This monograph on convection-coupled systems in the tropics was inspired by the life and career of Professor Michio Yanai, whose major contributions to the subject spanned more than five decades. From a distant perspective, Professor Yanai’s career can be understood in the context of Japanese scientists who immigrated to the United States in the decades of the 1950s and 1960s, enriching the meteorological research community in the United States as well as abroad (Lewis 1993). A closer look reminds us that the tapestry of scientific progress is created by the contributions of individual scientists with their unique backgrounds, motivations, and talents, and the serendipity of events that shape their lives.
1. Early encounters with typhoons
Michio Yanai was born 16 January 1934, in Tokyo, and grew up in Chigasaki, a relatively small town nearby. He was fascinated by weather—typhoons in particular—at an early age. Immediately after entering Shonan Middle School in 1946, he joined the meteorology club (otenkikai). The “meteorology boys” (Fig. 1) were most excited by the approach of typhoons that came every year, typically in September. The club issued storm warnings to the school principal. Teachers not only received the club’s “unauthorized” warnings with enthusiasm, but also dismissed classes based on the meteorology club’s judgment.
On one occasion, the meteorology club decided that they would determine the precise path of a typhoon that had just passed over Chigasaki (Fig. 2). Riding on a commuter train the next afternoon, a student got off at each station to investigate the direction of the fallen trees. In this way the club systematically determined the exact path of the typhoon’s eye. Later in his career, Yanai was convinced that the meteorology boys invented this method before Professor Tetsuya (Ted) Fujita used a similar strategy to determine tornado tracks. The club reported its results to the Yokohama Weather Station (currently, the Yokohama District Observatory). They received thanks and praise for their investigation, and made the news the next morning.
2. Typhoon research at the University of Tokyo
Michio Yanai arrived as an undergraduate at the University of Tokyo in 1952. All students in their first and second years were required to attend the “Culture School” on the Komaba campus. Third-year studies focused on classical mechanics, fluid dynamics, thermodynamics and statistical mechanics, electromagnetism, and quantum mechanics. In the fourth and final year Yanai and his fellow geophysics majors studied geodesy, seismology, geo-electromagnetism, oceanography, and meteorology.
In 1956, Yanai entered the University of Tokyo graduate school as a meteorology major. Professor Shigekata Syono suggested that he study typhoons, but gave no specific instructions. After publishing his master’s thesis on a decaying typhoon (Yanai 1958), Yanai decided to study the formative process of typhoons. An intensive literature survey showed him that typhoons form either in high sea surface temperature regions in the ITCZ, frequently called the “equatorial front” (Palmén 1948), or from preexisting disturbances such as waves in the easterlies (e.g., Dunn 1940; Riehl 1948). But he found that there was not a single illustration of the actual formation process in the existing literature, even in Riehl’s (1954) Tropical Meteorology.
The Japanese Meteorological Agency (JMA) loaned Yanai a set of teletyped station reports covering the period of Typhoon Doris (1958). Professor Syono assisted in the data collection by asking Lt. Col. Bundgaard (10th Weather Group, U.S. Air Force) to provide a more complete collection of teletyped reports. There were no copying machines in 1958 and the loaned data forms had to be returned promptly, so Yanai copied the data by hand, day and night. The drawing of streamlines, isotherms, cross sections, and so forth, and the analysis of the formation process, occupied two full years of his graduate work. To study the heat source of this typhoon, Yanai defined the Q1 (heat source) and Q2 (moisture sink) terms (Fig. 3) to show their similarity. Later diagnostics developed at UCLA would emphasize the difference between Q1 and Q2 as a measure of convective activity and radiative effects.
In 1959, another typhoon would have a great impact on Yanai’s life. High tides and the storm surge from the Ise-wan Taifu (Typhoon Vera) caused 5000 deaths along the coast of Ise-wan (bay) near Nagoya. The cabinet member in charge of Science and Technology (future Prime Minister Yasuhiro Nakasone) advocated the creation of the Typhoon Research Laboratory. In 1960, the Diet (legislature) instructed the government to form a typhoon research laboratory in the JMA’s Meteorological Research Institute (MRI), and Yanai was invited to join after graduation and passing the civil service exam.
Meanwhile, Taroh Matsuno and Yanai (Fig. 4) formed a research team of eight graduate students to study the precipitation bands of Typhoon Vera (1959). They met once a week to discuss strategy and analyze data. This activity was purely voluntary and not directly related to the students’ thesis research. The group was so thorough in collecting data from non-JMA stations such as schools and railway stations that they received official protests from the JMA when the agency found nothing left at these stations when they attempted to collect the data. Results from the graduate students were published much later in two papers (Hamuro et al. 1969, 1970).
In 1960, Dr. Yanai presented his case study of Typhoon Doris at the First International Symposium on Numerical Weather Prediction in Tokyo. There he met Jule Charney, Akira Kasahara, Hsiao-Lan Kuo, and other scientists interested in numerical modeling of typhoons. Charney then visited Michio at the University of Tokyo to examine the original analyses of the typhoon. This occurred as the first attempts at numerically simulating tropical cyclone formation were being made. Simulations by Charney’s group at the Massachusetts Institute of Technology (MIT), Kasahara’s at the University of Chicago, and Syono’s at the University of Tokyo were all failing to produce an organized tropical cyclone, producing model grid-size, Rayleigh-convection-like structures instead.
The main thrust of Yanai’s graduate research to this point was to show the detailed transition process of a cold core easterly wave into an incipient typhoon with an organized surface circulation and a relatively warm core (Yanai 1961a). This work had been well received, but Professor Syono commented “your paper has no equations.” Yanai’s good-humored response was “OK, I will write another paper with 100 equations!” Soon he produced a second paper based on inertial instability (Yanai 1961b), extending the work of Kleinschmidt (1951), and received his D.Sc. degree from the University of Tokyo (Fig. 5).
3. A visit to Fort Collins and the National Hurricane Research Project
Shortly after taking up his first position at the MRI’s Typhoon Research Laboratory, Yanai received the 1962 Meteorological Society of Japan Award for his study of typhoon formation, and was invited by Professor Herbert Riehl to visit Colorado State University (CSU) in Fort Collins for two years (Fig. 6) as a visiting assistant meteorologist. Although this period did not result in many publications, his 1964 review paper on tropical cyclone formation (Yanai 1964) served as the most comprehensive review of the field for over a decade. This work was the result of a direct meeting in 1963 with Professor Joanne Simpson. She had recently been appointed as a UCLA professor and was visiting her former advisor, Professor Riehl. She encouraged Riehl’s young postdoc to write a review paper on the formation of tropical cyclones. Without hesitation Yanai accepted her invitation.
Professor Riehl arranged for his postdoc to visit the National Hurricane Research Project (NHRP) in Miami. By an agreement between CSU and NHRP, Michio was stationed in Miami during the 1962 hurricane season. There, he met Ed Zipser, then a student at Florida State University (FSU), who was also visiting NHRP and helped Yanai find an affordable apartment. He met NHRP scientists Craig Gentry, Stan Rosenthal, and others, and took in as many experiences as possible, including a hurricane reconnaissance flight over the Gulf of Mexico. In August 1962, Hurricane Alma nearly hit Miami. Yanai studied this hurricane in depth and subsequently published the work as Yanai and Nitta (1967) and Yanai (1968).
Dr. Yanai would meet many scientists in the United States who would become valued colleagues throughout his career. These included of course his hosts Herb Riehl, Bill Gray (Fig. 7), Elmar Reiter, and Ferdinand Baer while he was at CSU. Yanai also attended meetings in New York (1963) and Mexico City (1963) where he met Professors Jordan and LaSeur (FSU), Krishnamurti [University of California, Los Angeles (UCLA)], Ooyama (New York University), and Miyakoda, Platzman, and Fultz (Chicago). On his way home to Tokyo in 1964, he visited Professors Yale Mintz at UCLA and James Sadler at the University of Hawaii.
4. Tropical wave and cloud cluster studies in Tokyo
In 1965, Yanai returned to the University of Tokyo as an assistant professor at the request of Professor Syono. This was to fill the position vacated by Professor Kikuro Miyakoda, who had moved to the Geophysical Fluid Dynamics Laboratory in Princeton. Yanai’s tenure at the University of Tokyo was very productive. Together with outstanding graduate students Taketo Maruyama, Tsuyoshi Nitta, Yoshikazu Hayashi, Tatsushi Tokioka, Masato Murakami, and Masanori Yamasaki, the group (Fig. 8) published 25 papers.
The discovery of the Yanai–Maruyama wave in the equatorial stratosphere (Yanai and Maruyama 1966) was motivated by Professor Richard Reed’s review paper on the quasi-biennial oscillation (QBO). Reed had suggested relating the time change of the stratospheric zonal wind to the convergence of the eddy transport of momentum. This led Yanai and Maruyama to examine the vertical structure of the wave to examine its potential for vertical momentum transport. Soon the Yanai–Maruyama wave was identified as the mixed Rossby–gravity wave in the theory of equatorially trapped waves (Matsuno 1966; Lindzen and Matsuno 1968).
Professor Yanai’s group extended their earlier work on tropical easterly waves by quantitative analysis of the structure and evolution of Caribbean wave disturbances, and examined the potential for initial growth of easterly waves by barotropic instability processes. They also applied spectral analysis to time series of atmospheric sounding data throughout the tropical Pacific to study the kinetic and potential energy of energetic wave disturbances as a function of their frequency, from the surface to the lower stratosphere. They used increasingly sophisticated methods of cross-spectral analysis to sort out the vertical and horizontal structure of the energetic wave modes and to obtain clues about their excitation mechanisms. As stated in the introduction of Madden and Julian (1971), the studies by the Tokyo group motivated the work that led to the discovery of the Madden–Julian oscillation (MJO).
In 1967, the Global Atmospheric Research Program (GARP) was initiated and Professors Yanai, G. Yamamoto, and Yoshi Ogura were invited to its first international study conference held in Stockholm. Michio was appointed as a member of the Study Group on Tropical Disturbances in 1968, together with Professors Pisharoty (India) and Fujita (University of Chicago). The group met over the period of a month at the University of Wisconsin to conduct a census of cloud systems in the global tropics using geostationary satellite pictures, and to formulate the scientific requirements for the tropical GARP subprogram. The group classified cloud systems into “cloud clusters,” “monsoon clusters,” and “popcorn cumulonimbi.” They recommended the western North Pacific as the site for a tropical GARP field program, in agreement with recommendations of several other groups. In December 1969, Michio attended the GARP Joint Organizing Committee (JOC) planning meeting held in Miami for the First Tropical Experiment.
At this critical moment in Professor Yanai’s research and international activity, student riots broke out at the University of Tokyo and lasted for a year. Radical students influenced by China’s Great Cultural Revolution and Berkeley’s radical student movement occupied many key buildings and soon the teaching and research functions of the university ceased. Offers of “rescue” from his American friends and colleagues began to arrive.
In 1968, Professor Akio Arakawa attended the WMO–International Union of Geodesy and Geophysics (IUGG) Numerical Weather Prediction Symposium at JMA and communicated an offer of a full professorship at UCLA to Professor Yanai that he quickly accepted. Yanai requested a one-year leave of absence from UCLA to take care of his students and finish the remaining work in Tokyo. He departed the University of Tokyo for UCLA in September 1970, with his wife Yoko and their two young sons, Takashi and Satoshi (Fig. 9).
5. Cloud cluster, tropical waves, and monsoon studies in Los Angeles
On his arrival at UCLA in the fall of 1970, Professor Yanai became immersed in the tasks of building a new research group and participating in the teaching activities of the department, then known as the Department of Meteorology. The new courses that he would develop in tropical meteorology, tropical waves, and tropical cyclones were encyclopedic in their scope. And it was time to consider the lines of research that would define his UCLA research group.
a. The Q1 and Q2 studies
Soon after his arrival in Los Angeles, Professor Arakawa persuaded Yanai to begin a new research effort based on the diagnosis of the large-scale heat source and moisture sink, Q1 and Q2. Arakawa was developing the Arakawa–Schubert cumulus parameterization based on the concept that cumulus clouds modify the large-scale temperature and moisture fields through detrainment and cumulus-induced subsidence in their environment. He felt that many other cumulus parameterization schemes were using rather arbitrary assumptions and he encouraged Yanai to develop an analysis scheme that could diagnose the cumulus mass flux and detrainment from observed Q1 and Q2, based on fundamental physical laws. They became aware that the difference between Q1 and Q2 is a measure of eddy vertical transport of total heat, and thus it contains information about cumulus activity.
Yanai started work on the diagnostic scheme immediately. On 18 November 1970, he prepared a table of all upper-air stations in the Marshall Islands that participated in a special observation program in 1956 in support of nuclear tests (Operation Redwing, Joint Task Force Seven). From the data books for all 11 stations, sounding data were punched on cards by an undergraduate student, and then written on magnetic tapes.
Meanwhile, Yanai designed line-integral (horizontal) and finite-difference (vertical) schemes to calculate mass, heat, and moisture budgets, and began to construct a bulk cumulus ensemble model to derive the cumulus mass flux, detrainment, cumulus-induced heating, and moistening. Graduate students Steve Esbensen and Jan-Hwa Chu were interested in what he was doing and joined the project. By September 1971, Professor Yanai was writing up the early results using the Marshall Islands data.
The most important conclusion from this initial study was that the cloud mass flux exceeds the mean vertical mass flux required by the large-scale convergence, thereby causing a compensating sinking motion between the active clouds. The detrainment of large amounts of cloud water acts to cool the environment, counteracting the effect of warming due to adiabatic compression in the sinking environmental air. The detrainment of cloud water and water vapor counteracts the drying due to the environmental sinking motion. The calculation further demonstrated the feasibility of using cumulus ensemble models and observed Q1 and Q2 to quantify the various effects of clouds on the large-scale thermodynamic fields. Some of these conclusions were also reached by Gray (1973), using an independent approach.
The research on the Marshall Island data by Professor Yanai’s group was taking place at the same time as his participation in a flurry of scientific planning and organizational activities that would become the GARP Atlantic Tropical Experiment (GATE). The interim Scientific and Management Group for GATE (ISMG) was established with Joachim Kuettner as the United States member. Yanai became a consultant to the ISMG and a member of the U.S. GATE panel, attending the second meeting of the JOC Study Group in Geneva in January 1971. By request of ISMG, Michio wrote “A review of recent studies of tropical meteorology relevant to the planning of GATE” in September 1971. This manuscript included the early results from the UCLA project’s Q1 and Q2 calculations and was widely distributed within the GATE community (Annex I, Experimental Design Proposal for the GARP Atlantic Tropical Experiment, GATE Report 1, June 1972). In December 1971, early results from the Q1 and Q2 study were presented at the AMS tropical meteorology meeting in Barbados. Dr. Kuettner nominated Professor Yanai to become the Chief Scientist of GATE. He was, however, unable to accept. Foreshadowing health issues that would periodically limit his activities for the remainder of his career, Yanai became ill in 1972 and was hospitalized due to exhaustion from overwork.
A year was needed to polish the Marshall Islands diagnostic work, refining computational schemes as well as the English writing and organizational structure of the paper with the help of Professor Yale Mintz. The Yanai, Esbensen, and Chu paper was finally submitted to the Journal of Atmospheric Sciences in December 1972. Between the editor’s receipt of the manuscript and the time the paper was sent to the technical editor, the authors heard nothing from the editor. Concerned, they inquired about its status, only to learn that the paper was “unconditionally” accepted; that is, no changes were required.
b. Beyond Q1 and Q2
The Yanai et al. (1973) paper inspired Yanai’s group to push the method further by diagnosing detailed properties of cumulus ensembles, not only in large-scale heat and moisture budgets (Yanai et al. 1976), but in large scale momentum, vorticity, and potential vorticity budgets as well. Yanai and his graduate students wrote a series of papers diagnosing cloud properties of the vorticity and momentum fields (Chu et al. 1981; Yanai et al. 1982; Sui and Yanai 1986; Sui et al. 1989; Tung and Yanai 2002a,b) and introduced refinements in the cumulus ensemble model that allowed for diagnosis of the roles of cumulus downdrafts and mesoscale effects (Cheng and Yanai 1989) and vertical wind shear (Wu and Yanai 1994). In addition, there was an ongoing effort to diagnose the vertical structure and spatial distribution of Q1 and Q2 variability over the tropical oceans (Johnson et al. 1987; Tung et al. 1999; Yanai and Tomita 1998).
Professor Yanai viewed opportunities to interact with visiting postdocs and scholars, attending professional meetings and research proposal writing as part of integrated graduate student training (Fig. 10). His former graduate student Tsuyoshi Nitta was invited to make two short but, as it turned out, extremely productive, visits to UCLA, shortly after Nitta had obtained his degree from the University of Tokyo. Professor Yanai suggested that Dr. Nitta think about converting the bulk cloud model in Yanai et al. (1973) to a spectral model that allows the diagnosis of cloud types categorized by the cloud-top heights. Within days, the problem was solved, and Yanai’s group used this formalism for all of its cloud diagnostic work over the following decade. Dr. Nitta also worked closely with Steve Esbensen on the analysis of the structure and variability of the trade wind boundary layer during BOMEX and its cloud ensemble properties (Nitta and Esbensen 1974a,b; Nitta 1975). Dr. Nitta later became a much respected professor at the University of Tokyo. Sadly, Professor Nitta would die at a young age, as would former University of Tokyo students Dr. Masato Murakami (MRI) and Dr. Yoshikazu Hayashi (GFDL).
c. Tropical waves and the Madden–Julian oscillation
Tropical atmospheric waves and the MJO continued to fascinate Professor Yanai after his arrival at UCLA, and became a major focus of his newly formed research group (Fig. 11). His first student to receive a Ph.D. at UCLA was Abraham Zangvil. Together they studied the dynamics of upper tropospheric waves and their association with clouds in the wavenumber–frequency domain (Zangvil 1975; Zangvil and Yanai 1980, 1981), extending the analysis beyond individual station data (e.g., Yanai et al. 1968; Wallace and Kousky 1968) to the entire tropical domain and applying Dr. Hayashi’s new space–time analysis techniques (Hayashi 1977, then at GFDL) to create wavenumber–frequency spectra of the latest 200-mb wind and satellite brightness fields. The same techniques were made even better known by Professor Nitta’s Ph.D. student Yukari Takayabu (Takayabu 1994a,b) in the study of convection-coupled equatorial waves.
A particular emphasis of the UCLA group (Fig. 12) was quantifying the importance of tropical–midlatitude interactions over the central and eastern Pacific. Professor Yanai and graduate students Mong-Ming Lu and Victor Magaña calculated lateral energy fluxes by large-scale disturbances in the upper troposphere and found evidence for selective forcing of equatorially trapped planetary waves in the upper troposphere depending on season and the characteristics of waves propagating energy into the tropics (Yanai and Lu 1983; Magaña and Yanai 1995). Magaña and Yanai (1991) also presented evidence for a summertime modulation of tropical–midlatitude interaction in the mid-Pacific by the MJO. And although analyses over the warm pool region of the western Pacific showed that heating perturbations were responsible for generating available potential energy in disturbances with 30–60-day periods, they found strong horizontal convergence of wave energy flux into the upper tropospheric energy duct over the central-eastern Pacific from extratropical latitudes, suggesting a link between the MJO and the midlatitudes (Yanai et al. 2000; Chen and Yanai 2000).
d. The Tibetan Plateau and Asian monsoon
In the 1980s, a decade after arriving at UCLA, Professor Yanai and his Chinese colleagues began a series of diagnostic studies that would quantify the heating effects of the Tibetan Plateau in the Asian monsoon. One can view this work as a logical extension of Yanai’s computations of Q1 and Q2, the large-scale heat source and moisture sink, but from a phenomenological point of view this was a new line of scientific research.
From the beginning, all of Professor Yanai’s monsoon research activities were conducted as cooperative research projects with Chinese scientists (Fig. 13), and he frequently went to China on official visits together with his American colleagues C.-P. Chang, T.N. Krishnamurti, Takio Murakami, Peter Webster, Pam Stephens, Gabriel Lau, and Bill Lau. Many of Yanai’s papers on this subject were written with invited Chinese scientists (Luo, He, Song, Liu) (Fig. 14) and Professor Guo-Xiong Wu.
The first set of monsoon papers (Luo and Yanai 1983, 1984; He et al. 1987; Yanai and Li 1994) quantified heat and moisture budgets, circulation, and precipitation patterns over the Tibetan Plateau during the period surrounding the 1979 summer monsoon onset, using gridded data from the First GARP Global Experiment. The results show intense heating of the air over the elevated plateau before the arrival of the monsoon rain, with a deep diurnally varying mixed layer extending almost to the tropopause, formed every day by the late afternoon. This suggests that the dry thermal convection originating near the heated surface of the plateau is responsible for atmospheric heating found there during the preonset phase. After the rains begin over the eastern plateau, the Luo and Yanai analysis shows that the heating due to the dry convection in that region is replaced by condensation heating associated with cumulus convection.
A second group of papers addresses the roles of heating over the Tibetan Plateau and land–sea contrast in the evolution and interannual variability of Asian summer monsoon (Yanai et al. 1992; Li and Yanai 1996). Other investigations included the relationship between Indian monsoon rainfall and tropospheric temperature over the Eurasian continent (Liu and Yanai 2001), and between Eurasian spring snow cover and Asian summer rainfall (Liu and Yanai 2002). Professor Yanai, graduate student Chih-Wen Hung (Fig. 15), and their Chinese colleague Xiaodong Liu identified factors contributing to the Australian summer monsoon and examined its symmetries and asymmetries with the Asian summer monsoon (Hung and Yanai 2004; Hung et al. 2004). The effects of the Tibetan Plateau on the Asian monsoon were summarized in a review article by Yanai and Wu (2006).
6. Postretirement activities
Professor Yanai’s seminal research contributions were recognized with two major awards, the American Meteorological Society’s (AMS’s) Jule Charney award in 1986 and the Fujiwara Award from the Meteorological Society of Japan in 1993. A few years later, in 1999, Professor Yanai formally retired from UCLA (Fig. 16) but remained an active and valued member of his department and the global tropical meteorological community. He was particularly fond of communicating by e-mail with his many friends and colleagues not only on scientific matters but also on the arts, especially opera (Fig. 17).
A great example of Yanai’s postretirement service to the tropical research community was his UCLA Tropical Meteorology and Climate Newsletter, an e-mail digest of announcements, paper abstracts, and short articles that he started in 1996. Originally named the UCLA Tropical Meteorology Newsletter, it was created as a means of communication among the past, present, and future members of his research group. However, the topics covered by the newsletter and the number of contributors and recipients expanded very rapidly, becoming a valuable forum for more than 800 scientists worldwide by the time of his passing. The last newsletter was issued on 8 October 2010.
A symposium in Professor Yanai’s honor was held at the American Meteorological Society’s annual meeting in January 2011 (Fig. 18). The occasion of the symposium led him to reminisce not only about his life and career but also about the histories and contributions of his colleagues, especially his fellow meteorologists who emerged from postwar Japan. Indeed, much of this epilogue is based on notes he wrote in the last year of his life, allowing us to tell and preserve his story. Professor Yanai was in the midst of collecting oral histories of the UCLA department from past and present members of the UCLA family when he passed away on 13 October 2010.
Those who had the pleasure to know him as a colleague, teacher, and friend will remember Professor Yanai’s passion and good humor, and his concern for all those around him. He was not only a superb scientist but also a renaissance man with an infectious interest in politics and the arts. This volume is dedicated to Professor Yanai and to the better understanding of tropical multiscale cloud systems that motivated so much of his research.
This epilogue is based on the narrative in the program booklet for the AMS Yanai Symposium held in Seattle, Washington, on 27 January 2011. Many friends offered their help to enhance the narrative, including C. P. Chang (Fig. 13), Baode Chen (Fig. 12), Bill Gray (Fig. 7), Chih-Wen Hung, Mong-Ming Lu (Figs. 10, 14, and 15), and Taroh Matsuno (Figs. 4 and 8). Most importantly, we thank the Yanai family, Yoko, Takashi, and Satoshi, who shared the most wonderful pictures of Michio (Figs. 1–6, 8, 9, 11, 12, 16, and 17) and helped get the factual details straight. Finally, we thank the amazing and supportive AMS organizers and participants at the symposium (Fig. 18), who completed the narrative.