SEARCH

SEARCH BY CITATION

Keywords:

  • mesoscale convective systems;
  • tropical cyclones

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methodology
  5. 3. Climatology of Outer-MCSs
  6. 4. Stratiform Precipitation Regions and Mesoscale Surface Jets Collocated With Outer-MCSs
  7. 5. Conclusion
  8. Acknowledgments
  9. References

[1] Long-lasting mesoscale convective systems (MCSs) may occur in the outer region of tropical cyclones in the western North Pacific, especially in conjunction with the southwest monsoon as in the case of Typhoon Morakot that caused great flooding and landslides in Taiwan. These “outer-MCSs” are linear convective systems that develop from distant rainbands, have a large cold cloud shield, and last more than six hours. These outer-MCSs are important for typhoon rainfall forecasting because of the torrential rainfall when they interact with land and terrain to produce serious flooding that is separate from the rainfall near the center. A total of 109 outer-MCSs that occurred during 1999–2009 are identified using infrared and passive microwave images. About 22% of all typhoons in the western North Pacific have at least one outer-MCS during their life cycle. In 85% of the QuikSCAT oceanic 10-m wind observations of outer-MCSs, positive shear vorticity on the left side of mesoscale surface jets below the stratiform precipitation regions may be contributing to the continuous formation of new convective cells.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methodology
  5. 3. Climatology of Outer-MCSs
  6. 4. Stratiform Precipitation Regions and Mesoscale Surface Jets Collocated With Outer-MCSs
  7. 5. Conclusion
  8. Acknowledgments
  9. References

[2] The record accumulated rainfall of over 3000 mm during 6–10 August 2009 due to Typhoon Morakot caused very serious damage to southern Taiwan. Although the inner-core convection had been weakened by passage over the Central Mountain Range, the outer circulation of Morakot interacted with the southwest monsoon [Chien and Kuo, 2011; Lee et al., 2011]. From 12 UTC 8 August to 03 UTC 9 August 2009, a long-lived and well-organized mesoscale convective system (MCS) existed in the outer region of Morakot (Figures 1a–1d) that caused extreme precipitation in southern Taiwan (Figure 1e), and is termed an “outer-MCS”. As observed by theSpecial Sensor Microwave Imager/Sounder(SSMI/S) at 1043 UTC 8 August 2009, a strong convective system crossing Taiwan from the Taiwan Strait to the Pacific Ocean existed under the cold cloud shield of the outer-MCS (Figure 1f). This outer-MCS produced rain over the land and then more extreme rain when it interacted with the steep Central Mountain Range. Even though this outer-MCS was about 300 km to the south of the typhoon center, it caused about 1500 mm of precipitation due to its long duration and the orographic enhancement, which accounted for 40–50% of the total precipitation during the typhoon passage.

image

Figure 1. (a–d) IR1 images at selected times from 1200 UTC 8 August 2009 to 2300 UTC 8 August 2009. (e) Positions each 2 h of Typhoon Morakot (2009) and the accumulated precipitation from 1200 UTC 8 August 2009 to 0000 UTC 9 August 2009. The red line indicates the TC track during the outer-MCS period. (f) The 91H brightness temperature of the SSMI/S observation at 1043 UTC 8 August 2009 (fromwww.nrlmry.navy.mil/tc_pages/tc_home.html).

Download figure to PowerPoint

[3] This kind of outer-MCS with “unexpected” heavy rainfall in the outer region has occurred in Typhoons Mindulle (2004), Kalmaegi (2008), and Bilis (2006) when they moved over land and interacted with the terrain to cause serious flooding, landslides, and debris flows.Chien et al. [2008] studied the severe rainfall in southern Taiwan associated with the interaction of the outer circulation of Typhoon Mindulle (2004) with the southwest monsoon. Chien et al.showed that the southwesterly monsoon contributed to the asymmetric rainfall distribution associated with the tropical cyclone (TC) due to the convergence of moist, unstable air with the TC outer circulation. If the low-level convergence provided enough lift, strong MCSs were triggered, and heavy rainfall ensued. Predicting the rainfall due to this kind of MCS and its interaction with the topography is a great challenge and such events are of interest to forecasters in the western North Pacific. Consequently, the climatology of these outer-MCSs and understanding the mesoscale self-sustaining mechanisms of them are important.

[4] In this paper, a climatology of outer-MCSs and some mesoscale features that are hypothesized to be the mechanisms sustaining them are described. First, these long-lived and well-organized MCSs in the western North Pacific Ocean are defined inSection 2using satellite observations. The climatology of outer-MCSs is presented inSection 3. A hypothesis of the self-sustaining mechanisms is provided inSection 4. Future work for a more detailed study is described in Section 5.

2. Methodology

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methodology
  5. 3. Climatology of Outer-MCSs
  6. 4. Stratiform Precipitation Regions and Mesoscale Surface Jets Collocated With Outer-MCSs
  7. 5. Conclusion
  8. Acknowledgments
  9. References

[5] A two-step procedure that first examined an hourly infrared channel-1 cloud-top temperature data set (IR1) and then passive microwave images (PMW) to detect convective features was applied to detect long-lived and well-organized MCSs in the outer regions of TCs in the western North Pacific from 1999 to 2009. The first step was automated and the second step was manual. Most of these MCSs closely resembled the outer-MCS in Typhoon Morakot (2009) and thus are termed “outer-MCSs” (OMCSs).

[6] The IR1 data set, which was obtained from a digital archive at Kochi University in Japan, was on a 5 km grid with a spatial coverage of 20°S–70°N, 70°–160°E. The corresponding PMW images were obtained from the Naval Research Laboratory tropical cyclone web page. These PMW images included ice-scattering channel images observed by theSpecial Sensor Microwave/Imager (SSM/I), SSMI/S, Tropical Rainfall Measurement Mission Microwave Imager (TMI), and Advanced Microwave Scanning Radiometer (AMSR-E).

[7] Even though the cloud-top temperatures are poorly related to precipitating areas due to non-raining anvil clouds [Liu et al., 2007], the advantage of the present IR1 approach is the continuous, high frequency infrared imagery [Maddox, 1980; Williams and Houze, 1987] for tracking long-lived convective systems associated with TCs every hour, whereas the PMW imagery is from polar-orbiters and thus is irregular in time. Thus, IR1 images were used to identify convective systems with a contiguous cold cloud shield (CCS) with IR1 brightness temperatures (TB) colder than −75°C that exceeded 36000 km2 in area.

[8] These candidate outer-MCSs had to be within a radius of 1000 km of the TC centers, which were interpolated from the Joint Typhoon Warning Center (JTWC) best track data. Two types of CCSs in the next image (one hour later) were used to define “active” MCSs: (i) a CCS that had TB colder than −75°C with an area larger than 36000 km2 (CCS−75); (ii) a CCS that had TB colder than −65°C with an area larger than 72000 km2 in size (CCS−65) that overlapped the previous CCS−75 area (or CCS−65 area, if the CCS−75was not available) by at least 50%. If the MCS then remained active for a period greater than six hours, it was included in the second analysis stage. This IR1 analysis identified 603 active and long-lasting CCSs from 1999 to 2009.

[9] The threshold TB values of −65°C and −75°C follow Mapes and Houze [1993], who regarded −65°C as a moderate estimate of radar echo area and −75°C as a conservative estimate of radar echo area in the tropical warm pool. Yuter and Houze [1998] found an 88% probability of precipitation beneath a cloud shield of −65°C in the region of “The Tropical Ocean Global Atmosphere Coupled Ocean Atmosphere Response Experiment” (TOGA COARE). Chen and Houze [1997]also defined MCSs in TOGA COARE by a cloud-top temperature threshold of −65°C and demonstrated that the timescales of the MCSs were positively correlated with the spatial scales. They defined a MCS with a diameter over 300 km as a “super convective system” and showed that these systems usually had a duration over six hours. Therefore, the threshold area of CCS−65 was defined as 72000 km2, which is the area of a circle with a 303 km diameter. Because the area that was colder than −65°C sometimes was too large to distinguish the active and long-lasting convection from the anvil region of the TC, an area colder than −75°C was selected to help identify the occurrence of long-lasting CCSs.

[10] Considering that the 603 active long-lasting CCSs included different kinds of convective structures that are associated with TCs, the objective of the second stage was to select active and long-lasting CCSs that were associated with distant rainbands. Many studies have described the convective features in TCs [Cecil et al., 2002; Cecil and Zipser, 2002; Willoughby et al., 1984; Houze, 2010]. Houze [2010]stated that “distant rainbands are composed of buoyant convective cells aligned with confluence lines in the large-scale, low-level wind field spiraling into the TC vortex and are radially far enough from the eye of the storm that the vertical structure of the convection within them is relatively unconstrained by the dynamics of the inner-core vortex of the cyclone.”Cecil et al. [2002] stated that the outer rainband regiontypically begins 150 – 200 km from the cyclone center, includes any rain features associated with the TC, and is usually bounded on the inside by a precipitation-free lane adjacent to an inner rainband.Cecil et al. [2002] defined a minimum outer rainband radius of 100 km and a median radius of 350 km for the outer rainband region. Corbosiero and Molinari [2002, 2003] defined “outer band regions” as being 100 to 300 km from the center of hurricanes.

[11] The magnitude of ice-scattering (e.g., 85 GHz scattering for TMI, 91 GHz scattering for SSMI/S) has been used to characterize the vigor and spatial extent of convective systems [Mohr and Zipser, 1996; Mohr et al., 1999; Cecil and Zipser, 1999; Nesbitt et al., 2000; Cecil et al., 2002; Cecil and Zipser, 2002]. In the second stage, three criteria were required for an outer-MCS. First, the outer-MCS must be independent of the inner-core convection, which required the −55°C brightness temperature (color yellow inFigure 1f) in the PMW was separate from the inner-core convection. Second, an outer-MCS had to be associated with convection in the distant rainbands. Third, the average distance from the center of the outer-MCS to the TC center must be not less than 200 km during its lifetime. The center of the outer-MCS was defined as the centerpoint of the convective line, which was usually the center of the closed contour of the minimum IR1 TB. In most cases, this contour enclosed only one area in the outer-MCS cold cloud shield. If the PMW images were available at that hour, they were used to confirm the outer MCS center positioning by IR1 Tb. When it was difficult to position the center with these methods, the geometric center of the CCS−65 or CCS−75was regarded as the center. This second stage identified 109 outer-MCSs with a total of 1122 hourly track positions.

3. Climatology of Outer-MCSs

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methodology
  5. 3. Climatology of Outer-MCSs
  6. 4. Stratiform Precipitation Regions and Mesoscale Surface Jets Collocated With Outer-MCSs
  7. 5. Conclusion
  8. Acknowledgments
  9. References

[12] Approximately 22% (71/322) of all TCs during 1999–2009 had at least one outer-MCS during their lifetime, and the total of 109 outer-MCSs indicates an average of 1.7 per TC. The mean duration of outer-MCSs was 10.3 h. The tracks and the probability density functions (PDFs) of outer-MCSs calculated using hourly outer-MCS track records in a 2° × 2° latitude-longitude box are shown inFigure 2a. Although the outer-MCSs were widely distributed in the western North Pacific, two PDF maxima are noted in the South China Sea and Philippine Sea. The outer-MCS PDF maximum in the South China Sea mainly occurred from June to September.

image

Figure 2. (a) Tracks of outer-MCSs and the PDF of frequency (contour interval of 0.1 12 hr/yr) calculated in a 2° × 2° latitude-longitude box. (b) Tracks of outer-MCSs in storm-relative coordinates. The average initial position (black circle) and translation velocities (black arrows) in four quadrants are indicated. Red tracks represent the outer-MCSs that occurred during June – September, and blue tracks for October – November, green tracks for other months. Small triangles indicate the terminal positions. (c) Histograms of the number of outer-MCSs for the orientation of outer-MCSs with respect to the vertical wind shear vector, where U.S.L., D.S.L., D.S.R and U.S.R stand for upshear left, downshear left, downshear right and upshear right.

Download figure to PowerPoint

[13] In a coordinate system moving with the TC centers (Figure 2b), the mean distance of the outer-MCS centers from the TC centers was 367 km. Whereas those outer-MCSs that occurred during June-September were most often to the southwest of the TC centers as in Typhoon Morakot, the outer-MCSs that occurred during October and November were often to the north of the TC centers. The synoptic environments associated with outer-MCSs that occurred south of the TC centers and outer-MCSs that occurred north of the TC centers were defined based on the FNL (final) operational global analyses of the National Centers for Environmental Prediction (NCEP) on a 1° x 1° latitude-longitude grid (not shown). The outer-MCSs that occurred south of the TC centers usually developed in environments influenced by southwesterly flows, as was the case with TY Morakot [Chien and Kuo, 2011] and TY Mindulle [Chien et al., 2008]. On the other hand, the outer-MCSs that occurred north of the TC centers usually developed in a northeasterly flow environment that was similar to the environment that interacted with TY Babs as described inWu et al. [2009]. Therefore, the outer-MCSs were classified according to their formation locations into the south-type (85 cases) and the north-type (24 cases). As indicated inFigure 2b, outer-MCSs in all quadrants tend to propagate outward at speeds ranging from 6 km h−1 to almost 16 km h−1, with the larger speeds being associated with the more numerous south-type.

[14] Numbers of outer-MCSs with various orientations with respect to the vertical wind shear vector are shown inFigure 2c. The 850–200 hPa shear vector was calculated in a radial ring of 200–800 km from the NCEP FNL analyses. Corbosiero and Molinari [2002, 2003] examined the orientation of lightning in TCs related to the vertical wind shear and showed that lightning strike maxima are favored in the downshear right quadrant for the “outer rainband region” and in the downshear left quadrant for the “inner-core region”. As indicated in Figure 2c, outer-MCSs during June-September were favored in the downshear right quadrant, and during other months were favored in both the downshear right and downshear left quadrants. Since these orientations of the shear vectors are related to the monsoonal flow, the shear orientation may be one of the important factors determining the azimuthal variations of outer-MCSs.

[15] Although both south-type and north-type outer-MCSs may occur with TCs of any intensity (Figure 3a), the south-type were slightly more frequent during the Tropical Storm stage. By contrast, the north-type outer-MCSs most commonly occurred in the Tropical Depression stage. Relative to the overall frequency of TCs in the JTWC best-track file, only a small percentage of north-type outer-MCSs occur in the weak Typhoon stage (65–94 kt) and the strong Typhoon stage (>94 kt). As indicated above, the south-type outer-MCSs occurred more frequently relative to the overall frequency in June, July, and August, while north-type outer-MCSs were more frequent in October and November (Figure 3b). The explanation is that these outer-MCSs occur in conjunction with the low-level southwesterly monsoon flows during June - August and with the northeasterly tradewinds or northeasterly monsoon flow during October and November.

image

Figure 3. Histograms of the percentages of the 8968 JTWC 6-h files during 1999–2009, the ratios of south-type (84) outer-MCSs relative to all (109) outer-MCSs and the ratios of north-type (25) outer-MCSs to all outer-MCSs for (a) TC intensity categories and for (b) months. (c) Histograms of the diurnal distributions of the center times in Local Standard Time (LST) of the 85 south-type outer-MCS and of 24 north-type MCSs. Values in Figure 3c have been smoothed with a 3-hourly running mean.

Download figure to PowerPoint

[16] Diurnal variability of south-type outer-MCSs and north-type outer-MCSs according to the midpoint in their lifecycle is shown inFigure 3c. Both types had an early morning diurnal maximum and an afternoon minimum. Following Gray and Jacobson [1977], the proposed explanation is that deep cumulus clouds over the tropical ocean are forced by the difference in nocturnal radiational cooling between cloudy and clear areas. The diurnal variability of south-type outer-MCSs and north-type outer-MCSs was also examined by the times that maximum areas of CCS−75 were developed (not shown). A similar diurnal variability was found as for the centerpoints of these convective system cycles.

4. Stratiform Precipitation Regions and Mesoscale Surface Jets Collocated With Outer-MCSs

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methodology
  5. 3. Climatology of Outer-MCSs
  6. 4. Stratiform Precipitation Regions and Mesoscale Surface Jets Collocated With Outer-MCSs
  7. 5. Conclusion
  8. Acknowledgments
  9. References

[17] The PMW images and the QuikSCAT oceanic 10-m wind observations of outer-MCSs embedded in Typhoon Fengshen (2008) and Typhoon Nuri (2008) are shown inFigure 4. These QuikSCAT winds were from the microwave scatterometer SeaWinds as archived at the Remote Sensing Systems website (http://www.ssmi.com). Cecil et al. [2002] and Cecil and Zipser [2002] examined 261 TRMM overpasses of 45 hurricanes and documented the radar reflectivity values, passive microwave ice scattering magnitudes, and total lightning. They showed that the 215 K brightness temperature threshold may represent a distinction between stratiform and convective rain. Hong et al. [1999] and Varma and Liu [2010] provided an algorithm for classifying rain types using multiple channels of satellite microwave observations. For this study, the large area of moderate PMW TB (215 K–230 K) and the areas of very low PMW (TB< 215 K) of each outer-MCS inFigure 4are referred to as the stratiform precipitation region and the convective precipitation region of the outer-MCS, respectively.

image

Figure 4. (a and c) PMW images and (b and d) QuikSCAT oceanic surface wind observations for the outer-MCSs embedded in Fengshen (2008) and Nuri (2008), respectively. The thick black lines indicate the contour of zero relative vorticity based on the QuikSCAT wind observations. The brown dashed lines indicate the −75°C cold cloud shield of the outer-MCSs from the IR1 images. Figures 4a and 4c are modified versions from NRL TC_PAGES Page (www.nrlmry.navy.mil/tc_pages/tc_home.html).

Download figure to PowerPoint

[18] Two examples of QuikSCAT oceanic surface wind observations on a 0.25° × 0.25° latitude-longitude grid that revealed strong low-level winds under the stratiform precipitation regions of these outer-MCSs are shown inFigures 4b and 4d. These mesoscale surface jets, which were defined as regions extending over 200 km in length with surface wind speeds greater than 17 m/s, were identified under the cold cloud shields of 80% of the south-type and 88% of the north-type outer-MCSs. Notice inFigure 4that the deep convective cells were located on the left side of the mesoscale surface jets, which is an area of positive relative vorticity. Thus, it is hypothesized that formation of new convective cells will be favored in the cyclonic shear vorticity environment on the left side of a mesoscale jet because of the frictional convergence in the boundary layer. The continued convective formations in conjunction with the mesoscale jet under the stratiform region may possibly contribute to the long duration of the outer-MCS.

[19] The mechanisms for the formation of these mesoscale jets in the outer-MCSs are not clear.Didlake and Houze [2012]used airborne Doppler radar to document that the inner-core rainbands of Hurricane Rita had a low-level jet, which was attributed to radial advection of tangential momentum. The rainbands at larger radii had low-level and/or midlevel jets that may be attributed to both radial and vertical advection, because the larger Convective Available Potential Energy (CAPE) at larger radii enhances upward advection. Since the low-level jet associated with the inner-core rainbands of Hurricane Rita had convective cells oriented along the rainband, theDidlake and Houze [2012]mechanism may not be applicable to the outer-MCSs in typhoons without further justification.

[20] Downward momentum transport many be another mechanism for the formation of the mesoscale surface jets. Moncrieff and Klinker [1997] and Mechem et al. [2002]have proposed that when the stratiform region subsidence becomes particularly strong and widespread, the subsiding middle-level flow transports momentum downward. Similarly,Houze [2004] found some of the very large MCSs over the tropical warm pool had downward transport of momentum. Further study is required to determine the applicability of this mechanism in the environment of the outer region of TCs.

5. Conclusion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methodology
  5. 3. Climatology of Outer-MCSs
  6. 4. Stratiform Precipitation Regions and Mesoscale Surface Jets Collocated With Outer-MCSs
  7. 5. Conclusion
  8. Acknowledgments
  9. References

[21] This study is the first-known systematic climatology of the “outer-MCS” in western North Pacific tropical cyclones. These long-lived, well-organized, and cold-topped MCSs develop from distant rainbands in the outer region of a tropical cyclone. When these outer-MCSs interact with the terrain, they may lead to unexpected torrential rainfall well outside of the TC inner core, as occurred in Typhoons Mindulle (2004), Kalmaegi (2008), and Morakot (2009) and caused severe damage in Taiwan. Thus, the asymmetric precipitation patterns (on both spatial and temporal scales) due to outer-MCSs associated with TCs are very critical to quantitative precipitation forecasting in the western North Pacific.

[22] This study demonstrates outer-MCSs are not rare events in the western North Pacific. Approximately 22% (71/322) of all TCs produced at least one outer-MCS during their lifecycle. These outer-MCSs occurred most frequently in the Tropical Depression and Tropical Storm stages. The mean duration and the mean distance from the outer-MCS to the TC center are 10.3 h and 367 km, respectively. The outer MCSs that occurred from June to September were usually located to the southwest of the TC center and those that occurred during October and November were usually located to the north of the TC center. The south-type outer-MCSs are attributed to the interactions between the TC outer circulation and the typical environmental flow associated with the southwesterly monsoon, and the north-type to interaction with the northeasterly tradewinds or northeasterly monsoon. As the stratiform precipitation region of an outer-MCS became larger, a surface wind jet was typically present. Formation of new convective cells appeared to be favored on the cyclonic shear (left) side of the mesoscale surface wind jet. However, the formation mechanism of the jet is unclear and will be the focus of a future study.

[23] Yuter and Houze [1998] and Houze [2004]suggested such large stratiform regions are sustained by the environmental flow. Future studies will focus on the environmental factors that lead to the development and maintenance of the stratiform precipitation regions of outer-MCSs, e.g., unstable and moist air transported by the southwesterly flows, and low-level to middle-level vertical wind shears. Better understanding and the ability to numerically model and forecast the outer-MCSs in TCs will lead to more accurate and timely warnings of these heavy rainfall events.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methodology
  5. 3. Climatology of Outer-MCSs
  6. 4. Stratiform Precipitation Regions and Mesoscale Surface Jets Collocated With Outer-MCSs
  7. 5. Conclusion
  8. Acknowledgments
  9. References

[24] Cheng-Shang Lee is supported by the National Taiwan University and the Taiwan Typhoon Flood Research Institute, Buo-Fu Chen is supported by the National Taiwan University, and Russell Elsberry is supported by the Office of Naval Research Marine Meteorology section. Penny Jones provided excellent manuscript preparation support. This research is supported by the National Science Council of the Republic of China (Taiwan) under grants NSC 98-2625-M-002-002 and NSC 99-2625-M-002-013-MY3.

[25] The Editor thanks the two anonymous reviewers for their assistance in evaluating this paper.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methodology
  5. 3. Climatology of Outer-MCSs
  6. 4. Stratiform Precipitation Regions and Mesoscale Surface Jets Collocated With Outer-MCSs
  7. 5. Conclusion
  8. Acknowledgments
  9. References