Annual Report on the Climate System 2016

Transcription

1 Annual Report on the Climate System 2016 March 2017 Japan Meteorological Agency

2 Preface The Japan Meteorological Agency is pleased to publish the Annual Report on the Climate System The report summarizes 2016 climatic characteristics and climate system conditions worldwide, with coverage of specific events including the effects of the summer 2014 spring 2016 El Niño event and notable aspects of Japan s climate in summer I am confident that the report will contribute to the understanding of recent climatic conditions and enhance awareness of various aspects of the climate system, including the causes of extreme climate events. Teruko Manabe Director, Climate Prediction Division Global Environment and Marine Department Japan Meteorological Agency

3 Contents Preface 1. Explanatory notes Outline of the Annual Report on the Climate System Climate in Japan Climate around the world Atmospheric circulation Oceano graphic co n ditions Sno w co ver and sea ice 5 2. Annual summaries o f the climate syste m Climate in Jap an Climate aro und the world Extratropical c ircula tio n Tr o p i c a l c ir c u l a t i o n a n d c o n v e c t i v e a c t i v i t y Oceano graphic co nd itio ns Stratospheric circulation in boreal winter Summary of the Asian summer monsoon Arctic sea ice conditions Snow cover in the Northern Hemisphere Analysis of specific events T he El Niño event end ing b oreal sp ring and its effects Extreme climate conditions in Japan in August

4 1. Explanatory notes 1.1 Outline of the Annual Report on the Climate System The Japan Meteorological Agency (JMA) has published the Annual Report on the Climate System on the Tokyo Climate Center website 1 since The aim of such provision is to share information on the climate system and recent related conditions with national meteorological services, research institutes, universities and other interested parties. This report summarizes 2016 climatic characteristics and climate system conditions worldwide, with coverage of specific events including the effects of the summer 2014 spring 2016 El Niño event and notable aspects of Japan s climate in summer For more detailed climate information, see the various products provided via the Tokyo Climate Center/JMA website at The following sections describe the data sources and analysis methods used in the compilation of this report. Climatological normals are averages for the period from 1981 to Unless otherwise noted, anomalies are deviations from normals. 1.2 Climate in Japan The descriptions in this section mainly relate to Section Average temperature over Japan Annual anomalies of the average surface temperature over Japan since 1898 are illustrated in Section The anomalies shown are calculated from temperatures recorded at 15 meteorological observatories (Abashiri, Nemuro, Suttsu, Yamagata, Ishinomaki, Fushiki, Iida, Choshi, Sakai, Hamada, Hikone, Miyazaki, Tadotsu, Naze and Ishigakijima) selected from among those deemed to be least influenced by the urban heat island phenomenon. Average temperatures over Japan are first derived based on temperature deviations from the average of the 15 stations, and are then adjusted to the baseline. The observatories at Miyazaki and Iida were relocated in May 2000 and May 2002, respectively. For these stations, any discontinuity in the temperature time series is adjusted to cancel out the influence of the moves Climatological normal and rank The seasonal characteristics of Japan s climate are summarized in Section 2.1, which reports temperature anomalies, precipitation ratios and sunshine duration ratios derived from daily observations made at 154 surface meteorological stations. Regional averages are calculated for the four divisions of northern Japan, eastern Japan, western Japan and Okinawa/Amami as well as for the eleven subdivisions of Hokkaido, Tohoku, Kanto-koshin, Hokuriku, Tokai, Kinki, Chugoku, Shikoku, northern part of Kyushu, southern part of Kyushu and Okinawa. For precipitation ratios and sunshine duration ratios, the divisions of northern, eastern and western Japan are further divided into the Pacific side and the Sea of Japan side (Fig ). Tables on regional climate conditions contain regional averages and rankings of temperature anomalies, precipitation ratios and sunshine duration ratios. The ranking categories are below normal, near normal and above normal, each of which has an equal relative frequency of occurrence (33%) for the period from 1981 to The bottom and top 10% of the below normal and above normal categories are defined as significantly below normal and significantly above normal, respectively

5 Fig Operational climatological regions and station locations 1.3 Climate around the world The descriptions in this section mainly relate to Section 2.2. The regions used in this report are defined as shown in Fig Global average temperature Annual anomalies of the global average surface temperature since 1891 are illustrated in Section The anomalies shown are derived from a combined dataset of near-surface air temperatures over land and sea surface temperatures (SSTs). The over-land air temperatures are based on on-site observation data derived from monthly CLIMAT reports for the period from 2001 onward, and from Global Historical Climate Network (GHCN) datasets produced by the National Oceanographic and Atmospheric Administration (NOAA) for the period before The SSTs are 1 1 grid values derived from COBE-SST datasets (JMA 2006), and values in areas partly covered by sea ice are excluded. In the calculation of global averages, land surface temperature anomalies and SST anomalies against the baseline are incorporated into 5 5 grid values, which are weighed in proportion to the area of the relevant grids, and the grid values are averaged over the globe. The global averages are adjusted to the baseline. The annual values are accompanied by 90% confidence intervals based on estimated errors attributable to the inhomogeneity of data availability (Ishihara 2007). 2

6 1.3.2 Data and climatological normals Figures on world climatic conditions are based on CLIMAT reports. Historical datasets are derived from GHCN datasets and CLIMAT reports (from June 1982 onward, prior to GHCN datasets). Data and information on disasters are based on official reports from the United Nations and national governments, and from databases of research institutes (Emergency Events Database (EM-DAT)). precipitation are calculated by dividing the total number of extreme events observed at stations by the total number of available observation data for each 5 5 grid box. Frequencies are represented by semicircles. For grid boxes where fewer than ten observations are available, no semicircle is shown. Since the frequency of extreme events is expected to be about 3% on average, occurrence is considered to be above normal when the figure is 10% or more Extreme climate events JMA defines an extreme climate event as a phenomenon likely to happen only once every 30 years or longer. For monthly/seasonal mean temperatures, extremely high (or low) temperatures are deemed to be those with an anomaly greater than 1.83 times the standard deviation based on the period For monthly/seasonal precipitation totals, extremely heavy (or light) precipitation is that above (or below) any value observed during the period Atmospheric circulation The descriptions in this section mainly relate to Sections 2.3, 2.4, 2.6, 2.7 and 2.8 and Chapter Data and climatological normals Atmospheric circulation data are based on the results of six-hourly global objective analysis conducted at 00, 06, 12 and 18 UTC using data from the Japanese 55-year reanalysis (JRA-55; Kobayashi et al. 2015). The normal is the average of JRA-55 data (JMA 2011) Annual figures For annual mean temperature anomalies shown in Section 2.2.2, categories are defined by the annual mean temperature anomaly against the normal divided by its standard deviation and averaged in 5 5 grid boxes. For annual total precipitation, categories are defined by the ratio of annual precipitation to the normal averaged in 5 5 grid boxes. For frequencies of extreme events based on monthly observations for the year, the ratios of extremely high/low temperature and heavy/light Atmospheric circulation and convection Wave activity flux (Takaya and Nakamura 2001) indicates the propagation of Rossby wave packets. Tropical convective activity is inferred from outgoing longwave radiation (OLR). It can be assumed that lower values of OLR indicate enhanced convective activity except in the mid- and high latitudes during the winter season and in high-altitude areas. The original OLR data are from observations conducted by NOAA s polar-orbiting satellites. 3

7 Fig Names of world regions According to Helmholtz's theorem in vector analysis, horizontal wind can be decomposed into rotational and divergent components. Using the stream function, the rotational component can be written as (ψ: stream function; u ψ, v ψ : rotational components of horizontal winds). A positive stream function anomaly indicates stronger (weaker) clockwise (counter-clockwise) circulation than the normal, and vice versa. Using the velocity potential, the divergent component can be written as (χ: velocity potential; u χ, v χ : divergent components of horizontal winds). A positive velocity potential anomaly indicates stronger (weaker) large-scale convergence (divergence) than the normal, and vice versa. The equatorial intra-seasonal variation associated with the Madden-Julian Oscillation (MJO) can be determined from the time-longitude cross section of the five-day mean 200-hPa velocity potential Atmospheric and oceanic monitoring indices for the tropics Section 2.4 describes the characteristics of atmospheric and oceanographic monitoring indices to support analysis of variations related to El Niño-Southern Oscillation (ENSO). The Southern Oscillation Index (SOI) is defined as the normalized difference in monthly mean sea-level pressure (SLP) anomalies normalized by their standard deviations between Tahiti and Darwin. SLP anomalies are calculated based on CLIMAT reports. OLR indices are defined as reversed-sign area-averaged OLR anomalies normalized by their standard deviations (Table 2.4.1). It should be noted that positive and negative OLR index values indicate enhanced and suppressed convective activity, respectively, compared to the normal. Equatorial zonal wind indices are defined as area-averaged zonal wind anomalies normalized by their standard deviations. Asian summer monsoon OLR indices (SAMOI) are derived from OLR anomalies from May to October. SAMOI (A), (N) and (W) indicate the overall activity of the Asian summer monsoon, its northward shift and its westward shift, respectively. SAMOI definitions are as 4

8 follows: SAMOI (A) = ( 1) (W + E) SAMOI (N) = S N SAMOI (W) = E W W, E, N and S indicate area-averaged OLR anomalies for the respective regions shown in Fig normalized by their standard deviations (JMA 1997). 1.6 Snow cover and sea ice The descriptions in this section relate to Sections 2.8 and 2.9. The number of days of snow cover is determined based on observations by the Special Sensor Microwave/Imager (SSM/I) and the Special Sensor Microwave Imager Sounder (SSMIS) on board the Defense Meteorological Satellite Program (DMSP) polar-orbiting satellites. The algorithm for the analysis was developed by JMA. The sea ice extent is based on observations conducted by SSM/I and SSMIS and by the Scanning Multichannel Microwave Radiometers (SMMRs) on board the Nimbus satellites. Fig Asian Summer Monsoon OLR Index (SAMOI) areas 1.5 Oceanographic conditions The descriptions in this section relate to Section 2.5. Sea surface temperatures (SSTs) are based on COBE-SST datasets in 1 1 grid boxes. The details of SST analysis and the SST normal have been described previously by JMA (2006). Ocean heat content is determined using five-day mean datasets of MOVE/MRI.COM-G2 (Multivariate Ocean Variational Estimation / Meteorological Research Institute Community Ocean Model - Global version 2; Toyoda et al. 2013), which was developed at the Oceanography and Geochemistry Research Department of the Meteorological Research Institute (MRI). References Ishihara, K., 2007: Estimation of standard errors in global average surface temperatures (in Japanese), Weather Service Bulletin, Vol. 74, JMA, 1997: Monthly Report on Climate System, June JMA, 2006: Characteristics of Global Sea Surface Temperature Data (COBE-SST), Monthly Report on Climate System, Separated Volume No. 12. JMA, 2011: JMA s New Climatological Normals for , Data Report on Climate System Kobayashi, S., Y. Ota, Y. Harada, A. Ebita, M. Moriya, H. Onoda, K. Onogi, H. Kamahori, C. Kobayashi, H. Endo, K. Miyaoka, and K. Takahashi, 2015: The JRA-55 Reanalysis: General Specifications and Basic Characteristics. J. Meteorol. Soc. Japan, 93, Takaya, K. and H. Nakamura, 2001: A formulation of a phase-independent wave-activity flux for stationary and migratory quasigeostrophic eddies on a zonally varying basic flow. J. Atom. Sci., 58, Toyoda, T., Y. Fujii, T. Yasuda, N. Usui, T. Iwao, T. Kuragano and M. Kamachi, 2013: Improved Analysis of Seasonal-Interannual Fields Using a Global Ocean Data Assimilation System, Theoretical and Applied Mechanics Japan, 61,

9 2. Annual summaries of the 2016 climate system 2.1 Climate in Japan - As temperatures were generally above normal all over Japan, annual mean temperatures were significantly above normal almost nationwide. - Four typhoons made landfall on northern Japan in August, bringing record heavy rainfall with storms in northern Japan. - In autumn, seasonal precipitation amounts were significantly above normal and seasonal sunshine durations were significantly below normal in western Japan Average surface temperature The annual anomaly of the average surface temperature over Japan (i.e., that averaged over 15 observatories confirmed as being relatively unaffected by urbanization) for 2016 was C above the average, making it the highest since On a longer time scale, annual mean temperatures over Japan have risen at a rate of about C per century since 1898 (Fig ). The record-high temperature recorded in 2016 was due to year-round warm climatic conditions associated with seasonal atmospheric circulation as described in (a) to (d) of Section Long-term climate change and tropospheric air temperature increase on a global scale due to the El Niño event that peaked strongly during boreal winter 2015/2016 are identified as background factors Annual characteristics (Table 2.1-1, Fig , Fig ) In 2016, temperatures were generally above normal all over Japan except for autumn in northern Japan. Annual mean temperatures were significantly above normal almost nationwide. In particular, the temperature for eastern Japan tied with 2004 as the highest since 1946 (+1.0 C above the normal). Temperature records were broken at 16 of 154 observatories in Japan (Table 2.1-2). Annual precipitation amounts were significantly above normal on the Pacific side of northern Japan, in western Japan and in Okinawa/Amami. This was attributable to the significant influence of low-pressure systems and fronts in western Japan and Okinawa/Amami in winter 2015/16 and autumn, and to the numerous typhoons that approached northern Japan in August. Muroto-misaki (Kochi Prefecture) recorded its highest rainfall amounts since In eastern Japan, annual precipitation amounts were near normal. Annual sunshine durations were below normal in western Japan and above normal in northern Japan and on the Sea of Japan side of eastern Japan due to high-pressure systems that tended to cover northern Japan in spring. Durations were near normal on the Pacific side of eastern Japan and in Okinawa/Amami Seasonal characteristics (a) Winter (December 2015 February 2016, Fig (a)) In association with a weak winter monsoon, seasonal temperatures were above normal all over Japan, especially in eastern and western parts. Seasonal snowfall amounts for the Sea of Japan side were generally below normal and significantly above normal in northern Kyushu due to considerably cold-air outbreaks at the end of January. Due to the significant influences of low-pressure systems and fronts, seasonal precipitation amounts were above normal all over Japan, with Okinawa/Amami experiencing record highs (188% of the normal) for winter since 1946/47. (b) Spring (March May, Fig (b)) Seasonal mean temperatures were significantly above normal due to warm southerly winds associated with dominant high-pressure systems to the east of Japan and the development of the subtropical high to the south of Japan. 6

10 Seasonal sunshine durations were significantly above normal on the Sea of Japan side of eastern Japan and above normal in northern and western Japan due to significant influences from high-pressure systems. Seasonal precipitation amounts were significantly below normal on the Sea of Japan side of eastern Japan and below normal on the Pacific side of northern Japan. Meanwhile, seasonal precipitation amounts were above normal on the Pacific side of western Japan and Okinawa/Amami due to the influences of low-pressure systems and fronts in April. (c) Summer (June August, Fig (c)) Seasonal mean temperatures and sunshine durations were above normal all over Japan. In Okinawa/Amami, the seasonal mean temperature was the highest on record for summer since 1946 (+1.1 C above the normal) in association with strong solar radiation accompanying high sunshine durations. Meanwhile, in northern Japan, seasonal precipitation amounts were significantly above normal. On the Pacific side of northern Japan, the figure was 163% of the normal (the highest on record for summer since 1946) in association with the frequent passage of cyclones around northern Japan in June and the approach of numerous typhoons around northern Japan in August. Four typhoons made landfall on the Hokkaido region and Iwate Prefecture, bringing significant rainfall with storms. Hokkaido and Iwate Prefecture experienced record heavy rainfall, which caused serious damage including river overflows and landslides. Seasonal precipitation amounts were also above normal on the Pacific side of western Japan in association with the active Baiu front, which brought frequent heavy rain to the Pacific side of western Japan (Table 2.1-3). At the end of June, Kyushu experienced intermittent heavy rain with flood damage and landslides. (d) Autumn (September November, Fig (d)) Western Japan and Okinawa/Amami experienced record seasonal mean temperatures (+1.3 and +1.2 C above the normal, respectively) for autumn since 1946 due to warm southerly winds. In association with significant influences from low-pressure systems and fronts, western Japan experienced significantly above-normal seasonal precipitation amounts and below-normal seasonal sunshine durations. Seasonal precipitation was 173% of the normal on the Sea of Japan side of western Japan (the highest on record for autumn since 1946), while seasonal sunshine durations were 74% of the normal on the Sea of Japan side of western Japan and 82% of the normal on the Pacific side of the region (both the lowest on record for autumn since 1946). Seasonal sunshine durations were also below normal in other regions. In northern Japan, seasonal mean temperatures were below normal for the first time since 2002, even with the high temperatures recorded in September, and temperatures remained low after October. 7

11 Fig Long-term change in the annual anomaly of average surface temperature over Japan Anomalies are deviations from the baseline (i.e., the average). The black line indicates the annual anomalies of the average surface temperature for each year. The blue line indicates the five-year running mean, and the red line indicates the long-term linear trend. Table Regional average and rank of annual mean temperature anomaly, annual precipitation ratio, and annual sunshine duration ratio for divisions and subdivisions (2016) 8

12 Fig Five-day running mean temperature anomaly for divisions (January December 2016) Table Number of observatories reporting record monthly and annual mean temperatures, precipitation amounts and sunshine durations (2016) From 154 surface meteorological stations across Japan. Temperature Precipitation amount Sunshine duration Highest Lowest Heaviest Lightest Longest Shortest January 5 3 February March 3 8 April 2 2 May 25 June 1 6 July 1 1 August September October November 1 1 December Year

13 Fig Annual climate anomaly/ratio for Japan in 2016 Table Onset/end of the Baiu (Japan s rainy season) for individual subdivisions (2016) Subdivisions Onset of rainy season* Average date of onset of rainy season ( ) End of rainy season* Average date of end of rainy season ( ) Area averaged precipitation ratio during rainy season (%) Okinawa 16 May 9 May 16 June 23 June 84 Amami 16 May 11 May 18 July 29 June 103 Southern Kyushu 24 May 31 May 18 July 14 July 147 Northern Kyushu 4 June 5 June 18 July 19 July 117 Shikoku 4 June 5 June 18 July 18 July 132 Chugoku 4 June 7 June 18 July 21 July 106 Kinki 4 June 7 June 18 July 21 July 103 Tokai 4 June 8 June 28 July 21 July 89 Kanto- Koushin 5 June 8 June 29 July 21 July 74 Hokuriku 13 June 12 June 19 July 24 July 91 Southern Tohoku 13 June 12 June 29 July 25 July 70 Northern Tohoku 13 June 14 June 29 July 28 July 91 * The onset/end of the rainy season normally has a transitional period of about five days. The dates shown in the table denote the middle day of this period. 10

14 (a) Winter (b) Spring (c) Summer (d) Autumn Fig Seasonal anomalies/ratios for Japan in 2016 (a) Winter (December 2015 to February 2016), (b) spring (March to May), (c) summer (June to August), (d) autumn (September to November). 11

15 2.2 Climate around the world Global average surface temperature The annual anomaly of the global average surface temperature for 2016 was ± 0.13 C above the average. This was the warmest year since records began in 1891, surpassing the previous record of 2015 (+0.42 C). On a longer time scale, global average surface temperatures have risen at a rate of about C per century since 1891 (Fig ). In 2016, monthly average air temperatures for January, February, March, April, June and July, and seasonal average air temperatures for boreal winter, spring and summer, were also the highest on record since High temperature deviations were seen over wide areas of Eurasia, North America, the Tropical Pacific and the Indian Ocean (Fig ). The high temperatures observed in recent years are thought to be associated with a global warming trend caused by increased atmospheric concentrations of carbon dioxide and other anthropogenic greenhouse gases. The global temperature is also affected by inter-annual to decadal-scale natural fluctuations intrinsic to the earth s climate. The record-high temperatures of 2016 are partially attributed to the El Niño event that continued until boreal spring 2016 and to global warming. Fig Long-term change in the annual anomaly of global average surface temperature Anomalies are deviations from the baseline (i.e., the average). The black dots indicate annual anomalies of the global average surface temperature for each year. The error bars indicate 90% confidence intervals. The blue line indicates the five-year running mean, and the red line indicates the long-term linear trend. 12

16 Fig Annual mean temperature anomalies in 2016 The circles indicate anomalies of surface temperature averaged in 5 x 5 grid boxes. Anomalies are deviations from the average Regional climate Annual mean temperatures were above normal in many parts of the world and below normal in the southwestern part of Eastern Siberia and in northern Argentina (Fig ). In particular, extremely high temperatures continued for most of the year in various places at low latitudes, and were also frequently observed from the northern part of Central Siberia to the Svalbard Islands, from the eastern part of Eastern Siberia to the western coast of Canada, and from northern to southeastern Australia (Fig ). Annual precipitation amounts were above normal in eastern China, Mongolia, Central Asia, southeastern Europe, Indonesia and southern Argentina, and were below normal in eastern Brazil and southern Chile (Fig ). Extremely high amounts were frequently observed in southeastern Europe, from the Midwest to the southern USA and in southeastern Australia, while extremely low amounts were frequently observed from southwestern France to northeastern Spain and in eastern Brazil (Fig ). Seasonal distribution maps for temperature and precipitation are shown in Figs and 2.2-8, respectively. Major extreme climatic events and weather-related disasters occurring in 2016 are shown in Fig , and related overviews are given below. (1) Heavy rain: the northeastern part of the Korean Peninsula (August September) (2) Cold: in and around eastern Mongolia (January, October November) (3) Heavy rain: China (April July) (4) Warm: the southern Kyushu region of Japan, to southeastern China (April June, October, December) (5) Warm: Southeast Asia (January May, July November) (6) Drought: Southeast Asia (January May) (7) Tropical storm: Sri Lanka, northeastern India and Bangladesh (May) 13

17 (8) Warm: southern India to Sri Lanka (January April, July August, October, December) (9) Heat wave (March May) and wet (July October): India (10) Heavy rain: Pakistan (July August) (11) Heavy rain: northern Pakistan to Afghanistan (March April) (12) Warm: the northern part of Central Siberia to the Svalbard Islands (February, April July, September) (13) Wet: southeastern Europe (February March, May June, October) (14) Dry: southwestern France to northeastern Spain (July August, October, December) (15) Warm: in and around northern Algeria (January February, October) (16) Warm: northeastern Saudi Arabia to the southern coast of the Red Sea (March, May July) (17) Warm: in the western part of Western Africa to the northwestern part of Central Africa (April June, August December) (18) Warm: Seychelles to the northeastern part of South Africa (January April, October) (19) Warm: the eastern part of Eastern Siberia to the western coast of Canada (April August, October) (20) Wet: the Midwest to the southern USA (March April, July August) (21) Warm: the eastern to the southern USA (March, June October) (22) Warm: the southwestern USA to northwestern Mexico (February March, October December) (23) Hurricane: Haiti and the southeastern USA (October) (24) Warm: southern Mexico to Colombia (January August, October) (25) Warm (February August) and Dry (February May): eastern Brazil (26) Warm: in and around central Chile (January February, August September, November) (27) Warm: Micronesia (March April, June, August) (28) Warm: northern to southeastern Australia (March July, September, November) (29) Wet: southeastern Australia (January, June, September) (30) Warm: in and around New Zealand (February, May, September) 14

18 Fig Annual mean temperature anomalies for 2016 Categories are defined by the annual mean temperature anomaly against the normal divided by its standard deviation and averaged in 5 5 grid boxes. The thresholds of each category are -1.28, -0.44, 0, and The normal values and standard deviations were calculated from statistics. Land areas without graphics represent regions for which observation data sample is insufficient or normal data are unavailable. Fig Frequencies of extreme high/low temperature for 2016 shown as upper/lower red/blue semicircles The size of each semicircle represents the ratio of extremely high/low temperature based on monthly observation for the year in each 5 5 grid box. As the frequency of extreme high/low temperature is expected to be about 3% on average, occurrence is considered to be above normal for values of 10 20% or more. 15

19 Fig Annual total precipitation amount ratios for 2016 Categories are defined by the annual precipitation ratio to the normal averaged in 5 5 grid boxes. The thresholds of each category are 70, 100 and 120%. Land areas without graphics represent regions for which observation data sample is insufficient or normal data are unavailable. Fig Frequencies of extreme heavy/light precipitation amounts for 2016 As same as Fig , but for monthly values of extremely heavy/light precipitation. 16

20 (a) Winter (December February) (b) Spring (March May) (c) Summer (June August) (d) Autumn (September November) Fig Seasonal mean temperature anomalies for (a) winter (December 2015 February 2016), (b) spring (March May), (c) summer (June August) and (d) autumn (September November) As same as Fig , but for seasonal mean temperature anomaly. (a) Winter (December February) (b) Spring (March May) (c) Summer (June August) (d) Autumn (September November) Fig Seasonal total precipitation amount ratios for (a) winter (December 2015 February 2016, (b) spring (March May), (c) summer (June August) and (d) autumn (September November) As same as Fig , but for seasonal total precipitation amount ratios. 17

21 Fig Extreme events and weather-related disasters observed in 2016 Schematic representation of major extreme climatic events and weather-related disasters occurring during the year. 18

22 2.3 Extratropical circulation In 2016, the warm conditions observed over wide areas of the Northern Hemisphere extra-tropics are presumed to have been associated with the El Niño event that continued until spring. This section outlines the seasonal mean characteristics of atmospheric circulation observed in these areas Zonal mean temperature anomaly calculated from thickness and zonal wind in the troposphere Tropospheric zonal mean temperature anomalies calculated from thickness are shown in Fig Temperature anomalies reached the peak of the warm conditions with five-month running means near +1 K. Temperature anomalies decreased during the second half of the year, but remained above normal. In the zonal mean zonal wind of the Northern Hemisphere (Fig , top), the westerly jet stream was generally stronger than normal until October and shifted southward in the second half of November. The stream over Japan shifted northward in April, October and December (Fig , bottom). Fig Time-series representation of zonal mean temperature anomalies calculated from thickness in the troposphere (2006 to 2016) The top (bottom) panel shows the temperature anomalies in the global mean (the Northern Hemisphere; 90 o N 30 o N), respectively. The thin and thick lines show monthly and five-month running mean values, respectively (unit: K). Fig Time-latitude cross section of five-day running mean 200-hPa zonal wind (December 2015 December 2016) The top panel shows zonal mean zonal wind, and the bottom panel shows zonal wind averaged over 120 o 150 o E. The black lines and shading show zonal wind at intervals of (top) 10 and (bottom) 15 m/s, respectively. The green lines indicate the normal at intervals of (top) 20 and (bottom) 30 m/s, respectively. 19

23 2.3.2 Winter (December 2015 February 2016) In the 500-hPa height field (Fig ), positive anomalies were seen over wide areas of the Northern Hemisphere, particularly in Western and Central Siberia. The polar vortex was weaker than normal. Negative anomalies were seen from Eastern Siberia to the seas south of Alaska and to the west of the UK. Wave trains were dominant over northern Eurasia with clear positive anomalies over Western and Central Siberia in January (Fig ). In the sea level pressure field (Fig ), the Icelandic Low and the Aleutian Low were stronger than normal over the eastern part of their normal extents. Positive anomalies were seen over eastern Eurasia, particularly in January (Fig ). In the lower troposphere, temperatures were above normal over wide areas of the Northern Hemisphere, particularly in the polar region, Alaska and Western and Central Siberia, and below normal over the southeastern part of Eastern Siberia (Fig ). In the upper troposphere, the jet stream meandered southward over and around China and northward to the east of Japan (Fig ). Fig Three-month mean 500-hPa height and anomaly (December 2015 February 2016) The contours show 500-hPa height at intervals of 60 m. The shading indicates its anomalies. Fig Three-month mean sea surface pressure and anomaly (December 2015 February 2016) The contours show sea level pressure at intervals of 4 hpa. The shading indicates its anomalies. Fig Three-month mean 850-hPa temperature and anomaly (December 2015 February 2016) The contours show 850-hPa temperature at intervals of 4 o C. The shading indicates its anomalies. The dot patterns indicate areas with altitudes exceeding 1,600 m. Fig Three-month mean 200-hPa wind speed and vectors (December 2015 February 2016) The black and brown lines show wind speed and its normal at intervals of 20 m/s and 40 m/s, respectively. Fig Monthly mean 500-hPa height and anomaly (January 2016) As per Fig , but for monthly mean. Fig Monthly mean sea level pressure and anomaly (January 2016) As per Fig , but for monthly mean. 20

24 2.3.3 Spring (March May 2016) In the 500-hPa height field (Fig ), positive anomalies were seen over wide areas of the Northern Hemisphere, particularly over high-latitudes, the northwestern part of North America, the eastern part of Northern Africa and Japan. Negative anomalies were seen to the southwest of Alaska and in northeastern Canada. In the sea level pressure field (Fig ), positive anomalies were seen over the Beaufort Sea, and negative anomalies were seen over wide areas of Eurasia. The Aleutian Low and the Pacific High were stronger than normal over the eastern and western parts of their normal extents. In the lower troposphere, temperatures were above normal over wide areas of the Northern Hemisphere, particularly over the northwestern part of North America, the eastern part of North Africa, Central Asia and from Western Russia to Western Siberia (Fig ). In the upper troposphere, the jet stream shifted southward from its normal extent over and around China and northward from Japan to the sea east of Japan (Fig ). Fig Three-month mean 500-hPa height and anomaly (March May 2016) As per Fig , but for March May Fig Three-month mean sea level pressure and anomaly (March May 2016) As per Fig , but for March May Fig Three-month mean 850-hPa temperature and anomaly (March May 2016) As per Fig , but for March May 2016 and with contour intervals of 3 o C. Fig Three-month mean 200-hPa wind speed and vectors (March May 2016) The black and brown lines show wind speed and its normal at intervals of 10 m/s and 20 m/s, respectively. 21

25 2.3.4 Summer (June August 2016) In the 500-hPa height field (Fig ), positive anomalies were seen over wide areas of the Northern Hemisphere except in the Arctic region, particularly from the Kamchatka Peninsula to the northeastern part of the North Pacific, from eastern Canada to Greenland and in Western Siberia. Clear positive anomalies around the Kamchatka Peninsula and negative anomalies to the southeast of Japan were seen in August (Fig ). In the sea level pressure field (Fig ), positive anomalies were seen over Eurasia, the mid-latitudes of the North Pacific and from northern Canada to Greenland. The westward extension of the Pacific High was weaker than normal. In the lower troposphere, temperatures were above normal over wide areas of the Northern Hemisphere, particularly from Eastern Siberia to the Aleutian Islands, in the northern part of North America and from Western Russia to Western Siberia (Fig ). In the upper troposphere, the jet stream shifted northward from its normal position over eastern Eurasia in association with the eastward extension of the Tibetan High (Fig ). Fig Three-month mean 500-hPa height and anomaly (June August 2016) As per Fig , but for June August Fig Three-month mean sea level pressure and anomaly (June August 2016) As per Fig , but for June August Fig Three-month mean 850-hPa temperature and anomaly (June August 2016) As per Fig , but for June August Fig Three-month mean 200-hPa wind speed and vectors (June August 2016) As per Fig , but for June August Fig Monthly mean 500-hPa height and anomaly (August 2016) As per Fig , but for August

26 2.3.5 Autumn (September November 2016) In the 500-hPa height field (Fig ), clear positive anomalies were seen over the eastern hemisphere side of the polar region and the eastern part of North America, and negative anomalies were seen over the mid-latitudes from Central Asia to the western North Pacific and to the west of North America. Clear negative anomalies in the mid-latitudes were seen in October (Fig ). In the sea level pressure field (Fig ), positive and negative anomalies were seen over the high latitudes from Europe to Central Siberia and China, respectively. Over the North Pacific, positive and negative anomalies were seen to the south and north of the latitude bands of 40 o N, respectively. In the lower troposphere, temperatures were above normal over the polar region and central and eastern parts of North America, and were below normal over the mid-latitudes from eastern Europe to the western North Pacific (Fig ). In the upper troposphere, the jet stream was stronger than normal over the latitude bands of 40 o N from Eurasia to the North Pacific (Fig ). Fig Three-month mean 500-hPa height and anomaly (September November 2016) As per Fig , but for September November Fig Three-month mean sea level pressure and anomaly (September November 2016) As per Fig , but for September November Fig Three-month mean 850-hPa temperature and anomaly (September November 2016) As per Fig , but for September November Fig Three-month mean 200-hPa wind speed and vectors (September November 2016) The black and brown lines show wind speed and its normal at intervals of 15 m/s and 30 m/s, respectively. Fig Monthly mean 500-hPa height and anomaly (October 2016) As per Fig , but for October

27 2.4 Tropical circulation and convective activity Tropical circulation was presumed to be influenced by the El Niño event that ended in spring This section outlines the seasonal mean tropical circulation and convective activity observed in 2016 with focus on links to El Niño Tropical indices Monthly indices related to tropical circulation are shown in Table and Fig (see Section for related definitions). The SOI was negative (indicating weaker-than-normal trade winds) until April and remained positive after May except in October and November. In the first half of 2016, OLR-PH (for the area around the Philippines) and OLR-MC (for the area around Indonesia) generally remained negative, indicating suppressed convection, and OLR-DL (for the area near the dateline) remained positive, indicating enhanced convection. In terms of equatorial zonal wind indices for the first half of the year, U200-CP (for the central Pacific in the upper troposphere) was generally negative, indicating easterly wind anomalies, and U200-IN (for the Indian Ocean in the upper troposphere), U850-WP, CP and EP (for the Pacific in the lower troposphere) were generally positive, indicating westerly wind anomalies. The variability of these indices indicates the impact of the El Niño event on tropical circulation in the first half of The active phase of equatorial intra-seasonal oscillation intermittently propagated eastward throughout the year. Clear eastward propagation events occurred during the period from December 2015 to August 2016 (Fig (a)). Enhanced convective activity on a seasonal timescale was observed over the central and eastern Pacific until April, and shifted toward the area from the Indian Ocean to the Maritime Continent in May (Fig (a)). In association with this convection, westerly wind anomalies also shifted from the central Pacific to the eastern Indian Ocean (Fig (b)). Table Tropical atmospheric and oceanographic indices (December 2015 December 2016) 24

28 Fig Time-series representation of tropical atmospheric and oceanographic indices from 2006 to 2016 Thin and thick lines indicate monthly and five-month running mean values, respectively. 25

29 Fig Longitude-time cross section of five-day running mean (a) 200-hPa velocity potential anomalies and (b) 850-hPa zonal wind anomalies (December 2015 December 2016) The contour interval is (a) m 2 /s and (b) 2m/s. The blue (red) shading of (a) indicates areas of divergence that are stronger (weaker) than normal. That of (b) shows easterly (westerly) wind anomalies. 26

30 2.4.2 Winter (December 2015 February 2016) Convective activity was enhanced from the area west of the dateline to the eastern equatorial Pacific and over the Indian Ocean, and was suppressed over and around the Maritime Continent (Fig ). In the upper troposphere, divergence anomalies were seen from the area west of the dateline to the eastern Pacific and convergence anomalies were seen over Africa and the Maritime Continent (Fig ). Anti-cyclonic and cyclonic circulation anomalies straddling the equator were seen over central to eastern parts of the Pacific and the Maritime Continent, respectively (Fig ). In the lower troposphere, cyclonic circulation anomalies straddling the equator were seen over western to central parts of the Pacific and anti-cyclonic circulation anomalies straddling the equator were seen over the Maritime Continent and the Atlantic (Fig ). Eastward propagation of the MJO was seen from the Maritime Continent to the Indian Ocean during the period from mid-december to mid-january and from the eastern Indian Ocean to the eastern Pacific in February (Fig ). Fig Three-month mean OLR anomalies (December 2015 February 2016) Original data provided by NOAA. Fig Three-month mean 200-hPa velocity potential and anomalies (December 2015 February 2016) The contours show the stream function at intervals of m 2 /s, and the shading shows its anomalies. D and C denote divergence and convergence, respectively. Fig Three-month mean 200-hPa stream function and its anomalies (December 2015 February 2016) The contours show the stream function at intervals of m 2 /s, and the shading shows its anomalies. H denotes the center of anti-cyclonic circulation. Fig Three-month mean 850-hPa stream function and its anomalies (December 2015 February 2016) The contours show the stream function at intervals of m 2 /s, and the shading shows its anomalies. H and L denote the center of anti-cyclonic and cyclonic circulation, respectively. 27

31 2.4.3 Spring (March May 2016) Convective activity was enhanced from the area west of the dateline to the central Pacific, and was suppressed from the Maritime Continent to the western Pacific (Fig ). In the upper troposphere, divergence anomalies were seen over the Indian Ocean and from the dateline to the eastern Pacific, and convergence anomalies were seen over the Maritime Continent and from the Atlantic to Africa (Fig ). Anti-cyclonic and cyclonic circulation anomalies straddling the equator were seen over the central Pacific and the Maritime Continent, respectively (Fig ). In the lower troposphere, cyclonic circulation anomalies were seen over the Indian Ocean and anti-cyclonic circulation anomalies were seen from the Bay of Bengal to the western Pacific (Fig ). Eastward propagation of the MJO was seen from the Indian Ocean to the Pacific in March and from Africa to the Indian Ocean during the period from early to mid-may (Fig ). Fig Three-month mean OLR anomalies (March May 2016) As per Fig , but for March May Fig Three-month mean 200-hPa velocity potential and anomalies (March May 2016) As per Fig , but for March May Fig Three-month mean 200-hPa stream function and anomalies (March May 2016) As per Fig , but for March May Fig Three-month mean 850-hPa stream function and anomalies (March May 2016) As per Fig , but for March May

32 2.4.4 Summer (June August 2016) Convective activity was enhanced over the eastern Indian Ocean and suppressed over the western Indian Ocean and western to central parts of the equatorial Pacific (Fig ). In the upper troposphere, divergence anomalies were seen from the eastern Indian Ocean to the southern part of the Maritime Continent and convergence anomalies were seen over the western Indian Ocean and the western Pacific (Fig ). Anti-cyclonic circulation anomalies straddling the equator were seen over the western Pacific. Wave trains were seen over Eurasia in the latitude bands of 40 N, and the northeastward extension of the Tibetan High was stronger than normal (Fig ). In the lower troposphere, cyclonic circulation anomalies straddling the equator were seen over the Indian Ocean (Fig ). The westward extension of the Pacific High was weaker than normal. Eastward propagation of the MJO was seen from Africa to the Maritime Continent in June, from the Pacific to the Indian Ocean in July and from the Maritime Continent to the Pacific in August (Fig ). Fig Three-month mean OLR anomalies (June August 2016) As per Fig , but for June August Fig Three-month mean 200-hPa velocity potential and anomalies (June August 2016) As per Fig , but for June August Fig Three-month mean 200-hPa stream function and anomalies (June August 2016) As per Fig , but for June August Fig Three-month mean 850-hPa stream function and anomalies (June August 2016) As per Fig , but for June August

33 2.4.5 Autumn (September November 2016) Convective activity was enhanced over the Maritime Continent and the latitude bands of 10 N 15 N in the Pacific, and was suppressed over western to central parts of the Indian Ocean and the equatorial Pacific (Fig ). In the upper troposphere, divergence anomalies were seen over the Maritime Continent and the Atlantic, and convergence anomalies were seen over western to central parts of the Indian Ocean and the equatorial Pacific (Fig ). Anti-cyclonic circulation anomalies straddling the equator were seen from the Indian Ocean to the Maritime Continent (Fig ). In the lower troposphere, cyclonic and anti-cyclonic circulation anomalies straddling the equator were seen from the eastern Indian Ocean to the Maritime Continent and the Pacific, respectively (Fig ). Eastward propagation of the MJO was seen from the eastern Indian Ocean to the Maritime Continent in September and from the Pacific to the Indian Ocean in November (Fig ). Fig Three-month mean OLR anomalies (September November 2016) As per Fig , but for September November Fig Three-month mean 200-hPa velocity potential and anomalies (September November 2016) As per Fig , but for September November Fig Three-month mean 200-hPa stream function and anomalies (September November 2016) As per Fig , but for September November Fig Three-month mean 850-hPa stream function and anomalies (September November 2016) As per Fig , but for September November

34 2.4.6 Tropical cyclones over the western North Pacific In 2016, 26 tropical cyclones (TCs) with maximum wind speeds of 17.2 m/s or higher formed over the western North Pacific (Table 2.4-2), which was near the normal of 25.6 ( average). The first named TC of 2016 formed over the western North Pacific on July 3, making it the second-latest since 1951 after the July 9 record of The late start to the typhoon season is attributed to the atmospheric circulation pattern over the tropics of the northwestern Pacific Ocean, which was unfavorable for TC formation as often seen in the year after El Niño peaks. The number of TC formations after July was higher than usual, making the eventual annual total normal. A total of 11 TCs came within 300 km of the Japanese archipelago, which was near the normal of 11.4 ( average). Six made landfall on Japan (against a normal of 2.7), which is the joint-second-highest number on record after the ten recorded in The tracks of TCs generated in 2016 are shown in Fig Table Tropical cyclones forming over the western North Pacific in 2016 Based on information from the RSMC Tokyo-Typhoon Center Maximum Number Date Name Category 1) wind 2) ID (UTC) (knots) 1601 Nepartak 7/3 7/9 TY Lupit 7/23 7/24 TS Mirinae 7/26 7/28 STS Nida 7/30 8/2 STS Omais 8/4 8/9 STS Conson 8/9 8/15 TS Chanthu 8/13 8/17 STS Dianmu 8/17 8/19 TS Mindulle 8/19 8/23 TY Lionrock 8/21 8/30 TY Kompasu 8/20 8/22 TS Namtheun 9/1 9/4 TY Malou 9/6 9/7 TS Meranti 9/10 9/15 TY Rai 9/12 9/13 TS Malakas 9/12 9/20 TY Megi 9/23 9/28 TY Chaba 9/29 10/5 TY Aere 10/5 10/10 STS Songda 10/8 10/13 TY Sarika 10/13 10/19 TY Haima 10/15 10/21 TY Meari 11/3 11/7 TY Ma-on 11/10 11/12 TS Tokage 11/25 11/28 STS Nock-ten 12/21 12/27 TY 105 1) Intensity classification for tropical cyclones (range of maximum wind speed) TS: Tropical Storm (34 47 knots) STS: Severe Tropical Storm (48 63 knots) TY: Typhoon (64 knots ) 2) Estimated maximum 10-minute mean wind speed 31

35 Fig Tracks of tropical cyclones in 2016 The lines indicate the tracks of tropical cyclones with maximum wind speeds of 17.2 m/s or higher. The numbers in circles indicate points where maximum wind speeds exceeded this value, and those in squares indicate points where they fell below it. 32

36 2.5 Oceanographic conditions Throughout 2016, the global average sea surface temperature (SST) was much higher than normal, especially until summer. This was partly attributable to a long-term trend of SST increase caused by global warming and to SST increase in tropical regions of the Pacific and the Indian Ocean in association with the El Niño event that ended in spring The annual mean anomaly was C, which was above the previous record of C observed in 2015 and was the highest since In the equatorial Pacific, remarkably positive SST anomalies were observed in central and eastern parts during winter 2015/2016. The positive SST anomalies in the eastern part weakened during spring. From summer onward, remarkably positive SST anomalies were observed in the western part, and negative SST anomalies were observed in central and eastern parts (Fig , Fig (left)). The SST deviation from the reference value (the climatological mean based on a sliding 30-year period) averaged for the NINO.3 region decreased from +3.0 C in December 2015 to 0.6 C in July 2016, and remained between 0.6 and 0.3 C from August onward (Fig (top)). The five-month running mean of the deviation remained at +0.5 C or more from June 2014 to April 2016 and fell below +0.5 C in May. The El Niño event that began in summer 2014 ended in spring Southern Oscillation Index (SOI) values were negative until April and remained positive from May onward except in October (Fig (bottom)). Positive ocean heat content (OHC) anomalies in the central part of the equatorial Pacific propagated eastward during winter 2015/2016 before negative anomalies in the western part propagated eastward during spring. Negative OHC anomalies were observed in most parts in spring and persisted in central and eastern parts until autumn (Fig (right)). In the North Pacific, remarkably positive SST anomalies were observed near the western coast of North America and from central to eastern parts of the tropical region. Pacific Decadal Oscillation (PDO) 1 index values were positive throughout the year (Fig ). In the South Pacific, positive SST anomalies were observed near the western coast of South America throughout the year, and remarkably positive SST anomalies were observed east of Australia from spring to summer and northeast of New Zealand in autumn. In the Indian Ocean, remarkably positive SST anomalies were observed in most regions of the tropical area until spring. Positive anomalies weakened in the western part during summer. Remarkably positive SST anomalies were observed in the eastern tropical region from summer onward. In the North Atlantic, positive SST anomalies were observed east of the USA throughout the year. Remarkably negative SST anomalies were observed south of Greenland from winter to spring and in autumn (Fig ). 1 For details, see the Pacific Decadal Oscillation (PDO) index information on the TCC website ( o.html). 33

37 (a) Winter (b) Spring (c) Summer (d) Autumn Fig Seasonal mean sea surface temperature anomalies (2016) (a) Winter (Dec Feb. 2016), (b) Spring (Mar. May), (c) Summer (Jun. Aug.), (d) Autumn (Sep. Nov.). The contours and shading show sea surface temperature anomalies at intervals of 0.5 C. Maximum sea ice coverage areas are shaded in gray. Fig Time-longitude cross sections of SST anomalies (left) and ocean heat content anomalies (right: vertically averaged temperature over the top 300 m) along the equator in the Pacific Ocean from 2014 to 2016 The contours and shading show SST anomalies (left) and ocean heat content anomalies (right) at intervals of 0.5 C. 34

38 El Niño monitoring index ( C) Southern Oscillation Index Fig Time-series of the El Niño monitoring index (top: NINO.3 SST deviation from a sliding 30-year mean) and the Southern Oscillation Index (bottom) from 2006 to 2016 The thin lines represent monthly mean values, and the thick lines represent five-month running mean values. The shading indicates El Niño (red) and La Niña (blue) events. Fig Time-series of the PDO index from 1901 to 2016 The red line represents annual mean values for the PDO index, the blue line represents five-year running mean values, and the gray bars represent monthly values. 35

39 2.6 Stratospheric circulation in boreal winter A stratospheric sudden warming (SSW) event is a phenomenon in which a rapid stratospheric temperature increase of several tens of Kelvin is observed over a period of a few days in the polar region during winter, and was identified by Richard Scherhag at the Free University of Berlin in It is caused by enhanced propagation of energy from the troposphere due to planetary-scale wave action (Matsuno 1971). According to the World Meteorological Organization (WMO) definition (WMO 1978), a minor SSW occurs when polar temperatures increase by 25 K or more within a week at any stratospheric level. In addition to this criterion, if the stratospheric zonal mean temperature increases in the poleward direction and net zonal mean zonal winds become easterly north of 60 N at 10-hPa or below, the event is classified as a major SSW. The stratospheric polar vortex was stronger than normal in winter 2015/2016, but a minor SSW event occurred between late January and mid-february. A major SSW event also occurred between late February and early April (Fig ). This section outlines the characteristics of stratospheric circulation seen in winter 2015/2016, including the period of the two SSW events (c)), corresponding to the SSW event observed from late January to mid-february. These characteristics were also seen in March (Fig (d)), corresponding to the major SSW event observed from late February to early April. Fig Time-series representation of 30-hPa temperatures over the North Pole from 1 September 2015 to 31 August 2016 The black line shows daily temperatures (unit: o C), and the gray line indicates the climatological mean Characteristics of stratospheric circulation In the three-month mean 30-hPa height field from December 2015 to February 2016 (Fig ), annular patterns with negative anomalies in the high latitudes and positive anomalies in the mid-latitudes were observed, indicating a stronger-than-normal polar vortex. The eastward extension of the Aleutian High was stronger than normal. In the monthly mean 30-hPa height field, the polar vortex was stronger than normal in December and January (Figs (a) and (b)). In February, the Aleutian High was stronger than normal and the center of the polar vortex shifted toward Eurasia (Fig. Fig Three-month mean 30-hPa height and anomaly (December 2015 February 2016) The contours show 30-hPa height at intervals of 120 m, and the shading indicates its anomalies. 36

40 (a) Dec (b) Jan (c) Feb (d) Mar (e) Apr Fig Monthly mean 30-hPa height and anomaly for (a) December 2015, (b) January 2016, (c) February 2016, (d) March 2016 and (e) April 2016 The contours show 30-hPa height at intervals of 120 m, and the shading indicates its anomalies SSW from late January to mid-february In the five-day mean 30-hPa field (Fig ), the stronger-than-normal polar vortex shifted toward the area from the Atlantic to Eurasia during the period from the latter half of late January to the first half of early February in association with the enhancement of the Aleutian High (Fig (a) (c)). Prior to a prevalent SSW event, 30-hPa temperatures over the North Pole rapidly increased in late January (Fig ). A value of 100-hPa Eliassen-Palm (E-P) flux 1 (Edmon et al. 1980) in the first half of late January shows enhanced upward propagation of planetary 1 E-P flux provides a useful framework for diagnosing interaction between eddies and mean flow in the Transformed Eulerian Mean (TEM) equation. Convergence (divergence) of E-P flux corresponds to deceleration (acceleration) of westerly winds in the zonal mean field. waves over and around the latitude bands of 60 N in the troposphere, contributing to deceleration of the polar night jet stream in the stratosphere (Fig (a)). Clear planetary wave propagation was seen from Western and Central Siberia in the troposphere to the central and eastern Pacific in the stratosphere (Fig (b)) in association with the enhancement of the Aleutian High (Fig (a) (c)). In the 500-hPa height field during the same period, an enhanced ridge was observed over Western and Central Siberia, indicating a related impact both on cold-air outbreaks in East Asia and on stratospheric circulation (Fig ). The polar vortex then exhibited enhancement over the polar region in the second half of mid-february (Figs (d) (f)), and temperatures over the North Pole were below normal (Fig ). 37

41 (a) Jan. (b) Jan. (c) 31 Jan. 4 Feb. (d) 5 9 Feb. (e) Feb. (f) Feb. Fig Five-day mean 30-hPa height and anomaly for (a) January, (b) January, (c) 31 January 4 February, (d) 5 9 February, (e) February, (f) February 2016 The contours show 30-hPa height at intervals of 120 m, and the shading indicates its anomalies. Fig (a) Latitude-height cross section of zonal mean zonal wind, E-P flux and zonal wind tendency in line with the divergence/convergence of the E-P flux and (b) longitude-height cross section of height anomalies from the zonal mean and wave activity flux averaged over 50 o N 70 o N averaged from 21 to 25 January In (a), the contours show zonal mean zonal wind at intervals of 10 m/s, the shading indicates divergence/convergence of the E-P flux (yellow: acceleration; green: deceleration) and the vectors denote E-P flux (units: 10 6 m 3 /s 2 (horizontal); m 2 /s 2 (vertical)) scaled using the square of pressure. The units of vertical axis are hpa. In (b), the contours show height anomalies at intervals of 100 m and the vectors denote wave activity flux with reference to Plumb (1985) (units: m 2 /s 2 (horizontal); Pa m/s 2 (vertical)). 38

42 2.6.4 Summary The occurrence of the two SSW events during the period from winter 2015/2016 to early spring 2016 was presumed to be associated with an enhanced ridge over and around Western and Central Siberia in the troposphere. Monitoring of temporal evolution for the ridge over Siberia is important to determine its impacts on the enhancement of the Siberian High and the stratospheric circulation. Fig Five-day mean 500-hPa height and anomaly for January 2016 The contours show 500-hPa height at intervals of 60 m, and the shading indicates its anomalies Major SSW from late February In the five-day mean 30-hPa field, the polar vortex that was stronger than normal in mid-february shifted toward Eurasia and the Aleutian High was enhanced from the northern part of North America to the polar region in early March (Fig (a) (d)). 30-hPa temperatures over the North Pole rapidly increased in late February again (Fig ) and zonal mean zonal wind turned from westerly to easterly wind in the stratospheric high-latitudes (not shown), and the major SSW event occurred. The wave activity flux fields indicate that the enhanced upward propagation of the planetary waves from Western Siberia in the troposphere to the central to eastern Pacific in the stratosphere (Fig ), corresponding to, the enhancement of the Aleutian High (Fig (a) (d)) as is the case with the SSW event from late January to mid-february. In the 500-hPa height field during the same period, an enhanced ridge was observed over Western Siberia, indicating its impact on the stratospheric circulation (Fig ). References Edmon, H. J., B. J. Hoskins and M. E. McIntyre, 1980: Eliassen-Palm cross sections for the troposphere. J. Atmos. Sci., 37, Matsuno, T., 1971: A dynamical model of stratospheric sudden warming. J. Atmos. Sci., 28, Plumb, R. A., 1985: On the three-dimensional propagation of stationary waves. J. Atmos. Sci., 42, WMO, 1978: Abridged final report of Commission for Atmospheric Sciences. WMO Rep., 509, 113pp. Fig Five-day mean 500-hPa height and anomaly for February 2016 The contours show 500-hPa height at intervals of 60 m, and the shading indicates its anomalies. 39

43 (a) Feb. (b) 25 Feb. 1 Mar. (c) 2 6 Mar. (d) 7 11 Mar. (e) Mar. (f) Mar. Fig Five-day mean 30-hPa height and anomaly for (a) February, (b) 25 February 1 March, (c) 2 6 March, (d) 7 11 March, (e) March, (f) March 2016 The contours show 30-hPa height at intervals of 120 m, and the shading indicates its anomalies. Fig (a) Latitude-height cross section and (b) longitude-height cross section averaged over 50 o N 70 o N during the period from 20 to 24 February 2016 The elements are the same as Fig

44 2.7 Summary of the Asian summer monsoon Asian summer monsoon monitoring is important because related fluctuations in convective activity and atmospheric circulation can influence the summer climate in Asia, including Japan. This section summarizes the characteristics of the Asian summer monsoon in Temperature and precipitation Four-month mean temperatures based on CLIMAT reports covering the monsoon season (June September) were more than 1 C above normal in many parts of East Asia, especially north of 30 N. Extremely high monthly temperatures were recorded in many areas from June to September as follows (Fig ): June: Okinawa Islands (Japan) to southern China July: northeastern China to central Mongolia, southern Borneo Island to southern Indonesia, in and around southern India August: Kyushu region (Japan) to central China, Malaysia to central Indonesia, southern India September: western Mongolia to Pakistan, western Borneo Island to northern Sumatra Island Four-month total precipitation amounts for the same period were more than 140% of the normal in Japan s Hokkaido region, eastern China, southern Mongolia, northern Lao PDR to northern Myanmar, in and around Pakistan, and in and around central Indonesia. The corresponding numbers were less than 60% of the normal on the Korean Peninsula and in northeastern and central China, southern India and southwestern Pakistan (Fig ). These amounts were mostly consistent with the distribution of outgoing longwave radiation (OLR) anomalies (Fig ). Extremely high precipitation based on monthly data was seen in the Hokkaido region of Japan in August and from the northern Kyushu region of Japan to eastern China and in southern Indonesia in September. A number of major weather-related disasters were reported from June to September. Heavy rain and flooding in northeastern parts of D.P.R Korea from late August to early September were brought by the remnants of Typhoon Lionrock, causing more than 130 fatalities, according to a report from the United Nations Office for the Coordination of Humanitarian Affairs (OCHA). China suffered more than 600 fatalities from June to August due to heavy rain over the Yangtze River Basin, over the Yellow River basin and in the southern part of the country, and was also hit by Typhoon Nepartak, according to the Chinese government. Monthly precipitation in July amounted to 352 mm ( average: mm) at Changsha in Hunan Province and 364 mm ( average: mm) at Shenyang in Liaoning Province. Heavy rain and flooding also affected eastern Nepal in July, causing more than 120 fatalities, according to OCHA. In India, heavy rain and flooding caused more than 120 fatalities from southern to northern parts and elsewhere in July and August, and more than 30 fatalities in central parts and elsewhere, according to the government of India. Monthly precipitation for September was 440 mm ( average: mm) at Hyderabad in central India. In Pakistan, heavy rain caused more than 140 fatalities in the northern part from June to August, according to the government of Pakistan. 41

45 Fig Four-month mean temperature anomaly ( C) for June September 2016 See Section for the data source. Fig Four-month total precipitation ratio (%) for June September 2016 See Section for the data source Convective activity and atmospheric circulation Convective activity (inferred from OLR) averaged for June September 2016 was enhanced from the eastern Indian Ocean to the Maritime Continent, over the area from the Bay of Bengal to the southwestern Indochina Peninsula and from southern China to the seas northeast of the Philippines, and was suppressed over the western part of the equatorial Pacific (Fig ). OLR index data (Table 2.7-1) indicate that the overall activity of the Asian summer monsoon (represented by the SAMOI (A) index) was near normal until August and above normal in September. The most active convection area was shifted westward of its normal position until July (see the SAMOI (W) index). In the upper troposphere, the Tibetan High was stronger than normal over its northeastern part (Fig (a)). In the lower troposphere, monsoon circulation over the Indian Ocean was stronger than normal and cyclonic circulation anomalies straddling the equator were seen over the area from the Indian Ocean to the Maritime Continent (Fig (b)). Zonal wind shear between the upper and lower troposphere over the North Indian Ocean and southern Asia (Fig ) remained above normal, indicating stronger-than-normal monsoon circulation from mid-may onward (except in the second half of July and the first half of late August). Convective activity over the Maritime Continent was enhanced throughout the summer monsoon season. It was suppressed to the east of the Philippines until July and enhanced after August (Fig ). In the lower troposphere, cyclonic circulation associated with a deep monsoon trough was clearly seen over the seas to the southeast of Japan in August in response to enhanced convective activity over the western North Pacific (figures not shown). Reference Webster, P. J. and S. Yang, 1992: Monsoon and ENSO: Selectively interactive systems. Quart. J. Roy. Meteor. Soc., 118,

46 (a) (b) Fig Four-month mean outgoing longwave radiation (OLR) and its anomaly for June September 2016 The contours indicate OLR at intervals of 10 W/m 2, and the colored shading denotes OLR anomalies from the normal. Negative (cold color) and positive (warm color) OLR anomalies show enhanced and suppressed convection compared to the normal, respectively. Original data are provided by NOAA. Table Summer Asian Monsoon OLR Index (SAMOI) values observed from May to October 2016 SAMOI is described in Year 2016 Summer Asian Monsoon OLR Index (SAMOI) SAMOI (A) Activity SAMOI (N) Northward- shift SAMOI (W) Westward- shift May Jun Jul Aug Sep Fig Four-month mean stream function and its anomaly for June September 2016 (a) The contours indicate the 200-hPa stream function at intervals of m 2 /s, and the colored shading indicates 200-hPa stream function anomalies from the normal. (b) The contours indicate the 850-hPa stream function at intervals of m 2 /s, and the colored shading indicates 850-hPa stream function anomalies from the normal. Warm (cold) shading denotes anticyclonic (cyclonic) circulation anomalies in the Northern Hemisphere and vice versa in the Southern Hemisphere. Oct

47 Fig Time-series representation of the zonal wind shear index between 200 hpa and 850 hpa averaged over the North Indian Ocean and southern Asia (pink rectangle in the right figure: equator 20ºN, 40ºE 110ºE) The zonal wind shear index is calculated after Webster and Yang (1992). The thick and thin pink lines indicate seven -day running mean and daily mean values, respectively. The black line denotes the normal, and the gray shading shows the range of the standard deviation calculated for the time period of the normal. (a) June 2016 (b) July 2016 (c) August 2016 (d) September 2016 Fig Monthly mean anomalies of outgoing longwave radiation (shading at intervals of 10 W/m 2 ) and 850-hPa stream function (contour at intervals of m 2 /s) for (a) June, (b) July, (c) August and (d) September

48 2.8 Arctic sea ice conditions The sea ice extent in the Arctic Ocean has recently shown a decreasing tendency that has been particularly marked in terms of the annual minimum extent (Fig ). The monitoring of Arctic sea ice conditions has become more significant because of their possible influence on the climate as a result of related changes in the radiation budget and heat exchange between the Arctic Ocean and the atmosphere. This section outlines the characteristics of the Arctic sea ice extent seen in 2016 along with those of atmospheric circulation. Fig Time-series representation of annual minimum (red line) Arctic sea ice extents from 1979 to 2016 The dashed line denotes the trend. The trend of the annual minimum for the period from 1979 to 2016 is km 2 /year Presence of sea ice in the Arctic in 2016 The Arctic sea ice extent 1 in 2016 reached its annual maximum on 23 March at million square kilometers and its annual minimum on 5 September at 4.09 million square kilometers (Fig ). Both the maximum and minimum sea ice extent was the second smallest since 1979 (Figs ) Arctic atmospheric circulation and melting of sea ice In September 2016, the center of a low-pressure system shifted toward North America and the northward extension of a high-pressure system over Eurasia was stronger than normal in contrast to circulation in August (middle panel, Fig ). In the lower troposphere, above-normal temperatures over a wide area of the high latitudes (Fig ) supported a reduction of sea ice extents for a lower-than-normal level. Fig Time-series representation of sea ice extents for the Arctic region for 2016 The black line indicates the normal, and the gray shading denotes the range of the standard deviation calculated for the period of the normal. 1 The sea ice extent is defined as the area in which ice concentration (i.e., the ratio of ice cover in a particular reference area) is %. Fig Sea ice concentration on 5 September 2016 (left) and sea ice extent climatology on 5 September (right) The blue scale in the panel on the left shows deciles of sea ice concentration, and the white areas in the panel on the right show sea ice extent climatology. 45

49 Fig Monthly mean sea level pressure and its anomalies over the Arctic region in July (left), August (middle) and September (right) 2016 The contours show sea level pressure at intervals of 4 hpa, and the shading indicates its anomalies. H and L denote the center of a high- and low-pressure system, respectively. Fig Monthly mean 925-hPa temperature and its anomalies over the Arctic region in July (left), August (middle) and September (right) 2016 The contours show 925-hPa temperature at intervals of 3 o C, and the shading indicates its anomalies. 46

50 2.9 Snow cover in the Northern Hemisphere Snow cover has a close and mutual association with climatic conditions. The albedo of snow-covered ground (i.e., the ratio of solar radiation reflected by the surface) is higher than that of snow-free ground. As a result, the variability of snow cover has an impact on the earth s surface energy budget and radiation balance. In addition, snow absorbs heat from its surroundings and melts, thereby providing soil moisture. The variability of atmospheric circulation and oceanographic conditions affects the amount of snow cover. This section outlines the characteristics of snow cover in 2016 as well as its long-term variability and related trends Related characteristics in 2016 In winter (December February) 2015/2016, there were fewer days of snow cover than normal in many parts of the Northern Hemisphere (Fig (a)), and this situation continued until April. In May 2016, there were more days of snow cover than normal in and around the southern part of Central Siberia (Fig (b)). In November 2016, there were more days of snow cover than normal in and around Central Asia and northeastern China and fewer than normal in western China and North America (Fig (c)). (a) February 2016 (b) May 2016 (c) November 2016 Fig Number of snow-cover days (top) and its anomaly (bottom) for (a) February 2016, (b) May 2016 and (c) November 2016 Statistics on the number of snow-cover days are derived using data from the Special Sensor Microwave Imager (SSM/I) and the Special Sensor Microwave Imager Sounder (SSMIS) on board US Defense Meteorological Satellite Program (DMSP) satellites based on an algorithm developed by the Japan Meteorological Agency. The base period for the normal is

51 2.9.2 Interannual variability and related trends Fig shows interannual variations in the total area of monthly snow cover in the Northern Hemisphere and Eurasia over the 29-year period from 1988 to The Northern Hemisphere exhibits a decreasing trend (with a 95% confidence level) for May, June and the period from September to December, while no trend (with a 95% confidence level) is seen for the period from January to April. In Eurasia there is a decreasing trend for April, May, June and the period from September to December, while no trend is seen for January, February and March. Fig Interannual variations in the total area of monthly snow cover (10 6 km 2 ) in the Northern Hemisphere (north of 30 N; left) and Eurasia (30 N 80 N, E; right) for February ((a) and (d)), May ((b) and (e)), and November ((c) and (f)) from 1988 to 2016 The blue lines indicate total snow cover area for each year and the black lines show linear trends (95% confidence level). 48

52 3. Analysis of specific events 3.1 The El Niño event 1 ending in boreal spring 2016 and its effects The characteristics of the El Niño event that occurred from boreal summer (June August) 2014 to spring (March May) 2016 are described in Section 3.1.1, and various related effects observed from boreal winter (December February) 2015/2016 to autumn (September November) 2016 are outlined in Section /15/16 El Niño event 2 (1) Overview The El Niño event starting in summer (June August) 2014 and ending in spring (March May) 2016 covered eight seasons, making it the longest since The monthly mean sea surface temperature (SST) deviation from the climatological reference 4 over the El Niño monitoring region (NINO.3 in Fig ) was +3.0 C in the mature stage of November December 2015, which was the third-highest on record after the +3.6 C of the 1987/98 event and the +3.3 C of the 1982/83 event. The amplitudes of SST variations in the monitoring regions of the tropical Indian Ocean (IOBW 5 ) and the tropical western Pacific Ocean (Fig ), which are important climate effect indicators for El Niño events, were also as large as those of the 1 JMA judges that an El Niño has begun when the five-month running mean sea surface temperature (SST) deviation for NINO.3 remains at +0.5 C or more for six months. El Niño periods are expressed in seasonal units. 2 Previous El Niño events are identified by their relevant periods (the full four-number expression for the first year and the final two numbers for subsequent years). By way of example, the 1997/98 El Niño event ran from boreal spring 1997 to spring The second-longest El Niño events after that of 2014/15/16 (eight seasons) were those of 1968/69/70, 1986/87/88, 1982/83 and 1991/92 (six seasons each). 4 SST climatological references are monthly averages of the latest sliding 30- year period for NINO.3, and are defined as linear extrapolations with respect to the latest 30-year period for NINO.WEST and IOBW in order to remove the effects of significant long-term warming trends observed in these regions. 5 Indian Ocean Basin-Wide 1997/98 El Niño event as seen in NINO.3 SST variations. Lower-than-normal temperatures were observed in western Japan throughout the boreal summers of 2014 and 2015, and higher-than-normal temperatures were observed in eastern Japan during boreal winter 2015/2016. These characteristics were consistent with common patterns observed in past El Niño events. The global average surface temperature anomaly in 1998 was the highest since records began in 1891, and this record was again broken in each year of the 2014/15/16 El Niño event. The formation of 2016 s first typhoon was also later than normal as similarly observed in the El Niño termination years of 1973, 1983 and 1998, when record-high NINO.3 SSTs were recorded. These climatic characteristics also relate to the descending (ascending) nature of SST anomalies in NINO.WEST (IOBW) regions in concurrence with (subsequent to) the rise in NINO.3 SST anomalies. The lifetime of the 2014/15/16 El Niño event in the course of life is described below. (2) SST deviation from climatological reference in individual monitoring regions Fig shows a time-series representation of NINO.3 SST deviation from its climatological reference in past El Niño events. The termination year for each event is set as Year0, and NINO.3 SST deviations are shown from January of Year 2 (two years before Year0) to January of Year+1 (the year after Year0). The black solid line indicates values for the 2014/15/16 El Niño event, and the dotted black line indicates the average of the 13 previous events. These deviations are referred to as NINO.3dev below. In the average of the 13 previous events, NINO.3dev is +0.5 C or above for boreal spring in Year 1, which results in the onset of an El Niño event that reaches its mature stage around November December of Year 1. The value falls below +0.5 C 49

53 around spring of Year0, resulting in the termination of the event. The 2014/15/16 El Niño event began in 2014 (Year 2), which was the year before its mature stage. NINO.3dev varied between +0.2 and +1.0 C, and did not show the signs of development commonly seen in past El Niño events. Meanwhile, five-month running averages of NINO.3dev remained at or above +0.5 C from June 2014 onward and between +0.5 and +0.6 C for eight of the ten months through to March 2015, thereby meeting the criteria for the definition of an El Niño event from boreal summer 2014 onward. After spring 2015 (Year 1), NINO.3dev increased at double the rate for the average of the previous 13 events, reaching its positive maximum of +3.0 C (the third-highest of the past events) in December The peak values of the four strongest El Niño events occurring in 1972/73, 1982/83, 1997/98 and 2014/15/16 considerably exceeded the peak of the average value of +1.7 C. These stand out from the corresponding values for the 10 other events, which were equal to or below the average. NINO.3dev decreased rapidly from January 2016 (Year0) onward and approached the average of +0.1 C in May, bringing about the end of the El Niño event. The value subsequently remained near the average (between 0.3 and 0.6 C) from July to November. Fig is the same as Fig except for the NINO.WEST region. NINO.WEST SST deviations from the climatological reference are referred to as NINO.WESTdev below. NINO.WESTdev for the average of the 13 previous events (shown by the dotted black line) turned negative around the summer of Year 1 immediately after the start of the averaged El Niño event, and exhibited two negative peaks around September of Year 1 and February of Year0. The negative values eased around boreal spring of Year0 as the event ended, and turned positive in the summer of Year0. During the El Niño event, distinctly negative NINO.WESTdev values continued from February 2015 (Year 1), in contrast to the average value for the same season. Three negative peaks distinctly below the average were observed in March, in July October 2015 and in February 2016 (Year0). Despite the prolonged nature of these below-average values, the negatives eased in boreal spring 2016 (Year0) along with the average and turned positive in summer 2016 after the end of the El Niño event. Fig is the same as Fig , but for the IOBW region. IOBW SST deviations from the climatological reference are referred to as IOBWdev below. In the average of the previous 13 events, IOBWdev tended to increase in association with elevated NINO.3dev values around spring of Year 1 when the averaged El Niño event began. Values reached their positive peak around January April of Year0 a few months after the mature stage of the El Niño event (coinciding with the NINO.3dev peak) around December of Year 1. In the Pacific Ocean, positive NINO.3dev values eased in boreal spring of Year0 resulting in the termination of El the Niño event, while positive IOBWdev values persisted in the Indian Ocean until boreal summer. This is an important factor in considering the climate over the western North Pacific during boreal summer (Xie et al., 2009; Du et al., 2011). Fig Locations of El Niño monitoring region, western tropical Pacific region, and tropical Pacific region NINO.3 indicates El Niño monitoring region (5 S 5 N, 150 W 90 W), NINO.WEST indicates the western tropical Pacific region (equator 15 N, 130 E 150 E), and IOBW indicates the tropical Indian ocean (20 S 20 N, 40 E 100 E). 50

54 Fig NINO.3 SST deviations from climatological references for past El Niño events Time-series representation of NINO.3 SST deviations from climatological references for January in past El Niño events. The termination year for each event is set to Year0, and NINO.3 SST deviations are plotted from January of Year 2 (i.e., two years before Year0) to January of Year+1 (i.e., the year after Year0). The solid black line represents the 2014/15/16 El Niño event, and the dotted black line represents the average of 13 previous events. The Year0 for each El Niño event is listed to the upper left of the figure. Fig Same as Fig except for NINO.WEST SST deviations Fig Same as Fig , but for IOBW SST deviations During the 2014/15/16 El Niño event, IOBWdev remained near zero after the onset of the event from boreal summer 2014 (Year 2) to around February 2015 (Year 1) before turning positive in spring 2015 (Year 1) in association with the rapid development of the event, and continued to rise before and after the event s mature stage (corresponding to the peak of NINO.3dev). Three months after the peak of NINO.3dev, IOBWdev peaked at C in March 2016 (Year0). This IOBWdev was the second highest on record after the C value of January 1998 (Year0), and was twice as high as the average. Values rapidly decreased thereafter, and the positive values mostly eased in June 2016 (Year0) a month after the disappearance of positive NINO.3dev values. As mentioned above, the considerably above-average positive IOBWdev values observed during the 2014/15/16 El Niño event continued, but disappeared earlier than average. During boreal summer 2016 (Year0), values were near zero and turned negative in autumn. 51

55 (3) Atmospheric and oceanic temporal changes To clarify the characteristics of air-sea interaction in the onset, development and termination of the 2014/15/16 El Niño event, time-longitude sections for areas along the equator (0.5 S 0.5 N) over the Indian and Pacific Oceans for SST anomalies and for depth averaged temperature anomalies from the ocean surface to 300 m are shown in Fig , and time-longitude sections for areas near the equator (5 S 5 N) for velocity potential anomalies in the upper troposphere (200 hpa) and for zonal wind anomalies in the lower troposphere (850 hpa) are shown in Fig Fig also shows three-month (seasonal) average latitude-longitude sections covering 14 seasons from boreal spring 2013 to summer 2016 for outgoing long radiation (OLR) and related anomalies and SSTs with related anomalies, along with longitude-depth sections at the equator for the uppermost 300-m subsurface temperatures and related anomalies. Most typical El Niño events, such as that described in Rasmusson and Carpenter (1982), emerge in boreal spring or summer and develop during summer and autumn, passing through the mature stage from late autumn to early winter and terminating in winter or spring the year after onset 6. Although the 2014/15/16 El Niño event continued for eight seasons from boreal summer 2014 to spring 2016, it did not start early or end late and was almost twice as long as typical El Niño events. Consequently, the phenomenon is viewed as having been separated into units of around a year from spring to spring, representing a cycle of development and decay. Its characteristics are described below for (a) spring 2014 spring 2015, (b) spring 2015 spring 2016, and (c) spring 2016 onward. 6 Five exceptional periods of past El Niño events were boreal spring 1953 autumn 1953, autumn 1968 winter 1969/1970, autumn 1986 winter 1987/1988, spring 1982 summer 1983 and spring 1991 summer 1992, whose start/end points were unusual. (a) Boreal spring 2014 spring 2015 Strong lower-troposphere westerly wind bursts over the western equatorial Pacific in mid-to-late January 2014 preceded the onset of the 2014/15/16 El Niño event. These bursts are illustrated in Fig (right) as strong westerly anomalies 7 of 9 m/s or more. Westerly bursts were again observed in late February and early March. Warm Kelvin waves below the ocean surface resulting from these bursts migrated eastward through the central equatorial Pacific from March to April 2014 to the eastern part (Fig , right). Eastward migration of weak warm Kelvin waves was subsequently observed, and increased subsurface water temperature anomalies in the uppermost 300 m were seen in the central and eastern equatorial Pacific from April to July 2014 (Fig , right; spring (MAM) 2014, Fig , right). In accordance with this increase, SST anomalies in the eastern equatorial Pacific increased from May to July 2014 (Fig , left; summer (JJA) 2014, Fig , center), and positive anomalies of +1.5 C emerged in the eastern part in June 2014, resulting in the onset of the 2014/15/16 El Niño event. The area of above-normal convective activity observed near Indonesia ( E) until boreal winter 2013/2014 moved to the western equatorial Pacific in boreal spring 2014, resulting in below-normal convective activity over Indonesia and above-normal convective activity over the western and central equatorial Pacific. However, the subsequent east-west contrast of convective activity 7 A westerly burst is an event in which westerly winds with speeds exceeding 5 m/s or so are observed for around 10 days in the lower troposphere over the western equatorial Pacific when easterly trade winds blow under normal conditions. Although several definitions of the term have been utilized in previous research, here it refers to westerly wind anomalies of 9 m/s or more. Easterly wind speeds in the lower troposphere (trade winds) average around 4 6 m/s near the date line over the equatorial Pacific, with strength on the eastern side and weakness on the western side of the date line. For strong westerly wind anomalies of 9 m/s or more, westerly winds blow in the central equatorial Pacific, resulting in the disappearance of trade winds. 52

56 between Indonesia and the central equatorial Pacific was unclear, and above-normal values in the western equatorial part did not persist (winter (DJF) 2014 summer (JJA) 2014, Fig , left; Fig , right). In June July 2014, easterly wind anomalies were observed in the central and eastern Pacific, and in July August eastward migration of equatorial cold Kelvin waves was observed in the ocean subsurface along with negative SSTs (Fig ). Displacement of above-normal convection area to the central equatorial Pacific as commonly observed in past El Niño events was not clearly seen, but above-normal convective activity was occasionally observed to the west of the date line, and westerly wind anomalies were seen over the western equatorial Pacific in July and September 2014 (Fig ). These effects stimulated two weak warm Kelvin waves that reached the eastern equatorial Pacific in October and December 2014, and positive SST anomalies persisted in the eastern and central Pacific (Fig ; Autumn (SON) 2014, Fig , center) Fig Time-longitude sections for SST anomalies (left), and subsurface temperature anomalies averaged from ocean surface to the depth of 300 m (right) along the equator (0.5 S 0.5 N) The data are from November 2013 to October

57 Fig Time-longitude sections of velocity potential anomalies in the upper troposphere (200 hpa) (left) and zonal wind anomalies in the lower troposphere (850 hpa) (right) along equatorial regions (5 N 5 S) Negative velocity potential anomalies (left) indicate stronger-than-normal divergence (i.e., above-normal convective activity), and positive values indicate weaker-than-normal divergence (i.e., below-normal convective activity). Positive zonal wind anomalies (right) represent westerly anomalies, and negative values indicate easterly anomalies. The data cover the period from November 2013 to October In November and December 2014, above-normal convective activity was observed near Indonesia, and easterly wind anomalies were seen in the western equatorial Pacific (Fig ; winter (DJF) 2015, Fig. 54

58 3.1-7, left). Cold Kelvin waves stimulated by easterly wind anomalies reached the eastern equatorial Pacific in January March 2015, and the SST anomalies there turned negative (Fig ; winter (DJF) 2015, Fig , center). Thus, from boreal spring 2014 to spring 2015, the El Niño event continued with no clear air-sea interaction (i.e., no sign of development), and neither developed nor decayed. During this period, SSTs were considerably higher than in average years over the entire tropical region in the North Pacific and over the entire tropical Indian Ocean, which contributed to the record-high global average SST recorded in At the same time, SSTs remained below normal over the central and eastern tropical Pacific in the Southern Hemisphere in contrast to the above-normal SSTs commonly observed in the same area during the development process in past El Niño events. (b) Boreal spring 2015 spring 2016 From boreal spring 2015 onward, the El Niño event developed with continued above-normal convective activity over the western equatorial Pacific along with westerly wind anomalies from around January 2015 and the onset of westerly wind burst activity in late March (Fig ). Ocean subsurface warm Kelvin waves excited by this activity reached the eastern equatorial Pacific in April May, and ocean subsurface temperature anomalies subsequently turned positive in the central eastern equatorial Pacific (Fig , right; spring (MAM) 2015, Fig , right). SST anomalies then rose near the western coast of South America in the eastern equatorial Pacific, and in this area positive anomalies expanded gradually westward in boreal summer autumn 2015 (Fig , left; spring (MAM) 2015 Autumn (SON) 2015, Fig , center). Meanwhile, the relative maximum positive SST anomaly was observed near the date line in the equatorial Pacific. Before the development of the El Niño event, the relative maximum was to the west of the date line until early boreal spring 2015, and slowly migrated eastward during boreal spring and summer 2015 in accordance with the development of the event, joining positive anomalies expanding westward from the eastern Pacific in boreal summer and autumn (Fig , left). This eastward migration of the relative maximum SST anomaly near the date line indicates displacement of water at temperatures of 28 C or more (referred to as warm pools) extending from the ocean surface to a depth of 100 m in the western equatorial Pacific (Fig , right). In the course of eastward warm-pool expansion from boreal winter 2014/2015 to autumn 2015, ocean subsurface water temperatures of 30 C or above and relative maximum water temperature anomalies migrated eastward. SST variations corresponded to those of the ocean subsurface (Fig , center). Ocean subsurface variations closely corresponded to those of atmospheric circulation. In May, June July, August and October 2015, four westerly bursts occurred in areas shifting from west to east of the date line with warm-pool eastward migration (Fig , right). In boreal spring 2015, above-normal convective activity areas were centered west of the date line and expanded to the central and eastern equatorial Pacific. The center of this activity gradually moved eastward and reached the central equatorial Pacific east of the date line during the mature stage of the El Niño event. Convective activity near Indonesia turned below normal with the displacement of the above-normal area. The clear contrast of convective activity with the above-normal levels near the date line persisted until boreal spring 2016 when the El Niño event ended (Fig , left; spring (MAM) 2015 spring (MAM) 2016, Fig , left). The positive SST anomalies in the central and eastern equatorial Pacific peaked in November 55

59 December 2015 before gradually easing from the eastern part (Fig , left). In January 2016, another westerly wind burst was observed over the central equatorial Pacific, and ocean subsurface warm Kelvin waves stimulated by this burst arrived at the eastern equatorial Pacific in January February No remarkable warm Kelvin waves were subsequently observed. Ocean subsurface cold waters in the western equatorial Pacific migrated eastward in March and April, and ocean subsurface water temperature anomalies in the uppermost 300 m turned negative over most of equatorial Pacific from the western part to the eastern part in April (Fig , right; spring (MAM) 2016, Fig , right). As a result, the thermocline 8 was shallower than normal over most of the equatorial Pacific, and negative SST anomalies expanded westward from the eastern equatorial Pacific where the thermocline was at its shallowest. In boreal spring 2016, the El Niño event ended with the easing of positive SST anomalies over the central and eastern equatorial Pacific (Fig , left; spring (MAM) 2016, Fig , center). (c) Boreal summer 2016 In boreal summer 2016, the relative minimum negative ocean subsurface temperature anomaly moved to the central equatorial Pacific (Fig , right; summer (JJA) 2016, Fig , right), and SSTs turned below normal from the central to eastern equatorial Pacific (Fig , left; summer (JJA) 2016, Fig , center). Meanwhile, ocean subsurface temperature anomalies turned positive and SSTs rose above normal over most of the western tropical Pacific, where SST areas of 30 C or more prevailed. SSTs turned remarkably above normal from the eastern Indian Ocean near Indonesia to the northeastern coast of Australia. The area of above-normal convective activity periodically varied in association with intra-seasonal oscillations from May to July 2016, and the positive/negative status of zonal wind anomalies in the lower troposphere changed periodically over the equatorial Pacific. Meanwhile, westerly wind anomalies in the lower troposphere persisted over the Indian Ocean (Fig ). From around August 2016, easterly wind anomalies were continually observed in the lower troposphere over the equatorial Pacific. The seasonally averaged OLR showed common characteristics of past El Niño events in boreal spring 2016, with convective activity being below normal near Indonesia and above normal near the date line over the equatorial Pacific. However, in boreal summer 2016, the area of above-normal convective activity near the date line disappeared, and convective activity fell below normal over most of the equatorial Pacific from western to eastern parts. Meanwhile, convective activity was above normal over the eastern Indian Ocean from boreal spring 2016, and the area of above-normal activity extended over the eastern Indian Ocean and Indonesia (summer (JJA), Fig , left). The 2014/15/16 El Niño event is described above in the context of year units running from spring to spring, representing the period from before the onset until after the end of the event. The atmospheric and oceanic processes observed in the period from boreal spring 2015 to spring 2016 (described in (b)) correspond to the stages of a typical El Niño event from development to decay as described in Rasmusson and Carpenter (1982), and are in contrast to the period from spring 2014 spring 2015 (described in (a)). 8 The ocean subsurface layer with its steep vertical temperature gradient indicated in C temperature layers with tight contours (Figure 3.1-7, right). 56

60 References Du, Y., L. Yang. and S.-P. Xie, 2011: Tropical Indian Ocean Influence on Northwest Pacific Tropical Cyclones in Summer following Strong El Niño. J. Climate, 24, Rasmusson, E. M. and T. H. Carpenter, 1982: Variations in Tropical Sear Surface Temperature and Surface Wind Fields Associated with the Southern Oscillation/El Niño. Mon. Wea. Rev., 110, Xie, S.-P., K. Hu, J. Hafner, H. Tokinaga, Y. Du, G. Huang, and T. Sampe, 2009: Indian Ocean Capacitor Effect on Indo-Western Pacific Climate during the Summer following El Niño. J. Climate, 22,

61 Fig Seasonally averaged latitude-longitude sections for outgoing longwave radiation (OLR) (left) and SST (center), and longitude-depth sections for ocean subsurface temperature along the equatorial Pacific (right) along with their anomalies (boreal spring (March May) 2013 autumn (September November) 2014) Blue and black contours indicate observed values, and shading with white contours indicates anomalies from the normal (i.e., the average). Contour intervals are 20 W/m 2 (OLR), 10 W/m 2 (OLR anomalies), 1 C (SST and ocean subsurface temperature) and 0.5 C (SST anomalies and ocean subsurface temperature anomalies). Contours for OLR are shown for values of 250 W/m 2 or less, with lower values indicating greater convective activity. Green shading indicates regions of above-normal convective activity, and brown shading indicates regions of below-normal convective activity. 58

62 Fig Continued (boreal winter (December February) 2014/2015 summer (June August) 2016) 59

63 3.1.2 Influences of the El Niño event on the global climate As described in the previous subsection, the El Niño event peaked in winter 2015/2016 and ended in spring SSTs in the Indian Ocean trailed the event by a couple of months and remained above normal toward spring/summer Influences from the resulting SST anomalies were extensively felt across the globe, with effects including dry conditions in Southeast Asia, extremely heavy precipitation along the Yangtze river basin, delayed formation of the first typhoon of the season in the western North Pacific, and far higher-than-normal temperatures over Japan in the first half of winter 2015/2016. (1) Development of the El Niño event and associated atmospheric circulation Atmospheric circulation anomalies associated with the event are briefly described here for the period from May to October 2015 (the Asian summer monsoon season) during the development phase and before the peak, and for the period from April to June (around the onset of the Asian summer monsoon), when SST anomalies in the Indian Ocean peaked in the wake of the event. Also shown are results from statistical analysis of atmospheric circulation observed during the past El Niño events and high-sst events in the Indian Ocean. Fig shows changes in the NINO.3 index and the IOBW index, which are defined as SST departures from the climatological mean based on the latest sliding 30-year period averaged over the eastern equatorial Pacific and the tropical Indian Ocean, Mar. to May 2015 Jun. to Aug Sep. to Nov Dec to Feb Mar. to May 2016 Fig NINO.3 and IOBW index fluctuations Thin lines indicate monthly values and thick lines indicate the five-month moving average. These indices are defined as SST anomalies averaged over the areas shown in the bottom panel. Fig month mean SST anomalies From top to bottom: boreal spring, summer, autumn 2015, winter 2015/2016 and spring Anomalies are represented with respect to the average. 60

64 respectively. The NINO.3 index turned positive around spring 2014 and began to increase rapidly in spring Values began to decline after peaking in winter 2015/2016, returned to near-normal in spring 2016 and turned negative in summer The IOBW index surged on the heels of NINO.3, peaking in spring 2016 before declining throughout summer. Fig. (a) (b) (c) indicates seasonal mean SST anomalies observed from spring 2015 to spring Fig shows stream function anomalies at 850 hpa composited over the three-month periods of May to July (early Asian summer monsoon), August to October (late Asian summer monsoon) and December to February (boreal winter) of El Niño years from based on JRA-55 (Kobayashi et al., 2015). The figures show that, during Asian summer monsoon periods, equatorial symmetric cyclonic and anticyclonic circulation anomalies tend to develop in the Pacific and in the area from the Indian Ocean to the Maritime Continent, respectively, in response to convection anomalies associated with El Niño events. This anomaly pattern leads to weaker-than-normal southwesterlies and suppressed monsoon precipitation over Southeast Asia. In winter, anticyclonic circulation anomalies extend over and to the east of Japan in association with a wave train pattern in the upper troposphere (figure not shown), indicating the mild winters experienced in Japan during El Niño events. A composite map of stream function anomalies at 850 hpa for the three-month periods of April to June in positive IOBW years based on JRA-55, as shown in Fig , indicates cyclonic circulation anomalies north of the equator in the Indian Ocean and Fig Composite map for stream function at 850 hpa during El Niño events Three-month mean for (a) early Asian summer monsoon (May to July), (b) late Asian summer monsoon (August to October) and (c) boreal winter (December to February). Anomalies are represented as deviations from the zonal mean. Contours are at intervals of 0.5 x 10 6 m 2 /s. Shading denotes statistical confidence. Fig Composite map for stream function at 850 hpa during warm IOBW events Three-month mean for April to June. Anomalies are represented as deviation from the zonal mean. Contours are at intervals of 0.5 x 10 6 m 2 /s. Shading denotes statistical confidence. 61

65 equatorial symmetric anticyclonic circulation anomalies over the area from Indochina to the western North Pacific. These anticyclonic anomalies are likely related to equatorial Kelvin waves, which propagate from the Indian Ocean where SSTs remain above normal in the aftermath of an El Niño event, toward the western Pacific and induce Ekman divergence north and south of the equator (Xie et al., 2009). Anomalies of outgoing longwave radiation (OLR) and stream function at 850 hpa for May to October 2015 are shown in Fig (a). The circulation pattern of this period is characterized by cyclonic circulation anomalies over the Pacific and anticyclonic circulation anomalies centered over Indochina, which is quite similar to the situation of anomalies observed in past El Niño summers as (a) May to Oct shown in Fig (a) and (b). Anomalies of OLR and stream function at 850 hpa for April to June 2016 (around the monsoon onset) are shown in Fig (b). The anomaly pattern closely resembles that for the positive IOBW shown in Fig , with cyclonic circulation anomalies in the Indian Ocean and anticyclonic anomalies and suppressed convection over the area from Indochina to the western tropical North Pacific. (2) Influences on the global climate Some pronounced influences on the global climate from atmospheric circulation anomalies associated with the El Niño event and positive SST anomalies in the Indian Ocean are described below. (a) Suppressed precipitation over Southeast Asia Southeast Asia experienced below-normal precipitation from spring 2015 to spring 2016, which adversely affected water resource management and agriculture. In addition to the worst drought conditions for 90 years in Viet Nam (United Nations Food and Agriculture Organization), a state of W/m 2 (b) Apr. to Jun W/m 2 Fig Anomalies of outgoing longwave radiation (shading) and stream function at 850 hpa (contours) (a) May to October 2015, and (b) April to June H and L denote anticyclonic and cyclonic circulation anomalies, respectively. Contours are at intervals of 0.5 x 10 6 m 2 /s. Fig Cumulative precipitation averaged over stations in Indochina Observation stations are shown on the inset map. The red, yellow and blue lines indicate cumulative precipitation for 12-month periods starting April 2015, April 2014 and April 2011, respectively. Grey lines indicate other years after All data are from SYNOP. 62

66 emergency was declared for the Mekong Delta in relation to damage caused by sea water running up the water-deprived river (Unite Nations Country Team Viet Nam). Wildfires were frequently reported in Indonesia and Malaysia (United States National Aeronautic and Space Administration). Daily cumulative precipitation calculated from Indochina observation station data is shown in Fig for the period from April to March along with the same period in recent years for comparison. In 2015, precipitation remained below normal from around May, and cumulative precipitation for the 12-month period ending March 2016 was the lowest since Precipitation totals for the 12 months from April 2015 to March 2016 were lower than 60% of the normal for some stations in Borneo and 60 70% for stations in Indochina (Fig ). Precipitation was also below normal for the southern part of the Philippines. As mentioned previously, southwest summer monsoon activity in Southeast Asia tends to be weak during El Niño events. The anticyclonic anomalies in the lower troposphere centered over Indochina, which are considered to be responses to the weak monsoon and similar to atmospheric characteristics seen in past El Niño events (Fig (a)), were a factor behind below-normal precipitation from 2015 to (b) Heavy precipitation in the Yangtze River basin Areas along the middle and lower Yangtze River experienced above-normal precipitation starting in April Cumulative precipitation from April 1 averaged over the stations in the basin was the highest since 1997 (Fig ). Amounts soared from late June onward in particular, with the highest cumulative 30-day precipitation among the stations for June 21 to July 20 exceeding 900 mm (Fig ). More than 200 fatalities were reported in relation to heavy rainfall and landslides from late June to early July, according to the government of China. Such an extended period of extremely heavy precipitation was caused by strong convergence of moist air flow from the South China Sea over the Yangtze River (Fig ). This was induced by anticyclonic circulation anomalies over the western tropical North Pacific associated with the high SSTs in the Indian Ocean (Fig (b)). This pattern of high SSTs in the Indian Ocean, the anticyclonic circulation anomalies over the western tropical North Pacific, moist air intrusion from the Fig month precipitation anomalies for April 2015 to March 2016 Anomalies are based on CLIMAT reports and represented as ratios against the normal. Fig Cumulative precipitation averaged over stations in the middle and lower Yangtze River basin Observation stations are shown on the inset map. The red, blue and green lines indicate cumulative precipitation for the periods starting on April 1 of 2016, 1998 and 1999, and grey lines indicate the same period for all other years since The dashed black line indicates the average over the 19 years from 1997 to

67 South China Sea and water vapor convergence over southern China resembled the conditions seen in 1998 another year when the Yangtze River basin was hit by heavy precipitation. Fig day precipitation in the middle and lower Yangtze River basin The map indicates 30-day precipitation for June 21 to July 20, 2016, when particularly heavy rainfall was recorded. Red dots denote stations recording the three highest precipitation amounts for the 30-day period (Anqing, Wuhan and Macheng) and the highest amount for April 1 to July 24 (Huangshan). its maximum wind speed of 17.2 m/s or higher. This was the second-latest since 1951, and slightly earlier than the July 9 date recorded in 1998 (Table 3.1-1). The top four records in Table coincide with typhoon seasons subsequent to winter when an El Niño event reached its peak and the IOBW index remained high (Fig ). During all these typhoon seasons, pronounced anticyclonic circulation anomalies developed in the lower troposphere and convection activity was suppressed over the western tropical North Pacific as per the pattern in Fig (b). In summary, suppressed convective activity over the western North Pacific in association with high SSTs in the Indian Ocean in the wake of the El Niño event was a factor in the delayed first TC formation of Table Top 10 years of delayed TC formation Rank Year Time of first TC formation (UTC) Z, July Z, July Z, July Z, June Z, June Z, June Z, May Z, May Z, May Z, May 7 Fig Water vaper flux (arrows) and normalized divergence (shading) anomalies at 850 hpa for April to June 2016 Warm and cool colors indicate divergence and convergence anomalies, respectively. (c) Delayed formation of the season s first typhoon The first tropical cyclone (TC) of 2016 over the western North Pacific basin formed on July 3, where a TC is defined as a tropical low pressures system with Fig IOBW index changes over the last 50 years (d) Mild 2015/2016 winter in Japan In winter 2015/2016, particularly early in the season, significantly above-normal temperatures (Fig ) and below-normal snowfall were observed 64

68 across Japan. The monthly mean temperature for December averaged over eastern Japan was the highest since It was reported that the extremely low snowfall amount adversely affected the winter sports industry. Its influence extended to spring and summer, when restrictions on river water usage were put into effect because earlier-than-normal snow disappearance led to low water reserves. In the first half of winter 2015/2016, convective activity was suppressed over the Maritime Continent and anticyclonic circulation anomalies extended from the South China Sea to the seas east of Japan (Fig (a)). This anomaly pattern closely resembled the composite map in Fig (c), which depicts circulation anomaly characteristics seen in past El Niño events. Meanwhile, negative sea level pressure anomalies were seen across Eurasia, indicating a weaker-than-normal Siberian High (Fig (b)). The EU index, which is closely correlated with the intensity of the Siberian High, remained in a negative phase throughout most of December (Fig (a)). The negative phase of the EU index (the reverse of the anomaly pattern shown in Fig (b)) is consistent with the weak Siberian High and a weak cold air mass over the Eurasian continent. The thermal balance over and around Japan shown in Fig corroborates the above as factors involved in Japan s mild winter that is, southerly warm air advection associated with anticyclonic anomalies to the east of the country (Fig (a)) and temperature anomaly advection associated with the weak cold air mass over the continent (Fig (b)). It can therefore be concluded that influences from the El Niño event and the internal variability of the high-latitude atmosphere (a negative EU phase) were factors behind the higher-than-normal temperatures recorded in Japan in the first half of winter 2015/2016. Any possible relationship between the polarity of ENSO and EU still needs to be clarified. References Du, Y., L. Yang. and S.-P. Xie, 2011: Tropical Indian Ocean Influence on Northwest Pacific Tropical Cyclones in Summer following Strong El Niño. J. Climate, 24, Kobayashi, S., Y. Ota, Y. Harada, A. Ebita, M. Moriya, H. Onoda, K. Onogi, H. Kamahori, C. Kobayashi, H. Endo, K. Miyaoka and K. Takahashi, 2015: The JRA-55 Reanalysis: General Specifications and Basic Characteristics. J. Meteorol. Soc. Japan, 93, Rasmusson, E. M. and T. H. Carpenter, 1982: Variations in Tropical Sear Surface Temperature and Surface Wind Fields Associated with the Southern Oscillation/El Niño. Mon. Wea. Rev., 110, Xie, S.-P., K. Hu, J. Hafner, H. Tokinaga, Y. Du, G. Huang, and T. Sampe, 2009: Indian Ocean Capacitor Effect on Indo-Western Pacific Climate during the Summer following El Niño. J. Climate, 22, Fig Five-day running mean of area-average temperature anomalies for winter 2015/

69 (a) (a) (b) (b) (c) Fig (a) Anomalies of OLR (shading) and stream function at 850 hpa (contours) and (b) sea level pressure anomalies for Dec to Jan Arrows in (a) indicate wave activity flux at 850 hpa in units of m 2 /s 2. Contours in (a) are at intervals of 10 x 10 6 m 2 /s (thick) and 2.5 x 10 6 m 2 /s (thin). Fig (a) Daily EU index for Nov to Feb (b) Geopotential height anomalies at 500 hpa regressed onto EU indices (contours) and correlation coefficients (shading) (c) Geopotential height at 500 hpa (contours) and anomalies (shading) for Dec

70 (a) (b) Fig (a) Climatological temperature advection associated with wind anomalies, and (b) temperature anomaly advection associated with climatological winds at 925 hpa (K/day) for Dec to Jan

71 3.2 Extreme climate conditions in Japan in August 2016 Western Japan experienced hot summer conditions in August 2016, especially in the middle of the month. Meanwhile, primarily due to the approach of typhoons, monthly precipitation was the highest on record on the Pacific side of northern Japan. This section reports on surface climate characteristics and atmospheric circulation observed in August Surface climate conditions, SSTs and typhoon activity in and around Japan (1) Surface climate conditions Fig shows temperature, precipitation and sunshine duration for Japan in August 2016 as deviations from or ratios against the normal (i.e., the average). Monthly mean temperatures and sunshine durations were generally above normal all over the country. Western Japan experienced hot summer conditions, especially in mid-august, with monthly mean temperatures +0.9 C above the normal and the second-highest 10-day mean temperature for mid-august since 1961 (+1.6 C above the normal). Monthly sunshine durations against the normal on the Sea of Japan side and the Pacific side of western Japan were 131% (the second-highest since 1946) and 126% (the third-highest since 1946), respectively. Monthly precipitation amounts were below normal on the Pacific side of western Japan and in Okinawa/Amami. Meanwhile, due to rainfall from typhoons, fronts and moist air inflow, values were significantly above normal in northern Japan. The total on the Pacific side of northern Japan was the highest on record at 231% of the normal since Fig Temperature anomalies, precipitation ratios and sunshine duration ratios for August 2016 Fig day mean sea surface temperature (top) and its anomaly (bottom) for August 2016 Sea surface temperatures (unit: C) are based on the MGDSST dataset. The aqua rectangle indicates the northern part of the East China Sea (30 35 N, E). 68

72 (2) Sea surface temperature around Japan 1 As with the hot conditions in western Japan, SSTs in the northern part of the East China Sea were much higher than normal in association with greater-than-normal solar radiation and weak surface winds. Areas with SSTs exceeding 31 C were seen in mid-august (Fig ). The 10-day mean sea surface temperature in the northern part of the East China Sea in mid-august was the highest since 1982 at 29.9 C. (3) Typhoon activity in the western North Pacific Seven tropical cyclones (TCs) with maximum wind speeds of 17.2 m/s or more formed over the western North Pacific in August 2016 (Fig ). Four of them (Chanthu (T07), Mindulle (T09), Lionrock (T10) and Kompasu (T11)) made landfall on Japan in rapid succession. This was the country s highest monthly landfall total since records began in 1951 (tying with August 1962 and September 1954). Several TCs affected Hokkaido and other parts of northern Japan. Chanthu (T07) made landfall around Cape Erimo in Hokkaido on 17 August, Kompasu (T11) made landfall on Kushiro City in Hokkaido on 21 August, and Mindulle (T09) made landfall on Tateyama City in Chiba Prefecture on 22 August before moving over mainland Japan and making landfall again on the Hidaka district of Hokkaido on August 23. This was the first year in which multiple TCs made landfall on Hokkaido since Hokkaido was also affected by Conson (T06), which passed the region s Nemuro Peninsula. Lionrock (T10) was the first typhoon to make landfall on the Tohoku region from the Pacific side since Based on the Merged satellite and in-situ data Global Daily Sea Surface Temperature (MGDSST; Kurihara et. al, 2006) of JMA. Climatological normal (i.e., the average) are calculated from MGDSST and COBE-SST (JMA, 2006) datasets. Fig Tracks of tropical cyclones in August 2016 T05 T11 are TC identification numbers. The solid lines show the tracks of TCs with maximum wind speeds of 17.2 m/s or more, and the dashed lines show the tracks of tropical depressions or extratropical cyclones Atmospheric conditions (1) Hot summer conditions in western Japan The active phase of the Madden-Julian Oscillation (MJO) propagated eastward from the Maritime Continent to the Pacific during the period from the end of July to mid-august 2016 (not shown). The time-latitude cross section for OLR anomalies averaged over the E area (Fig ) indicates that an enhanced convection phase, which started to propagate northward in mid-july (Boreal Summer Intraseasonal Oscillation; BSISO), reached the area around the Philippines in August. Convective activity from this area to the sea east of the Philippines was enhanced in association with MJO and BSISO (especially in mid-august). This enhancement was also probably due in part to higher-than-normal sea surface temperatures over the same area (Fig ). Fig shows 200-hPa stream function anomalies and divergent wind anomalies, along with latitude-height cross section data for meridional wind/vertical pressure velocity anomalies averaged over the E area for 8 to 17 August In the upper troposphere, outward flow from the area over the Philippines is clearly seen in association with enhanced convective activity. The Tibetan High was stronger than normal over its northeastern part, and anticyclonic circulation anomalies were seen over 69

73 northeastern China. These two flows converged over the area from eastern China to the East China Sea (approx N), and downward flows were seen in the mid-troposphere. Fig shows vertical temperature advection at 925 hpa. This advection and greater-than-normal solar radiation were considered to be factors behind the hot summer conditions observed from eastern China to western Japan. (a) (b) Fig Time-latitude cross section for OLR anomalies averaged over the E area Fig (a) 200-hPa stream function anomalies (shading; unit: 10 6 m 2 /s) and divergent wind anomalies (vectors; unit: m/s) (b) Latitude-height cross section for meridional wind/vertical pressure velocity anomalies averaged over the E area for 8 to 17 August 2016 The green rectangle in (a) indicates the area of E and N, and the shading in (b) shows vertical pressure velocity anomalies (unit: Pa/s). Positive (negative) values denote downward (upward) flow anomalies. Vectors for the meridional wind/vertical pressure velocity anomaly are magnified x 100 vertically. Fig Monthly mean sea surface temperature anomalies for August 2016 (unit: C) Based on the MGDSST dataset Fig Advection of normal temperatures due to vertical pressure velocity anomalies at 925 hpa for 8 to 17 August 2016 (unit: K/day) 70

74 (2) Record precipitation in northern Japan Fig (a) shows the 500-hPa height field for August The westerly jet stream meandered over a wide area of the Northern Hemisphere, and was displaced northward of its normal position over and around the Kamchatka Peninsula and southward over Japan and the central Pacific. Blocking highs over western Siberia (around 60 E) were seen throughout the month, and also developed over and around the Kamchatka Peninsula from mid-august onward (Fig ). In the upper troposphere, propagation of quasi-stationary Rossby wave packets along the subtropical jet stream from the cyclonic circulation anomalies located to the south of the blocking high over western Siberia was seen, with anticyclonic circulation anomalies over northern China and the Kamchatka Peninsula (Fig (b)). Over and around western Siberia and the Kamchatka Peninsula, positive anomalies of 500-hPa height tendency associated with eddy vorticity flux were seen in the areas where anticyclonic circulation anomalies were observed (Fig (c)). This suggests that eddy-related feedback may have contributed to the development and maintenance of these highs. Fig shows stream function anomalies, wave activity flux and OLR anomalies in the upper and lower troposphere for August Convective activity was enhanced from the western North Pacific to the area near the dateline around 20 N. In response to this enhancement (a Rossby wave response), massive cyclonic circulation associated with a deep monsoon trough was seen over a wide area from the South China Sea to the south of Japan in the lower troposphere. Convective activity over the seas to the southeast of Japan ( E, N) in August 2016 was enhanced to record levels (Fig ). Intrusions of high potential vorticity (PV) air associated with the trough over the mid-latitude central Pacific (the mid-pacific trough) contributed to the enhanced convective activity. Fig shows how high PV air intruded equatorward in a southern or southwestern direction over the central Pacific. Such air also frequently intruded southward from the mid-latitudes of the central Pacific (not shown), and propagated westward over the subtropical Pacific (Fig ). Cyclonic circulation in the lower troposphere was enhanced, and tropical depressions formed west of the dateline. In this way, high PV migrating from the mid-latitudes contributed to enhanced convective activity and the formation of more tropical cyclones than normal in the central Pacific. The westerly jet stream meandered and southerly winds prevailed over the sea to the east of Japan. The Pacific High was displaced far eastward of its normal position and extended toward the south of the Kamchatka Peninsula in August 2016 in association with a persistent wave train pattern in the upper troposphere extending from Eurasia to the mid-pacific (Fig ). In the lower troposphere, propagation of quasi-stationary Rossby wave packets from cyclonic circulation anomalies over the sea to the south of Japan was seen. This may have been related to the expansion of the Pacific High toward the south of the Kamchatka Peninsula (Fig (b)). Tropical depressions forming over the sea to the southeast of Japan were upgraded to named tropical cyclones that moved northward over the sea to the east of Japan and approached or hit the northern part of the country. Lionrock (T10) followed a peculiar path, first moving southwestward over the sea south of the Kanto region and then making a U-turn over the Pacific Ocean and moving northwestward in association with the meandering westerly jet stream (Fig ). This was the first typhoon to make landfall on the Tohoku region from the Pacific side since These TCs brought a series of heavy precipitation events and serious damage to northern Japan, especially on the Pacific side. 71

75 (a) (b) Fig Time-longitude cross section showing maximum geopotential height anomalies at 500 hpa in the latitude bands between 40 and 80 N for June to August 2016 (a) (c) (b) Fig (a) 500-hPa height (contours at intervals of 60 m) and anomalies (shading) (b) 300-hPa wave activity flux (vectors; unit: m 2 /s 2 ) and stream function anomalies (contours at intervals of m 2 /s) (c) 500-hPa height tendency anomalies associated with eddy vorticity flux (shading; unit: m/day) and 500-hPa height anomalies (contours at intervals of 60 m) for August 2016 H and L in (b) represent anticyclonic and cyclonic circulation anomalies, respectively. In (c), eddies are defined as two- to eight-day band-pass-filtered fields. Fig (a) 200-hPa and (b) 850-hPa stream function anomalies (contours at intervals of (a) m 2 /s and (b) m 2 /s) and wave activity flux (vectors; unit: m 2 /s 2 ) for August 2016 Shading indicates OLR anomalies (unit: W/m 2 ). The green rectangle in (b) indicates the area of E and N. 72

76 Fig Time-series representation of OLR (unit: W/m 2 ) averaged over the area to the southeast of Japan ( E, N) for August from 1979 to 2016 Fig Monthly mean sea level pressure (contours at intervals of 4 hpa) and anomalies (shading) for August 2016 Fig Monthly mean OLR (shading; unit: W/m 2 ) and potential vorticity on the 360-K isentropic surface (contours at intervals of 1 PVU) for August 2016 Fig Time-longitude cross section for potential vorticity on the 360-K isentropic surface averaged over the N area (shading; unit: PVU) and relative vorticity at 850 hpa averaged over the N area (contours at intervals of 10-6 /s; shown for /s or more) for August 2016 The blue dots represent genesis points of tropical depressions later upgraded to named TCs. T16xx expresses TC identification numbers Summary The atmospheric circulation conditions discussed here are summarized in Fig In the upper troposphere, propagation of quasi-stationary Rossby wave packets from cyclonic circulation anomalies located to the south of the blocking high over western Siberia was seen. The westerly jet stream meandered over a wide area, ridges were seen over north China and the Kamchatka Peninsula, and troughs were seen over Japan and the central Pacific. In association with intrusions of high PV air from the trough over the mid-latitude central Pacific, convective activity was enhanced from the area southeast of Japan to the area near the dateline at around 20 N. In response to this enhancement (a Rossby wave response), massive cyclonic circulation was seen over the sea to the south of Japan in the lower troposphere. The Pacific High was displaced far eastward of its normal position and extended toward the south of the Kamchatka Peninsula. The blocking highs over and around the Kamchatka Peninsula as well as the propagation of quasi-stationary Rossby wave packets from cyclonic circulation anomalies over the sea to the south of Japan in the lower troposphere may have contributed to this extension. 73

77 Tropical depressions forming over the sea to the southeast of Japan were upgraded to named tropical cyclones (TCs) that moved northward over the sea to the east of Japan and brought a series of heavy precipitation events and serious damage to northern Japan. In association with enhanced convective activity over and around the Philippines and the stronger-than-normal Tibetan High over northeastern China, downward flows were seen from eastern China to western Japan in the mid-troposphere. This vertical advection and greater-than-normal solar radiation brought hot summer conditions to western Japan. Fig Characteristics of atmospheric circulation associated with extreme climate conditions in Japan in August 2016 References JMA, 2006: Characteristics of Global Sea Surface Temperature Data (COBE-SST), Monthly Report on Climate System, Separated Volume No. 12. Kurihara, Y., Sakurai, T., and Kuragano, T., 2006: Global daily sea surface temperature analysis using data from satellite microwave radiometer, satellite infrared radiometer and in-situ observations (in Japanese), Weather Service Bulletin, Vol. 73, S1 S18. 74