the cloud showed that 90% of the IR emission at cloud base came from the same layer as that defined by the attenuated lidar pulse. A correction was made for the 10% emitted from above that layer. A further investigation of some anvils studied previously at Darwin by Platt et al. (1984) indicated identical phenomena. Thus, the process seems to be a common feature of isolated tropical storm anvils. These anvils are advected outward at high altitudes and fall subsequently into drier air beneath. At some altitude, the atmosphere is sufficiently dry to allow evaporation of all the remaining ice water content at the anvil base. This condition defines the height of the cloud base. A model based on the above premise yielded calculated values of cloud cooling that were comparable, or slightly lower, than calculated observed values. The measured radar reflectivity at 9.05 mm also yielded values of ice water content and, thus, cloud-base cooling. The radar data indicated rapid evaporation or sublimation of the ice water content near cloud base, in agreement with the model. The lidar/radar method was used to obtain the mean volume diameter Do of the cloud particle size distribution, necessary for calculating the ice water content. By varying Do, the evaporation and subsequent cooling from radar data were varied. When this cooling equaled that from the radiometric method, the values of Do agreed, within experimental error, with those obtained from the radiometric method.
The agreement, within experimental uncertainty, between observed and calculated cooling with the simple model of evaporation confirms the early hypotheses of Wagner (1948) and Ludlam and Scorer (1953) that evaporation of the anvil particles and consequent saturation of the layers at cloud base play a large part in the subsequent cloud dynamics. The cooling destabilizes the atmosphere and allows formation of mammatus. Although we did not look closely for such phenomena, there was evidence of their presence in the lidar data. The more recent observations of Martner (1995) and Winstead et al. (2001) are also consistent with evaporation of cloud particles as a dominant factor. Martner states that the radar data provide ample evidence that evaporation was significant in the mammatus region of the anvil that they studied. The decrease in reflectivity indicated shrinking particle sizes from evaporation. Martner (1995) further found that reflectivity and velocity undulations occurred to almost 1 km above cloud base. He considered that air would be saturated at these levels, but the results in the present paper were consistent with cooling to at least 1 km above cloud base. The present data thus confirm many of the features found or hypothesized by others in the past and add to our knowledge by allowing the measurement of cooling at and above cloud base by the lidar radiometer (LIRAD) method.
Acknowledgments
The authors wish to acknowledge the following: The Tiwi Islands Land Council for permission to make observations on Melville Island; our colleagues at Aspendale for assistance, particularly in the Engineering and Electronics Workshops; the Bureau of Meteorology, Melbourne; and Monash University for supplying radiosonde data. Stephen Marsden made a preliminary analysis of the data. U. S. Department of Energy, Office of Health and Environmental Research, Grants DE-FG02-92ER61373 and DE-FG03-94ER61748 funded the field observations and part of the salaries. We also acknowledge the contributions of two reviewers whose valuable suggestions and criticisms have made this a better paper.
REFERENCES
Austin, R. T., S. A. Young, C. M. R. Platt, G. R. Patterson, S. M. Sekelsky, and R. E. McIntosh, 1999: Retrieval of tropical cirrus cloud properties from ground-based lidar and millimeter-wave radar sensing at the Maritime Continent Thunderstorm Experiment. Proc. Seventh Atmospheric Radiation Measurement (ARM) Science Team Meeting, San Antonio, TX, U.S. Department of Energy, 389–393.
Gates, W. L., and Coauthors. 1999: An overview of the results of the Atmospheric Model Intercomparison Project (AMIP I). Bull. Amer. Meteor. Soc., 80 , 29–55.
Keenan, T. D., B. R. Morton, Xu Shu Zhang, and K. Nyguen, 1990: Some characteristics of thunderstorms over Bathurst and Melville Islands, near Darwin, Australia. Quart. J. Roy. Meteor. Soc., 116 , 1153–1172.
Keenan, T. D., and Coauthors. 1994: Science plan for the Maritime Continent Thunderstorm Experiment (MCTEX). BMRC Research Rep. 44, 61 pp. [Available from Bureau of Meteorology, BMRC, P. O. Box 1289k, Melbourne, Victoria 3001, Australia.].
Keenan, T. D., and Coauthors. 2000: The Maritime Continent Thunderstorm Experiment (MCTEX): Overview and some results. Bull. Amer. Meteor. Soc., 81 , 2433–2455.
Ludlam, F. H., 1980: Clouds and Storms: The Behavior and Effect of Water in the Atmosphere. Pennsylvania State University Press, 405 pp.
Ludlam, F. H., and R. S. Scorer, 1953: Convection in the atmosphere. Quart. J. Roy. Meteor. Soc., 79 , 317–341.
Martner, B. E., 1995: Doppler radar observations of mammatus. Mon. Wea. Rev., 123 , 3115–3121.
Platt, C. M. R., 1979: Remote sounding of high clouds. I: Calculation of visible and infrared optical properties from lidar and radiometer measurements. J. Appl. Meteor., 18 , 1130–1143.
Platt, C. M. R., 1981: Remote sounding of high clouds. Part III: Monte Carlo calculations of multiple scattered lidar returns. J. Atmos. Sci., 38 , 156–167.
Platt, C. M. R., 1997: A parameterization of the visible extinction coefficient of ice clouds in terms of the ice/water content. J. Atmos. Sci., 54 , 2083–2098.
Platt, C. M. R., A. C. Dilley, J. C. Scott, I. J. Barton, and G. L. Stephens, 1984: Remote sounding of high clouds. V: Infrared properties and structure of tropical thunderstorm anvils. J. Climate Appl. Meteor., 23 , 1296–1308.
Platt, C. M. R., J. C. Scott, and A. C. Dilley, 1987: Remote sounding of high clouds. VI: Optical properties of midlatitude and tropical cirrus. J. Atmos. Sci., 44 , 729–747.
Platt, C. M. R., S. A. Young, P. J. Manson, G. R. Patterson, S. C. Marsden, R. T. Austin, and J. Churnside, 1998: The optical properties of equatorial cirrus from observations in the ARM Pilot Radiation Observation Experiment. J. Atmos. Sci., 55 , 1977–1996.
Platt, C. M. R., R. T. Austin, S. A. Young, and G. R. Patterson, 2002: LIRAD observations of tropical cirrus clouds in MCTEX. Part I: Optical properties and detection of small particles in cold cirrus. J. Atmos. Sci., 59 , 3145–3162.
Sekelsky, S. M., and R. E. McIntosh, 1996: Cloud observations with a polarmetric 33 GHz and 95 GHz radar. Meteor. Atmos. Phys., 58 , 123–140.
Sekelsky, S. M., W. L. Ecklund, J. M. Firda, K. S. Gage, and R. E. McIntosh, 1999: Particle size estimation in ice-phase clouds using multifrequency radar reflectivity measurements at 95, 33, and 2.8 GHz. J. Appl. Meteor., 38 , 5–28.
Wagner, F., 1948: Mammatusform als anzeichen Absinkbewegung in Wolkluft. Ann. Meteor., 1 , 336.
Warner, C., 1973: Measurements of mamma. Weather, 28 , 394–397.
Winstead, N. S., J. Verlinde, S. T. Arthur, F. Jaskiewicz, M. Jensen, N. Miles, and D. Nicosia, 2001: High-resolution airborne radar observations of mammatus. Mon. Wea. Rev., 129 , 159–166.
Data from a dissipating storm anvil on 27 Nov 1995: (a) 532-nm lidar time–height image of attenuated backscatter, and (b) 9.05-mm radar time–height image of reflectivity
Citation: Journal of the Atmospheric Sciences 59, 22; 10.1175/1520-0469(2002)059<3163:LOOTCC>2.0.CO;2
Data from a dissipating storm anvil on 27 Nov 1995: (a) measured IR radiance at 10.86 ± 0.25 μm (large symbols) and water vapor path (small symbols) from the ARM microwave radiometer; (b) retrieved IR cloud radiance; (c) effective midcloud temperature of the lidar-detected cloud layer; (d) cloud emittance, ϵa; and (e) integrated attenuated backscatter, γ′(π)
Citation: Journal of the Atmospheric Sciences 59, 22; 10.1175/1520-0469(2002)059<3163:LOOTCC>2.0.CO;2
Dissipating anvil on 28 Nov 1995: (a) time–height image of lidar backscatter; (b) cloud emittance, ϵa; and (c) cloud integrated attenuated backscatter, γ′(π)
Citation: Journal of the Atmospheric Sciences 59, 22; 10.1175/1520-0469(2002)059<3163:LOOTCC>2.0.CO;2
Dissipating anvil on 6 Dec 1995: (a) time–height image of lidar backscatter; (b) cloud emittance, ϵa; and (c) cloud integrated attenuated backscatter, γ′(π)
Citation: Journal of the Atmospheric Sciences 59, 22; 10.1175/1520-0469(2002)059<3163:LOOTCC>2.0.CO;2
Dissipating anvil on 7 Dec 1995: (a) time–height image of lidar backscatter; (b) cloud emittance, ϵa; and (c) cloud integrated attenuated backscatter, γ′(π)
Citation: Journal of the Atmospheric Sciences 59, 22; 10.1175/1520-0469(2002)059<3163:LOOTCC>2.0.CO;2
Number of observations for various values of emittance of anvils obtained over Darwin in Mar 1981 (Platt et al. 1984)
Citation: Journal of the Atmospheric Sciences 59, 22; 10.1175/1520-0469(2002)059<3163:LOOTCC>2.0.CO;2
Retrieved cloud backscatter coefficient Bc(π, z) (solid line) and calculated IR emission (broken line) for anvil cloud layer at 1730 LT 27 Nov 1995
Citation: Journal of the Atmospheric Sciences 59, 22; 10.1175/1520-0469(2002)059<3163:LOOTCC>2.0.CO;2
Same as Fig. 7, but with artificial values of Bc(π, z) introduced from 8.0 km to fixed cloud top at 9.8 km
Citation: Journal of the Atmospheric Sciences 59, 22; 10.1175/1520-0469(2002)059<3163:LOOTCC>2.0.CO;2
Profile of cloud radar reflectivity (9.05 mm) at 1730 LT during the 27 Nov 1995 anvil, showing the rapid decrease in reflectivity near cloud base
Citation: Journal of the Atmospheric Sciences 59, 22; 10.1175/1520-0469(2002)059<3163:LOOTCC>2.0.CO;2
Measured values of cloud-base cooling ΔTm4 as a function of time for the anvil of 27 Nov 1995
Citation: Journal of the Atmospheric Sciences 59, 22; 10.1175/1520-0469(2002)059<3163:LOOTCC>2.0.CO;2
Calculated values of cloud-base cooling, ΔT, for various cloud-base temperatures and atmospheric humidity below cloud base (solid lines). Values were calculated from radiosonde data observed at 1158 LT (before advent of the anvil) 27 Nov 1995. Measured values of cloud-base cooling, ΔTm4, for various cloud-base temperatures are shown (single points)
Citation: Journal of the Atmospheric Sciences 59, 22; 10.1175/1520-0469(2002)059<3163:LOOTCC>2.0.CO;2
Various values of emittance, radiance, and associated quantities (1730 LT 27 Nov)
Estimates of cooling at 1730 LT 27 Nov
Estimated cloud-base cooling for each storm anvil
Comparison of values of mean volume diameter from radar reflectivity and observed IR radiometric values with that retrieved from the lidar/radar method on 27 Nov 1995
