Estimating the Peak Intensity of Super Typhoon Hagibis in 2019

Super Typhoon Hagibis nearing peak intensity on October 7, 2019. Image courtesy of The Naval Research Lab.

On August 15th, 1987, the 54th Weather Reconnaissance Squadron flew its final mission in Typhoon Cary, marking the end of an era. Since the 1940s, flying aircraft reconnaissance into typhoons was a regular occurrence all across the Western Pacific, but aside from the occasional research mission, they would be flown no longer. All was not lost though. Unlike when recon was first flown in the basin, typhoons could now be tracked by both geostationary and polar orbiting satellites, and their intensity could be estimated with reasonable confidence through application of the Dvorak Technique. As time went on, other ways to remotely estimate the intensity of a tropical cyclone were developed, including the automated objective Advanced Dvorak Technique (ADT) and the satellite consensus SATCON, which uses primarily microwave data provided by polar orbiting satellites.

There is no substitute for direct observation though. While remote sensing offered by the Dvorak Technique, ADT, SATCON, and others is often a reasonable placeholder, the true intensity of a system without direct observation will never be actually known. These intensity estimation methods can struggle with various extremes as well. Small overall TC size, small core size, large core size, and rapid fluctuations in intensity can all give these methods fits. The Tropical Western Pacific is a land of extremes. Aircraft reconnaissance investigated fourteen different typhoons with a minimum pressure of 900 mb or lower in the ten year span from 1978-1987, a number that exceeds all other recorded occurrences during observed history anywhere else in the world. Seven of the eight deepest observed tropical cyclones have called the confines of this basin home, including the deepest of them all: Super Typhoon Tip from 1979. Because of a lack of direct observations, surpassing the 870 mb measured minimum pressure from Tip is a tall order, whether it deserves to be or not.

Super Typhoon Tip nearing peak intensity on 2117Z October 11, 1979, as seen by DMSP imagery. Aircraft reconnaissance  would record the global minimum sea level pressure of 870 mb in the eye of Tip on the next flight into the system about 6 hours later. Image courtesy of Mariner’s Weather Log, Vol. 42, no. 2 August 1998.

Early in October 2019, a disturbance began to gain organization in the eastern portions of the Tropical Western Pacific, not too far removed west of the International Date Line, which separates the Pacific Ocean into eastern and western halves. As the disturbance traversed westwards, it encountered a region of low latitude westerly winds, aiding vorticity and helping to establish the disturbance’s identity amidst the background trade winds. While in a low vertical wind shear environment, an anticyclone aloft quickly developed in response to thunderstorm convection associated with the disturbance, hastening the development process along even further. On October 5th, it had become apparent that the disturbance had gained sufficient organization to be classified as a tropical cyclone. Not long afterwards, the Japan Meteorological Agency (JMA) gave the tropical cyclone the next name on the naming list: Hagibis.

Hagibis developing into a tropical cyclone northwest of the Marshall Islands early on October 5, 2019, as seen by Himawari-8 band 3 visible imagery. Of note are trade winds to the north and east of the system and low latitude westerlies to the system’s southwest. Image courtesy of RAMMB SLIDER.

Upon development, Hagibis wasted little time intensifying. As the system traversed westwards towards the Northern Mariana Islands, it was passing over the warmest and most heat laden waters anywhere in the world at the time, mostly undisturbed this season by abnormally low typhoon activity. Additionally, because of atmospheric changes associated with seasonal shifts, the upper troposphere had already begun to cool. This temperature difference between ocean waters and upper troposphere is what drives a tropical cyclone, and because a tropical cyclone is a Carnot Heat Engine, the bigger the temperature difference between the two results in a higher maximum potential intensity. In the case of Hagibis, maximum potential intensity was truly eye-popping. The system was already taking advantage of these favorable conditions due to basically non-existent vertical wind shear and a well-established upper level anticyclone with good outflow, but in order to actually approach the potential provided by the environment, Hagibis would need to establish an inner core, a chaotic process that currently is not fully understood.

Top Left: Oceanic Heat Content (OHC, kJ/cm**2) in advance of newly developed Hagibis, valid 12Z October 5, 2019. Image courtesy of RAMMB. Top Right: Atmospheric sounding from Guam displayed on a Skew-T, valid 00Z October 7, 2019. Data from University of Wyoming, formatted for use with SHARPpy. Bottom Left & Right: Maximum Potential Intensity (MPI) valid 06Z October 5th, 2019 expressed in terms of maximum one minute sustained winds and MSLP, respectively. Both images courtesy of CIMSS.

During the course of October 6th, it became apparent that based on passive microwave data from polar orbiting satellites, an inner core was indeed developing and consolidating at a rapid pace. The first hints of a possible eye were observed early on October 6th, but was soon obscured by intense convection. A SSMIS F-16 pass at 0506Z corroborated the brief appearance of the eye feature on conventional imagery with a complete ring on 91H. Over the course of the following twelve or so hours, each following microwave pass showed continued organization overall and a decreasing eye diameter, and by the time the next F-16 pass came in at 1742Z on October 6th, the now developed core had morphed into a configuration supportive of a pinhole eye. Hagibis had already been strengthening rapidly, but the stage now appeared set for explosive intensification.

Selected 89-91 GHz horizontal pol microwave passes of Hagibis on October 6th. Noted organization and core contraction can be noted with each new pass. Images courtesy of The Naval Research Lab.

As the sun rose over Hagibis late on October 6th and early on October 7th UTC, the pinhole eye emerged in the center of the cold CDO the typhoon had developed the previous day. Eye diameter was extremely small, even for pinhole standards, and reminiscent of the eyes from Hurricane Wilma in 2005 and Super Typhoon Parma in 2009. The eye size was small enough to challenge even the spatially impressive 2 km resolution infrared data provided by the geostationary satellite Himawari-8. Impressively, Hagibis maintained a ring of very cold cloudtops of about -80ºC around the pinhole eye throughout the entirety of the daylight hours, when sunshine has a warming effect on cloud tops.

Eye size comparison between Super Typhoon Hagibis (top left), Super Typhoon Parma (right two), and Hurricane Wilma (bottom left). Both Parma images are from the same time (2157Z September 30, 2009), but included twice for direct comparison between both Hagibis and Wilma despite image source limitations. All three have very similar eye sizes. Top two images courtesy of RAMMB and bottom two courtesy of the Naval Research Lab. Images from the same source are to scale.

Hagibis maintained its pinhole eye through sunset and into the overnight hours. As the sun disappeared beneath the horizon, the region of -80ºC or colder cloud tops expanded within the CDO. Gravity waves could also clearly be seen emanating radially outwards from the eye across the CDO. The extremely rapid strengthening appeared to continue until roughly 12Z October 7, at which point satellite presentation appeared to level off or even slightly degrade. Microwave imagery would reveal that an outer eyewall developed at about this time, completely encircling the tiny pinhole eye feature and initiating the beginning of an eyewall replacement cycle and a weakening phase.

Infrared loop of Super Typhoon Hagibis rapidly intensifying and approaching apparent maximum intensity from 0500Z to 1100Z October 7th. Note in particular the marked increase of -80ºC or colder cloud tops surrounding the pinhole eye after 0600Z, coinciding with sunset over the system. Loop courtesy of RAMMB.

At the apparent intensity maximum at 12Z October 7th, JMA and the Joint Typhoon Warning Center (JTWC), the two main agencies in the Western Pacific, finally assigned Hagibis their T7.0/category 5 equivalent intensity. This intensity is based mostly on fairly sound subjective Dvorak analyses that adheres to some degree of constraints, but tropical cyclones that undergo extreme rapid intensification don’t nicely follow these constraints that help accurately assess most normal systems. SATCON and other microwave estimates are often particularly useful in assessing the intensity of high end systems, but the pinhole eye and associated small radius of maximum winds plays havoc on their resolution and ability to produce an accurate estimate. The objective ADT is also a useful tool, but has automated constraints that limit intensity estimates in extreme rapid intensification scenarios like the subjective Dvorak Technique. Additionally, the extremely small pinhole eye of Super Typhoon Hagibis was able to challenge even the very good resolution of Himawari-8, and it is possible that the actual warmth of the eye (which is used in the ADT intensity calculation) isn’t fully captured. With all these limitations in more traditional intensity estimation techniques, perhaps a better intensity estimate can be derived for this case of extreme rapid intensification based on past data from similar systems.

Using aircraft reconnaissance along with satellite and other data dating back to 1970, four suitable analog tropical cyclones based on eye size, structure, and environment were found. These four tropical cyclones in chronological order are Super Typhoon Irma (1971), Super Typhoon Forrest (1983), Hurricane Wilma (2005), and Hurricane Patricia (2015). Several other tropical cyclones were considered but eliminated for various reasons. Super Typhoon Nora (1973), Super Typhoon Tip (1979), Super Typhoon Wynne (1980), and Super Typhoon Vanessa (1984) had impressive extreme rapid intensification episodes as well, but none of these systems had an eye that reached as small as 5 nm in diameter. Eye size appears to be an important determinant in deepening rate with extreme rapid intensification. Super Typhoon Megi (2010) met eye size criteria, but the early appearance of an outer eyewall while deepening eliminated it on structural grounds. Super Typhoon Vera (1979) and Typhoon Agnes (1984) were also eliminated structurally due to small overall storm and anticyclone size. Super Typhoon Kim (1980) and Hurricane Maria (2017) were crossed off on an environmental basis due to land interaction. Super Typhoon Judy (1979) also failed environmentally due to moderate northerly shear impinging upon the system. Two notable omissions include Super Typhoon June (1975) and Hurricane Gilbert (1988), a pair of storms whose recon data is difficult to find.

Top two images, from left to right: Typhoon Hagibis 1800Z October 6th, 2019 & Hurricane Patricia 1215Z October 22nd, 2015. Bottom two images, from left to right: Hurricane Wilma 1915Z October 18th, 2005 & Typhoon Forrest 0000Z September 22nd, 1983. All images are infrared images from geostationary satellites on the Basic Dvorak (BD) temperature curve. All images are near the point at which extreme rapid intensification began. Each TC can be characterized by embedded centers when using the Dvorak Technique as well as expansive outflow. Images courtesy of The Naval Research Lab and Digital Typhoon. Not all images are to the same scale.

With four reasonable analog tropical cyclones with recon data selected, their deepening rates can be examined. All four tropical cyclones had about an 18 hour period where deepening rates approached or even exceeded an average of 5 mb/hr. Geostationary imagery was not available in 1971 to view Super Typhoon Irma, but at the beginning of the period of extreme deepening rates, Super Typhoon Forrest, Hurricane Wilma, and Hurricane Patricia all featured a cloud pattern that would be classified with the embedded screening when assessing a tropical cyclone with the Dvorak Technique. The same can be said for Typhoon Hagibis at 18Z October 6th, 2019, which appears to be when it began a similar 18 hour period of extreme rapid intensification. Based on a center embedded at least 30 nm in a shade black or colder on BD imagery, Typhoon Hagibis is estimated to have an intensity of 90 kt at 18Z October 6th. Minimum pressure at the same time is estimated at 965 mb using the Knaff-Zehr-Courtney Wind/Pressure Relationship (KZC) with JTWC operational best track data. This intensity estimate appears reasonable based on a SATCON estimate of 87 kt/968 mb at about the same time.

Interpolated deepening rates of four analog tropical cyclones for Super Typhoon Hagibis when referenced against the initial pressure at the beginning of 18 hour extreme rapid intensification. An average of the four is also provided. Pressure data used courtesy of Tropical Atlantic/East Pacific/West Pacific and the JTWC Annual Tropical Cyclone Reports.

The average deepening rate of the four analog tropical cyclones was then applied to Typhoon Hagibis with 18Z October 6th used as the initial point of extreme rapid intensification. The average 88 mb drop from the four tropical cyclone average applied to the estimated 965 mb pressure results in a minimum pressure estimate of 877 mb at 12Z October 7th. Using JTWC data in KZC to solve from wind from pressure results in somewhat disparate estimates of 170 kt when using the radius of tropical storm force winds (r34) of 204 nm as the size parameter and 179 kt when using the radius of the outermost closed isobar (rOCI) of 210 nm as the size parameter instead. This required further investigating.

When examining ASCAT data from near the time of maximum intensity, JTWC’s r34 estimate of 204 nm appears to be reasonable. However, after cross-referencing surface analysis data at the time, the rOCI of 210 nm at 1002 mb appears far too small and at a pressure too low. Based on JMA surface analysis, it is estimated that the rOCI is actually closer to 420 nm at a pressure of 1006 mb. Re-running KZC with the modified data resulted in a new estimate of 173 kt using r34 as the size parameter and 175 kt with the new rOCI data. With much better agreement between the two KZC methods, the maximum intensity of Super Typhoon Hagibis is estimated to be 175 kt and 877 mb at 12Z October 7th, 2019.

ASCAT and surface analysis data for Super Typhoon Hagibis near the time of maximum intensity. ASCAT data courtesy of NOAA STAR Center for Satellite Application and Research. Surface analysis courtesy of Digital Typhoon.

Due to the lack of direct data observed with Super Typhoon Hagibis, such an intensity estimate is of low confidence. Many assumptions and extrapolations had to be made to arrive at this estimate, and this estimate is not official. However, the author feels that this estimate is reasonable. Given limitations with other intensity estimation methods previously discussed, this estimate may provide a better idea of Hagibis’s true intensity.

Following peak intensity, Super Typhoon Hagibis would weaken some due to eyewall replacement, briefly weakening to category 4 before once again regaining category 5 intensity. The typhoon would then move north and make landfall in Japan as a significantly weaker typhoon, but still as one of the two most impactful typhoons (along with Typhoon Jebi in 2018) in Japan in over a decade.

Best track file for Super Typhoon Hagibis modified based on the author’s intensity estimation through October 7th.
KZC intensity estimation worksheet for Super Typhoon Hagibis created and used by the author. The worksheet is generated using Python and the Hagibis best track data to help make the new estimate. The final estimates are also provided in the BT columns.
Top two images, from left to right: Super Typhoon Hagibis, 1100Z October 7th, 2019 & Hurricane Patricia, 0730Z October 23rd, 2015. Bottom two images, from left to right: Hurricane Wilma, 1045Z October 19th, 2005 & Super Typhoon Forrest, 1800Z September 22nd, 1983. All four tropical cyclones were near completing their period of extreme rapid intensification and approaching maximum intensity. Images courtesy The Naval Research Lab and Digital Typhoon. Not all images are to the same scale.

Reanalyzing the Pacific Typhoon Seasons: 1982

Even with the current best track data, the 1982 Pacific Typhoon Season is an above average one in terms of metrics like ACE and PDI. However, the season occurred alongside one of the strongest El Ninos on record, in the ballpark with the 1997 and 2015 El Ninos. When also including the very strong El Nino year of 1972, it is notable that 1982 is the only of these four very strong El Nino seasons to fail and reach 400 units of ACE in the best track data. Furthermore, 1982 featured a rather low number of super typhoons. JTWC data only has two for the season, Bess and Mac. This seems somewhat unusual for an El Nino season, especially one as strong as 1982’s. For these reasons, the 1982 Pacific Typhoon Season is the next season I decided to reanalyze. My summary tables for the 1982 best track data is below.

While reanalyzing the systems of the 1982 Pacific Typhoon Season, I began to have some issues with my wind radius estimates derived from JMA data. Even after reducing them via the method explained in my reanalysis introduction entry, the derived 34 kt wind radii were still very large and likely too large in many cases. In order to combat this, I began to also use a version of KZC that derives the storm size parameter ‘S’ from the radius of the outermost closed isobar rather than the 34 kt wind radius. This ROCI method of using KZC is laid out by Knapp et al. However, I began to have some issues with some of the more intense estimates above 155 kt or so where winds appeared to begin increasing more than they should. After some playing around and testing things (data from Patrica ’15 was particularly helpful) I came up with a tweak that appeared to iron the higher intensities out. For intensities above 127 kt, I would use the ROCI derived S parameter back-solve for a dummy 34 kt wind radius at a 127 kt intensity. Because the operational KZC that uses 34 kt wind radii to derive the S parameter involves using a modified rankine vortex in the original Knaff & Zehr, maximum sustained winds have an influence on the S parameter. The dummy 34 kt wind radius is then used with the Vmax and other data in the operational version of KZC to derive a modified S parameter. This helped greatly with estimating both wind speeds and pressures at the higher intensities.

After reanalysis (summary tables above), the 1982 Pacific Typhoon Season gets a notable upward bump in most metrics, including pushing the season’s ACE above 400 units. The difference between the best track data and the reanalysis data can be found here. Perhaps the most notable difference between the two is the huge increase in the number of super typhoons in the reanalysis data. I upgraded six additional systems to 130 kt or higher (Nelson, Andy, Cecil, Ellis, Nancy, and Pamela), bringing the total number of super typhoons up to eight. The number of tropical storm days for several systems went down slightly in response to me analyzing them to have transitioned to an extratropical system or a remnant low earlier than best track data indicates, but like my findings for the 1979 Pacific Typhoon Season, there are many intensity increases over the largely AH77 derived intensities in JTWC best track. Some of the more interesting systems will now be discussed in more detail.

02W Nelson: Vmax = 130 kt (C4), ACE = 23.33, PDI = 20.056

Nelson was a small early season straight running typhoon that rapidly intensified prior to landfall on the Philippine island of Leyte on March 25th. Just before landfall and near the time of the image above, a minimum pressure of 934 mb was recorded by aircraft recon between 07Z and 08Z on March 25th. I ended up using the 934 mb as the 06Z pressure and overall lowest pressure for the system, although the next pass less than two hours later had risen to 938 mb as Nelson began interacting with the Philippines. Combined with the near 24 hour gap in recon coverage prior to the 934 mb pass, it isn’t impossible that Nelson’s pressure was a little lower in the few hours previous to recon’s arrival, perhaps in the neighborhood of 930 mb or so. Satellite presentation for the small system was very impressive just prior to landfall with instantaneous DTs of 7.5 and approaching 8.0 at times. KZC output for the system with the 934 mb pressure was about 130 kt, which is the basis for my peak intensity. Nelson is the first of my eight analyzed super typhoons for 1982.

10W Andy: Vmax = 130 kt (C4), ACE = 20.88, PDI = 19.47475

Andy was the first in a string of several intense typhoons beginning in late July and continuing through the following months. Andy developed from the same monsoon trough setup that spawned Super Typhoon Bess, and both existed together for most of their lives, with Andy to the west and Bess to the east. For the most part, Andy was a large and sloppy storm and had a slow, steady pressure drop for most of its development. However, between 00Z and 12Z July 27th, a tiny pinhole core developed within the sprawling circulation, and in response the pressure rapidly dropped from 944 mb to 915 mb in that timeframe. Unfortunately, the 915 mb recon pass at 0921Z July 27th was the last mission flown into the typhoon. I maintained the 915 mb pressure for the 12Z July 27th best track point as well as used it for the system’s overall deepest pressure. Satellite presentation at the time (image above) was the most impressive over the system’s life. Andy appeared to enter eyewall replacement almost immediately afterwards. Owing to the somewhat unusual structure, the maximum sustained winds at the time are a little more uncertain than usual. When using KZC, the maximum sustained winds expected are about 130 kt. Additionally, AH77 and the ROCI version of KZC output winds of 125 kt and 135 kt, respectively. I ended up going with 130 kt for peak intensity, making Andy my second super typhoon of the 1982 Pacific Typhoon Season. Andy ended up making landfall in Taiwan a little over a day later with an intensity I assessed at 105 kt. Despite increasing the Vmax over the best track data, Andy is one of the few systems I analyzed to have less ACE and PDI than best track.

11W Bess: Vmax = 150 kt (C5), ACE = 39.6975, PDI = 44.860875

As mentioned above, Bess developed out of the same monsoon trough setup as Andy, emerging a ways to Andy’s east. Bess is one of the two super typhoons originally assessed by JTWC, and that doesn’t change here. Like it’s contemporary Andy, Bess was a large tropical system that only deepened gradually initially. Also like Andy, Bess ended up going through a phase of rapid intensification, although Bess ended up doing so on July 28th, about 24 hours after Andy did so. Unlike Andy, Bess deepened with a much more stable core structure, allowing the system to drop as low as 901 mb by 00Z July 29th, which is around the time of the satellite image above. Bess is one of the systems in which the wind radii derived from JMA data appeared to be too large. With an average wind radius of about 300 nm derived at peak intensity, KZC expected a maximum sustained wind of 145 kt. This is a very low Vmax for a 901 mb pressure, and while Bess was a large system, a 300 nm wind radius is really probably pushing it. I also derived a 450 nm outermost closed isobar radius and ended up with an expected Vmax closer to 155 kt when using that version of KZC. I ended up settling on a compromise intensity of 150 kt since the T7.0 satellite presentation did not strike me as one that would belong to a large 155 kt system. Following peak intensity, Bess began to steadily weaken as it approached Japan. Bess ended up landfalling on the main Japanese island of Honshu as a minimal typhoon (estimated 65 kt). Despite this, Bess became one of the most impactful typhoons in Japan for the time.

12W Cecil: Vmax = 135 kt (C4), ACE = 23.3025, PDI = 23.603375

As Bess was impacting Japan, Cecil was beginning to organize over the Philippine Sea. Unlike with the previous two systems, Cecil was a small to normal sized system. While slowly meandering to the east of the Luzon Strait, Cecil began to rapidly intensify, dropping 50 mb in 24 hours on August 7th. Based on a 700 mb height of 2369 m, Cecil’s minimum pressure is estimated at 917 mb at 06Z August 8th. Such a pressure is usually low enough to indicate a category 5 system, especially for one of a more typical size, but the slow and erratic movement and low background pressures were working against Cecil in this case. Additionally, satellite imagery near peak intensity (above) did not match typical category 5 presentation. KZC analysis at the time of the estimated 917 mb pressure yielded outputs near 135 kt, and this value is chosen for the lifetime maximum intensity (LMI). This makes Cecil the fourth of my eight analyzed 1982 super typhoons. Cecil then traveled north up the East China and Yellow Seas, eventually making landfall in North Korea as a greatly diminished storm.

14W Ellis: Vmax = 145 kt (C5), ACE = 30.5825, PDI = 33.036125

A brief gap existed between Cecil’s demise and Ellis’s development, but Ellis continued the string of notable typhoons that began with Andy and Bess in July. Like Andy and Bess, Ellis is another large storm that developed from the monsoon trough east of the Mariana Islands. Ellis began a steady to rapid deepening early in its life and continued to do so until reaching pressures below 915 mb late August 22nd. The lowest extrapolated pressure of 913-914 mb came near 22Z August 22nd, but it came following a ~12 hour gap in recon coverage, and the next several passes all came in a couple millibars higher than the preceding one, so it is possible if not likely that the pressure was a few millibars lower in the recon coverage gap. Like with Bess, my derived wind radii appeared too large at times, so I leaned much more heavily on the version of KZC that uses the radius of the outermost closed isobar to calculate the S parameter at the higher intensities. Based on a 140 kt intensity estimated with an analyzed 915 mb pressure at 00Z August 23rd, the LMI is set 5 kt higher to 145 kt at 18Z August 22nd. Using the LMI to backsolve for pressure, it is estimated that Ellis achieved a minimum pressure of 911 mb coincident with the LMI. This certainly appears reasonable based on the prior pressure falls and following pressure rises. The 145 kt/911 mb intensity estimate also appears reasonable based on satellite imagery near peak intensity (above). Following peak intensity, Ellis begin to steadily weaken, continuing to do so until a Kyushu landfall several days later as an analyzed category 1 typhoon. Ellis is the fifth of my eight analyzed super typhoons for the season.

15W Faye: Vmax = 95 kt, ACE = 13.3525, PDI = 9.671125

Typhoon Faye is not significant because of the intensity it achieved. I only assessed a peak intensity of 95 kt before making landfall in western Luzon. After the landfall, Faye nearly dissipated, recon missions into what was left of the system were halted. It is at this point when things became interesting. During the time absent of recon missions, Faye developed and maintained a small area of convection. Then on August 28th, an eye appeared right in the center of this small area of convection. After nearly dissipating after crossing Luzon, Faye reinvented itself as a true micro-typhoon with one of smallest, if not the smallest, convective footprints I have observed in a tropical cyclone. Recon missions into Faye then resumed, but not until slightly after peak intensity. Still, pressures a little below 980 mb were found in the very small system. I estimate a peak intensity of 75 kt/977 mb at 12Z August 28th (around the time of the image above), but there is some uncertainly with this estimate.

23W Mac: Vmax = 160 kt, ACE = 39.98, PDI = 51.53075

Mac is the deepest typhoon observed during the 1982 Pacific Typhoon Season, so it isn’t a surprise that it is the strongest I have analyzed. Mac is also the sixth of the eight super typhoons analyzed. As is common in El Nino years, Mac developed out of the monsoon trough to the east of Guam. LMI is determined to be 160 kt based on the measured 895 mb pressure between 18Z October 4th and 00Z October 5th and both versions of KZC. After weakening from peak intensity, I have assessed that Mac became a category 5 a second time at 12Z October 6th, reaching a secondary minimum pressure six hours later at 911 mb. Aside from passing near Guam early in the system’s existence, Mac did not impact any land areas, threading the needle between a few groups of islands such as the Bonin Island while recurving into the extratropical North Pacific.

24W Nancy: Vmax = 130 kt, ACE = 23.5425, PDI = 23.418375

Nancy was a straight tracking westerly typhoon that spent almost its entire existence between 16ºN and 18ºN. During the 36 hours prior to landfall in north Luzon, the typhoon began to rapidly deepen, reaching a measured 933 mb just before landfall. Both versions of KZC estimate a 130 kt intensity at the time, and this is chosen as the LMI, making Nancy the seventh super typhoon of the year. Somewhat unusually, Nancy maintained good core structure following the system’s crossing of Luzon, allowing the system to reach an estimated secondary peak intensity of 120 kt by 18Z October 16th. No recon missions were flown into Nancy near the time of this secondary intensity peak, but I derived a 939 mb at the time using KZC.

27W Pamela: Vmax = 130 kt, ACE = 25.48, PDI = 22.549

Pamela, my final super typhoon of 1982, is a system that featured several rapid swings in intensity both up and down during its primarily westward track across the Tropical Western Pacific in late November and early December. Pamela originated at the far eastern reached of the basin near the International Date Line and began it long journey west at the beginning of November’s final week. While moving west fairly rapidly, Pamela began to rapidly deepen, featuring a 19 mb drop in 6 hours and 30 mb drop in 12 hours between November 26th and 27th down to a measured 940 mb pressure. Such a pressure is fairly high for a super typhoon, but the combination of faster than average movement, low latitude, and high background pressures mean that KZC derived intensity at the time of the 940 mb pressure was 130 kt. Considering the satellite signature (above) also supported such an intensity, I set the 130 kt to the 06Z November 27th best track point with the 940 mb pressure. After a little weakening and then some oscillating between category 3 and 4 intensity through November 28th, Pamela then began a period of rapid weakening, with pressures rising just as fast as they fell during the initial episode of rapid intensification. Pamela spent the last few days of November and first few days of December as a weak generally westward moving storm with pressures hovering around 1000 mb until intensification began again. A second bout of rapid intensification began on December 4th, ending with a 958 mb pressure and assigned 105 kt intensity at 00Z December 5th. This peak was short lived though, falling back to tropical storm intensity by December 6th before making one last brief stint at minimal typhoon intensity before crossing southwestward through the central Philippines and dissipating over the South China Sea.

Reanalyzing the Pacific Typhoon Seasons: 1979

Based on a 1970-2016 time period, the 1979 Pacific Typhoon Season currently stands as near average to marginally below average Pacific Typhoon Season. Per JTWC data, the 22 tropical storms is four below that average of 26, 14 typhoons below the 16.8 average, and 8 category 3+ typhoons near the 8.8 average. Beyond storm count, ACE and PDI (278.0175 and 255.981375, respectively) also fall a little below the averages for my time period (294.4839 and 268.450441, again respectively). A set of tables with different types of storm data for the 1979 Pacific Typhoon Season is provided below. For a full context a full breakdown of the 1970-2016 data is provided here.

As mentioned in my previous entry, a few tropical cyclones had winds considerably lower than what would be expected for their recon measured pressures even using the Atkinson & Holliday pressure/wind relationship (AH77). The three biggest offenders are Typhoons Hope, Judy, and Owen. Hope and Judy were each rated as category 4 typhoons despite pressures measured at 898 mb and 887 mb, respectively. In a similar vein, Owen was rated as a category 3 typhoon despite a 918 mb pressure. Straight AH77 outputs for these three in order would be 142 kt, 151 kt, and 125 kt. It is systems like these that were largely the spur for my efforts at reanalysis. Pressures for these three typhoons and the rest of the 1979 Pacific Typhoon Season were lifted from the 1979 JTWC Annual Tropical Cyclone Report. An outline in the main methodology for the reanalysis is provided in the introductory post. With that said, the results of my 1979 Pacific Typhoon Season reanalysis begin now.

By just about every metric available, reanalysis data jumps the 1979 Pacific Typhoon Season to an above average season. Storm count rises to 27 tropical storms, 18 typhoons, 11 category 3+, and 5 category 5s (four of which with an intensity of 155 kt or greater). Changes resulted in ACE substantially increasing to 354.31 and PDI to 353.63975. Common themes across the majority of reanalyzed systems include earlier initial classification, quicker intensification, and higher/earlier peak intensities with respect to best track data. 10 classified systems came up 20 kt or higher maximum wind velocities in reanalysis. 10 systems also had an ACE and PDI total over 3 higher than best track. In contrast, only four systems went down in maximum wind velocity in any way, only three lost ACE, and only four lost PDI. Additionally, one system not in best track was added (X1W). A brief numerical summary of changes in reanalysis (in the same table format above) can be seen here. Some of the more interesting individual systems will now be discussed in more detail.

01W Alice: Vmax = 135 kt (C4), ACE = 42.16, PDI = 42.368

The 1979 Pacific Typhoon Season began exceptionally early with Alice. In fact, Alice technically could be a 1978 system, gaining sufficient organization to become a tropical depression by 12Z December 31, 1978. However, I do not believe that Alice became a tropical storm until 00Z January 1, 1979, and while I would argue that systems the occur prior to the annual Western Pacific tropical cyclone minimum in early to middle February are more closely associated with the previous season than the calendar year season, I have kept Alice here with the 1979 storms. Regardless, Alice was already a fairly intense tropical storm (I estimate 60 kt) by the time recon first investigated the system at 0115Z January 2 and found a 986 mb pressure. My graphical pressure trace for Alice can be found here.

Despite forming from a very large and convectively blessed disturbance likely enhanced by an upward pulse of Madden Julian, Alice matured into a smaller than average system. Alice featured two separate peaks of category 4 intensity. The first, and overall peak intensity (pictured above), occurred late on January 7 with a pressure of 928 mb. Using the 928 mb pressure, 14 kt forward speed, 98 nm average TS wind extent, 12.2ºN latitude, and 1006 mb outermost closed isobar, KZC actually output 137 kt, which does technically meet category 5 criteria. However, intensification was halted at this point by the onset of southwesterly shear and the eye had already begun to cool on IR imagery. Because of this, I erred on the conservative side with a 135 kt intensity, which would make Alice a top-end category 4. This is still considerably higher than the 110 kt best track intensity.

Secondary peak intensity was preceded by a period where Alice’s pressure rose all the way up to 974 mb by January 9, which I analyzed as an 80 kt category 1. However, Alice reintensified back down to 938 mb by January 11th as a borderline midget tropical cyclone. My reanalysis yields an intensity of 125 kt for this peak, much higher than the 100 kt listed in best track at this time. Alice quickly weakened after this secondary peak however, becoming a remnant low by January 14.

05W Unnamed: Vmax = 65 kt (C1), ACE = 2.2325, PDI = 1.237375

05W was a slightly more tricky system to reanalyze since it did not have any recon data. However, it is clear based on satellite imagery that it was more intense than JTWC’s 30 kt, which is apparently based on AH77 output from a 998 mb ship report near the system from an unspecified time and location in the JTWC report. Peak DTs were 4.5 for an off-white eye embedded in medium grey. Constraints came into play some due to the rapidly emerging banding eye feature as the system accelerated to the northeast through the Luzon Strait and into the subtropics, but I did estimate a peak intensity of 65 kt at 06Z, May 23. KZC output a 983 mb pressure at that time, which appears reasonable enough to me.

09W Hope: Vmax = 155 kt (C5/T7.5), ACE = 25.33, PDI = 30.56275

Hope is one of three systems from 1979 that deepened to below 900 mb. Despite this, JTWC’s best track only lists a peak intensity of 130 kt. I found a more intense system. Between July 29 and July 31, Hope is a system that can be characterized by rapid deepening, as seen in the pressure trace. The deepening was particularly extreme prior to 12Z on July 31 and featured a 14 mb drop from 912 mb to 898 mb in 2 hours 22 minutes between 0648Z and 0910Z. KZC outputs for the 898 mb pressure, 15 kt forward speed, 222 nm average TS wind radius, 19.4ºN latitude, and 1003 mb outermost closed isobar a 151 kt intensity. Based on the rapid intensification taking place at the time and warming of the eye on IR imagery up to 12Z, I estimated a peak intensity of 155 kt at 12Z, July 31. At that time, KZC output an 894 mb pressure, which appears reasonable to me.

12W Irving: Vmax = 80 kt (C1), ACE = 11.9325, PDI = 7.850875

Irving is one of the few systems that I found to be less intense than best track. The system is characterized by steady gradual deepening. Based on satellite imagery, Irving frequently appeared to be sporting multiple incomplete eyewall structures, resulting in a sloppy, sprawling structure. When coupled with very low background pressures around 1000 mb, KZC estimates actually undercut the standard AH77 estimates and ran almost identically to a variable environmental pressure version of AH77 that I run on occasion [6.6 * (oci + 2 – p) ** 0.65]. The deepest pressure of 954 mb measured by recon yielded an 81 kt intensity (which I rounded to 80 kt) when coupled with a 7 kt forward speed, 222 nm average TS wind radius, 24.6ºN latitude, and 999 mb outermost closed isobar. This is about 10 kt lower than the 90 kt in best track.

X1W Unclassified: Vmax = 45 kt (TS), ACE = 1.135, PDI = 0.4335

As mentioned previously, I added one additional system not in JTWC’s best track. I originally spotted this little system in the subtropics off the Japan coast while looking at satellite imagery for Typhoon Irving. In fact, the only agency that has this system classified is the China Meteorological Agency. Because of this though, it has thankfully been included in IBTrACS as 1979224N32143, meaning that archived imagery of the system is available through HURSAT. Based on my own use of the Dvorak Technique, I estimate that this system, which I have personally classified as X1W (essentially extra one), achieved a peak intensity of 45 kt at 06Z, August 15. This is near the time of the image above. No wind radii data was available, so I somewhat spitballed a 40 nm average TS wind radius estimate to use KZC with. The 988 mb pressure that ended up resulting from the use of KZC looks a little low to me, but without much to go on, that is the pressure I stuck with.

13W Judy: Vmax = 165 kt (C5/T7.5), ACE = 37.6575, PDI = 44.769375

It’s hard to find a more classic pinhole eye tropical cyclone than Super Typhoon Judy. Not long after initially developing, Judy dropped about 100 mb in 48 hours, reaching a minimum pressure of 887 mb. Judy also possesses the warmest 700 mb temperature recorded in a tropical cyclone of which I am aware, a 34ºC temperature from the 1931Z center pass on August 19. Despite this, JTWC only listed a 135 kt maximum intensity in best track. Why? Well, I really can’t say. Regardless, the 887 mb pressure resulted in a 165 kt intensity estimate when put into KZC, and that’s even with an average TS wind radius derived from JMA data that looks like it could possibly be too big (186 nm). I did leave it alone since Judy did appear to be embedded in a surge of the southwest monsoon, which did give the system a rather striking monsoon tail at the time of peak intensity, but it wouldn’t surprise me if Judy were perhaps a few knots even more intense than my reanalysis indicates.

19W Owen: Vmax = 140 kt (C5), ACE = 26.0975, PDI = 26.639875

Typhoon Owen is another system that received a generous intensity bump. Based on a measured pressure of 918 mb, KZC returned a 138 kt intensity with a 6 kt forward speed, average TS wind radius of 149 nm, 23.1ºN latitude, and 1007 mb outermost closed isobar. I rounded this up to 140 kt, a 30 kt increase over the 110 kt category 3 maximum intensity in best track. As is often the case with intense systems, Owen was yet another rapid intensifier. The recorded pressures did oscillate near the time of peak intensity, so I tried to smooth it out as best as I could.

23W Tip: Vmax = 175 kt (C5/T8.0), ACE = 60.7375, PDI = 76.782875

A system that needs little introduction, Super Typhoon Tip is the world’s deepest recorded tropical cyclone, with an 870 mb pressure recorded at 0350Z, October 12. Unlike many of the most intense tropical cyclones however, Tip did not rapidly intensify early in its lifetime. Tip developed well to the southeast of Guam in early October, east of an exceptionally messy monsoon trough/gyre/slop setup that involved Tropical Storm Roger. Because of this it took Tip several days before any significant intensification could take place. However, on October 9, Tip began to strengthen appreciably, and at about 12Z, October 10, Tip began its first ~12 hour bout of explosive intensification. At the end of this bout, I analyzed Tip to have achieved a 155 kt based on the measured 900 mb pressure and KZC. Tip weakened slightly after this episode, but then began a second round of explosive intensification around 12Z October 11. By the time recon arrived back to the system early on October 12, Tip had intensified to an 870 mb system. Using KZC, an 870 mb pressure, 8 kt forward speed, average TS radius of 204 nm, 16.7ºN, and outermost closed isobar of 1005 mb yields a 175 kt maximum sustained wind estimate on the nose. As far as comparisons to other studies is concerned, this is in close agreement to the 173 kt/873 mb ADT estimate from Veldon et al.

Following peak intensity, Tip remained a rather deep system (although nowhere as deep as near peak), but grew to become an absolute behemoth with poor core structure. Much like with Irving, KZC estimated wind values sank down to and even occasionally below those from both versions of AH77 I ran for the system. Because of this, I tugged down a few post-peak intensities from the best track values, but usually only by 5 kt or so here or there.

24W Vera: Vmax = 160 kt (C5/T7.5), ACE = 23.85, PDI = 30.27075

Vera is the system where I departed from my methodology the most. The first involved the TS wind radii. I ended up determining that they were too large for the small typhoon. For peak intensity at 12Z, November 4 (image above), the average wind radii I was deriving was 149 nm, which is an average to slightly above average size. When using the minor axis of 30 kt winds from JMA rather than the average, I ended up with a 113 nm average for the same time, which is similar in size to Hurricane Patricia from 2015. Deciding that this was a much more reasonable estimate, I ended up using this methodology for estimating the average extent of TS winds for Vera’s lifetime.

The second revolved around a recorded flight level wind of 170 kt slightly before peak intensity at 0507Z, November 4. When using a fairly standard 90% reduction from 700 mb, a 155 kt wind velocity could then be derived. This is in spite of the extrapolated pressure of 921 mb at the time, which would only result in an estimate of about 145 kt using KZC. Keeping this in mind, the minimum pressure of 915 mb at 12Z, October 4 would usually result in a 150 kt intensity from KZC. However, when adjusted with the ratio of flight level wind reduction from the previous fix, maximum winds are estimated at a higher 160 kt. Because of the 170 kt flight level wind, reanalyzed winds are a little above the KZC estimates near peak intensity.


Reanalyzing Pacific Typhoon Seasons: An Introduction

With 2018 just hours away and the Northern Hemisphere entering its annual tropical cyclone minimum period. During times like these, I often look to the past and try to learn more about tropical cyclone histories and climatologies. As I have studied though, I’ve noticed some things seem a little off in the Western Pacific records. Since 1970, storm count independent statistics like Accumulated Cyclone Energy (ACE) per Tropical Storm and ACE per Tropical Storm Day have increased. Perhaps the most significant trends are observed with the Power Dissipation Index (PDI)/TS Day and the ratio of PDI to ACE (both graphed below).

Both metrics involve PDI, which correlates strongly to intense tropical cyclone activity with winds at or above the Saffer-Simpson Hurricane Wind Scale (SSHWS) category 3 threshold. In the middle of these trends lies 1987, a year significant in marking the end of regular Western Pacific aircraft reconnaissance. Knowing this, it may be logical to assume that the post-recon storms were perhaps overestimated in terms of intensity. However, Knaff & Zehr and Knaff & Sampson have both concluded that the intensity estimates, particularly for the more intense systems SSHWS category 3 and higher, rely too heavily on the Atkinson & Holliday Wind/Pressure Relationship (AH77), which has a low maximum sustained wind (Vmax) bias for deeper observed pressures in most cases. In order to more accurately asses the Tropical Western Pacific typhoon climatology, it appears a reanalysis needs to be conducted, particularly for years in the 1970s and 1980s where it was common practice to plug recon recorded pressures into AH77 and output Vmax. Beginning with this entry, I will try to do so, beginning with the 1979 Typhoon Season.

Nowadays, the favored wind/pressure relationship of most agencies is the Knaff, Zehr, & Courtney (KZC) relationship. This is because it accounts for multiple variables, unlike a traditional relationship like AH77 or the Dvorak relationship previously used in the North Atlantic basin. This will be my starting point in most cases for reanalysis. In order to use KZC operationally, either a pressure or Vmax is needed (pressure in this case), as is the storm’s overall speed, latitude, pressure of the outermost closed isobar (OCI), and an average radius of tropical storm force winds. Knapp et al. also demonstrated a use of the average radius of the outermost closed isobar as a substitute for average TS force winds, but given JMA’s well kept wind radii archive available at Digital Typhoon dating back to 1977, I felt it easier to use the average TS wind method. However, JMA’s wind radii are only kept 30 kt and 50 kt, meaning some conversions would need to be made for the data to be meaningful to my application. For that, I turned to the 2016 Typhoon Season, where a relationship between JMA’s 30 kt wind radii and JTWC’s 34 kt (TS force) wind radii was established (below). I wanted to use more seasons for a larger sample size, but given JTWC’s statement that 34 kt wind radii were not quality controlled prior to 2016, I decided to forego quantity in favor of quality.

System speed and latitude are both taken from JTWC best track coordinates, while the OCI is interpreted from the JMA weather chart archive provided by Digital Typhoon. When combined with the recon recorded pressures interpolated at best track points, a first guess for wind speeds are estimated using KZC, and any necessary adjustments are done from that point. For times in which recon data is unavailable, the Dvorak Technique is the primary basis for intensity, which is then used with KZC to derive a pressure.

I chose 1979 as my first year for reanalysis for a few reasons. The primary reason is that 1979 is the first year with near complete geostationary satellite coverage (provided by JMA’s GMS-1). This satellite imagery provides added insight to storm structure, size, and intensity, which can be used in combination with the KZC outputs for a greater degree of confidence in the new intensity estimates. 1979 also features several storms that on a glance appear to obviously be underestimated. In the current JTWC best track, two category 4s have pressures below 900 mb (Hope and Judy) and three category 3s have pressures recorded below 930 mb (Alice, Owen, and Sarah). The final reason is of course Super Typhoon Tip, the deepest observed tropical cyclone on record, occurred during the course of the 1979 Typhoon Season.

The next several of my entries will be dedicated to the reanalysis of Pacific Typhoon Seasons, beginning with 1979.

Western Pacific Tropical Analysis: October 18, 2017

Over the Tropical Western Pacific, the most potent monsoon trough setup of the year thus far has allowed Typhoon Lan to take center stage. Lan is a particularly large, sprawling system, and with favorable background conditions, is a system with a very high intensity ceiling. How much of this potential becomes realized depends largely on notoriously difficult to forecast inner core dynamics. Lan will move to the north, slowly at first but with a gradual increase in speed, and then to the northeast into the mid-latitudes. Prior to becoming a likely significant extratropical cyclone, Lan is expected to impact Japan with either a pass just offshore or direct landfall in about five days. An additional system or two may accompany Lan, but the impacts from Lan’s vast circulation should keep any additional systems in check concerning intensity.

***NOTE: While I would consider myself well-learned in meteorology, I am still a student with more to learn before becoming a degreed meteorologist. This forecast is not from an official source and should not be treated as such. For official information, please refer to your local weather agency.***

The 2017 Pacific Typhoon Season has largely been accompanied with a very weak or absent monsoon trough, so the season is unsurprisingly running well below seasonal activity. However, a high amplitude upward pulse from Madden Julian has ejected east from the Maritime Continent and overspread the western Pacific. Accompanying it has been a very impressive surge in monsoon southwesterlies extending into the Philippine Sea. As mentioned in the post’s lead paragraph, this is by far the most robust monsoon trough setup so far this year. Taking shape at the center of the Philippine Sea on the monsoon trough is Typhoon Lan. As of 03Z October 18, Typhoon Lan was located at 12.1ºN, 132.5ºW. Both JMA and JTWC assessed maximum sustained winds of 65 kt in their latest advisory packages.

As can be seen in the loop above, Lan is not the most visually striking typhoon. That is because it has yet to put together a respectable inner core, despite very favorable conditions. To me, this appears to be partly because of Lan’s size. Lan originally developed as a more normal sized system initially, but an additional surge of monsoon southwesterlies from the parent monsoon trough wrapped around and into the circulation, greatly expanding the system’s size. Lan featured an impressive CDO 12 or so hours ago and looked to be in the midst of core building, but the convection over the center has since collapsed. Conditions will remain particularly favorable for the next 72 hours prior to Lan’s passage of 20ºN, but considering the current state of the core (seen in a METOP-B pass from 0054Z October 18), it will probably take at least 24 hours for Lan to build a core and take advantage of the favorable conditions. Once a core is established, intensification is likely to be of the rapid variety. Most intensity guidance take Lan to category 4 intensity in 48 hours. Some guidance like the HWRF is probably consolidating the core too quickly, but forecasting a category 4 sometime in the next 48-72 hours does not appear unreasonable to me.

Lan will be moving north into a slight weakness in zonal subtropical ridge. Poleward outflow from Lan may actually help enhance the subtropical ridging to the east, which in turn will help draw Lan to the north with a gradual increase in speed. Lan should then begin to take on an easterly component to accompany the present northward motion once it begins to round the axis of the eastern ridge near 25ºN in 3-4 days. It is around this time when Lan will begin to interact with the right entrance region of a standing jet streak. The jet streak may temporarily increase upper divergence, but it will also begin shearing Lan as well as introducing dry air aloft into the circulation. The approach towards the mid-latitude flow in which the jet streak is embedded within will result in Lan’s acceleration towards the northeast. Lan is expected to pass near or just clip Honshu as a weakening, but still possibly significant system in slightly over 5 days. Lan could bring the potential for strong winds and impressive wave action along the coast, but rainfall may be the biggest threat to Japan from Lan. Due to Lan interacting with the right entrance region of the jet streak, heavy rain may extend well ahead of the main system.

Due to the track largely parallel with Japan, only a small difference in trajectory will exist between a direct strike and an offshore pass. At the moment, I am favoring a direct strike. Most guidance members are leaning towards the left of the two options. Probably the highest regarded member of guidance keeping Lan offshore is the HWRF. However, as mentioned earlier, the HWRF is probably consolidating Lan’s core too quickly, in turn ramping up Lan earlier than is likely. This may be impacting downstream depictions of subtropical riding and the jet streak, in turn steering the system further to the east. It’s a solution worth watching, but more the reliable guidance members remain to the west, landfalling in Honshu. Keep in mind that all the normal caveats for a 5+ day track forecast apply here. A Honshu encounter is just beyond JMA’s and JTWC’s 5 day forecasts, but an extrapolation from their 5 day forecast points appear to favor a direct strike as well.

Due in part to a busy schedule balancing college courses and a part time job, this is regrettably my first entry in over a month. I’m not sure I can keep a regular blog schedule like I was able to over the summer, but I will try to post when I can. Thank you all for reading! Until the next entry is up, updates will come in this entry’s comments section.

North Atlantic Tropical Analysis: August 24, 2017 – Harvey Special

After dissipating in the eastern Caribbean Sea, Harvey has managed to redevelop in the Gulf of Mexico. Over the past 24 hours, Harvey has managed to strengthen from a tropical depression to a SSHWS category 1 hurricane. Harvey is continuing to strengthen as it heads towards the Texas coast, where it will bring the threat for strong winds, storm surge, and highly excessive rainfall.

***NOTE: While I would consider myself well-learned in meteorology, I am still a student with more to learn before becoming a degreed meteorologist. This forecast is not from an official source and should not be treated as such. For official information, please refer to your local weather agency.***

As of 18Z, Hurricane Harvey was located at 24.4ºN, 93.6ºW per NHC coordinates. Winds of 75 kts and a pressure of 976 mb were measured by aircraft recon, and this is the initial intensity to which Harvey is set. Harvey is located over very warm waters and in a good upper air environment, which has allowed a brisk strengthening over much of the past day. Final reports from the last recon mission indicate that Harvey may be leveling off in the near term. I do think this is temporary as the system transitions from a banding type structure to one more centralized around an eyewall. An eye has started to become more apparent on all bands of imagery over the past few hours.

Aside from the warm waters, a favorable upper level environment has been paramount to Harvey’s strengthening thus far. The system has developed an impressive divergent poleward outflow jet, allowing for good mass removal from the top of the tropical cyclone. The divergent nature of this outflow jet also extends just to the northeast of Harvey’s core where spiral bands have been able to consistently fire deep convection. Some light southwesterly shear has helped to keep outflow to the south and southwest of the system restricted, but the vigorous nature of the poleward outflow channel has been able to offset this up to this point.

Thus far, Harvey’s main steering mechanism has been the subtropical ridge to the system’s east. This ridge will push Harvey on a general northwestward trajectory for the next 48 hours, during which time it is expected to make landfall along the middle Texas coast. Because of the angle of approach towards the Texas coast, only a small change in angle could result in a landfall many miles further down the coast. At this time, I see little reason to differ from the NHC’s forecast, which has the track centerline making landfall not far north of Corpus Christi in just under 36 hours.

Every indication I am seeing is that Harvey will continue to strengthen up to landfall. Waters remain warm and heat laden, the upper air environment looks good, and the structure as revealed by a recent F-16 pass continues to improve (shown below). Harvey is also beginning to come into the range of the Brownsville radar, where an impressive eyewall is already showing up. Based on this, I think it is about 90% certain that Harvey will become a major hurricane. The NHC official forecast takes Harvey up to 110 kt just prior to landfall, which is a high end category 3 hurricane. That appears to be a good intensity estimate to me, although it is not impossible for the system to strengthen even slightly beyond that. With landfall, along with the high winds that will in all likelihood be above major hurricane intensity in the eyewall, the threat for a double digits of feet storm surge exists near and just to the right of the center at the landfall point.


Even after landfall, Harvey will remain an impactful tropical cyclone, perhaps even more than at landfall itself. Once landfall occurs, steering is expected to break down. Harvey could be stuck drifting around near or just inland the Texas coast for a couple of days. Guidance begins to diverge at this point where exactly it will drift and how long it will remain quasi-stationary, but one thing that has been consistent is the monumental modeled rainfall totals. The WPC QPF forecast places a swath of over 20″ of rain along much of the Texas coast, and that is within an absolutely massive area that is expected to receive over 10″ of precipitation accumulation. As is the case with QPF forecasts, the rainfall will not be evenly distributed as the forecast graphic appears to indicate, but such mesoscale variations are nearly impossible to predict even in the short range. Regardless, much of south and southeast Texas should prepare for large scale flooding.

While there certainly is considerable spread, it does appear that the majority of guidance begins to trek Harvey to the northeast near or just off the coast after a few days of nearly stationary motion. Some guidance that brings the system offshore indicate that Harvey could intensify fairly quickly once more, but I expect the core to have taken a beating after moving inland, and do not think such solutions are very plausible. It is possible that it could take over five days from the time of this entry for Harvey to finally leave the vicinity of Texas.

For official, up to date information on Harvey and its expected impacts, please check out The National Hurricane Center. Updates from me will likely come in the comments section of this blog and/or on Twitter.

I usually post about the Western Pacific in my blog. For those interested, check out my post from earlier today, Western Pacific Tropical Analysis: August 24, 2017

Western Pacific Tropical Analysis: August 24, 2017

Typhoon Hato is now moving inland across southern China after making landfall near Macau as a category 3 typhoon. In its wake, Tropical Depression 16W has developed a little south, but near where Hato developed. It could end up taking a similar track as well. Elsewhere in the basin, no strong signal exists for tropical cyclone development over the next week.

***NOTE: While I would consider myself well-learned in meteorology, I am still a student with more to learn before becoming a degreed meteorologist. This forecast is not from an official source and should not be treated as such. For official information, please refer to your local weather agency.***

This is going to be a shortish entry today with a lot going on close to home.

Tropical Depression 16W has been designated today east of Luzon. As of 12Z, JMA placed the system at 16.4ºN, 126.5ºE. JTWC’s position was also nearby, but not exactly the same, which is not surprising considering the current state of organization. At this early point in development, 16W is experiencing some moderate easterly shear. That said, both JMA and JTWC are expecting 16W to gain enough organization to be classified as a tropical storm in about 24 hours. When this happens, it will be named “Pakhar.”

With a similar genesis and similar steering pattern to Hato, it should be unsurprising that 16W is expected to take a similar track to the one Hato took, towards Guangdong. However, unlike with Hato, 16W’s more southerly starting position means it will have to cross the Philippine island of Luzon. Because of this, I expect 16W to be a weaker storm than Hato, with lower odds that it will reach typhoon intensity. JTWC’s landfall intensity of around 65 kt appears reasonable to me, but I might hedge slightly lower. Regardless, 16W will be moving into the same regions recently affected by Typhoon Hato, bring additional impacts to the already storm weary areas.

Aside from 16W, not much else appears to have much development potential across the Tropical Western Pacific. Some potential development does appear possible along the weakening monsoon trough in the vicinity of the northern Marianas in about five days, particularly with American guidance. Any developing storm will likely move north and quickly into the mid-latitudes based on the active extratropical pattern.

My next Western Pacific entry will likely arrive Sunday. I am also doing a rare North Atlantic entry focusing on Harvey later today. Until the next WPac entry is up, updates will come in this entry’s comments section.

Western Pacific Tropical Analysis: August 17, 2017

Despite generally unfavorable conditions across most of the basin over the past week, a well organized disturbance was able to develop into Typhoon Banyan over the eastern portion of the basin. In Banyan’s wake, the Tropical Western Pacific is once again quiet, but it will likely not remain that way for long. A surge in the Southwest Monsoon is expected to sharpen the monsoon trough over the western Philippine Sea over the next couple of days, assisting in the development of a system to the east of Luzon and southeast of Taiwan. Additional tropical development may occur after the first system moves west.

One invest is currently designated in the Tropical Western Pacific. Invest 92W is located in the basin’s subtropical waters southeast of Japan. As of 18Z August 17, JTWC analyzed 92W’s center to be at 32.1ºN, 144.5ºE. 92W originates from a surface shear axis, which itself is the remnants of a decayed cold front. I was initially impressed with 92W’s convective cluster, but closer inspection revealed that the circulation center was well removed from this convection. I would be surprised if 92W ends up becoming a classifiable tropical cyclone. 92W is expected to move off to the northeast and into a baroclinic zone, where it will contribute to a new frontal wave in a couple of days.

No other invests are currently designated in the Tropical Western Pacific, but there is one more disturbance of note. An inverted surface trough exists just to the west of the Marianas in the eastern Philippine Sea. This disturbance is convectively active but poorly organized at the moment. Near-term development if any is expected to be slow. However, this disturbance will begin to interact with a tongue of monsoon southwesterlies that extends east of the Philippines in a couple of days. I expect this newfound monsoon trough environment to aid in the organization of a tropical storm in the western Philippine Sea by this time four days from now at the latest. The next name for a tropical storm on JMA’s naming list is “Hato.”

From its expected development near 20ºN to the northeast of Luzon and southeast of Taiwan, stout subtropical ridging to the north should keep this probable tropical storm on a primarily westward heading. Some easterly shear may be present to the south of this ridge, but a westward movement and a good outflow dump into the Tropical Easterly Jet may help offset the shear to a degree. At the moment, I am expecting the system to pass through the Luzon Strait and into the South China Sea, which is in good agreement with the ECMWF/EPS and UKMET solutions. However, a slight break between two cells in the subtropical riding could allow the system to track slightly more northerly and possibly impact Taiwan, which some of the GEFS and GEPS solutions indicate. I am favoring the European guidance at the time largely due to the consistency these solutions have displayed over the past few days. After entering the South China Sea, this system should continue to move primarily west towards the Chinese province of Guangdong. Depending on how much the system interacts with the higher terrain of Luzon and Taiwan, it is possible that it could intensify into a lower intensity typhoon.

In the wake of the previously discussed system, the monsoon trough is expected to remain established around 20ºN in the western Philippine Sea. Guidance is already alluding to a second developing system in 7-8 days in a very similar location to the first system. It’s too early to talk specifics on a system that might develop in the medium range, but the signal for such a system is fairly strong in spite of the range. Climatologically, activity seriously ramps up this time of year across the Tropical Western Pacific as well, which adds credence to the idea of additional development.

This will be my last blog for a week or so as I embark on a road trip to see the Solar Eclipse on August 21st. Until the next entry is posted, analyses and updates in forecast philosophy will arrive in the comments section.

Western Pacific Tropical Analysis: August 9, 2017

After about two and a half weeks, the Noru saga has come to a close at last. In the wake of the prolifically long lived typhoon, not much is left to be discussed as the basin has gone back into its slumber. Upper level subsidence has once again spread over the basin, and the monsoon southwesterlies have retracted back towards Southeast Asia, leaving little available focus for tropical development. The next period of significant tropical development is likely 10-14 days down the road, but as we approach the seasonal activity peak, it will become more and more difficult to keep down at least weak development for an extended period of time.

One coherent disturbance does currently exist over the Tropical Western Pacific. Located well to the east at approximately 14ºN, 174ºE, the disturbance features decent vorticity but lacking convective activity. The area this disturbance is traversing features high 500 mb heights and may be struggling with large scale subsidence. Considering the system’s present consolidation, tropical cyclogenesis cannot be ruled out with the system as it moves northwest and then north into a slight weakness between subtropical ridging, but subsidence and then shear north of 20ºN will likely prevent the disturbance from meeting requirements for a classifiable tropical cyclone. The system will likely be tagged as an invest, however.

Aside from the aforementioned disturbance, it is rather difficult to identify something with tropical cyclone potential over the next week or so. The belt of monsoon southwesterlies associated with last week’s reverse-oriented monsoon trough is now very far north and in the process of cutting off from the parent southwest monsoon. Large scale subsidence has also returned to the Tropical Western Pacific, and 500 mb heights are once again running abnormally high. Overall, conditions do not look favorable for tropical development.

In the medium range, there are some early hints that the monsoon trough may re-establish itself in the basin as subsidence begins to relax a little. Most guidance except American guidance show low level westerlies returning around day 10. I wouldn’t get hung up on American guidance failing to show a similar solution either, since it appears it gets Madden-Julian caught up in the Western Hemisphere, keeping a powerful downward pulse over the Tropical Western Pacific. This is a well-known bias in American guidance. Taking this into account, I’d say chances are rather good that the monsoon trough will make a return to the Philippine Sea, perhaps kicking off the next wave of activity around day 10.

Apologies for the missed post this past weekend as I was feeling rather poor. My next post is slated for this weekend. Until the next entry is posted, analyses and updates in forecast philosophy will arrive in the comments section.

Western Pacific Tropical Analysis: August 2, 2017

Typhoon Noru continues to spin across the Tropical Western Pacific, and the system has been joined by Tropical Storm Nalgae well to the east. Nalgae appears to be no threat to land, but the same cannot be said about Noru. Noru appears to be headed on a collision course for southern Japan and possibly South Korea. Noru has come back down to earth after a period of explosive intensification a few days ago, and some guidance solutions showing similar strengthening are likely overdone, but some strengthening remains possible before final landfall. Aside from the two active storms, no additional tropical activity appears likely over the next several days.

***NOTE: While I would consider myself well-learned in meteorology, I am still a student with more to learn before becoming a degreed meteorologist. This forecast is not from an official source and should not be treated as such. For official information, please refer to your local weather agency.***

The ever persistent Typhoon Noru is now approaching two weeks as a named system, an impressive feat. Noru has also generated the most ACE of any Pacific Typhoon since Goni and Atsani of 2015, and it will likely be passing them soon as well. As of 21Z August 2nd, the center of Noru’s now very large eye was located at 27.1ºN, 135.3ºE per JMA. JMA and JTWC’s latest intensities for the system are 85 kt and 95 kt, respectively. These are at or just below (respectively again) each agency’s T5.5 intensity. However, considering the particularly large eye and somewhat lackluster convective appearance, I have a suspicion that these intensities are perhaps a little generous. Regardless, Noru has come back down to earth after explosively intensifying into the first SSHWS category 5 tropical cyclone of 2017 a few days ago.

The defining feature of Typhoon Noru at the moment is the system’s massive eye. The eye both on infrared and microwave imagery measures about a degree and a half, or 90 nautical miles, in diameter. While not the largest tropical cyclone eye observed, it is in the upper percentile. This eye size is even more rare considering it lacks a remnant inner eyewall within it. With such a large core, the upper level outflow needs to be working overtime to properly vent the system. This has not been the case recently. Analyses have shown outflow to the system’s northwest restricted by a passing trough and associated PV streamer. This trough is now departing though, and there are already some early signs that the outflow is beginning to recover.

The aforementioned trough has been a major source of uncertainty in previous forecasts. Noru’s primary steering mechanism for the past several days has been a subtropical ridge to the north. The trough that is in the process of passing by weakened the ridge, and it was initially uncertain if enough of a weakness would be created to allow Noru to escape poleward. However, it now appears that Noru will not be escaping poleward. As the mid-latitude trough departs, the subtropical ridge should restrengthen some to Noru’s north, forcing the system slowly westwards towards the northern portion of Japan’s Ryukyu Islands. This west motion should continue for at least the next two days and at most three. After that, Noru is expected to be picked up by the next trough and recurve to the northeast. The exact recurve point is not yet known, but it does appear like it will occur either in the vicinity of the northern Ryukyus or in the East China Sea just to the west. This means landfall could still come in a variety of locations from anywhere on Kyushu to eastern South Korea, but considering that I expect Noru to continue maintaining an exceptionally large core, the worst of the typhoon could spread unusually far from the centerline. At the moment though, I expect a western Kyushu landfall, near the center of the guidance envelope.

As has been the case with Noru for all its life, the intensity forecast remains far from a walk in the park. Much of guidance has been blowing Noru back up into a very powerful system in the vicinity of the northern Ryukyus. American guidance in particular has been insistent on a system well below 900 mb. However, I continue to remain skeptical of such solutions, perhaps even more now than ever before. As mentioned previously, I expect Noru to maintain an exceptionally large core for the remainder of its life, which will need near perfect outflow for any strengthening to occur. Additionally, upwelling could become an issue for the large, slow moving storm. A strenghening signal does appear to be nearly unanimous in the vicinity of the Ryukyus, and this does seem to be in response to improving upper air conditions. The degree of strengthening is almost certainly overdone though. The HWRF more or less maintains the large cored Noru as a higher end SSHWS category 2 up until landfall, which appears to be a reasonable intensity forecast to me.

The other active system is Tropical Storm Nalgae, with formed from the reverse oriented monsoon trough well to the east of Typhoon Noru. In fact, the location of Nalgae’s birth is almost exactly the same as Noru’s. Nalgae is expected to strenghen some, perhaps nearly or just meeting the criteria for a minimal typhoon before being picked up by the same trough that just missed picking up Noru, sending the system north into the mid-latitude North Pacific.

Development over the rest of the Tropical Western Pacific looks unlikely over the next several days. There is an outside shot that Invest 98W, located to Tropical Storm Nalgae’s east, could organize enough to become classified, but considering the hostile conditions currently assaulting the invest, I do not expect this outcome to transpire.

My next post is slated for either Saturday or Sunday. Until the next entry is posted, analyses and updates in forecast philosophy will arrive in the comments section.