High Ice Water Content at Low Radar Reflectivity Near

Open Access Atmos. Chem. Phys. Discuss., 15, C7168–C7175, 2015 Atmospheric www.atmos-chem-phys-discuss.net/15/C7168/2015/ Chemistry ACPD © Author(s) 2015. This work is distributed under the Creative Commons Attribute 3.0 License. and Physics 15, C7168–C7175, 2015 Discussions

Interactive Comment

Interactive comment on “High ice water content at low radar reﬂectivity near deep convection – Part 1: Consistency of in situ and remote-sensing observations with stratiform rain column simulations” by A. M. Fridlind et al.

A. M. Fridlind et al. [email protected]

Received and published: 24 September 2015 Full Screen / Esc

General Comment. This paper is part of a two-part series focused on the meteoro- Printer-friendly Version logical conditions responsible for jet engine power loss. While it falls in the category of applied-aviation research, it includes enough topics of basic atmospheric physics to Interactive Discussion be of interest to readers of this journal. There are at least two major concerns with this (type of) study. First the ice water content conditions that result in engine failure are Discussion Paper inadequately known, in terms of magnitude and horizontal extent (duration). Second,

C7168 our ability to accurately measure the ice water content is inadequate, especially when the wide range of possible atmospheric values is considered and considering that sizes ACPD and masses of ice particles ranging from about 10 microns to several millimeters must 15, C7168–C7175, 2015 be sampled properly. Together, these two problems limit the utility of this paper and the authors should try to reduce these unknowns as much as possible and, if that is not possible, they should at least try to provide more detailed information than is Interactive presently included in the paper. Below, some suggestions are made in this regard Comment and some problems that should be corrected or addressed are described. If proper improvements to the paper can be made, it should be published. General response: We thank the reviewer for his time and comments. We fully agree that the conditions that result in engine power loss are not adequately known (see response 1 below) and that our ability to measure ice water content is inadequate (already discussed at length in the introduction). More detailed information is provided per point-by-point responses below. Speciﬁc Comments. Section 1. The introductory material provides a very useful overview of the problem and a history of the current situation. However, it would be improved by providing more complete information on what amount of ice ingestion is likely to cause a problem in modern engines. While this may not be known for all engines, there must at least be some minimum value that the authors believe would not be a problem and this information would help the reader better understand which of the sampled clouds are likely to be of interest. A second area which might provide more Full Screen / Esc context for the present study is engine certiﬁcation requirements (FAA, 2014), especially part 33, appendix D, which includes both length scales and magnitude scales for Printer-friendly Version water content, yet these requirements are not referenced in the paper. Interactive Discussion Response 1: It is impossible to name such an IWC amount at least in part because it is

"impossible to rule out the sufficiency of either long or short duration exposure events" Discussion Paper (see p. 16509, line 26); it is also beyond the scope of this paper to discuss the many factors that affect engine response to ice ingestion, which may vary with engine type, C7169 operating conditions, and sequence of exposure. We can, however, usefully add the context of appendix D, revision to p. 16514, line 13: "As described by Grandin et al. ACPD (2014), an Airbus objective was to obtain preliminary measurements to evaluate newly 15, C7168–C7175, 2015 proposed standards for engine performance in glaciated icing conditions (Mazzawy and Strapp, 2007). Conditions sampled included..." Section 2. The origin of eq. (1) is not clear. Contrary to what is stated the text, it does Interactive not appear to be in either Baker and Lawson (2006) or in Lawson et al. (2010). Lawson Comment et al. (2010) used the area-mass relationship from Baker and Lawson (2006), not eq. (1). Response 2a: Equation 1 is correctly transcribed from the first relations shown in Ta- bles 1 and 2 of Baker and Lawson (2006), but we stand corrected re Lawson et al. (2010). Correction to p. 16512, lines 12–16: "mass-dimension" should be "area- dimensional" and delete remainder of sentence and equation after "Dmax". Clarifi- cation to Fig. 1 caption: replace "following Eq. (1) (Baker and Lawson, 2006)" with "following Baker and Lawson (2006, their Table 1)". The focus of this paper is on stratiform anvils associated with deep convection, yet they do not partition the in situ measurements to determine which ones were made in convective regions versus stratiform regions (e.g. in Fig. 2). This is a surprising omission, since regions with active updrafts would likely exhibit the highest water contents and the stated purpose of the measurements was to find and stay within the highest IWC regions possible, within the limits of safety. While the aircraft might not have been Full Screen / Esc equipped for vertical wind measurements, aircraft performance and flight conditions Printer-friendly Version (e.g. vertical accelerations) might be useful proxies to segregate the data into convective and stratiform segments. It would also be helpful to state the safety limits or safety Interactive Discussion criteria, since they are significant experimental constraints on the reported data. Response 2b: Re classification, please see response 1 to reviewer 2. Safety was not Discussion Paper a leading concern during sample legs with this aircraft at the elevations and locations

C7170 targeted. On p. 16514, line 17: ", within the limits of safety" to be deleted. ACPD Section 3. This section, on the Airbus measurements, has several problems that should be addressed before the paper is published: 1. The nephelometer was designed for 15, C7168–C7175, 2015 measurement of liquid droplet size distributions but used here for measurement of ice mass size distribution, yet the performance of the instrument for sampling ice is not discussed. Unlike most airborne cloud physics instruments, details on this instrument Interactive performance are not widely available in the literature, so the authors should devote Comment more effort to describing its performance (e.g. see items 2 and 3 below). Response 3.1: For the lead authors of this manuscript it is a new experience to use data from a proprietary industry instrument for publicly funded research science. We agree that the level of documentation does not match that of most airborne cloud physics instruments, but do not believe that gap can be ﬁlled with a few sentences here. See response 3.2 for additional references reporting this data. See response 3.3 for reference to new measurements from more widely used instruments. 2. The reference on the nephelometer (Roques, 2007) does not adequately describe the instrument, as it does not include information on the sample volume, the effects of out of focus particles, etc. Sample volume is of particular concern, since many of the clouds in this study contained large ice particles, which typically require a large sample volume (e.g. compared to liquid water measuring instruments) to sample effectively.

Response 3.2: Following on response 3.1, at p. 16515, line 13: replace "(Roques, Full Screen / Esc 2007)" with "(Roques, 2007; Dezitter et al., 2013; Grandin et al., 2014)". 3. Of particular concern is that the effects of particle shattering and breakup for the Printer-friendly Version nephelometer are not addressed in this study. Several recent studies (including those Interactive Discussion by the paper’s co-authors!) have documented that particle breakup during sampling has a very signiﬁcant effect on the measured particle size distribution. This has been Discussion Paper demonstrated for traditional optical array instruments (Jackson et al., 2015, and references therein, provides a recent overview of the problem), but it is a problem for other C7171 optical instruments. In particular, larger particles, such as those greater than about 0.5 mm, which were likely present in these clouds, often break into fragments, and these ACPD fragments are likely to contribute signiﬁcantly to the 100 to 500 micron-mode particles 15, C7168–C7175, 2015 which are a major topic of this paper. Without an understanding of the shattering and breakup of particles in the instrument, it is impossible to determine if the observed “self similarities” of particles in the 100 to 500 micron mode are real features of the Interactive clouds or instrumental artifacts of the sampling in conditions where large ice particles Comment are present. This is also a subject of part II of this study (Ackerman et al., 2015), so this is a particularly relevant concern for these papers.

Response 3.3: Owing to the objective of investigating IWC and Ze, here we are able to focus analysis on mass-weighted diameter measures, where shattering effects are relatively reduced. Also following response 3.1, to be added at p. 16520, line 18: "Leroy et al. (2015) more recently report MMDeq typically 250–500 µm and weakly decreasing with increasing IWC over 0.5–3 g m 3 at 36 C in a system extensively sampled near Darwin during the recent High Altitude Ice Crystals / High Ice Water Content campaign, but similar measurements in another system yielding typical MMDeq of 400– 600 µm instead weakly increasing with IWC. Aside, we note that shattering artifacts that may contaminate airborne probe measurements are relatively reduced for higher- order moments of the size distribution such as mass [Korolev et al., 2013; Jackson and McFarquhar, 2014]. Since the Korolev et al. [2013] study was performed for probes with different inlet conﬁgurations, we expect that that general conclusion can be extended Full Screen / Esc to the Airbus nephelometer. It has been found that size distribution measures such as

MMDmax may be subject to roughly 20% uncertainty owing to shattering artifacts, for Printer-friendly Version instance [Jackson and McFarquhar, 2014]." 4. As the author’s point out, the uncertainties in the Robust probe severely limit the Interactive Discussion interpretation of the collected data. The authors also make an assumption that liquid Discussion Paper water contributions at temperatures below -20C are negligible. While that may be true for many cloud conditions (such as the stratiform regions studied in sections 4

C7172 and 5), it is a doubtful assumption for deep convection, where several studies have documented the importance of homogeneous freezing. This assumption further adds ACPD to the uncertainties of this study which could be reduced by partitioning of the data into 15, C7168–C7175, 2015 convective and stratiform regions and perhaps using different assumptions for the two cases. Response 3.4: Clarification re PSD measurements used in Fig. 6 and all later analysis Interactive to be added at p. 16519, line 16: "Although Robust probe data are available at warmer Comment temperatures (see Fig. 2), we note that Cayenne and Darwin size distributions are available only for flight legs at temperatures colder than 40 C (e.g., see Fig. 5), aside from several short segments of flight 1422 reaching 33 C, which are not qualitatively different." Also at p. 16516, line 13: "here we consider only measurements taken at temperatures colder than 20 C, where liquid contributions are considered negligible" to be replaced with "in Fig. 2 we include only measurements taken at temperatures colder than 20 C, where airframe icing was non-existent or negligible; size distribution measurements are limited to temperatures colder than 33 C, as discussed further below." 5. The Locatelli-Hobbs relationship (eq. 2), might not be the best choice for these clouds, compared to, for example, the Baker and Lawson (2006) area-mass relationship. It would be helpful to have an explanation of why eq 2 was chosen over other methods and a better explanation of the uncertainties in computed mass content associated with these types of assumptions would certainly be worth considering for the Full Screen / Esc revised paper. As the author’s point out, their IWC measurements are roughly a factor of two greater than measurements documented in the scientific literature to date, so it Printer-friendly Version is important for the authors to demonstrate why they believe their measurements offer an improvement over previously reported IWC measurements in similar clouds. Interactive Discussion

Response 3.5: Clarification re choice of eq. 2 to be added at p. 16517, line 24: "as- Discussion Paper suming a widely used relationship" to replace "assuming a relationship". Clarification re associated uncertainty to be added at p. 16518, line 13: "As discussed above, C7173 roughly a factor of two uncertainty in calculated IWC may be associated with uncertainty in the validity of Eqn. 1 or another such relationship (e.g., McFarquhar and ACPD Heymsfield, 1986)." We do not believe that these measurements offer an improvement 15, C7168–C7175, 2015 over previously reported IWC measurements, but rather that "a large database of such measurements is not yet available" (see p. 16528, line 20). 6. The paper could be improved by including more information on the ice particle Interactive morphology. The nephelometer appears to provide excellent imagery of ice particles Comment (e.g. as in Figure 1, Ackerman et al., 2015). This type of imagery has traditionally been used together with size distributions to explain the microphysical characteristics of ice, yet the authors have not utilized this technique, which might offer significant insights into the nature of the ice environments that were sampled. Response 3.6: Clarification to be added at p. 16517, line 19: "While capped columns are commonly present (see Part II), the majority of crystals appear irregular, as found in CEPEX anvils (cf. McFarquhar and Heymsfield, 1996), and the nephelometer images do not commonly produce images of sufficient clarity to distinguish rime or other morphological details." We consider our current approach adequate because IWC calculated from the size distributions without habit-dependent analysis agrees within uncertainty of Robust probe IWC over an order of magnitude in dynamic range, as shown in Figure 5. Minor Comments. Line 51-2. “industry concluded..” This seems out of place without a reference. Full Screen / Esc

Response: Reference to be added: "(e.g., Dezitter et al. 2013)". Printer-friendly Version Section 7 first sentence. “power less” should read “power loss”. Interactive Discussion Response: Correction to be made. References Discussion Paper Dezitter, F., A. Grandin, J.-L. Brenguier, F. Hervy, H. Schlager, P. Villedieu, and G. Za- C7174 lamansky (2013), HAIC - High Altitude Ice Crystals, In Fifth AIAA Atmospheric and Space Environments Conf., American Institute of Aeronautics and Astronautics, Re- ACPD ston, Virginia. 15, C7168–C7175, 2015 Jackson, R. C., and G. M. McFarquhar (2014), An Assessment of the Impact of An- tishattering Tips and Artifact Removal Techniques on Bulk Cloud Ice Microphysical and Optical Properties Measured by the 2D Cloud Probe, Journal of Atmospheric and Interactive Oceanic Technology, 31(10), 2131–2144, doi:10.1175/JTECH-D-14-00018.1. Comment Korolev, A., E. Emery, and K. Creelman (2013), Modification and Tests of Particle Probe Tips to Mitigate Effects of Ice Shattering, Journal of Atmospheric and Oceanic Tech- nology, 30(4), 690–708, doi:10.1175/JTECH-D-12-00142.1. Leroy, D., E. Fontaine, A. Schwarzenboeck, J. W. Strapp, L. Lilie, J. Delanoe, A. Protat, F. Dezitter, and A. Grandin (2015), HAIC/HIWC Field Campaign-Specific Findings on PSD Microphysics in High IWC Regions from In Situ Measurements: Median Mass Diameters, Particle Size Distribution Characteristics and Ice Crystal Shapes, No. 2015- 01-2087, SAE Technical Paper. Mazzawy, R. S., Strapp, J. W. (2007), Appendix D-An Interim Icing Envelope, No. 2007-01-3311, SAE Technical Paper.

Interactive comment on Atmos. Chem. Phys. Discuss., 15, 16505, 2015. Full Screen / Esc

Printer-friendly Version

Interactive Discussion

Discussion Paper

C7175 Open Access Atmos. Chem. Phys. Discuss., 15, C7176–C7180, 2015 Atmospheric www.atmos-chem-phys-discuss.net/15/C7176/2015/ Chemistry ACPD © Author(s) 2015. This work is distributed under the Creative Commons Attribute 3.0 License. and Physics 15, C7176–C7180, 2015 Discussions

Interactive Comment

A. M. Fridlind et al. [email protected]

Received and published: 24 September 2015 Full Screen / Esc

General Comment: As the ﬁrst part of a serial study of “high IWC-low Ze” condition Printer-friendly Version that coincide with jet engine power loss, this paper utilized the microphysical properties including IWC, mass size distribution, and derived Ze sampled during 2010-2012 Interactive Discussion Airbus campaign to reveal the possible meteorological scenario for that condition. Sur- prisingly, consistency was found among documentation, in situ measurements, satellite Discussion Paper retrieval, surface radar observations, and model simulations, which validate the com-

C7176 mon occurrence of “high IWC-low Ze” condition. This paper is definitely suitable for publication because of its meaningful inspiration in aviation safety. However, a couple ACPD of improvements should be made in order to promote the quality of this study. 15, C7176–C7180, 2015 General response: We appreciate the reviewer’s time and comments, which are addressed point-by-point below. Several points were raised almost verbatim by reviewer 1, and where those occur, we point to those responses to be concise. Interactive Comment Major comments 1. In the title, the authors clarify that the “high IWC-low Ze” condition happens in “near deep convection” regions. As a result, in the abstract and following discussion of Airbus sampling, “stratiform rain” becomes the major focus. The ques- tion then arises, how does the cloud classification work in this study? Figs. 13 and 14 demonstrate the separation of convective and stratiform on radar images, however without a detailed description of classification algorithm. Even for the same condition that Ze less than 30 dBZ at 11 km (I assume this 30 dBZ value serves as the threshold of classification because it appears on 6 times related to cloud classification discussion throughout the entire paper), it could either be authentic stratiform or the vertical ex- tension of weak convection, and only the examination of entire radar reflectivity column could separate those. The following figure gives an example of radar cross-section associated with UND CSA classification results sampled during MC3E campaign. Clearly, cloud classification can’t separate weak convective and stratiform regions based on only near the cloud top Ze values, because there is no significant discontinuity in Ze between those two cloud types. Furthermore, the microphysical properties (even at Full Screen / Esc temperatures < 20 C) between convective and stratiform regions are quite different, and the readers could argue that the identification of “high IWC-low Ze” condition could Printer-friendly Version attribute to the weak convection was mistakenly classified as stratiform due to the lack of 3D cloud information. Interactive Discussion

Response 1: Two clarifications are needed. The 30 dBZ at 11 km is only a point-wise Discussion Paper definition of "low Ze" based on aircraft radar sensitivity, without reference to classification. Clarification to be added to p. 16508, line 28: "and the Ze threshold is based on C7177 radar sensitivity, without any specified relation to convective or stratiform structures." The S-band observations and 1D modeling are focused stratiform columns only so that ACPD poorly constrained vertical motion can be neglected. Clarification to be added to p. 15, C7176–C7180, 2015 16521, line 7: "Here we will focus on extended stratiform regions where the mean contribution of vertical winds to MDV can be neglected relative to the reflectivity-weighted fall speed of ice at elevations circa 11 km." Interactive 2. The morphology should be further discussed, because it is the key explanation of Comment the “high IWC-low Ze” condition. This study could be carried out from two aspects: (1) By examination of OAP images; (2) Rather than directly applying LH74 relationship to calculate IWC and Smith [1984] method to calculate Rayleigh radar reflectivities, various mass-dimensional relationships and corresponding area-dimensional relationships were developed for different ice crystal habits [Mitchell, 1996], and those relationships should be adopted for better estimation of IWC and Ze values. Response 2: Please see responses 3.5 and 3.6 to reviewer 1. 3. Contrary to the assumption that liquid contributions are considered negligible for temperatures colder than 20 C, Rosnefeld et al. [2013] found the common occurrence of highly supercooled drizzle and rain near the coastal regions of the western United States even at colder temperature. There is still possibility of supercooled liquid droplets anywhere warmer than 40 C, so rather than a fixed temperature threshold, phase separation is suggested if proper instruments are available (like icing detector, CDP, or hot-wire King LWC probe, etc.). Thus, the contamination from supercooled liq- Full Screen / Esc uid droplets could be eliminated, because as the authors mentioned in the introduction, Printer-friendly Version the role of supercooled liquid water was caused confusion. Response 3: Please see response 3.4 to reviewer 1. Interactive Discussion 4. Instead of number size distribution, mass size distribution is intensively investigated Discussion Paper in this study, and the solid conclusion is derived that particles within the size range from 150 to 600 micron contribute a large portion of IWC, which is a very interesting feature C7178 associated with “high IWC-low Ze” condition. However, mass size distribution is still a derived bulk property, and in this case it is roughly the second moment representation ACPD of original number size distribution. It would be nice that the authors could release the 15, C7176–C7180, 2015 number size distribution information without any assumption imposed, because after gamma or exponential functions were fitted to observed number distribution, the fitted parameters could serve as indicators to testify the cloud classification algorithm used Interactive in this study, and comparison with previous stratifrom studies becomes easier. Comment Response 4: Although mass concentration is a derived quantity with an uncertainty of roughly a factor of two (see response 3.5 to reviewer 1), the uncertainty in number (the measured quantity) can easily be a factor of five or more (e.g., Fridlind et al., 2007). We have reasonable agreement of IWC derived from size distributions and Robust probe over a reasonably large dynamic range (cf. Fig 5.), but we have no source of confidence in number concentrations and therefore prefer not to promote their use here.

5. If the “high IWC-low Ze” condition commonly exists, does it mean there will be bright band of high IWC near the cloud top based on retrieval from radar observation? Sur- prisingly, even the IWC values were lower than this study, the discontinuity in Ze-IWC relationship studies were found by Heymsfield [2005] (Figure 10) and further discussed by Wang [2015] based on in situ measurements of stratiform rain. From the following figure, the jump in IWC clearly takes place at the Ze value range from 12 to 15 dBZ, and it was caused by the drastic changes in the overall shape of number size distribution as discussed by Wang [2015] in section 3.4 and Figure 10. This is another reason Full Screen / Esc why detailed investigation of number size distribution is necessary. Printer-friendly Version Response 5: We agree that future detailed investigation is necessary, and furthermore that it will require robust observations at more elevations, whereas the size distribu- Interactive Discussion tion data here were intentionally gathered almost exclusively at cruise elevations near 40 C (see response 3.4 to reviewer 1). Discussion Paper Minor comments Page 16509, line 9, ‘identifed’ should be ‘identified’.

C7179 Response: Correction to be made. ACPD Page 16521, line 26, ‘Reflecitivity’ should be ‘Reflectivity’. 15, C7176–C7180, 2015 Response: Correction to be made. Reference Interactive Fridlind, A.M., A.S. Ackerman, G. McFarquhar, G. Zhang, M.R. Poellot, P.J. DeMott, Comment A.J. Prenni, and A.J. Heymsfield, 2007: Ice properties of single-layer stratocumulus during the Mixed-Phase Arctic Cloud Experiment (M-PACE): Part II, Model results. J. Geophys. Res., 112, D24202, doi:10.1029/2007JD008646.

Interactive comment on Atmos. Chem. Phys. Discuss., 15, 16505, 2015.

Full Screen / Esc

Printer-friendly Version

Interactive Discussion

Discussion Paper

C7180 Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | , 3 , J. W. Strapp 2 , M. Weber 2 X class copernicus.cls. E T A , F. Dezitter 2 1 5 , A. Grandin 1 , A. S. Ackerman 1 , and C. R. Williams 4

Cloud Physics and Severe Weather ResearchCooperative Section, Institute Environment for Canada, Research Toronto, in Ontario, Environmental Canada Sciences, University of Colorado Boulder, and NASA Goddard Institute for SpaceAirbus Studies, Operations 2880 S.A.S., Broadway, New 316 York, route NYMet de 10027, Analytics Bayonne, USA Inc., 31060 Aurora, Toulouse Cedex Ontario, 03, Canada France NOAA/Earth System Research Laboratory, Boulder, Colorado, USA Correspondence to: A. M. Fridlind ([email protected]) 4 5 2 3 A. M. Fridlind A. V. Korolev 1 stratiform rain column simulations in situ and remote-sensing observations with near deep convection – Part 1: Consistency of with version 2014/05/30 6.91 Copernicus papersDate: of 25 the September L 2015 High ice water content at low radar reflectivity Manuscript prepared for Atmos. Chem. Phys. Discuss. Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 27 regions e Z , the measured km AMSL and temperatures of km 2 . Power loss events commonly occur during flight C 50 10 to ) less than 20–30 dBZ and no more than moderate turbulence, of- e Z . Given substantial and poorly quantified measurement uncertainties, here m (Lawson et al., 1998; Mason et al., 2006). During these incidents the engines lost , mean Doppler velocity, and retrieved rain rate. The results of these consistency e C Z 33 through the melting level. Crews reported no more than trace occurrences of hail or graupel. power slowly prior to abruptas uncommanded rollback, reduction and of in most power cases to engine idle, power commonly authority referred was to automatically returned after descent power loss near thunderstormsto at elevations of 8.5–9.5 1 Introduction Between 1990 and 1998, commuter transport jets experienced at least ten incidents of engine the observed size distribution features in Part 2. files of checks motivate an examination of the microphysical pathways that could be responsible for in situ size distribution measurements.size We distributions also as find an that upper column boundary simulations condition using are the generally in consistent situ with observed pro- Airbus-sampled locations. We find that profiler-observedvelocities radar at reflectivities Airbus and sampling mean temperatures Doppler are generally consistent with those calculated from eters of 100–500 µ we evaluate the consistencyradar of observations the obtained Airbus under in quasi-steady, heavy situ stratiform measurements rain with conditions ground-based in one profiling of the the highest IWC regions encountered, atice typical size sampling distributions exhibit elevations a circa notably 11 narrow concentration of mass over area-equivalent diam- flight tests seeking to characterizeof the large, highest cold-topped ice storm water systems content in (IWC) the in vicinity such of low- Cayenne, Darwin, and Santiago. Within glaciated deep convection at through radar reflectivity ( ten overlying moderate to heavy rain near the surface. During 2010–2012 Airbus carried out Occurrences of jet engine power loss and damage have been associated with flight through fully Abstract Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | oc- % length without ::::::: km sensitivity, :::::::::: that the inability of radar ::::: on ::: or more over a 10 over Alabama, induced by en- 3 C based ::::: gm is 28 :: : , less than 20–30 dBZ) or more than e Z and structures threshold km (e.g., Dezitter et al., 2013) ::::::::: ::::::::: :::::::::::::::::::::::: e 3 Z ::: the ::: stratiform ::::::::: and :::: , or :: . ” conditions, where the IWC designation remains qualitative or e Z convective :::::::::: to :: relation :::::::: occurred within an oceanic mesoscale convective system (MCS), where pilots typically specified at flight altitude and no identification otherwise of flight-level convective cores; 34 :::::::: % Outside of these commonalities, events occupied a rather wide envelope of storm conditions: After a period of confusion regarding the potential role of supercooled liquid water owing to e moderate turbulence at flight level.ditions Hereafter we as will “high refer IWC–low torelative, event-related as microphysical discussed con- further below Z 38 flew directly through the central and deepest clouds owing to a complete lack of detectable any ::: vection, under hypothesized andabsence sometimes of unambiguously pilot-reported fully equivalent glaciated radar conditions, reflectivity ( in the indicated several patterns in2010; the Mason meteorological and conditions Grzych, encountered 2011). (Grzych Namely, events and occurred Mason, almost exclusively near deep con- ing a dozen enginedocumented types events (Mason of and Grzych, jet 2011). engine As power of loss 2011, or a damage database connected of with roughly the 100 engine core had the appearance of streaming meltwater2003 on the heated connection cockpit between windshields bothand during the commuter some ingestion and of events, large by copious transport ice jet under engine glaciated power conditions loss was events established, ultimately involv- Mason et al., 2006). current instruments to reliably andto accurately addressing measure similar such hazards high being identified IWC in posed other a aircraft major types obstacle (Strapp et al., 1999, 2005; modified engine simultaneously experiencedported no rollback, on and that no particular2006). further aircraft events On since have the such been other modifications re- were hand, made industry standard concluded (Mason et al., scale, based on estimates obtained indirectlyMason owing to et primary probe al., failure 2006). (Strapp et On al., the 1999; one hand, the flight tests were deemed a success since a second, downwind of continental deep convection athanced 8.8 ice water content (IWC), sustained at likely 0.5–1 In 1997, industry-sponsored flight tests successfully reproduced a rollback event while circling Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | ), 1 Jkg occurred in ), marginal to % cm ) in length, making it im- km (10 O of the total) occurred beneath a cir- % conditions during some events has suggested e 4 Z occurred in classical continental anvils, where cores were in maximum dimension. However, an analysis of temperature % frequency, based on statistical comparison of reanalysis model e km Z and avoided on the downwind side of storms, in this case without signif- : :::::::: identified It is notable that the MCS class events documented (80 To evaluate event distribution geographically and seasonally, Grzych and Mason (2010) de- water content possible in convection was undertaken by the Royal Aircraft Establishment in 2 Microphysical conditions From the perspective of industry, the most recent effort to rigorously establish the condensed loss events led to onlycurrent one cockpit instrumentation: consistently use recommended the hazardlevel aircraft for detection radar’s moderate tilt action to and available heavy gain using rain to (Grzych scan and below Mason, the 2010; freezing Mason and Grzych, 2011). possible to rule out the sufficiency(Mason of et either long al., or 2006). short An duration exhaustive exposure analysis to of cause conditions such encountered events with and without power instrument errors caused bythat high enhanced IWC–low IWC regions might commonly be as short as rus anvil shield of at least 185 and modest wind shearflight (generally traffic frequency, sufficient these to mapsof support generally the explained squall equator the under lines). concentration preferentially When moist of conditions also events seasonally. within considering 45 fields with the following event conditions: highmodest precipitable atmospheric water instability (median (median of convectively available 6 potential energy of 1400 rived maps of high IWC–low icant rain immediately below; andelevated warm the frontal, remainder vigorous occurred winter variouslyvection (Mason within lake-effect, and or tropical Grzych, developing, multicell, 2011). immature, continental con- flying over heavy rain regions;identifed 6 but maneuvered to avoida them, typically continental passing MCS, over where regions pilots of again heavy rain; detected 8 and diverted around cores at flight altitude, curred within a strong tropical MCS, where pilots detected convective cores at flight altitude, Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | m of all reported estimated at that % % resolution over a tem- from storm locations of km , with a suggested factor-of- km , the diameter of the smallest 3 max D gm or degree of glaciation encountered. e Z , with an accuracy of 15 3 at a distance of 50 5 , respectively, and more than 5 at flight level from these scientific research cam- 3 e gm km Z gm , with little reference to C 26 , a vicinity commonly reported for initiation of commuter and large transport jet en- e over length scales of 10 and 100 (Heymsfield and Palmer, 1986; Heymsfield, 1986). From CEPEX, IWC was calculated Z 3 attributed to particle counting uncertainty (McFarquhar and Heymsfield, 1996). The 2DP % % Motivated by the 1990’s commuter jet engine power loss problem, Lawson et al. (1998) rean- gm highest paigns, Lawson et al. (1998) reportedtwo peak uncertainty, anvil IWC falling of to 1–3 well below 1 circle enclosing two-dimensional projectionssitu). of Working ice in particles the imaged absence at of recorded random orientation in 50 and 2DC probes respectively measurein particles maximum dimension that of are randomly nominally oriented projected 300–9600 area and ( 30–1800 µ from a PMS two-dimensional cloudtor particle probe of (2DC), two, with owing an largely estimated to accuracy of sensitivity a to fac- mass-dimension relationships applied, plus roughly and scanning radar reflectivityof fields 50 colocated with flight tracks, with an estimated accuracy paigns, it was calculatedculated from from the a most Particle suitable Measuring data Systems (PMS) available. From two-dimensional precipitation CCOPE, (2DP) IWC probe was cal- alyzed more limited in situ measurements(Knight, collected 1982) within and anvils near over Tiwi Montana duringand during CEPEX Heymsfield, CCOPE (Heymsfield 1996). and Since McFarquhar, no 1996; direct measurements McFarquhar of IWC were obtained during these cam- time. 5 measurements at each location exceeded 2 but the flight paths werelenge also adjusted engine to performance. find Based andSingapore, maintain on and high roughly water Darwin, 200 content maximum in cloud total order traverses to water in chal- contents the encountered vicinity were of in Entebbe, excess of 6 and perature range of 0 to Regions of the highest expected turbulence and water content were avoided for reasons of safety, controlled heated tube and delivered1959; it to Mason a et measuring al., system 2006). inside Measurements the were aircraft made (McNaughton, at roughly 10 the 1950’s using a pitot-type tube that collected ice and water condensate in a thermostatically- Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | . , 10 m km max mm with m (Lawson (MMD m max D C and then growing 12 of 100–400 µ max D graupel over a wide temperature range ( 6 less than 20 dBZ if present in monomodal size mm e Z , Fig. 1 illustrates that such relatively small mass me- 3 , where convection was approached as closely as the pi- C gm 65 15 to , whereas turrets contained 1–2 max ). Considering this diversity of new data with strikingly similar microphysical fea- above and below which half of mass resides) decreasing from 500 to 100 µ D C , calculated as described above). They reported typical mass median . 3 in max 50 max D D Based on measurements during the TC4 campaign over Central America (Toon et al., 2010), In a more detailed analysis of ice properties in three CEPEX anvils at elevations of 7–14 mm gm by vapor deposition, whilemented, smaller respectively. and larger outflow ice preferentially evaporated and sedi- tures, Lawson et al. (2010)cloud attributed droplets the freezing predominance heterogeneously of at several temperatures hundred colder micron than particles to relatively compact irregular crystals and2 irregular aggregates that wereto generally smaller than alike were consistently found toet be al., dominated 2010). During by both crystals TC4 with and NAMMA, fresh anvils contained columns, capped columns, et al., 2010), andgov/namma/), the the NAMMA ice campaign mass over size West distributions Africa in (http://airbornescience.nsstc.nasa. fresh anvils, aged anvils, and convective turrets reportedly the most extensive tropical anvil measurement campaign since CEPEX (Lawson distributions with typical assumptions regardingcalculations ice do crystal depend properties strongly (aside on weare assumed note not ice that directly crystal such measured by mass-dimension any characteristics, current which instruments, as well as size distribution properties). Even if IWC weredian as diameters high would as be 4 associated with increasing height. Prominent peaks in the mass size distribution were found at 150–600 µ the lots considered safe, McFarquhar and Heymsfieldplate-like, (1996) compact reported spatial a crystals predominance of1 and irregular, irregular aggregates in the highest IWC regions (0.1– and temperatures of except to report the largestin particles encountered, which were commonly on the order of 2 gine events (Mason et al., 2006). Ice size distribution properties were not extensively analyzed, Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | ) from two 8 mass-dimension on average at IWC conditions were not e % Z N) from 7 July to 16 August resolution) showed rare occur- hygrometer; IWC uncertainty is mm: ↵ S–25 5km in . 0 ⇥ max (Twohy et al., 1997). Lawson et al. (2010) D 3 4km . and 1 7 gm ⇥ mg in elevation (frequency of less than 10 in 8km ) . km 1 m ( ⇠ was calculated in one convective turret, maxima were less 3 at 10 N, mass , the estimated saturation of that instrument under TC4 measure- 3 3 gm S–25 gm , and the spread of algorithm results diverged sharply at IWC values gm 3 particle for values larger than 0.2 . in the tropics using CloudSat and CALIPSO measurements and retrievals 3 gm with % relation derived in Baker and Lawson (2006) to measurements from an un- , : 0 . km 2 max gm max are very rare (Delanoë et al., 2014). Aside, we note that no research aircraft in- in fresh and aged anvil regions. D 3 D 3 021 in . gm m gm -averaged aircraft data from TC4, NAMMA, and other campaigns, IWC values greater =0 s Motivated directly by the engine power loss problem, a more recent satellite-based study Comparison of IWC retrievals using CloudSat data (25 In a data set compiled for development of satellite retrieval algorithms, comprising thousands During TC4, IWC was calculated by applying the habit-independent m Calculated IWC was shown to agree with direct measurements from a counterflow virtual im- elevation of 10 greater than 1 aimed to establish IWC conditions conceivably encountered by commercial aircraft flying at an out of three algorithmsIWC (Wu exceeding et 10 al., 2009). However, a third algorithm showed occurrences of 2006 over the tropics (25 rences of greater than 2.5 specifically avoided, to our knowledge. of 5 than 2 volved encountered engine power loss events although high IWC–low than 1 do not offer an estimate ofAlthough uncertainty in IWC IWC exceeding calculated 2 or measured during TC4 or NAMMA. ment conditions (Lawson et al.,into a 2010). dry Within a airstream, heated whichestimated to CVI is be inlet, in 10 impacted turn sampled ice by is a evaporated Lyman- pactor (CVI) inlet mountedconcentrations on below 0.5 the fuselage of the aircraft to within 20 . : 3000 µ area-dimensional ::::::::::::::: derwing mounted PMS two-dimensional stereo (2DS) probe over a size range of roughly 10– Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | to km ::: (e.g., m conditions :::::::::: –100 µ -footprint peaks icing 25 measurements ::::::::::::: :::::: km ⇠ and ice mass mixing ratio. glaciated ::::::::: max preliminary in ::: :::::::::::: , Fig. 1 illustrates the relationship of obtain 3 :::::: to commonly less than ::: gm performance max :::::::::::: was :::: D below 18 dBZ, including 2.5 8 e Z engine ::::::: objective GHz for ::::::::: :::: and 94 Airbus ::::::: 3 standards :: an ::::::::: , cirrus formed in situ are most commonly characterized by a notably in clouds associated with deep convection. gm e max Z D proposed ::::::::: conditions (Heymsfield et al., 2013). Holding size distribution properties fixed, denser ice , with a large uncertainty owing to unknown ice morphological and size distribu- 1 3 to reflectivity for the simplest case of monodisperse ice particles with mass following newly ::::::: cm gm max In this two-part study, we examine a set of measurements gathered by Airbus during 2010– Focusing briefly on observed ice morphology, the McFarquhar and Heymsfield (1996) and evaluate ::::::: 2012 in deepet convection al. near (2014), Cayenne, Darwin, and Santiago. As described by Grandin size distribution, and freezing of abundant cloud droplets incal the activity strong, (e.g., mixed-phase Connolly updrafts et associatedmorphology with al., contributes electri- 2005). substantially As to uncertainty discussed in further the below, relationship the between complexity IWC, of ice crystal mass is attributed to theas presence well as of sufficiently electric many fields such small sufficient crystals to to enhance aggregate, perhaps the owing aggregation to process, homogeneous small frozen droplets orConnolly faceted et al., crystals 2005; with Gayet et al., 2014; Stith et al., 2014). The occurrence of chain aggregates TC4 or NAMMA, deep convection that isdominance electrically of active chain may also aggregates be (Stith distinguished et by a al., pre- 2004), which have been found to be composed of tain a mixture of habitshundred that microns include in irregular and ratherlarger compact fraction appearing of particles of rosettes several (Lawson et al., 2010, per their Fig. 2). Although not reported from Lawson et al. (2010)deep studies convection both outflow vs. identify those a seen in sharp cirrus distinction that between is formed crystal in shapes situ. seen Although both in may con- particles would lead to increasing reflectivity for the same MMD MMD Eq. (1), and for1–1000 exponential size distributions with slopes over an observation-based range of of 2–4 tion properties. For an ice mass mixing ratio of 4 torial coast of Brazil, they concludedwith IWC it greater likely than that 1 the storm included an area wider than 55 (Gayet et al., 2014). After identifying and analyzing a well-observed tropical MCS off the equa- Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | , . and Doppler fall speed e Z relative to reflectivity-weighted ice fall speed. 9 km sampled included oceanic and continental conditions, mon- . Here in Part 1, we first survey the IWC, mass size distribution, and Conditions :::::::::: safety conditions. We compare the flight test data with previous literature and with of e :::::::::::::::::::::::::::::::::::::::::::::::::: Z (Roques, 2007; Dezitter et al., 2013; Grandin et al., 2014) limits the derived from measurements; these are the factors immediately relevant to characterizing e (Roques, 2007) et al.,uid 2005) droplet and size an distribution Airbus but imaging used nephelometer here designed for for estimation measurement of ice of mass liq- size distribution The Airbus flightity campaigns under included high twoRobust in IWC solid situ conditions: probes a hot-wire intended newly probe to designed used maintain Science functional- for Engineering Associates estimation (SEA) of ice mass mixing ratio (cf. Strapp 3 Airbus in situ measurements wind can be neglected at elevations circa 11 Houze, 1992). Here simulationsuse are of limited Doppler to fall heavy speed stratiform information under rain relevant regions quiescent in conditions order where mean to vertical make convective and stratiform rain regionsin are large commonly tropical well systems separated, owing and to is the more complicated variety of convective region structure (e.g., Mapes and staff and Houze, 1991), commonly associatedet with al., a 1995), radar or reflectivity bright somecult band type to (e.g., of objectively Steiner define transitional in regime. a A mid-latitude transitional squall regime line region (cf. is Biggerstaff already and diffi- Houze, 1993), where a region classified asclear stratiform whether power rain loss by events are some associated measure with (Grzych classical and stratiform rain Mason, columns 2010), (Bigger- it is not simulations of heavy stratiform rain columns,profiles and with compare observations. predicted Aside, we note that although many power loss events may be within high IWC–low remote-sensing measurements of a well-observed MCSpaign over Darwin (May during et the al., TWP-ICE cam- 2008). Finally, we use Airbus observations to initialize steady-state column within Z soonal and sea breeze convection, and(1959) a range campaigns, of the storm purpose size extents. was Similar to to the find McNaughton and stay within the highest IWC regions possible (Mazzawy et::::::::::::::::::::: al., 2007). Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | or as :: :: :: C, 33 ::::: non-existent than :::::::::::: ::::: was ::: colder :::::: icing ::::: airframe :::::::: temperatures :::::::::::: to :: negligible (assuming the primary source of error to be % limited ::::::: 10 are :::: only measurements taken at temperatures colder than considered : are include ::::::: measurements we ::: ::::::::::::: 2 :: . contributions Fig. :::: in :: below :::::: distribution liquid ::::::::::: size :::: further ::::::: consider , where C we The accuracy of applying a 0.4 collection efficiency to estimate IWC from the Robust probe Since about 2005, it has been realized that hot-wire devices using trapping geometries are 20 negligible; ::::::::: discussed :::::::: here Recent flight tests have provided a datain set the that near is future hoped (Grandin will improve etis efficiency al., close characterization 2014). to unity, Because such the that probe’s liquid efficiency and for ice liquid contributions water to content mass cannot be accurately determined, an accurate lower limit on IWC),anything we better can offer than no the further evidence raw of factor accuracy of here to two establish uncertainty owing to overall collection efficiency. the IWC accuracy was initiallythe estimated loss of as re-entrained 30 mass, such that measurements uncorrected for efficiency should provide (Grandin et al., 2014).Airbus Comparisons flight tests with has IWC indicated estimated smaller efficiences from of the 0.2–0.3 (Grandin imaging et nephelometer al., in 2014). Although simple installation, to provide real-time guidance during flight tests. under natural ice particle conditions remains unknown and will require further investigation was estimated at about 0.4, theto value Airbus assumed on here. the The probe basis was of subsequently durability recommended and its simple principle of operation, requiring a relatively to deduce absolute IWC andefficiency ice of the mass capture, size melting, distributions. and During evaporation the by course the of Robust probe calibrations, for the shaved ice particles (Strapp et al., 2008),or conditions suspected where to essentially failRobust all owing probe available to measurements hot-wire particle were probes impact intended were damage to (Strapp known be et used al., alongside 1999). ice In ingestion a measurements wind tunnel, et al., 2013b). The Robustments probe for was the developed calibration primarily of to provide a surrogate wind IWC tunnel measure- producing high IWC at relatively high wind speed subject to relatively low captureticles efficiencies that owing to bounce re-entrainment out of of only the a capture fraction volume of (Emery par- et al., 2004; Strapp et al., 2005; Korolev Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | are 3 ) in an gm km (1 . Data include O km length scale reached at km is roughly twice the maximum ), 2.5, and 10 3 1 ms gm ,2 km ), consistent with the pattern seen here across 11 km was only reported at a length scale of 3 -mean condensate mixing ratios in excess of 2 km gm , whereas the Santiago flights targeted warmer temperatures; the (Lawson et al., 2010). TC4 IWC at a 10 1 km ms in fresh anvils, and mean anvil IWCs reported appear generally consistent with 3 data as reported, at typical airspeed of 200 Hz gm (1 km We next consider in greater detail the measurements obtained during a flight that encoun- Taken together, the envelope of measurements is qualitatively similar to that reported by In Fig. 2 the total condensed water content (CWC, liquid plus ice) reported by the Robust tered relatively high IWC. From the track of Cayenne flight F1423 (Fig. 3), with spirals in poorly quantified uncertainty of both measured and calculated IWC values. and to prioritize theof repeated such measurement measurements of does the not highest currently IWC exist. found Other there. differences A may be large within database the large and CEPEX data (McFarquhar and Heymsfield, 1996;results Lawson may et seem al., statistically 2010).be On inconsistent. explained the However, by face at the of least objective it, of some these the degree Airbus of flight tests difference to could sample very large, cold-topped storms updraft exceeding 20 least 0.25 value shown across the threecf. CEPEX their anvils Fig. examined by 3). McFarquhar During and TC4, Heymsfield 2 (1996, McNaughton (1959) insofar as 10 found in all three locations. At a length scale of 10 does exhibit a sharper decreasea in narrow CWC updraft, for with instance. averaging distance, as might be expected within scale over two ordersall of magnitude three (0.2–20 locations in Fig. 2. They also show that a leg with statistically extreme peak CWC and Darwin Robust probe dataGrandin as et al. a (2014) continuous report function that of 99th percentile length CWC scale is within a identified surprisingly legs, weak function of length cruise elevations circa 10 concentration of IWC as a functionmarily of as elevation at an each indication location of should sample therefore frequency. be Based viewed on pri- a more detailed analysis of the Cayenne eight flights from Cayenne (F1419–F1426),flights six from flights Santiago from (F1539–F1547). Darwin (F1403–F1409), We note and nine that the Cayenne and Darwin flights targeted probe during each campaign is0.2 plotted as a function of elevation over three averaging lengths: Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | ; to as :: m (1) µ mm m ::::: While images ::::::: 100 (based on calculated ::::::::: ⇠ of 1–3 irregular, relationship relationship ::::::::::: : ::::::::: in eq ::: morphological ::::::::::::: mm D used such :::: ::::: appear :::::: other ::::: nephelometer :::::::::::: or :: of 1–12 :::::: widely uncertainty the :::::::::: another ::::::: :::: is smaller than eq crystals ::::::: rime or :: ::::: D two measured) and to aggregates eq and :::: ::: of 1 :: ::: D of defined by Locatelli and Hobbs ::: mm eq Eq. , with the exception that effective :::: D eq of :: in altitude, Fig. 4 shows ice particle factor ::::: majority D distinguish :::::::::: :::::::: a :: to km :: of 2–12 the ::: eq validity 2), :::::::: ::: D clarity 12 roughly ::::::: :::::: the ::: Part :::: Also shown in Fig. 4 are Rayleigh radar reflectivities in ::: is slightly different from the average of chord lengths (see above, :::: eq ::::::: :::::::: sufficient D of region at circa 11 ::: present e ::::::: uncertainty Z :::::::::: discussed ::::::::: in mm derived from measurements of natural crystals sampled in the images :::::: :: As with eq (cf. McFarquhar and Heymsfield, 1996), ::::: :::::::::::::::::::::::::::::::::::: relation was derived from particles with D commonly eq :::::::::: D produce anvils :::::: :::::::: are – ::: 9 m . associated ::::::::: in mg to 1 eq D be m ::: CEPEX ), defined as the diameter of a circle with area equal to the two-dimensional projection during the flight segment analyzed (Fig. 5), the sampled cloud is fully glaciated. :::::::: The nephelometer size distributions are reported as a function of area-equivalent diam- columns eq 037 :::::::: : commonly in . C . Also shown is the mass size distribution calculated assuming a :: :::::::::: may D ::::: not =0 Here we apply this relation using the same dimension ::: 43 mm ::::::::::::::::::::::::::::::::::::: (e.g., McFarquhar and Heymsfield, 1996). number size distributions derived from the nephelometer measurements. At temperatures circa quantifiable at this time. IWC :::: such small particles dothough not contribute bin-wise substantially uncertainty to in the mass mass size size distributions distributions shown calculated here, using this approach is not roughly fifty particles total),density is it limited is to applied that here of bulk at ice, all which is relevant where parallel and perpendicularWhereas to this a photodiode array substituted in Brown and Francis (1995). radiating assembles of dendritesof or unrimed dendrites radiating ( assemblages ofmeasured). plates, Aside, side we planes, note bullets, that and this columns ( (1974) for their single-particle measurements, applicable both to aggregates of densely rimed m of particle Cascade Mountains, Washington (Locatelli and Hobbs, 1974): of an ice particle imaged3 at random orientation in situ, over a bin midpoint range of 67.5 µ do :: details. :::::: eter ( capped :::::: found ::::: an identified high IWC–low Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | , m (see over :::: m size dis- e Z at the highest % temperatures :::::::::::: similar to 250 µ eq warmer :::::::: at :: reported, with the mass calculated 3 in all three mass ranges shown, so this (Smith, 1984). available e ::::::::: scarcely deviates from 200–300 µ Z 93 . gm are eq 0 :::: / data 13 208 :::: . 0 probe in Fig. 4. Although the peak is narrower at higher IWC, :::::: m ) and then normalized by their total mass for Flight 1423, 3 Robust obtained from the measured size distributions. Here it is shown values are below 15 dBZ, although the sharp drop in ::::::: gm e e conditions. Figure 5 shows the time series of integrated IWC, mass Z Z e Z size distributions; although they must also be truncated to some degree :::::::: Although ), and e of roughly 250 µ eq Z eq D is low by an unknown amount. However, mass size distributions do not appear to (MMD e Z eq D The most striking feature of the median size distributions shown is the consistent peak in As mentioned above, here we focus primarily on the mass and radar reflectivity that con- Part 2 of this study. flight, which sampled lower elevations (seelarger Fig. particles 2); owing contributed also more to substantially the likelihood to that that unsampled flight, Santiago results are omitted from in each flight testalthough location more show smaller similar and behavior, with larger calculated particles MMD appear likely to have been present in the Santiago (1–2, 2–4, and greater thanthey appear 4 roughly self-similar insofara as they factor exhibit of similar four width in and shape mass across concentration. roughly Three other flights with the highest IWC encountered more than an order ofthe calculated magnitude mass variability size in distributions IWC. are averaged Figure within 6 the three further mass demonstrates ranges that shown in when Fig. 4 particle mass at the time series in Fig. 5 show that calculated MMD particles at lower elevations than sampled during this flight. be as truncated as by an unknown amount, agreementmay of not integrated be IWC a with great the amount. We Robust might probe expect suggests a that greater mass this contribution from unsampled larger include the largest particles contributing tocalculated calculated values reported. Calculated tributions at the largest particles sizes indicates that the nephelometer size distributions do not that IWC obtained from theculated Robust probe mass closely size tracks IWC distributionsfrom obtained over from the roughly integration nephelometer of 0.1–4 exceeding cal- that from the Robust probe by roughly 25 tribute to high IWC–low median diameter multiplied by a dielectric factor of calculated from the ice mass size distributions as the integral of the sixth moment of the melted Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | at :: 1422 ::::: legs :::: flight ::::: flight ::::: of :: for ::: only ::::: segments ::::::::: diameter (Heymsfield et al., short ::::: available :::::::: km altitude in the Cayenne and Dar- are :::: several :::::: km from ::::: were also noted in all three of the CEPEX distributions max aside ::::::::::: ::::: 14 different. ::::::::: range in both fresh and aged anvils around Central 5), ::: in the densest anvil clouds in different parts of the size . Lawson et al. (2010) also reported a consistent con- :::: max eq m Fig. :::: D m Darwin ::::::: (e.g., ::::: qualitatively ::::::::::: C and :::: not ::: 40 :::: are :::: of roughly 250 µ eq Cayenne :::::::: than ::::: which :::::: , that :::: C colder :::::: note :::: 33 :::: we :::: 2), ::: Encountering a consistency of MMD Within any single flight through dense anvil at circa 11 could explain such a pattern. The size range of mass concentration found during the Airbus flight America and West Africa; a particularly stablewithin peak turrets, of leading mass them in to that suggest size a range microphysical was pathway also of observed ice formation in updrafts that tion of mass at MMD centration of mass in the 100–400 µ geted central regions of the largest andshortly coldest-topped after clouds in electrical all activity locations ceased), (incesses and some that cases that very lead the to coupled the dynamical highest and IWC microphysical in pro- such clouds do so consistently by producing a concentra- world may also be surprising givenVarble the et relatively wide al., range 2015). of One updraft strengths possible expected explanation (e.g., is that the Airbus flight tests systematically tar- 2002). that exponential fits tocept ice but size little distributions in exhibited slope order-of-magnitude variability within in individual inter- spiral loops of 5–10 anvils examined by McFarquhar andsize sorting Heymsfield must (1996), not leading be the thestudy only of authors dominant precipitating to process deep in conclude the convection that CEPEX spanning anvils several they TRMM examined. field A later campaigns also reported sistent with size sortinghorizontal documented regions by of Garrett relatively et uniform al. MMD (2005) within Florida anvils. However, 2010). Consistent with that reasoning,ice McFarquhar crystals and Heymsfield were (1996) found reported preferentially that larger nearer to convective cores and lower in the anvil, con- be considered a surprisingflow from result convective updrafts because and it subsequent sedimentationlarger is consistently particles leads generally and to thought preferentially more that mass the occurring closer process to of updraft ice sources out- (e.g., Lawson et al., 1998, win flights, a stable self-similarity of mass size distributions over a wide range of IWC can temperatures ::::::::::: reaching :::::::: Fig. ::: Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | e Z we the ::: size will 20% :::: :::: :::: been ::::: ::::: higher :::::: instead weakly we ::: ::::::: : the sampled has :::: :::::::: Since m campaign, :::: µ of It :::::::::: ::: :::::: and contaminate ::::::::::: ::: :: Here roughly ::::: m :::::::: µ to may ::::: ::: configurations, Content ::::::: :::::::::::::: 400-600 extensively :::::::: :::::::::: moments that :::: ::::::::: of inlet ::: 250–500 subject ::::: Water eq nephelometer. :::::: ::::::: ::::::::: ::::::::::::: system at relevant volumetric scales ::::::: be Ice e ::: ::: a :: artifacts Z :::::::: : MMD in ), but variability in ice hydrom- :::::::: different 3 Airbus :::::::: :::::: may typically C High higher-order ::::: ::::: :::::::: / :::::::::::: km eq :: the with max :::: 36 for (1 typical ::: ::::: to ::::::: :: :::: (Jackson and McFarquhar, 2014). O shattering :::::::::::::::::::::::::::::: at :: :::::::::: MMD :::::::: , it provides the valuable benefit of an inde- MMD 3 Crystals e ::::::: probes that :::::: :::::::::: Z :::: reduced as 15 yielding Ice ::::::: :::::::: :: for instance :::: ::: gm extended report ::::::::: :::::::: :::::: note :::: for be such ::: ::: :::::: we 0.5–3 system :::::: ::::::: :::: Altitude can ::::::: relatively :::: ::::::::: recently :::::::: performed over :::::::::: ::::: are artifacts, Aside, High :::::::: ::::: :::::: :::: measures another was :::: :::::::: ::::::::: (Korolev et al., 2013a; Jackson and McFarquhar, 2014). IWC :::: in :: :::::::::::::::::::::::::::::::::::::::::::::::::: IWC. conclusion recent :::::: :::::: :::::::::: mass shattering ::::: ::::::::: the with :::: to as :: ::::: distribution ::: general increasing measurements :::::::::: ::::::: ::::::::::: :::::::::::::: during ), in addition to mean Doppler velocity (MDV). Since MDV represents a such :::::: size that 3 owing measurements ::::: :::: :::::: ::::::::::::: with ::::: ::::::::::::::::::::: ::::: Leroy et al. (2015) more m probe increasing :::::: that that :::::::::: :::: :::: Darwin (100 ::::::: similar ::::::: within dense anvil regions. Ground-based scanning weather radars can offer stormwide O e pendent measurement constraint that is also sensitive to mass size distribution. of order moment of the ice size distribution than measurements at relevant volumetriceteor resolutions single-crystal of and size-distributiontime. properties Vertically pointing preclude radars reliable can also retrievals provide of non-attenuated IWC at this neous stormwide coverage, forZ instance, but include no direct information on either IWC or than in situ flightwith legs in and situ in probes part and because measurement they techniques. are Satellite free infrared of images the can various offer limitations instanta- associated Owing to the uniquenesshow of with the independent measurements. Airbus It flightpossible, is test in also part data, desirable because to it they make is commonly use desirable offer of the remote-sensing to advantage data corroborate of if it better spatiotemporal some- coverage 4 Profiling radar measurements uncertainty :::::::::: expect :::::: found ::::: distribution :::::::::: Korolev et::::::::::::::::::::::::: al. (2013a) study but ::: weakly :::::: airborne ::::::: decreasing ::::::::: near :::: less throughout previous scientificsampled. campaigns, even within graupel-containing updraft turrets tests is also roughly consistent with that reported from CEPEX, although IWC was substantially Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | km. :::: MDV 11 :::::: ::: to :: circa :::: winds , where melting :::::: km , MDV and Doppler and Doppler spectral 1 elevations vertical ::::::: :::::::::: at . Below the bright band, :: of ::: ms km ice ::: :::::::::: Reflectivity of :: . Where a bright band is present, km speed :::::: contribution ::::::::::: fall , where soundings available from Darwin ::: Reflecitivity in excess of 10 dBZ commonly still extends , close to that during Cayenne flight F1423. e km mean :::::: C Z 16 the is greater than 25 dBZ. The beginning of the first ::: 43 downward. The strongest drafts are colocated with e 1 Z , where as the meltwater undergoes conversion to faster-falling :::::: km ms e Z reflectivity-weighted is greater than 5 dBZ, MDV is 0–2 ::::::::::::::::::: regions ::::::: in excess of 30 dBZ. e . At 4.7 e the Z ::: 1 ) to the last panel. Also shown in Fig. 9 are retrievals of vertical ) corresponds to the earliest panel in Fig. 7 and the end of the second Z , to :: km ms UTC UTC stratiform :::::::::: relative ::::::: at this elevation nears 30 dBZ (during convective periods), it does not exceed , with some detectable echoes extending above 15 e Z upward and exceeding 10 km extended :::::::: 1 where, at 11.7 neglected ::::::::: on ::: be ::: min ms The S-band MDV curtain in Fig. 8 shows two broad periods of stratiform rain passing over Figure 9 shows time series extracted from the S-band profiles and surface rain rate measured wind speed obtained from the profilers at the highest elevation where retrievals were usually period (12:17–14:07 period (20:13–21:49 15 width is less than 2.5 on 23 January indicateAlthough a temperature of that. The two stratiform time periods within dashed lines are the contiguous periods of at least by a colocated NOAA Joss Waldvogel disdrometer. spectral width are shown at an elevation of 11.7 heavy rain is indicated by owing to the melting ofabove ice into 11 stratiform rain, rain. During the turbulent convective period,10 Doppler velocities reach localized peaksthe of strongest nearly deep echoes observed in the storm above 15 ized by a sharpice decrease in also MDV creates at a the bright melting level band height in of roughly 5 the profiler on 23throughout January the 2006, observed column bisected at by roughly a 23.65–23.83. several-hour The period stratiform periods of are intermittent character- turbulence of a tropical MCS nearof Darwin these (Fig. measurements 7). in The comparison remainder with of the this Airbus flight work test is data. devoted to examination compared with a scanningcampaign, radar the volume. NOAA However, S-band on profiler 23 obtained January a time 2006 series during of the measurements TWP-ICE during evolution can ::: The chief disadvantage of profiling radars is the severely reduced spatial coverage of a profile focus ::::: Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | , ) e 1 Z km mm h frequently e Z near the midpoints conditions beneath . At the surface, rain 1 1 e Z under conditions similar ; omission of the full mh 1 3 mm h m during each period, consistent ms 6 1 below the melting level exceeds e mm Z ms reaches nearly 4 km 17 are likely indicative of gravity waves, and these generate , it appears possible that MDV may be a weak function of km in these regions, consistent with rough self-similarity in size e km in linear coordinates during each stratiform period. Mean values Z e Z -weighted fall speeds of roughly 1 e Z ) and near-surface rain rate where retrievals can confidently be made (2.5 and in MDV at 11.7 km km legs aloft. Colocated retrievals of rain rate indicate precipitation exceeding 3 can be associated with relatively stable MDV of circa 1 e e Z Z , as suggested by MDV calculated from the Airbus data, as expected in the case that size Overall, the S-band measurements and retrievals appear to generally confirm that regions of Figure 10 shows MDV vs. e low- consistent with heavy rain expected beneath high IWC conditions. distribution shape found in theexceeds in situ 30 Airbus dBZ measurements. Figure below 9 these shows that stratiform regions, consistent with high- low to those sampled onwith flight a factor F1423. of MDV, four which change includes in air motions, does not vary strongly the contribution of the largest particles to MDV in the Airbus data. speeds do appear fasterspectrum where apparent in reflectivity the exceeds Airbus 20–30 measurements, as discussed above, may substantially truncate Z distributions are not strongly dependent on IWC. During the second period in Fig. 10, fall a significant component ofconsidered scatter beyond around the the scope meanto of in lack this Fig. of work. 10; retrieved Although extraction winds it of at is gravity 11 difficult waves to is draw conclusions owing Heymsfield and Westbrook (2010). In the profilerwinds data, at coherent 9 oscillations in retrieved vertical of MDV indicate with those calculated from Cayenne flight F1423 measurements (Fig. 5) using the method of rates measured by disdrometer tend toof read the sustained stratiform peaks periods circa rather 2–4 than at the endpoints. 40 dBZ and maximum rain rate retrieved at 2.5 and no strong updrafts. During each period, maximum (Williams and May, 2008;spectral Williams width indicate and that Gage, the 2009). two Retrieved stratiform vertical periods are winds quiescent, and with S-band minimal turbulence possible (9 Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | on UTC , heavy stratiform measured between e Z 3 gm range at temperatures below max D m vertical grid spacing that is initialized that were shown to agree with the Robust m eq D . At the melting level relative humidity is roughly 18 km and MDV are shown with blue asterisks in Fig. 11. Here e Z shown in Fig. 6 represent a mean IWC of 4.3 3 . The calculated gm km . This exercise therefore also serves to test how well a column simulation with size- C , increasing above and below with respect to ice and water, respectively (not shown). % We initialize the column with Airbus ice size distribution measurements obtained over Dar- We begin with an atmospheric column with 250 40 11 and 11.5 we assume that the particlesprobe have measurements. the For mass and simplicity, we assume that the ice particles are spherical in shape. win during flight F1405. Thegreater mass than size distributions 4 averaged over regions with reported IWC 90 rain indicated in Figs. 8 andsecond 9. half Aside of we 23 note that Januaryrespect no owing to sounding to ice is balloon at available relevant failures. at elevations The Darwin above 8 during atmospheric the column is saturated with 23 January 2006, obtained from a variationalarray analysis (Xie of conditions et over al., the TWP-ICE 2010). sounding This time is selected to be representative of high- thermodynamically with a mean of soundings observed around at Darwin at 21:00 observation-based estimate of bin-wise ice properties,dominant and processes when of limited sedimentation to and the aggregation relatively in simple the stratiform region. with concentration of mass narrowly in theresolved 100–500 microphysics µ can perform when initialized with the observed size distributions and an heavy rain rates retrievedtions below, for generally fail instance. to As reproduce shown the in observed Part ice 2, size distribution three-dimensional features simula- in deep convection, during TWP-ICE with the profiling radarevaluation observations. of This whether comparison the is intended ice to mixing allow ratios an measured in high IWC regions are consistent with generally consistent with other recentand in with situ profiling measurements radar into measurements the compare vicinity in one-dimensional of a simulations deep large of convection tropical the MCS. quasi-steady Here stratiform we rain take regions observed a step further As shown above, the ice size distribution features found in Airbus flight tests are found to be 5 Column model simulations Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | ), i E monoton- e Z and MDV represent the e , the projected area is re- Z eq D is equal to max and MDV) or retrieved with a combination of D e 19 Z , Doppler velocity, precipitation rate, and CWC (liquid plus e Z , we therefore assume that eq D by definition, and the overall particle aspect ratio is unity. Particle fall speeds and eq D In Fig. 11 are shown simulated For these simulations, we adopt a steady-state simulation approach, considering that large mixed-phase particles from this simplified model, ice instantly melts when the temperature is and ice morphological propertiestionally in (e.g., a Kienast-Sjögren manner et that al., is 2013, and not references generally therein). well Owing established to observa- the omission of defined here as the probabilityis taken that here an to ice-ice be collision fixed, results although it in may aggregation. be The a probability substantial function of environmental conditions ically increases. In the simulations, the rate of increase scales with the sticking efficiency ( periods shown in Figs. 8 andsteady-state 9. boundary At condition. the As top of the the ice simulated aggregates columns, during sedimentation, the S-band, UHF and VHFretrieved data data below at the each elevation melting represent level means (rain of the rate available and values CWC). over the The stratiform observed time and ditions (the observed sounding orare a profiles saturated measured by version) the after S-band reaching profiler steady ( state. Also shown ice) for two values of ice-ice sticking efficiency (0.01 and 0.005) and two environmental con- cooling at the melting level,remain which stable would as otherwise observed. require Melting application ofice of ice to is divergence melted terms treated equivalent in to drops a when rudimentary the fashion, column by temperature transferring allows. all regions, treated as described in Fridlindwe et fix al. temperatures (2012b) and and discussed water further vapor in mixing Part 2. ratios However, to the observed profile, thereby neglecting tions above the meltingwith level with hydrometeors the is same permitted observed to mass represent size evaporation distribution. of Vapor exchange ice and liquid in the unsaturated and long-lived stratiform regions areroughly changing equal quantities slowly of over mass coming hourslarge-scale from in above vertical and time, motion. exiting therefore To below speed at exhibiting each up level. simulation We neglect time to steady state, we initialize all eleva- following Heymsfield and Westbrook (2010) as described by Avramov et al. (2011). lated to pairwise collision efficiencies are calculated following Böhm (1999), with fall speeds modified Given mass and Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 1 is e Z or the e mm h Z at the melting and the surface. % . Thus, aggregation km km , with the ice morphologi- km with height above the melting e Z introduced at 11 3 20 gm decrease in IWC between 11 and 5 % corresponds to a reduction of IWC as ice sediments faster. e under saturated conditions. Figure 12 shows that half as much IWC Z 1 mm h Unlike the constant precipitation rate profile in a saturated column at steady state, the sim- Simulations are shown using an observed sounding (with RH of roughly 90 MDV above the melting level is usually similar to observed or is underestimated by a few level also exhibit a roughly 30–50 consistent with increasing MDV), which is accompanied by aprecipitation rapid rate. decrease Above in the CWC melting toimentation; level, maintain IWC simulations a undergoes uniform that a (steady-state) roughly more gentle reproduce decrease the during trend sed- in ulated CWC profile decreasesMost monotonically dramatically, and melting drastically ice between causes 11 a rapid increase in fall speed (consistent with observed throughout a saturated column. rate of roughly 8 aloft, without a change in ice size distribution or properties, corresponds to roughly 4 sounding demonstrate that in a steady-stateis environment constant without with evaporation, height. precipitation Thus rate ancal IWC and of size 4 distribution properties consistent with Airbus measurements, corresponds to a rain level, as described above) andice with and the water same above sounding and that below has the been melting saturated level, with respectively. respect Simulations to with the saturated which could enhance aggregation. face, except in one simulationstantially with increase a MDV. However, saturated underestimation sounding, ofulation indicating MDV could that below also evaporation the be may melting attributable sub- level to in missing a mixed-phase sim- processes (e.g., Phillips et al., 2007), tenths of a meter persimilar second. to MDV observed, below including the reproducing melting a reduction level is of MDV similar with to distance observed towards when the sur- smooth increase in MDV seen in the observations. warm enough, and simulations therefore do not reproduce the melting band feature in Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | km km can 1 in alti- km mm h and rain rate from C- e Z does not exceed 30 dBZ at 11 there remain only several tiny areas e Z km cross-sections of km , although higher rain rates of 4–8 1 21 band scanning polarimetric radar at Darwin (C-POL, mm h C during roughly an hour-long period containing convective 1 northwest of the profiler. Thus maximum stratiform rain rates ms km ; these areas are periodically slightly larger during some periods earlier and later, e Z Although convective rain areas that contain some amount of hail are found north and south but are never extensive duringfrom this the profilers event reach (not 4–10 shown). Vertical wind speeds retrieved at 9 of hail according to C-POLof definitions. 30-dBZ However, at 11 of the strongest convective rain overwithin the a S-band, linear by band contrast. of Here convection the containing S-band a is continuous now embedded elongated region with some amount of the stratiform rainanywhere area within the over C-POL the domain at S-band that in time. Fig. Figure 14 13, shows the C-POL fields at the time from mesoscale ascent concentrated atrain lower rate locally elevations by or other other means, processes rather that than could high IWC enhance far aloft. menting under saturated conditions, inThe our calculations, locally but stronger such stratiform rain rates rain are rates not could widespread. also be attributable to enhanced precipitation be found locally roughly 50 in these observations appear roughly consistent with the maximum measured IWC aloft, sedi- using the Steiner et al. (1995)are classified textural as algorithm stratiform. as At in the Fridlind timeiform et shown, area the al. where S-band most (2012a); is rain surrounding embedded rates areas within are an 2–4 extensive strat- POL measurements near the time oflocation maximum is stratiform indicated rain with rate an at asterisk. the S-band Areas profiler, outlined whose in black are those identified as convective Fig. 8, the secondseen turns by out disdrometer to (Fig. 9). be Figure far 13 shows more 2.5 extensive horizontally, with higher peak rain rates Keenan et al., 1998)tude offers following a Bringi wider et view(Keenan, al. 2003; of (2001), May rain and and rates, a Keenan, which 2005). three-dimensional are Of view the retrieved of two at hydrometeor quiescent 2.5 class stratiform aloft periods identified in The Australian Bureau of Meteorology 6 Meteorological context Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | conditions encountered e Z , the Airbus flights primarily and damage events described C : loss ::: 26 less 22 were not rare at any of the three global regions sampled, and maximum IWC 3 gm Similar to CWC measurements reported by McNaughton (1959), IWC measurements of Separation of rainy areas into convective and stratiform areas has an origin in the first-order roughly 2 2010). IWC regions, within the limits(1959). of Whereas safety, that in earlier a study mannersampled measured similar at CWC to colder at that temperatures, 0 described where many to by such McNaughton events have occurred (cf. Grzych and Mason, a new SEA Robustsize hot-wire distributions. A probe chief and objective of an the imaging flight tests nephelometer was used to find to and report remain within ice the particle highest almost exclusively in theturbulence. vicinity During 2010–2012 of Airbus deep flight convection tests from under Cayenne, conditions Darwin, of and Santiago weak carried to moderate Roughly a hundred well-documentedin jet the engine literature power to date have been characterized by high IWC–low 7 Conclusions commonly scattered within regions identified asvertical stratiform, winds consistent and with low otherwise rain quiescent rates. a large variety (Mapesvection and is Houze, sometimes 1992, organized cf. into their linear Fig. features, 15). but individual In convective the turrets system are studied also here, con- unique consequences for atmospheric circulationZeng (e.g., et Houze, al., 2004; 2013). Schumacher However, the et structure al., of 2004; convective regions within a tropical MCS present particle sink regions dominatedSubsequent by analyses ice usefully particle extended sedimentation thisdiffering (Biggerstaff framework properties to and consideration of Houze, of convective 1991). tropical and storms, stratiform the region energy and water budgets, and their division of mid-latitude squall-lines into icemonly exceed particle ice source crystal regions, fall where speeds vertical (thus winds transporting com- particles to detrainment regions), and ice moderate turbulence. cells over the S-band, including the time of Fig. 14. These would presumably be associated with Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | , which is consistent and mean Doppler ve- km e Z at 11 e Z . e Z , consistent with that calculated from Airbus ), perhaps unrealistically enhancing the stabil- 1 23 mm ms size distributions from the nephelometer data as shown in length scales (see also Grandin et al., 2014). These IWC e is less than 30 dBZ at typical Airbus sampling elevations of Z km e Z indicate that mass size distributions were dominated by particles in the 100– nominally smaller than 2–3 km max D range of area-equivalent diameter consistently across all locations. The mass size distri- and MDV is found to be roughly 1 m , complicating analysis of MDV relationship to km An examination of the Airbus nephelometer data in the regions of highest IWC at eleva- Owing to the great uncertainty associated with both IWC and size distribution measurements, km 9 with self-similar ice size distributions. However, gravityof waves with the periods stratiform of perhaps rain one-half duration appear present in vertical wind speed retrievals available up to 11 flight test data. MDV appears relatively weakly correlated with remote-sensing instruments over Darwintions on are 23 neglected January to 2006,heavy first where stratiform order large-scale rain vertical compared periods, mo- with reflectivity-weighted ice fall speeds. During distribution properties over thelocity full atmospheric (MDV). column We focus observed: on regions of heavy stratiform rain within an MCS observed with here we describe asurements comparison that of offer the two measurements indepdendent with measurements ground-based sensitive profiling to radar ice mea- morphological and size also been reported in previous data sets and analyses. Fig. 4 make apparent that(limited the to largest particles present wereity not of mass measured median by diameter the over nephelometer the size range seen. On the other hand, similar features have butions also consistently exhibited an unexpected self-similarity,changes or in stability of IWC. shape, Calculations over of large tions circa 10 500 µ ground truth (Strapp et al., 2008;difference Baumgardner across et various al., direct 2011), and could indirect also measurement account approaches. for some of the of such measurements is notassigned yet to available. IWC In measurements addition, by the any factor-of-two means, uncertainty in generally the absence of past test facilities to provide measurements are roughly agions factor documented of in the two scientific greater literature than to date. measurements However, from as noted similar above, cloud a large re- database exceeded twice that value at 10 Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | undersaturated with decreasing % e Z reduced IWC by the melt- % is found to correspond to a rain rate of 3 gm 24 ice properties, albeit with large uncertainties. The objective of Part 2 is e in the 23 January event, but polarimetric retrievals of rain rate around the TWP-ICE soundings and S-band radar data were obtained from the Atmospheric . This is roughly twice the maximum rain rate observed by disdrometer or Z 1 km mm h The objective of Part 1 of this two-part study has been to vet the Airbus size distribution mea- Column model simulations initialized with ice morphological and particle size distribu- ence, Office of Biological and Environmental Research, Climate and Environmental Sciences Division. Acknowledgements. Radiation Measurement (ARM) Program sponsored by the U.S. Department of Energy, Office of Sci- to examine microphysical pathwaysrelevant by updraft which conditions. such properties could likely be achieved under relating IWC aloft withclude rain that rates size below. distribution Given and aof IWC lack high measurements of IWC–low appear gross to inconsistencies, give here a reasonable we representation con- surements to the degree possible inremote-sensing comparison data with past set. measurements A and simplified a modeling uniquely framework relevant served to extend the comparison by stratiform rain rates and ice properties observed during the 23 January 2006 event over Darwin. such conditions, IWC is expectedhighest to Airbus be IWC higher measurements aloft near than Darwin at do the appear melting grossly level. consistent In with summary, the the highest steady-state simulations with aggregationelevation sufficient owing to to aggregation explain alone increasing ing also level exhibit owing roughly to 30–50 increasing mass-weighted fall speed. Thus, in a steady-state column under equally possible. Model simulations alsoin indicate the that vicinity of in the a melting column level, that precipitation can is be 10 rapidly reduced by evaporation. Finally, profiler did report such rateslocally enhanced locally rain within rates can the be stratiform attributed region. to ice It aloft is or some unknown mesoscale whether ascent, such which seems rate is independent ofroughly height, and 8 an IWC ofretrieved at 4 2.5 precipitation appears consistent withwith respect the to measurements. ice and If water above the and column below the is melting level, assumed respectively, then saturated precipitation tions consistent with Airbustest whether Robust the probe properties and of nephelometer a steady-state measurements stratiform are rain intended column under to conditions of heavy Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 25 Environment Safety Technologies Project. Wesearch. thank Airbus for providing their data to support this re- Weather and Climate Research, Australian Bureauand of Ron Meteorology. Colantonio The for authors logistical thank andfor Thomas programmatic valuable Ratvasky assistance, discussion. and This Jeanne work Mason was and supported Matthew by Grzych the NASA Aviation Safety Program’s Atmospheric C-POL radar measurements and retrieval products were supplied by Peter May, Centre for Australian Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | G.: ::: Conf., :::::: Zalamansky, ::::::::::: and Environments :::: :::::::::::: P., :: Space ::::: and :::: Villedieu, ::::::::: H., ::: Atmospheric ::::::::::: Schalger, :::::::: F., 26 ::: AIAA :::::: Fifth Hervy, ::::: :::::: in: :: J.-L., ::::: Crystals, :::::::: Ice Brenguier, :::: ::::::::: A., ::: Altitude ::::::: Grandin, High :::::::: ::::: - F., :: : HAIC ::::: distribution for remote sensing application, J.2014. Geophys. Res., 119, 1–24, doi:10.1002/2013JD020700, Whiteway, J., and Lawson, R. P.:gation Aircraft of observations ice of crystals, the Q. influence J. of Roy. electric Meteor. fields Soc., on 131, the 1695–1712, aggre- doi:10.1256/qj.03.217, 2005. water probe, J. Atmos. Ocean. Tech., 12, 410–414, 1995. tial propagation phase: evaluation usingAtmos. a Ocean. Tech., C-band 18, radar 1810–1818, and 2001. a dense gauge network in the tropics, J. physics”, Atmos. Res., 52, 167–176, 1999. line, Mon. Weather Rev., 119, 3034–3065, 1991. June 1985 squall line, J. Atmos. Sci., 50, 3091–3091, 1993. aerosol particles, clouds and radiation: a102, cook’s 10–29, tour of doi:10.1016/j.atmosres.2011.06.021, mature 2011. and emerging technology, Atmos. Res., mann, M., Heymsfield, A., Korolev, A.,Twohy, C., Krämer, Wendisch, M., M., Bachalo, Petzold, W., A., and Strapp, Chuang, W., P.: Airborne Pilewskie, instruments P., Taylor, to J., measure atmospheric particle imagery. Part I: Image-to-mass relationships, J. Appl. Meteorol. Clim., 45, 1282–1290, 2006. M.: Toward ice formationGeophys. closure Res., 116, in D00T08, Arctic doi:10.1029/2011JD015910, 2011. mixed-phase boundary layer clouds during ISDAC, J. Korolev, A. V., Strapp, J. W., McFarquhar, G. M., Jackson, R., Brooks, S. D., Glen, A., and Wolde, ::::::: Dezitter, Delanoë, J. M. E., Heymsfield, A. J., Protat, A., Bansemer, A., and Hogan, R. J.: Normalized particle size Connolly, P. J., Saunders, C. P. R., Gallagher, M. W., Bower, K. N., Flynn, M. J., Choularton, T. W., Brown, P. R. and Francis, P. N.: Improved measurements of the ice water content in cirrus using a total- Bringi, V. N., Huang, G.-J., Chandrasekar, V., and Keenan, T.: An areal rainfall estimator using differen- Böhm, J. P.: Revision and clarification of “A general hydrodynamic theory for mixed-phase micro- Biggerstaff, M. and Houze Jr., R. A.: Kinematics and microphysics of the transition zone of the 10–11 Biggerstaff, M. and Houze Jr., R. A.: Kinematic and precipitation structure of the 10–11 June 1985 squall Baker, B. and Lawson, R. P.: Improvement in determination ofBaumgardner, ice D., water Brenguier, content J.-L., from two-dimensional Bucholtz, A., Coe, H., DeMott, P., Garrett, T. J., Gayet, J. F., Her- Avramov, A., Ackerman, A. S., Fridlind, A. M., van Diedenhoven, B., Botta, G., Aydin, K., Verlinde, J., References Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | AIAA :::::: 2013, ::::: June :::: 24–27 ::::: California, :::::::::: Diego, ::::: San :::: 27 Astronautics, ::::::::::: and :::: Aeronautics :::::::::: of :: : 2013. ::::: Institute ::::::: 2424–2451, 1996. microphysical interpretations from CEPEX and from Kwajalein, Marshall islands, J. Atmos. Sci., 53, Range, and Aerospace Meteorology, Atlanta, Georgia, USA, 17–21 January 2010,Sci., 6.8, 43, 2010. 2463–2478, 1986. doi:10.2514/6.2014-2753, 2014. due to ingestion of ice particles: What we’ve learned through 2009, in: 14th Conference on Aviation, water content regions, in:Institute 6th of AIAA Aeronautics Atmospheric and and Astronautics, Space Environments Reston, Conference, Virginia, American 16–20 June 2014, AIAA 2014–2753, tems and potential implications899–912, for doi:10.5194/acp-14-899-2014, 2014. commercial aircraft at flight altitude, Atmos. Chem. Phys., 14, physical properties and high ice water content in continental and oceanic mesoscale convective sys- Herman, R. L., Heymsfield, A.62, J., 2352–2372, and doi:10.1175/JAS3495.1, Lawson, 2005. P.: Evolution of a Florida cirrus anvil, J. Atmos. Sci., Zuidema, P., and Shupe,layer M. clouds: D.: Entrainment A rate FIRE-ACE/SHEBA limitations69, case 365–389, on study doi:10.1175/JAS-D-11-052.1, rapid 2012b. of primary mixed-phase ice Arctic nucleation boundary processes, J. Atmos. Sci., doi:10.1029/2011JD016595, 2012a. T. R., Khaiyer, M.J. M., P., Liu, Schumacher, C., G., Shipway, Minnis,of B. P., TWP-ICE J., Morrison, observational Varble, H., data A. with C., Nguyen, cloud-resolving Wu, L., model X., Park, results, Xie, S., J. S., Petch, Geophys. and Res., J. Zhang, 117, M.: C., D05204, A Pinty, comparison Aeronautics and Astronautics, Reston, Virigina, 5–8 January 2004, AIAA 2004-731 ,2004. 2013-2674, ::::::::: tent instrumentation, in: 42nd AIAA Aerospace Sciences Meeting and Exhibit, American Institute of American :::::::: Heymsfield, A. J. and McFarquhar, G. M.: High albedos of cirrus in the tropical Pacific warm pool: Heymsfield, A. J.: Ice particle evolution in the anvil of a severe thunderstorm during CCOPE, J. Atmos. Grzych, M. L. and Mason, J. G.: Weather conditions associated with jet engine power loss and damage Grandin, A., Merle, J.-M., Weber, M., Strapp, J., Protat, A., and King, P.: AIRBUS flight tests in high total Gayet, J.-F., Shcherbakov, V., Bugliaro, L., Protat, A., Delanoë, J., Pelon, J., and Garnier, A.: Micro- Garrett, T. J., Navarro, B. C., Twohy, C. H., Jensen, E. J., Baumgardner, D. G., Bui, P. T., Gerber, H., Fridlind, A. M., van Diedenhoven, B., Ackerman, A. S., Avramov, A., Mrowiec, A., Morrison, H., Fridlind, A. M., Ackerman, A. S., Chaboureau, J.-P., Fan, J., Grabowski, W. W., Hill, A. A., Jones, Emery, E., Miller, D., Plaskon, S., Strapp, W., and Lillie, L.: Ice particle impact on cloud water con- Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | : cloud ::::: effects artifact :::::: :::::: 2013. 2D ::: ::::: , and :::: C, J. Atmos. the ::: mitigate tips by ::::::: ::: :: 86 2014. ::::: to , :: : to tips ::: measured ::::::::: probe ::::: antishattering :::::::::::: of :: particle properties ::::::: :::::::: of :: impact :::::: 10.1175/JTECH-D-12-00142.1 tests the optical :::: ::: :::::: doi: :::: of :: and and ::: ::: 10.1175/JTECH-D-14-00018.1 28 doi: ::: 690–708, :::::::: assessment 30, ::: :::::::::: Modification ::::::::::: An microphysical ::: :::::::::::: K.: 2131–2144, Tech., ::: ::::: :::::::::: ice M.: ::: ::: 30, ::: G. ::: cloud ::::: Ocean. ::::::: Creelman, :::::::: Tech., :::::: bulk :::: and on :::: :: Atmos. :::::: E., McFarquhar, :: J. Ocean. :: :::::: ::::::::::: and ::: Strapp, J. W., Isaac, G. A., and Emery, E.: Improved airborne hot-wire measurements of ice :::::: Emery, C., Atmos. ::: :::::: techniques :::::::::: J. R. A., ::: :: :: shattering, ::::::::: ice :: and possible weather threat to aviation, J. Aircraft, 35, 113–121, 1998. water content in clouds, J.2013. Atmos. Ocean. Tech., 30, 2121–2131, doi:10.1175/JTECH-D-13-00007.1, B. Am. Meteorol. Soc., 63, 386–398, 1982. of :: 13-9021-2013, 2013. polarimetric (C-Pol) radar system, J. Atmos. Ocean. Tech., 15, 871–886,for 1998. two-moment bulk microphysics schemes, Atmos. Chem. Phys., 13, 9021–9037, doi:10.5194/acp- 31, 2003. removal :::::: probe, ::::: doi:10.1029/2004RG000150, 2004. dependent terminal velocities from inSci., situ 70, observations 4123–4154, at doi:10.1175/JAS-D-12-0124.1, temperatures 2013. from 0 stratiform precipitating clouds: results from inSci., situ 59, observations 1–35, in 2002. TRMM field campaigns, J. Atmos. C. A.: Observations and parameterizations of particle size distributions in deep tropical cirrus and storm water-budget estimates, J. Clim. Appl. Meteorol., 25, 691–702, 1986. tory and field measurements, J. Atmos. Sci., 67, 2469–2482, 2010. ::::::: ::::::: Lawson, R. P., Angus, L. J., and Heymsfield, A. J.: Cloud particle measurements in thunderstorm anvils Korolev, Korolev, A., Knight, C.: The Cooperative-Convective-Precipitation-Experiment (CCOPE), 18 May–7 August 1981, Kienast-Sjögren, E., Spichtinger, P., and Gierens, K.: Formulation and test of an ice aggregation scheme Keenan, T., Glasson, K., Cummings, F., Bird, T., Keeler, J., and Lutz, J.: The BMRC/NCAR C-band Keenan, T.: Hydrometeor classification with a C-band polarimetric radar, Aust. Meteorol. Mag., 52, 23– Houze Jr.,Jackson, R. A.: Mesoscale convective systems, Rev. Geophys., 42, RG4003, Heymsfield, A. J., Schmitt, C., and Bansemer, A.: Ice cloud particle size distributions and pressure- Heymsfield, A. J., Bansemer, A., Field, P. R., Durden, S. L., Stith, J. L., Dye, J. E., Hall, W., and Grainger, Heymsfield, A. J. and Westbrook, C.: Advances in the estimation of ice particle fall speeds using labora- Heymsfield, A. J. and Palmer, A. G.: Relationships for deriving thunderstorm anvil ice mass for CCOPE Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | A., PA, :::: :::: Protat, distribution :::::::::: :::::: microphysics ::::::::::: 2007-01-3311, J., :::::::::::: ::: size Warrendale, , :::: :::::::::: : ë : PSD ::::: Rep. ::::: on ::: particle :::::: Delano :::::: Tech. ::::: L., International, ::::::::::: : :::: findings ::::::: diameters, SAE ::::::::: :::: Lilie, envelope, 2007. :::::::: ::::: ::::: , W., mass icing ::::: :::: :::: J. ::: Median interim campaign-specific ::::::: ::::::: ::::::::::::::: 2015-01-2087, ::::::::::::: Strapp, 29 An ::: ::::::: - field : :::: Rep. :::: A., D :: ::: 10.4271/2007-01-3311 Tech. ::::: doi: ::: measurements: :::::::::::: : Appendix PA, ::::::::: :::: HAIC/HIWC situ shapes, :::::::::::: :::: :::::: W.: in 2015. ::: :: ::::: A.: , ::: J. :: Schwarzenboeck, ::::::::::::::: crystal from :::::: ::::: E., Warrendale, ice :::::::::: :::: ::: Strapp, :::::: Grandin, :::::::: and and ::: :::: Regions and ::: ::::::: S. ::: Fontaine, F., :::::::: ::: R. IWC :: :::: International, D., ::::::::::: 10.4271/2015-01-2087 :::: high :::: 1784–1788, doi:10.1088/0964-1726/16/5/032, 2007. doi:10.1175/JAS3832.1, 2007. sphere of the equatorial zone, Tech. Rep. MECH. ENG. 283,and Farnborough, 1959. precipitation production in maritime and continental storm clouds, J. Atmos. Sci., 64, 338–359, the Central Equatorial Pacific Experiment, J. Atmos. Sci., 53, 2401–2423, 1996. SAE :::: profiler data, J. Appl. Meteorol., 44, 827–838, doi:10.1175/JAM2230.1, 2005. The tropical warm pool – international cloud experiment, B. Am. Meteorol. Soc., 89, 629–646, 2008. uary 2006, Reno, Nevada, 2006. 2011. Aerospace Sciences Meeting, American Institute of Aeronautics and Astronautics (AIAA), 9–12 Jan- convective systems. I: Horizontal structure, Q. J. Roy. Meteor. Soc., 118,tal 927–963, icing, 1992. Tech. Rep. 2011-38-0094, SAE International, Warrendale, PA, doi:10.4271/2011-38-0094, 79, 2185–2197, 1974. characteristics ::::::::::: doi: ::: Dezitter, ::::::: in :: tive properties of tropical cloudsdoi:10.1029/2009JD013017, investigated 2010. in TC4 and NAMMA, J. Geophys. Res., 115, D00J08, :::::::: ::::: Roques, S.: An airborne icing characterization probe: nephelometer prototype, Smart Mater. Struct., 16, Phillips, V. T. J., Pokrovsky, A., and Khain, A.: The influence of time-dependent melting on the dynamics McFarquhar, G. M. and Heymsfield, A. J.: Microphysical characteristics ofMcNaughton, three I. anvils I.: sampled The during analysis of measurements of free ice and ice/water concentrations in the atmo- Mazzawy, May, P. T., Mather, J. H., Vaughan, G., Jakob, C., McFarquhar, G. M., Bower, K. N., and Mace, G. G.: May, P. T. and Keenan, T. D.: Evaluation of microphysical retrievals from polarimetric radar with wind Mason, J. G., Strapp, J. W., and Chow, P.: The ice particle threat to engines in flight, in: 44th AIAA Mason, J. G. and Grzych, M.: The challenges identifying weather associated with jet engine ice crys- Mapes, B. and Houze Jr., R. A.: An integrated view of the 1987 Australian monsoon and its mesoscale Locatelli, J. D. and Hobbs, P. V.: Fall speeds and masses of solid precipitation particles, J. Geophys. Res., Leroy, Lawson, R. P., Jensen, E., Mitchell, D. L., Baker, B., Mo, Q., and Pilson, B.: Microphysical and radia- Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 30 Shipway, B., and Williams, C.: Evaluation of cloud-resolving and limited area model intercomparison doi:10.1175/1520-0426(1997)014<0197:MOCWCI>2.0.CO;2, 1997. Coupling Experiment (TC4), J. Geophys. Res., 115, D00J04, doi:10.1029/2009JD013073, 2010. and ice clouds using an airborne counterflow virtual impactor, J. Atmos. Ocean. Tech., 14, 197–202, P. O., Wofsy, S. C.,L., Kurylo, Rosenlof, K. M. H., J., Selkirk,ing, Maring, H. K. H., E.: B., Planning, Jucks, Colarco, implementation, P. K. and R., first W., Kawa, results Craig, S. of the R., M. Tropical Mace, S., Composition, G. Cloud Vasques, and G., M. Climate Minnis, F., P., Pfister, and Picker- Mexico, P13.1, 2008. 2005, AIAA 2005-0860, 2005. cell facility, in: 15th International Conference on Clouds and Precipitation, 7–11 July 2008, Cancun, measured by hot wire instruments ofing varying and configuration, in: Exhibit, 43rd American AIAA Institute Aerospace Sciences of Meet- Aeronautics and Astronautics, Reno, Nevada, 10–13 January AIAA Aerospace Sciences Meeting and Exhibit,Reno, American Nevada, 11–14 Institute January of 1999, Aeronautics and A99-16381, Astronautics, 1999. doi:10.5194/acp-14-1973-2014, 2014. physical measurements in thunderstorm outflow regions during Allied/BAE 1997 flight trials, in: 37th Rogers, D. C., Rutledge, S.,continental and thunderstorms: Toohey, D. the W.: case Ice for particles frozen-drop in aggregates, the Atmos. upper Chem. anvil Phys., regions 14, of 1973–1985, midlatitude updrafts in clean conditions, J. Appl. Meteorol., 43, 779–794, doi:10.1175/2104.1, 2004. structure from operational radar and rain gauge data, J. Appl. Meteorol., 34, 1978–2007, 1995. estimates derived from the TRMM precipitation radar, J. Atmos. Sci.,1258–1260, 61, 1984. 1341–1358, 2004. Varble, A., Zipser, E. J., Fridlind, A. M., Zhu, P., Ackerman, A. S., Chaboureau, J.-P., Fan, J., Hill, A., Twohy, C. H., Schanot, A. J., and Cooper, W. A.: Measurement of condensed water content in liquid Toon, O. B., Starr, D. O., Jensen, E. J., Newman, P. A., Platnick, S., Schoeberl, M. R., Wennberg, Strapp, J. W., MacLeod, J., and Lilie, L.: Calibration of ice water content in a wind tunnel/engine test Strapp, J. W., Lilie, L. E., Emery, E. E., and Miller, D.: Preliminary comparison of ice water content as Strapp, J. W., Chow, P., Maltby, M., Bezer, A. D., Korolev, A., Stromberg, I., and Hallett, J.: Cloud micro- Stith, J. L., Haggerty, J. A., Heymsfield, A., and Grainger, C. A.: MicrophysicalStith, characteristics of J. tropical L., Avallone, L. M., Bansemer, A., Basarab, B., Dorsi, S. W., Fuchs, B., Lawson, R. P., Steiner, M., Houze Jr., R. A., and Yuter, S. E.: Climatological characterization of three-dimensional storm Smith, P.: Equivalent radar reflectivity factors for snow and ice particles, J. Clim. Appl. Meteorol., 23, Schumacher, C., Houze, Jr., R. A., and Kraucunas, I.: The tropical dynamical response to latent heating Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 31 A comparison of the water487–503, budgets doi:10.1175/JAS-D-12-050.1, 2013. between clouds from AMMA and TWP-ICE, J. Atmos. Sci., 70, structures and diabaticdoi:10.1175/2009JCLI3071.1, 2010. heating and drying profiles during TWP-ICE, J. Climate, 23, 57–79, Waliser, D. E.: ComparisonsGeophys. of Res., global 114, cloud D00A24, doi:10.1029/2008JD009946, ice 2009. from MLS, CloudSat, and correlative data sets, J. 2008. J. L., Livesey, N. J., McFarquhar, G. M., Pittman, J. V., Stephens, G. L., Tanelli, S., Vane, D. G., and to rainfall uncertainties, J. Atmos. Ocean. Tech., 25, 1881–1887, doi:10.1175/2008JTECHA1038.1, 919–13–945, doi:10.1002/2013JD021371, 2015. tics, Ann. Geophys., 27, 555–567, doi:10.5194/angeo-27-555-2009, 2009. simulations using TWP-ICE observations: 2. Precipitation microphysics, J. Geophys. Res., 119, 13– Zeng, X., Tao, W.-K., Powell, S. W., Houze Jr., R. A., Ciesielski, P., Guy, N., Pierce, H., and Matsui, T.: Xie, S., Hume, T., Jakob, C., Klein, S. A., Mccoy, R. B., and Zhang, M.: Observed large-scale Wu, D. L., Austin, R. T., Deng, M., Durden, S. L., Heymsfield, A. J., Jiang, J. H., Lambert, A., Li, Williams, C. R. and May, P. T.: Uncertainties in profiler and polarimetric DSD estimates and their relation Williams, C. R. and Gage, K. S.: Raindrop size distribution variability estimated using ensemble statis- Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | (1) Eq. ) as a func- e Z 0 100 and mass median area-equivalent di- m) 800 µ ( ) (Locatelli and Hobbs, 1974). : 1 eq max eq 2 max eq ) for ice particle properties following 600 max with a monomodal ice number size distribution 32 3 or MMD max 400 gm MMD 200 Monodisperse MMD Monodisperse Monodisperse MMD Exponential MMD Exponential MMD 0

10 30 20

e (dBZ) Z :::::::::::::::::::::::::::::::: Baker and Lawson (2006, their Table 1) ) for ice particle properties following Eq. ( eq For an ice water content of 4 ameter (MMD tion of mass(Baker median and Lawson, maximum 2006) dimension (MMD Figure 1. that is either monodisperse or exponential in shape, equivalent radar reflectivity ( Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | km measurements corresponding to Hz ,1 1 ms , where liquid water contributions are expected C 33 20 (top row) slightly exceed those averaged over 2.5 and 10 km Total condensed water content (CWC) measurements by the Robust probe during flight tests to be negligible. Consideringa horizontal a length typical scale of airlength roughly speed scales. 0.2 of 200 from Cayenne (left column), Darwinincluded (middle only column), for and air Santiago temperatures (right column). were Measurements below are Figure 2. Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 34 on 16 May 2010. Cayenne flight F1423 track over brightness temperature from Meteosat 9 channel 10 at UTC Figure 3. 12:45 Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | (right column). Measure- 3 4gm > on 16 May 2010, during spirals within an MCS 35 UTC (middle column), and 3 gm (left column), 2–4 3 gm Cayenne flight F1423 ice number size distribution (top row), mass size distribution (middle (cf. Figs. 3 and 5). Red line indicates bin-wise median size distribution. ranges: 1–2 ments from the analyzed time period 11:30–13:20 Figure 4. row), and equivalent radar reflectivity size distribution (bottom row) in three ice water content (IWC) Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | ), ice mass me- e Z 13.0 13.0 13.0 13.0 13.0 ) during Cayenne flight F1423. Z V 12.5 12.5 12.5 12.5 12.5

UTC (h) 36 -weighted fall speed ( e Z ), and eq 12.0 12.0 12.0 12.0 12.0 Nephelometer Robust probe

11.5 11.5 11.5 11.5 11.5

0 0 4 2 0 8 6

) IWC (g m (g IWC 30 20 10

1.0 0.5 0.0 2.0 1.5

-10 -35 -30 -55 -50 -45 -40

-3 100 400 300 200

) s (m V Reflectivity (dBZ) Reflectivity Temperature (C) Temperature

eq µ MMD m) ( -1 Static air temperature, ice water content (IWC), equivalent radar reﬂectivity ( dian area-equivalent diameter (MMD Figure 5. Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 3 4gm > ). 3 4gm > m m m m 1000 1000 1000 1000 µ µ µ µ m) µ

( eq D Darwin F1405 MMD = 260 MMD = 280 MMD = 240 MMD = 250 Santiago F1546 Cayenne F1423 Cayenne F1422 100 100 100 100

0.06 0.04 0.02 0.00 0.08 0.06 0.04 0.02 0.00 0.08 0.06 0.04 0.02 0.00 0.08 0.06 0.04 0.02 0.00 0.08

eq eq eq eq dM/dlogD (-) dM/dlogD (-) dM/dlogD (-) dM/dlogD (-) 37 1000 1000 1000 1000 m) µ

( eq D Darwin F1405 Santiago F1546 Cayenne F1423 Cayenne F1422 100 100 100 100 8 6 4 2 0 8 6 4 2 0 8 6 4 2 0 8 6 4 2 0

14 12 10 12 10 14 12 10 14 12 10 14

eq eq eq eq (g m (g dM/dlogD ) m (g dM/dlogD ) ) m (g dM/dlogD ) m (g dM/dlogD

-3 -3 -3 -3 (left panels) and normalized by IWC (right panels). Mass median area- 3 ) listed for the highest concentration range ( eq gm Ice mass size distributions in three ice water content (IWC) ranges (1–2, 2–4 and shown in cyan, blue andMass red, concentrations respectively) in based on bin-wiseequivalent diameter median (MMD measurements during four flights. Figure 6. Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | on 23 January UTC , overlying a region that includes the Tiwi Islands K 38 Cloud top temperature derived from MTSAT-1R at 12:00, 16:33 and 22:33 2006 over Darwin. Darkest shadein is each colder image. than 200 Figure 7. Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 39 ) and mean Doppler velocity curtain measured by the S-band e Z Equivalent radar reflectivity ( described in text. Figure 8. profiler on 23 January 2006. Dashed lines indicate duration of two stratiform rain periods defined as Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | just below the e Z ; 24.0 24.0 24.0 24.0 24.0 24.0 24.0 km 2.5 km Surface 11.7 km 11.7 km 11.7 km 9.0 km 4.7 km 23.9 23.9 23.9 23.9 23.9 23.9 23.9 ; and precipitation rate measured at the 23.8 23.8 23.8 23.8 23.8 23.8 23.8 km

23.7 23.7 23.7 23.7 23.7 23.7 23.7 40 Day of January (UTC) ; retrieved vertical wind speed at 9 ), mean Doppler velocity (MDV; negative downward), and 23.6 23.6 23.6 23.6 23.6 23.6 23.6 e km Z 23.5 23.5 23.5 23.5 23.5 23.5 23.5 ; retrieved precipitation rate at 2.5 8 6 4 2 0 8 6 4 2 0 8 6 4 2 0 8 6 4 2 0 0 5 0 23.4 23.4 23.4 23.4 23.4 23.4 23.4

-2 -5

10 10 10 50 40 30 20 10 30 20 10 10 10

-10 -15 -10

Rate (mm h (mm Rate h (mm Rate Width (m s (m Width s (m Wind Reflectivity (dBZ) Reflectivity ) ) )

) km

MDV (m s (m MDV Reflectivity (dBZ) Reflectivity )

-1 -1 -1 -1

-1

Precipitation Vertical Precipitation Spectral Time series of observations and retrievals from colocated S-band, UHF, and VHF proﬁlers surface. Vertical dashed lines indicate stratiform rain periods as in Fig. 8. Doppler velocity spectral widthmelting at level 11.7 at 4.7 Figure 9. on 23 January 2006: radar reﬂectivity ( Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 40 ) -3 m 6 30 10.9 dBZ 20 10 23.842-23.909 Reflectivity (mm 0

1 0

-1 -2 -3

) s (m MDV -1 41 40 ) -3 m 6 30 10.1 dBZ 20 10 23.512-23.588 Reflectivity (mm 0

1 0 -3 -1 -2

MDV (m s (m MDV

) Mean Doppler velocity vs. equivalent radar reflectivity measured by S-band profiler during -1 two stratiform periods shown inmean reflectivity Figs. reported 8 in and dBZ 9. units at Dashed lower lines left. indicate mean values within each sample; Figure 10. Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 5 4 ) -3 3 2 CWC (g m 1 0

5 0

15 10

20 ) -1 15 10 5 Precipitation Rate (mm h 0

5 0

10 15

) 12 -1 42 10 8 6 4 Ei=0.005 Ei=0.01 Ei=0.005 sat. Ei=0.01 sat. 2 Mean Doppler Velocity (m s 0

0 5

15 10

50 40 30 over Darwin (blue). Simulations use ice-ice sticking efﬁciencies of 0.01 or 0.005, 20 10 km Reflectivity (dBZ) 23.512-23.588 23.842-23.909 0 -10

0 5

10 15 Profiles of reflectivity, mean Doppler velocity (positive downward), precipitation rate (liquid Altitude (km) Altitude with steady-state column modelflights results at initialized 11–11.5 with sizeand the distributions baseline measured sounding during or a Airbus saturated test sounding (see text). equivalent), and condensed water contentband, UHF, (CWC) and VHF observed radars by during the the stratiform S-band periods profiler identified in or Fig. retrieved 8 (pink from and S- red) are compared Figure 11. Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 5 4 ) -3 3 2 CWC (g m 1 0

5 0

15 10

20 ) -1 15 10 5 Precipitation Rate (mm h 0

5 0

10 15

) 12 -1 43 10 8 6 4 Ei=0.02 sat. Ei=0.01 Ei=0.02 Ei=0.01 sat. 2 Mean Doppler Velocity (m s 0

0 5

10 15

50 40 30 20 10 Reflectivity (dBZ) 23.512-23.588 23.842-23.909 0

-10 0 5

10 15 Same as Fig. 11 except that the initial ice size distribution is scaled by a factor of 0.5. Altitude (km) Altitude Figure 12. Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | C-POL 23.8958 ﬁeld, where they enclose 1 = drizzle 2 = rain 3 = low-density snow 4 = high-density snow 5 = melting snow 6 = dry graupel 7 = wet graupel 8 = small hail 9 = large hail 10 = rain-hail mix e

altitude, and column maximum km 9 8 7 6 5 4 3 2 1 0 8 6 4 2 0

) 300 300 km -1

250 250 at 11

200 200 e Z

150 150 X (km) X (km)

100 100

44 50 50 altitude, 2.5-km Rain Rate (mm h 2.5-km Rain Rate (mm

0 0

km 50 50

300 250 200 150 100 300 250 200 150 100

Column Maximum Hydrometeor Class Y (km) Y Y (km) Y

24 18 12 6 0 36 30 24 18 12 6 0 54 48 42 36 30 54 48 42

300 300

250 250

200 200

and rain rate at 2.5 150 150 X (km) X (km) e

Z 100 100

50 50 2.5-km Reflectivity (dBZ) 11.0-km Reflectivity (dBZ)

0 0

50 50

100 250 200 150 150 100 300 300 250 200 Cross-sections from C-POL radar ﬁelds when the S-band experienced stratiform rain at

Y (km) Y Y (km) Y outlined in black are identiﬁedareas as greater convective than (see 30 text) dBZ except (only in present the in 11 Fig. 14). Figure 13. 23.8958 (cf. Fig. 8): hydrometeor class (see legend). The location of the S-band proﬁler is marked with an asterisk. Areas Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | C-POL 23.7639

1 = drizzle 2 = rain 3 = low-density snow 4 = high-density snow 5 = melting snow 6 = dry graupel 7 = wet graupel 8 = small hail 9 = large hail 10 = rain-hail mix

9 8 7 6 5 4 3 2 1 0 8 6 4 2 0

) 300 300 -1

250 250

200 200

150 150 X (km) X (km)

100 100

45 50 50 2.5-km Rain Rate (mm h 2.5-km Rain Rate (mm

0 0

50 50

150 100 300 250 200 150 100 300 250 200

Column Maximum Hydrometeor Class Y (km) Y (km) Y

24 18 12 6 0 54 48 42 36 30 24 18 12 6 0 54 48 42 36 30

300 300

250 250

200 200

150 150 X (km) X (km)

100 100

50 50 2.5-km Reflectivity (dBZ) 11.0-km Reflectivity (dBZ)

0 0

50 50

100 300 250 200 150 150 100 300 250 200

Same as Fig. 13 except when the S-band experienced convective rain at 23.7639. Y (km) Y Y (km) Y Figure 14.