What caused the cool summer over northern Central Asia, East Asia and central North America during 2009?

Cool and wet weather conditions hit northern Central Asia, East Asia and central North America during the 2009 summer in concert with a strong jet stream and a prominent meandering upper-level circulation in the Northern Hemisphere mid-latitudes despite the fact that the year 2009 is the fifth warmest year globally in the modern record. It is found that the conspicuous atmospheric variability in the entire Northern Hemisphere mid-latitudes during the summer of 2009 was caused by a combination of teleconnections associated with significant tropical thermal forcings, strong polar forcing, and interaction between high-frequency weather events and climate anomalies. The strong negative circumglobal teleconnection pattern associated with the deficient Indian summer monsoon rainfall and developing El Niño condition was the major contributor to the cool and wet summer in June. On the other hand, the July weather conditions were attributable to the high-latitude impact of the unprecedented negative Arctic Oscillation, together with the Rossby wave response to the subtropical heating generated by convective activities over the Western North Pacific summer monsoon region. It is also noted that enhanced storm track activity and frequent cold surges from high-latitudes may have played a role in the cool and wet summer over the regions of interest.


Introduction
The year 2009 has been ranked the fifth warmest year globally since the beginning of instrumental records in 1850, and the first decade of the 21st century  was the warmest decade on record (WMO 2009, Willett et al 2010. In spite of the warming trend, many parts of the Northern Hemisphere (NH) mid-latitude including northern Central Asia, East Asia and central North America experienced a cool and chilly summer accompanied by above-normal precipitation in 2009 ( figure 1(a)).
India suffered one of its worst droughts with a seasonal deficit of 23% (Ratnam et al 2010) and most of Europe was hit by a heat wave setting new maximum temperature records (WMO 2009). On the other hand, several regions in the central United States (US) experienced their coolest July in 115 years of records and many parts of Asia had heavy rains and floods along with below-normal temperatures (Arndt et al 2010). Unfortunately, most of the global climate models that participate in the real-time seasonal climate prediction at the Asia-Pacific Economic Cooperation Climate Center failed to predict the extreme climate anomalies over most of the NH mid-latitude even one month ahead. According to a report on the APEC Climate Center (APCC) Multi Model Ensemble (MME) prediction, the MME system failed to capture the El Niño developing during JJA 2009, but captured, to some extent, the dipole pattern of the observed precipitation anomaly characterized by enhanced precipitation over the Western North Pacific and Philippine Sea and reduced precipitation over the Indian monsoon region and the South China Sea (Wang et al 2010).
The extreme climate anomalies and weather events resulted in costly reparation for agriculture, transportation and other socio-economic activities. Investigation of the causes and underlying processes of the phenomena would improve our understanding of their origins and sources of predictability, and then eventually reduce future socio-economic losses by managing weather and climate risk. In particular, the cold summer of 2009 occurred against a global warming trend. Since most climate models tend to project continuous warming and increase of monsoon precipitation in the 21st century , it is of particular importance to understand the sources of the cold summer of 2009.
This study investigates the characteristics of the cool and wet summer over many parts of the NH extratropics in 2009 with particular attention to the following questions. (1) What are the dominant climate factors responsible for the cool and wet summer? (2) How does each factor contribute to the regional climate anomalies and extreme events over Asia and North America?
The observed data used are obtained from the Climate Prediction Center (CPC) Merged Analysis of Precipitation (CMAP) data set (Xie and Arkin 1997) for precipitation (PRCP) and from the NCEP/Department of Energy (DOE) reanalysis II (NCEP R2) data (Kanamitsu et al 2002) for 2 m air temperature (2 m TMP), geopotential height at 200 hPa (Z200), zonal wind at 200 (U200) and 850 hPa (U850), meridional wind at 850 hPa and temperature at 850 hPa, from 1979 to 2010. The Nino 3.4 (5 • S-5 • N, 170 • -120 • W) SST index is obtained from the NCEP/Climate Prediction Center website. The AO index, defined as the first leading EOF mode of monthly mean height anomalies at 1000 hPa over the NH extratropics, comes from the NOAA web site: www. esrl.noaa.gov/psd/data/correlation/ao.data. The CGT index is defined as the interannual variability of Z200 averaged over the northwest of India (35 • -40 • N, 60 • -70 • E) using NCEP R2 (Ding and Wang 2005). The ISM index (ISMI) and WNPSM index (WNPSMI) are defined as follows Fan 1999, Wang et al 2004): where the areas in parentheses denote the regions over which U 850 is averaged. The anomalies of all variables were calculated based on 1979-2010 climatology. shows that a conspicuous wave train was dominant over the entire NH mid-latitude all through the troposphere during June-July (JJ) in 2009. Since the climate anomalies during August were significantly different from those during JJ, we focus on the JJ climate anomalies. The wave pattern had a zonal wavenumber-5 and barotropic structure associated with a waveguide induced by the enhanced westerly jet stream. In the mid-latitude continents, three prominent variability centers of barotropic cyclonic circulation are located over northern Central Asia

Characteristics of the 2009 summer
along with cold temperature and positive precipitation anomalies ( figure 1(a)). In many parts of the regions, the temperature anomaly was lower than −2 standard deviation (SD) and the precipitation anomaly was higher than 2 SD of its interannual variation for the 32 years of 1979-2010.
The normalized time series of the temperature anomalies averaged over the three individual regions reveal that 2009 summer tied for the second coldest year over the nEA and the fourth coldest year over the nCA and ranked as the third coldest year over the cNA since 1979 ( figure 2(a)). In particular, the cold temperature anomaly over nEA was The aforementioned climate anomalies in the NH mid-latitudes were concurrent with equatorial Central Pacific warming (El Niño), and strong negative phases of AO and CGT (figure 2(b)), suggesting that tropical thermal forcing, high-latitude impact via AO and mid-latitude teleconnection were very strong during the time of interest. The El Niño developed from early summer of 2009 and the Nino 3.4 SST index anomaly was 1 SD in JJ. The AO and CGT anomalies reached −2.8 and −2.2 SD, respectively. The strong negative phase of AO represents above-normal temperature over the Arctic but below-normal temperature over the mid-latitudes. The CGT has been demonstrated as a dominant summertime teleconnection pattern having a zonal wavenumber-5 structure and geographically phase locked to preferred locations including the northeast Atlantic, western Europe, western Central Asia, East Asia, the Gulf of Alaska and northern North America. Ding and Wang (2005) considered the ISM as a primary forcing to CGT. An enhanced ISM initially generates an upper-level anomalous high to its northwest over west-central Asia, and then excites successive downstream cells along the waveguide through Rossby wave dispersion. The negative CGT can be excited by deficient ISM rainfall and deficient ISM is often induced by developing El Niños (DW05, Lee et al 2011). Table 1 shows the monthly anomalies of Nino 3.4, CGT, AO and two monsoon circulation indices normalized by their own standard deviations from May to August 2009. While moderate to strong Pacific warming persisted during whole summer, the CGT and AO indices exhibited large month-to-month variation. The CGT was the lowest (−2.5 SD) in June but the AO was the lowest (−3.0 SD) in July. In August, the CGT and AO were near normal. The ISMI was −2.3 SD in June, indicating a severely deficient summer monsoon, but it was near normal in July. On the other hand, the WNPSM was very active in July with 1.1 SD of its circulation index.

Possible causes
The previous section indicates that the 2009 summer climate was symphonized by strong tropical thermal forcing, monsoonal heat sources, high-latitude impact via AO and mid-latitude wave activities along with an above-normal jet stream. This section is devoted to understanding how those climate factors were coordinated with each other to result in the cold summer over Asia and North America in 2009. Composite analysis is adopted to better understand the relative role of the developing El Niño and the negative phases of the AO and CGT (figure 3). The developing El Niño composite is obtained from five summers in 1982, 1991, 1997, 2002and 2006, the negative CGT composite from seven summers in 1981, 1982, 1987, 1989, 2002, and the negative AO composite from six summers in 1982, 1987, 1993, 1997, 1998and 2007 figure 2(b).
During the last few decades, El Niño developing summers were characterized by zonally elongated negative temperature anomalies over EA and the North Pacific and positive ones over mid-latitude NA ( figure 3(a)). The Pacific East Asian teleconnection pattern through the bridge of the WNP anomaly (Wang and Zhang 2002) suggests that the developing El Niño contributes, in part, to nEA and some part of nCA but not to cNA.
The negative CGT composite pattern obtained in this study ( figure 3(b)) is, to a certain extent, different from the one shown in DW05 that was obtained from the 56 yr data without the long-term trend and decadal variation. It is interesting to note that the 2009 summer pattern ( figure 1(a)) more resembles DW05's pattern rather than figure 3(b) for the following two reasons. First, DW05's negative CGT pattern mainly emerged due to deficient ISM rainfall and a developing El Niño which is the case for summer 2009 while the negative CGT composite in this study mainly emerged due to deficient ISM without an El Niño. This is because during recent decades the negative correlation between ISM and ENSO has weakened (Kumar et al 1999). Second, the weaker trough over EA and NA and the stronger ridge over the North Pacific and the Gulf of Alaska in figure 3(b) (compared to DW05's result) reflect the global warming effect on the CGT pattern during recent decades that is not the case for the summer of 2009 (Lee et al 2011. Taken together, it is suggested that the cool summer over nCA, nEA and cNA is, in part, attributable to the negative phase of CGT emerging due to deficient ISM and developing El Niño. By definition, the negative AO composite is characterized by a zonally symmetric pattern of polar warming and mid-latitude cooling ( figure 3(c)). It indicates that the negative AO mainly contributes to cooling of EA and western and central NA but not to nCA during the last few decades.
Further analysis on the monthly anomaly fields reveals that the negative CGT pattern associated with deficient ISM rainfall and El Niño developing was a major contributor to the cold summer in June but the western North Pacific-East Asia-North America teleconnection associated with the abundant WNPSM rainfall and negative AO contributed to the cold summer in July (see table 1). In August, the El Niño

Discussion
The 2009 summer climate anomaly was a unique case caused by a combination of a developing El Niño, the negative phases of CGT and AO, severely deficient ISM and enhanced WNPSM rainfall in recent decades. That is because the negative relationship between the ISM rainfall and ENSO has weakened during recent decades and the CGT and AO have no significant correlation (DW05, Wang et al 2012). Since 1979, just the summer of 1982 had similar climate conditions. However, during that summer, El Niño and AO effects were dominant, resulting in a cool summer in NA but a normal to hot summer over CA and NA (not shown).
We also note that the interaction between weather and climate played an important role in the cool and chilly summer of 2009 over nCA, nEA and cNA. The interactions of the mean flow with high-frequency transients make an environment that favors development of synoptic events, moisture convergence, more convective storms and summer rainfall (Hu and Feng 2010). Figure 4 shows the climatology and anomaly in JJ 2009 of 200 hPa storm track activity and 850 hPa meridional heat flux by synoptic eddies. The storm track activity is defined as the root mean square of variance of 8-day filtered Z200. It is shown that an upper-level storm track was active along with frequent cold surges over nCA, nEA and cNA since a northerly was prevailing over the regions of interest. The impact of frequent cold surge and storm track activity seems to make the summer of 2009 unique because the impact was very weak in the summer of 1982 (not shown). The impact of interaction between weather and climate on extreme climate events deserves further study.
The monthly anomaly analysis indicates that the June-July mean state instead of the JJA mean is more robust to the described cool summer climate in 2009, which might be linked to the distinct changes of AO and CGT between June-July and August.
In 1992, the temperature patterns were similar to those in 2009 but the climate indices showed different anomaly patterns. It is found that the negative CGT pattern was most dominant during the 1992 summer among other climate factors. Associated with the negative CGT pattern, most of Asia and North America experienced cold temperatures.
However, the impact of high latitude seemed to be very weak in 1992.
Most of the global climate models that participated in the real-time seasonal climate prediction at the Asia-Pacific Economic Cooperation Climate Center failed to predict the extreme climate anomalies over most of the NH mid-latitude even one month ahead (Wang et al 2010). Further study will address deficiencies and possibilities of the global climate models in predicting the conspicuous climate anomalies addressed in this study.