• Cornman, L. B., , Morse C. S. , , and Cunning G. , 1995: Real-time estimation of atmospheric turbulence severity from in-situ aircraft measurements. J. Aircr., 32 , 171177.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ellrod, G. P., 1985: Detection of high level turbulence using satellite imagery and upper air data. NOAA Tech. Memo. NESDIS 10, 30 pp.

  • FAA, cited. 2008: Turbulence: Staying safe. Federal Aviation Administration. [Available online at http://www.faa.gov/passengers/fly_safe/turbulence/].

    • Search Google Scholar
    • Export Citation
  • Hogan, T. F., , and Rosmond T. E. , 1991: The description of the Navy Operational Global Atmospheric Prediction System’s Spectral Forecast Model. Mon. Wea. Rev., 119 , 17861815.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kidwell, K. B., 1995: NOAA polar orbiter data user’s guide. (TIROS-N, NOAA-6, NOAA-7, NOAA-8, NOAA-9, NOAA-10, NOAA-11, NOAA-12, NOAA-13, and NOAA-14). U.S. Department of Commerce, NOAA, NESDIS, National Climatic Data Center, Satellite Data Services Division, Washington, DC, 410 pp.

    • Search Google Scholar
    • Export Citation
  • King, M. D., , Kaufman Y. , , Menzel W. P. , , and Tanre D. , 1992: Remote sensing of cloud, aerosol, and water vapor properties from the Moderate Resolution Imaging Spectrometer (MODIS). IEEE Trans. Geosci. Remote Sens., 30 , 127.

    • Search Google Scholar
    • Export Citation
  • Knox, J. A., , Bachmeier A. S. , , Carter W. M. , , Tarantino J. E. , , Paulik L. C. , , Wilson E. N. , , Bechdol G. S. , , and Mays M. J. , 2009: Transverse cirrus bands in weather systems: A grand tour of an enduring enigma. Weather, in press.

    • Search Google Scholar
    • Export Citation
  • Menzel, W. P., , and Purdom J. F. , 1994: Introducing GOES-I: The first of a new generation of Geostationary Operational Environmental Satellites. Bull. Amer. Meteor. Soc., 75 , 757781.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rabin, R. M., , Corfidi S. F. , , Brunner J. C. , , and Hane C. E. , 2004: Detecting winds aloft from water vapour satellite imagery in the vicinity of storms. Weather, 59 , 251257.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Schmit, T. J., , Gunshor M. M. , , Menzel W. P. , , Gurka J. J. , , Li J. , , and Bachmeier A. S. , 2005: Introducing the next-generation Advanced Baseline Imager on GOES-R. Bull. Amer. Meteor. Soc., 86 , 10791096.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Sharman, R., , Cornman L. , , Williams J. K. , , Koch S. E. , , and Moninger W. R. , 2006: The FAA AWRP turbulence PDT. Preprints, 12th Conf. on Aviation Range and Aerospace Meteorology, Atlanta, GA, Amer. Meteor. Soc., 3.3. [Available online at http://ams.confex.com/ams/pdfpapers/104014.pdf].

    • Search Google Scholar
    • Export Citation
  • Velden, C., and Coauthors, 2005: Recent innovations in deriving tropospheric winds from meteorological satellites. Bull. Amer. Meteor. Soc., 86 , 205223.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • View in gallery

    (top left) AVHRR IR window image at 0905 UTC, (top right) GOES IR window image at 0915 UTC, and (bottom) GOES WV image at 0915 UTC 16 Jun 2005 over MO. Time matched (±30 min from image timestamp) EDR turbulence observations of light (blue squares) and moderate (green squares) intensity and null turbulence observations (gray squares) are overlaid upon the imagery.

  • View in gallery

    (left) BT transects through transverse bands on the (top) northern and (bottom) southern edges of the cirrus anvil. (right) MODIS IR window imagery from 0800 UTC 2 Aug 2006. Locations of cross sections are identified in the MODIS image by the black lines.

  • View in gallery

    (left) GOES visible and (right) IR window imagery at (top) 1315 and (bottom) 1445 UTC 19 Jul 2006 over MN.

  • View in gallery

    Temporal evolution of a transverse band event occurring on 2 Aug 2006. Images are taken every 30 min from 0515 through 0845 UTC over MN and WI from GOES IR window imagery.

  • View in gallery

    Geographic distribution of transverse band events. (top) Black dots represent the locations of the storm centers at the times when transverse bands first appeared. (bottom) Dark gray ovals represent the locations of the anvils, light gray ovals represent the locations of the bands, and the orientation of the black arrows shows the direction of the storm propagation, respectively.

  • View in gallery

    Bar graphs showing the (top left) temporal distribution of transverse bands, (top right) storm development, (bottom left) time lag between storm initiation and transverse band formation, and (bottom right) transverse band duration.

  • View in gallery

    The 300-hPa (left) relative vorticity and (right) divergence plotted on GOES water vapor imagery at (top) 1245, (middle) 1545, and (bottom) 1915 UTC 1 Jul 2006. Dashed lines indicate negative relative vorticity and convergence.

  • View in gallery

    The 300-hPa (left) relative vorticity and (right) divergence plotted on GOES water vapor imagery 2006 at (top) 1745 and (bottom) 1945 UTC 6 Aug. Dashed lines indicate negative relative vorticity and convergence.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 49 49 12
PDF Downloads 41 41 14

Convectively Induced Transverse Band Signatures in Satellite Imagery

View More View Less
  • 1 Department of Atmospheric and Oceanic Sciences, University of Wisconsin—Madison, Madison, Wisconsin
  • | 2 Cooperative Institute for Meteorological Satellite Studies, University of Wisconsin—Madison, Madison, Wisconsin
  • | 3 Department of Atmospheric and Oceanic Sciences, and Cooperative Institute for Meteorological Satellite Studies, University of Wisconsin—Madison, Madison, Wisconsin
© Get Permissions
Full access

Abstract

Transverse cirrus bands have commonly been observed in the outflow of thunderstorms, though little literature exists on the subject. The primary objective of this paper is to characterize the transverse band signature in satellite imagery with references to storm location, movement, and life cycle. The transverse band signature was observed in nearly half of all convective systems analyzed between May and August 2006, commonly in the mature and decay stages of the system. Storm size and propagation did not appear to influence transverse bands, though the bands did appear to be associated with negative 300-hPa relative vorticity and positive divergence. Transverse bands lasted an average duration of 9 h and generally occurred during the nighttime hours. The satellite analysis was combined with eddy dissipation rate (EDR) atmospheric turbulence observations collected by commercial aircraft. At least one observation of light (moderate) turbulence was found within transverse bands for 93% (44%) of events, indicating that the presence of transverse bands in satellite imagery is a strong indicator for aviation turbulence.

Corresponding author address: Kristopher M. Bedka, 1225 West Dayton St., Madison, WI 53706. Email: krisb@ssec.wisc.edu

Abstract

Transverse cirrus bands have commonly been observed in the outflow of thunderstorms, though little literature exists on the subject. The primary objective of this paper is to characterize the transverse band signature in satellite imagery with references to storm location, movement, and life cycle. The transverse band signature was observed in nearly half of all convective systems analyzed between May and August 2006, commonly in the mature and decay stages of the system. Storm size and propagation did not appear to influence transverse bands, though the bands did appear to be associated with negative 300-hPa relative vorticity and positive divergence. Transverse bands lasted an average duration of 9 h and generally occurred during the nighttime hours. The satellite analysis was combined with eddy dissipation rate (EDR) atmospheric turbulence observations collected by commercial aircraft. At least one observation of light (moderate) turbulence was found within transverse bands for 93% (44%) of events, indicating that the presence of transverse bands in satellite imagery is a strong indicator for aviation turbulence.

Corresponding author address: Kristopher M. Bedka, 1225 West Dayton St., Madison, WI 53706. Email: krisb@ssec.wisc.edu

1. Introduction

In-flight turbulence is the leading cause of injuries to airline passengers and flight attendants: approximately 58 people are seriously injured and more than 1000 minor injuries occur as a result of turbulence in the United States each year (FAA 2008). Areas of strong atmospheric turbulence over large regions are often characterized by extensive cloud cover, which sometimes contain well-defined transverse cirrus bands that can be observed in visible or infrared satellite imagery (Ellrod 1985). The occurrence of transverse band (or “radial cirrus”) signatures in satellite imagery is not an uncommon feature in the life cycle of a thunderstorm, but they are also present within jet streams, tropical cyclone outflow, and the warm conveyor belt of midlatitude cyclones (Knox et al. 2009). Knox et al. provide a comprehensive literature review and present detailed examples of the transverse band signature in high-resolution satellite imagery, but also indicate that “there is no consensus on why transverse bands form.”

This paper focuses on the characteristics of transverse cirrus banding observed within thunderstorm anvils in satellite imagery. The spatial resolution of current imaging systems on both geostationary (GEO) and low-earth orbiting (LEO) satellites allows for detailed analysis of features such as transverse bands. Multispectral satellite imagery from both GEO and LEO instruments during the 2006 convective season are used to identify the signature, estimate the frequency of occurrence, and aid in the characterization of the time of occurrence and temporal persistence of the bands. Some upper-airflow patterns associated with transverse bands are identified in an attempt to further understand this signature. Objective aircraft turbulence observations are also used to examine the frequency of turbulence when aircraft fly in or near transverse bands.

2. Data

a. Geostationary satellite observations

This study utilizes data from the Geostationary Operational Environmental Satellite-12 (GOES-12), which is positioned over the equator at 75°W and provides imagery over the eastern United States every 15–30 min. The relatively high temporal resolution allows one to monitor the emergence, evolution, and dissipation of transverse bands. GOES-12 collects data in five spectral channels in the visible and infrared (IR) portions of the electromagnetic spectrum (Menzel and Purdom 1994). For this study, imagery from the following three channels is used: channel 1 (0.65 μm or visible channel), channel 3 [6.5 μm or the water vapor (WV) channel], and channel 4 (10.7 μm or the IR window channel). GOES-12 visible channel imagery has two distinct advantages over the IR window and WV channels. First, visible channel data are collected at a higher spatial resolution than are those of the WV and IR window channels (1 km versus 4 km at nadir), allowing for a clearer depiction of detailed band features in a convective cloud top. Second, visible imagery depicts cloud-top texture through shadowing, allowing for the visualization of subtle banded cloud features that are not large enough to induce a variation in an IR or WV brightness temperature field (see top panels in Fig. 3). The IR window channel is essential for monitoring mature thunderstorm anvil evolution during both day and night and serves as the primary image type used in this study. The WV channel is used to identify transverse bands that do not induce significant fluctuations in the IR window channel temperature pattern, such as those that are present on the outermost periphery of the thunderstorm cirrus outflow.

b. Polar-orbiting satellite observations

The Moderate Resolution Imaging Spectroradiometer (MODIS) instrument has 36 spectral channels and flies aboard the National Aeronautics and Space Administration (NASA) satellites Terra and Aqua, which are LEO satellites that orbit the earth at 705 km (King et al. 1992). The Advanced Very High Resolution Radiometer (AVHRR) has six channels and is the primary sensor aboard the National Oceanic and Atmospheric Administration (NOAA) Polar Operational Environmental Satellites (POES) that orbit about 850 km above the earth (Kidwell 1995). These polar-orbiting satellite instruments provide better spatial resolution views of the earth’s atmosphere and clouds than geostationary satellites, enabling visualization of features that are often undetected in the GOES satellite observations.

c. Eddy dissipation rate turbulence observations

Eddy dissipation rate (EDR) turbulence observations from commercial aircraft provide an objective measure of the vertical accelerations induced by turbulent atmospheric phenomena (Cornman et al. 1995). This dataset is produced through the combination of existing sensors, avionics, and communication networks, resulting in a “state of the atmosphere” turbulence diagnostic (Sharman et al. 2006). United Airlines Boeing 737 and 757 aircraft began collecting this EDR data in 1997, recording an observation every minute during flight at altitudes above 20 000 ft and automatically transmitting it to the ground. The EDR observing system was designed to provide routine and quantitative measurements of atmospheric turbulence intensity levels, including null observations, addressing many of the deficiencies with pilot reports (PIREPs) (Sharman et al. 2006). EDR observations provide a unique dataset that can be used to study atmospheric turbulence associated with the transverse band signature.

3. Methodology

The 2006 Storm Database (information available online at http://cimss.ssec.wisc.edu/snaap/edr/Summer2006StormDatabase.html) was developed by identifying transverse banding events from May through August 2006 occurring over the continental United States. The time evolution of the outflow pattern was monitored for all convective events in the time period described above using 4-km GOES-12 IR window channel imagery. A convective event was defined as a localized region of convective storms having IR window brightness temperatures (BTs) <230 K present for longer than 1 h. The largest convective events contained numerous individual storms and covered entire regions of the United States (i.e., the entire southeast sector of the country) for a duration of up to 18 h. All identified convective events were then analyzed more thoroughly using the GOES-12 visible and IR window imagery to confirm the presence of transverse bands. In the case of multistorm events, each individual storm was analyzed for the existence of transverse bands. Previous studies of the transverse band feature (Ellrod 1985; Knox et al. 2009) and other unpublished Web-based satellite feature interpretation guides/examples were used as visual aides to subjectively identify the presence of bands in visible and IR window imagery. Convective events in which transverse bands did form and were sustained for a significant period of time (at least 30 min) were entered into the 2006 Storm Database, where the life cycle of the storm and its transverse band evolution were documented. The EDR turbulence dataset was then analyzed for each case in the 2006 Storm Database to determine the frequency of atmospheric turbulence present during flight in and near transverse bands.

4. Analysis of the transverse band signature

a. Identification of bands in satellite imagery

For the 2006 Storm Database, 131 convective events were identified between 1 May and 31 August. Of these events, 54 were associated with transverse bands: 41% of the total convective events produced banded structures. Transverse bands can be seen in the visible, IR window, and WV imagery. The appearance of transverse bands in the IR window and water vapor channels is illustrated in Fig. 1 and the visible channel is illustrated later (see Fig. 3). In the visible, bands appear as high cirrus clouds that vary in thickness. The clouds are arranged in long, thin bands that extend away from the storm center, often perpendicular to the storm edge and convectively generated gravity waves (if present). These cloud signatures are also apparent in the IR window and WV imagery, providing good temporal coverage when visible imagery is not available.

Transverse bands are sometimes slightly more difficult to visualize in the GOES IR window imagery due to the lower spatial resolution of GOES. Figure 1 depicts the difference between levels of GEO and LEO satellite image resolution for a transverse band case. While the major features of the convective system are depicted sufficiently in the lower-resolution GOES imagery, the transverse band signature is much less apparent. For instance, the structures to the east of the system do not appear to be transverse bands at all in the GOES imagery, but are very clearly banded structures when viewed in the AVHRR imagery. Thus, higher-resolution IR window data are required to accurately view and identify subtle transverse band signatures in satellite data in the absence of visible imagery. The more robust transverse band signatures can be sufficiently identified in the lower-resolution IR window data, as shown on the northern edge of the storm in Fig. 1.

Figure 2 shows brightness temperature transects through two different sets of transverse bands in MODIS IR window imagery. In both transects, the brightness temperature minima representing individual transverse band clouds do not extend more than a few kilometers in length, indicating that the width of an individual transverse band is a meso-γ-scale (1–10 km) feature. Band separation can vary; spacing between bands can be as small as 5 km and as large as 50 km (analysis not shown), though in most cases bands are generally observed with 10–15-km spacing.

Figure 3 illustrates a comparison of GOES visible and IR window imagery from two different times in a transverse band event. At 1315 UTC the newly developing bands can be detected in the visible imagery but are extremely subtle, if not nonexistent, in the IR window imagery. An 1.5 h later, at 1445 UTC, the bands are still clearly evident in the visible imagery and are beginning to appear in the IR window imagery, specifically along the northern edge of the anvil. The bands are difficult to see in the IR window imagery while they are within an optically thick anvil cloud due to the small gradient in temperature between the bands and the anvil. When the anvil becomes diffuse along its northern periphery, the temperature contrast between the clear air and the bands is much larger and the bands are more easily seen in the IR window imagery.

b. Examples of the evolution and emergence of transverse bands

No single factor appeared to be directly related to the formation or lack of transverse bands, though interactions between multiple convective systems appeared to inhibit band sustainability. The ideal case for band production appeared to be a strong, isolated convective system that developed, matured, and dissipated without interacting with any other convective storm. A few cases showed signs of transverse bands earlier in the life cycle of the convective system, near the end of the vertical growth stage and at the beginning of the mature stage, though transverse bands were more commonly observed to emerge near the end of the mature stage of the system and persist through the decay stage of the system. Rapid anvil expansion can sometimes be associated with the production of transverse cloud bands along the periphery of the anvil cirrus cloud. However, transverse bands are not exclusively associated with rapid anvil expansion and rapid anvil expansion does not necessarily guarantee transverse band development.

Figure 4 illustrates the evolution of banded structures on both the northwest and southeast sides of a mesoscale convective system (MCS). Prior to 0515 UTC, the system was intensifying (i.e., a general cooling in minimum cloud-top IR window brightness temperatures) and was propagating toward the east. At 0515 UTC, very small transverse bands begin to emerge from the anvil along the west-northwestern edge. The cloud-top brightness temperatures continue to decrease until 0545 UTC, when the bands along the northwestern edge of the anvil become more apparent. After 0615 UTC, cloud-top brightness temperatures begin to warm, indicating the system is beginning to decay. As the system decays, bands become more prominent and extend farther away from the storm core, though the initial burst between 0515 and 0545 UTC was the greatest increase in band size and intensity over the entire storm life cycle. An anticyclonic flow emerges in the anvil cloud pattern as the MCS begins to develop its own outflow environment, which acts to govern the rotation and orientation of the bands as the storm begins to rapidly decay around 0745 UTC. The bands begin to show signs of decay between 0815 and 0845 UTC as the parent storm continues to dissipate.

Most transverse band cases observed in the 2006 Storm Database involved the emergence of banded structures along the periphery of the storm; however, a few cases exhibited slightly different patterns of behavior. Figure 3 depicts a case in which banded structures appeared to be connected to the region near the overshooting top, similar to spokes on a wheel. Figure 3 illustrates how these “spokelike” bands can be traced all the way through the anvil of the storm from the overshooting top out beyond the edge of the anvil in the visible channel imagery. As the storm develops, the bands are evident farther away from the core of the storm and no longer appear connected to the overshooting top region. The spatial resolution of the GOES IR window imagery is not fine enough to distinguish these banded structures from the rest of the anvil, as previously discussed, so it is possible this type of band morphology exists more often than has been observed within this study.

c. Geographic distribution

Transverse bands were most often observed in convective systems originating in the central United States. As illustrated in Fig. 5, the majority of transverse band events in the 2006 Storm Database occurred in the Great Plains across the midwestern United States. These events were generally associated with convective systems that were initiated in the lee of the Rocky Mountains and moved eastward as they intensified and later dissipated. Figure 5 also depicts a secondary region of transverse band events along the coast of the Gulf of Mexico. A few of the events occurring in the southeastern United States and along the Gulf Coast were associated with northward-moving tropical cyclones, though many were associated with non–tropical cyclone isolated or multistorm events.

Time sequences of the GOES imagery were also used to estimate storm movement when bands were present. The geographic distribution of the banded structures with respect to the parent storm is not easily generalized. Analysis of Fig. 5 shows that the orientation of transverse bands is not clearly related to storm propagation or the geographic location of the convective system. The majority of storms in the midwestern United States were propagating in an eastward direction, though banded structures seem to form along all sides of the anvil edge, without an apparent preferred direction relative to the parent storm motion. The orientation of banded structures also did not appear to be influenced by the size or shape of the convective system, nor by the speed of storm propagation (not shown).

d. Temporal distribution

The temporal distribution of transverse band presence in 15–30-min IR imagery was completed by summing the number of times that bands were present during 1-h time intervals. For example, a storm with transverse bands present from 0345 to 0415 UTC would contribute to the total number of band cases in both the 0300–0359 and 0400–0459 UTC time periods. The results of this analysis are shown in Fig. 6. Through the method described above, the total number of cases over the entire graph in Fig. 6 is much larger than the number of cases analyzed. Figure 6 depicts a general peak in transverse band presence from approximately 0300 until 1559 UTC, with two local maxima at 0800–0859 and 1400–1449 UTC. A general minimum in band presence can also be inferred from Fig. 6, beginning around 1800 UTC and lasting through 0259 UTC.

Transverse bands were observed at all hours of the day, though they were most common during the nighttime hours in this study period. Since most transverse bands are observed in the mature and decay stages of convective systems, it follows that bands would most often be observed overnight, as many thunderstorms in this geographic region (see Fig. 5) develop in the late afternoon/evening and generally persist into the early morning hours of the next day. Figure 6 shows that convection initiation occurs for these band-producing convective systems between the hours of 1700 and 0359 UTC, corresponding with the minimum in band presence. In fact, 69% of the transverse band events from the 2006 Storm Database resulted from storms that developed between the hours of 1800 and 0259 UTC. Many instances were observed in which the banded structures persisted for several hours after the original convective system had completely dissipated. In these cases, the banded structures generally detached from the main system during the decay stage and were advected away by the ambient flow, maintaining the banded cloud pattern despite the dissipation of the parent convective system.

The time lag between storm development and the formation of transverse bands ranged from as short as 15 min to as long as 14 h and 45 min. Figure 6 illustrates the amount of time that passed between storm initiation and the first emergence of bands, showing an apparent peak in band development 5–9 h after storm initiation. This analysis in Fig. 6 disregards the two cases in the 2006 Storm Database that appeared to be tropical cyclones, as these storms developed before moving on shore. The average time lag between convection initiation and the first observation of transverse bands was 7 h. From this analysis, one would expect to see the peak in the temporal distribution of transverse bands to occur 5–9 h after the peak in the temporal distribution of the storm initiation. The peak time range of convective initiation was 1800 to 2100 UTC and the peak in transverse bands began around 0300 UTC, very close to the 7-h mean initiation-band formation time interval.

Figure 6 also illustrates the distribution of transverse band duration, peaking around 4–5 h and again at 8–10 h. The case in which banded structures lasted 24 h was a tropical cyclone that made landfall along the southeastern United States; however, the case with bands lasting 20 h was a convective system over the midwestern United States. Thus, it is possible for long-lived band events in cases other than tropical cyclones. Such a large variability in transverse band duration, along with a fairly long average duration, accounts for such a broad peak in the temporal distribution depicted in Fig. 6.

e. Relationship to upper-level divergence and vorticity

The relationship to upper-level divergence and vorticity was analyzed for only 33 cases from the 2006 Storm Database (out of 54), limited by the availability of numerical weather prediction model-derived divergence and vorticity data in the GOES Mesoscale Satellite Winds Web page (located online at http://cimss.ssec.wisc.edu/mesoscale_winds/archive.html). To objectively analyze the divergence and vorticity fields, the U.S. Navy’s Operational Global Atmospheric Prediction System Model (NOGAPS; Hogan and Rosmond 1991) forecast wind fields were combined with satellite-derived water vapor winds from the University of Wisconsin—Madison’s Cooperative Institute for Meteorological Satellite Studies (UW-CIMSS) algorithm (Rabin et al. 2004; Velden et al. 2005) to provide the best estimate of 300-hPa flow conditions at the time of the satellite image. Derived 300-mb divergence and vorticity fields are plotted atop GOES water vapor channel imagery, as these were the fields available at this online archived image resource. The maximum time separation between the NOGAPS fields and the underlying satellite imagery is 3 h.

Transverse bands have generally been observed in regions of divergence at 300 hPa, an example of which is illustrated in Fig. 7. From the cases with model data described above, 91% had bands form in a region of divergence. Of these cases, 76% of cases had bands form in a gradient of divergence, which was usually due to a local maximum of divergence located over the storm center, and 24% of the cases had bands form in a local maximum of convergence.

Transverse bands were nearly always observed in regions of negative relative vorticity at 300 hPa; 97% of cases had bands form in a region of negative relative vorticity. The most common location for the bands to form was in a region of a strong relative vorticity gradient, which occurred in 82% of the cases, 21% of cases had bands form in a local minimum of relative vorticity, and 9% of cases had bands form parallel to lines of constant negative relative vorticity. This analysis was done using all sets of bands, as some cases had two or more sets of bands developing in different relative vorticity fields (i.e., one case formed a band in a local minimum of relative vorticity and in a gradient of negative relative vorticity). In the event of more than one set of bands in different relative vorticity fields, cases were counted twice—once in each category.

The case illustrated in Fig. 7 represents the majority of cases studied, as there is a local minimum of relative vorticity positioned over the center of the storm with the bands forming in the strongest gradient. The bands also formed in a gradient of divergence due to the largest values of divergence also occurring at the center of the storm. The bands illustrated in Fig. 7 persist hours after the storm dissipates, extending in a north–south direction initially and tilting toward a northeast–southwest orientation over time. The bands appear to follow the relative vorticity gradient, changing orientation in order to remain parallel to the relative vorticity gradient vector and propagating to remain within the region of the strongest gradient. From Fig. 7, it is apparent that the bands remain in a region of divergence throughout their life cycle, though the structure and distribution of the divergence are not temporally consistent.

The majority of banded structures form in a region of a strong relative vorticity gradient, though the bands do not always favor the strongest gradient present (analysis not shown). Also, the existence of a strong relative vorticity gradient does not guarantee the maintenance and persistence of transverse bands. Figure 8 depicts a case in which a convective system produces transverse bands over Wisconsin around 1145 UTC. The bands move east more quickly than the decaying system and finally dissipate over New York. These bands do not form in a strong gradient of relative vorticity, but eventually move into the gradient depicted in Fig. 8. However, once the bands move into the gradient, they remain in the gradient and are oriented parallel to the gradient vector until they dissipate. Despite the increase in the gradient between 1745 and 1945 UTC, the bands almost completely dissipate by 1945 UTC. Also of interest is the fact that the bands dissipate in a region of divergence.

From this analysis it is apparent that, while generalizations can indicate possible behavior for transverse bands, the environmental characteristics associated with banded events are highly variable. Due to the variable nature of transverse bands, generalizations can guide our initial analysis but will not always accurately predict the behavior of transverse bands. A strong gradient of relative vorticity does not guarantee the development or sustainability of transverse bands. Divergence appears to be well correlated with band development, but does not guarantee the sustainability of transverse bands either.

f. Turbulence

Transverse bands are also often correlated with incidents of aviation turbulence. For example, Fig. 1 shows that three individual aircraft flew in or near a region with transverse band clouds. All three flights experienced light to moderate turbulence upon entering and throughout the entire duration of flight within the bands. No turbulence was observed when the aircraft were outside the bands.

Through comparison of satellite imagery and time-matched (±15 min of imagery) EDR turbulence observations, 41 of the 54 total transverse band events included at least one flight through the transverse bands. Of these events, 38 (93%) included at least one flight that observed light turbulence while in or near the bands, though some of these events also included flights through the bands that experienced no turbulence at all. Moderate or greater turbulence was observed at least once in 18 of the banded events: 44% of the events with a flight in or near the bands (analysis not shown).

Some care must be taken, however, in interpreting these results. Satellite cloud-top height retrievals were not directly compared to the aircraft altitude measurements to truly evaluate whether or not the aircraft flew “through” the bands. Though it may appear that an aircraft flew through the band features through comparison of the aircraft track and satellite imagery, it is very difficult to objectively determine the vertical proximity of the aircraft relative to the bands without precise knowledge of the cloud height and band vertical thickness. Nevertheless, the mere presence of transverse bands within satellite imagery appears to be a strong indicator that turbulence is likely in or near the banded cloud region.

5. Conclusions

Transverse bands were observed to be long-lived features and occurred in the mature and decaying stages of nearly half of all convective systems investigated during May–August 2006. The bands have been observed most often during the nighttime hours in the Great Plains and Midwest, though they were also not uncommon along the Gulf Coast. The average lag time between storm initiation and band formation was about 7 h and the average duration of the bands was about 9 h. In most cases, transverse bands appeared to originate at the storm edge in the outflow of the convective system, though some cases exhibited evidence for the production of banded structures earlier in the storm life cycle closer to the storm core.

The orientation of transverse bands did not appear to be influenced by storm size, storm movement, or geographic location, though the bands did appear to be associated with the upper-level dynamics as represented by a combination of NWP model and satellite wind output. Almost every case of transverse bands originated in a region of negative relative vorticity and divergence. Most cases produced bands in the gradient of relative vorticity, which appeared to sustain the bands beyond the life of the storm in some cases. Nearly every case of transverse bands was associated with at least light turbulence and just less than half of the cases were associated with moderate or greater turbulence, indicating that the presence of bands in satellite imagery may be used to nowcast a high likelihood for aviation turbulence.

The increased spatial and temporal resolutions of future satellite imagery such as the GOES-R Advanced Baseline Imager (Schmit et al. 2005) will greatly enhance our understanding of mesoscale atmospheric features, especially transverse bands. As this paper has focused on documenting the appearance and frequency of the transverse band signature in current-generation satellite imagery, the mechanism of band formation is unresolved in current study as is the relationship of the bands to aviation turbulence. As other airlines collect EDR observations, the EDR dataset will become more complete and thus more useful for the study of turbulence associated with this banded cloud signature. These future tools will prove to be a valuable resource for the ongoing research of transverse bands in the outflow of thunderstorms.

Acknowledgments

This research was supported under NASA decision support (CAN Award AGR DTD 5/18/06). The authors thank Robert Sharman, John Williams, and the NCAR RAL support staff for the EDR datasets and the exchange of ideas on convectively induced turbulence processes. The authors also thank John Knox for his support and insight on the transverse band phenomena.

REFERENCES

  • Cornman, L. B., , Morse C. S. , , and Cunning G. , 1995: Real-time estimation of atmospheric turbulence severity from in-situ aircraft measurements. J. Aircr., 32 , 171177.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ellrod, G. P., 1985: Detection of high level turbulence using satellite imagery and upper air data. NOAA Tech. Memo. NESDIS 10, 30 pp.

  • FAA, cited. 2008: Turbulence: Staying safe. Federal Aviation Administration. [Available online at http://www.faa.gov/passengers/fly_safe/turbulence/].

    • Search Google Scholar
    • Export Citation
  • Hogan, T. F., , and Rosmond T. E. , 1991: The description of the Navy Operational Global Atmospheric Prediction System’s Spectral Forecast Model. Mon. Wea. Rev., 119 , 17861815.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kidwell, K. B., 1995: NOAA polar orbiter data user’s guide. (TIROS-N, NOAA-6, NOAA-7, NOAA-8, NOAA-9, NOAA-10, NOAA-11, NOAA-12, NOAA-13, and NOAA-14). U.S. Department of Commerce, NOAA, NESDIS, National Climatic Data Center, Satellite Data Services Division, Washington, DC, 410 pp.

    • Search Google Scholar
    • Export Citation
  • King, M. D., , Kaufman Y. , , Menzel W. P. , , and Tanre D. , 1992: Remote sensing of cloud, aerosol, and water vapor properties from the Moderate Resolution Imaging Spectrometer (MODIS). IEEE Trans. Geosci. Remote Sens., 30 , 127.

    • Search Google Scholar
    • Export Citation
  • Knox, J. A., , Bachmeier A. S. , , Carter W. M. , , Tarantino J. E. , , Paulik L. C. , , Wilson E. N. , , Bechdol G. S. , , and Mays M. J. , 2009: Transverse cirrus bands in weather systems: A grand tour of an enduring enigma. Weather, in press.

    • Search Google Scholar
    • Export Citation
  • Menzel, W. P., , and Purdom J. F. , 1994: Introducing GOES-I: The first of a new generation of Geostationary Operational Environmental Satellites. Bull. Amer. Meteor. Soc., 75 , 757781.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rabin, R. M., , Corfidi S. F. , , Brunner J. C. , , and Hane C. E. , 2004: Detecting winds aloft from water vapour satellite imagery in the vicinity of storms. Weather, 59 , 251257.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Schmit, T. J., , Gunshor M. M. , , Menzel W. P. , , Gurka J. J. , , Li J. , , and Bachmeier A. S. , 2005: Introducing the next-generation Advanced Baseline Imager on GOES-R. Bull. Amer. Meteor. Soc., 86 , 10791096.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Sharman, R., , Cornman L. , , Williams J. K. , , Koch S. E. , , and Moninger W. R. , 2006: The FAA AWRP turbulence PDT. Preprints, 12th Conf. on Aviation Range and Aerospace Meteorology, Atlanta, GA, Amer. Meteor. Soc., 3.3. [Available online at http://ams.confex.com/ams/pdfpapers/104014.pdf].

    • Search Google Scholar
    • Export Citation
  • Velden, C., and Coauthors, 2005: Recent innovations in deriving tropospheric winds from meteorological satellites. Bull. Amer. Meteor. Soc., 86 , 205223.

    • Crossref
    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

(top left) AVHRR IR window image at 0905 UTC, (top right) GOES IR window image at 0915 UTC, and (bottom) GOES WV image at 0915 UTC 16 Jun 2005 over MO. Time matched (±30 min from image timestamp) EDR turbulence observations of light (blue squares) and moderate (green squares) intensity and null turbulence observations (gray squares) are overlaid upon the imagery.

Citation: Weather and Forecasting 24, 5; 10.1175/2009WAF2222285.1

Fig. 2.
Fig. 2.

(left) BT transects through transverse bands on the (top) northern and (bottom) southern edges of the cirrus anvil. (right) MODIS IR window imagery from 0800 UTC 2 Aug 2006. Locations of cross sections are identified in the MODIS image by the black lines.

Citation: Weather and Forecasting 24, 5; 10.1175/2009WAF2222285.1

Fig. 3.
Fig. 3.

(left) GOES visible and (right) IR window imagery at (top) 1315 and (bottom) 1445 UTC 19 Jul 2006 over MN.

Citation: Weather and Forecasting 24, 5; 10.1175/2009WAF2222285.1

Fig. 4.
Fig. 4.

Temporal evolution of a transverse band event occurring on 2 Aug 2006. Images are taken every 30 min from 0515 through 0845 UTC over MN and WI from GOES IR window imagery.

Citation: Weather and Forecasting 24, 5; 10.1175/2009WAF2222285.1

Fig. 5.
Fig. 5.

Geographic distribution of transverse band events. (top) Black dots represent the locations of the storm centers at the times when transverse bands first appeared. (bottom) Dark gray ovals represent the locations of the anvils, light gray ovals represent the locations of the bands, and the orientation of the black arrows shows the direction of the storm propagation, respectively.

Citation: Weather and Forecasting 24, 5; 10.1175/2009WAF2222285.1

Fig. 6.
Fig. 6.

Bar graphs showing the (top left) temporal distribution of transverse bands, (top right) storm development, (bottom left) time lag between storm initiation and transverse band formation, and (bottom right) transverse band duration.

Citation: Weather and Forecasting 24, 5; 10.1175/2009WAF2222285.1

Fig. 7.
Fig. 7.

The 300-hPa (left) relative vorticity and (right) divergence plotted on GOES water vapor imagery at (top) 1245, (middle) 1545, and (bottom) 1915 UTC 1 Jul 2006. Dashed lines indicate negative relative vorticity and convergence.

Citation: Weather and Forecasting 24, 5; 10.1175/2009WAF2222285.1

Fig. 8.
Fig. 8.

The 300-hPa (left) relative vorticity and (right) divergence plotted on GOES water vapor imagery 2006 at (top) 1745 and (bottom) 1945 UTC 6 Aug. Dashed lines indicate negative relative vorticity and convergence.

Citation: Weather and Forecasting 24, 5; 10.1175/2009WAF2222285.1

Save