Pacific and Atlantic Ocean influence on the spatiotemporal variability of heavy precipitation in the western United States

Global and Planetary Change 109 (2013) 38–45

Contents lists available at ScienceDirect

Global and Planetary Change journal homepage: www.elsevier.com/locate/gloplacha

Pacific and Atlantic Ocean influence on the spatiotemporal variability of heavy precipitation in the western United States Peng Jiang a, Zhongbo Yu a,⁎, Mahesh R. Gautam b a b

Department of Geoscience, University of Nevada Las Vegas, Las Vegas, NV 89154, USA Division of Flood Management, California Department of Water Resources, Sacramento, CA 95821, USA

a r t i c l e

i n f o

Article history: Received 12 November 2012 Revised 6 June 2013 Accepted 8 July 2013 Available online 16 July 2013 Keywords: Spatiotemporal variability Heavy precipitation Pacific and Atlantic Ocean EOF analysis Western United States

a b s t r a c t In this study, we test our hypothesis that no single index such as El Niño-Southern Oscillation (ENSO), Pacific Decadal Oscillation (PDO), Atlantic Multi-decadal Oscillation (AMO) or North Atlantic Oscillation (NAO) derived from the Pacific and Atlantic Oceans can explain the multi-scale temporal variability and spatial distribution of heavy precipitation in the western United States. Instead, it may be possible to utilize a characterization of their integrated effect or some other unidentified factors which reflects the combined physical oceanic–atmospheric processes that occur. For this purpose, Empirical orthogonal function (EOF) analysis is performed on summer (April–September) and winter (October–March) heavy precipitation expressed as total precipitation when daily precipitation is larger than 95th percentile (R95) to indentify the leading modes of variability during the period 1948–2009. The correlation between the principle components (PCs) of each EOF mode with Sea Surface Temperature (SST) anomalies is evaluated. The analysis has shown that the leading modes of R95 variability and the connections between local R95 and SST over western United States are seasonally dependent. The first EOF mode of summer R95 is associated with AMO. The first two EOF modes of winter R95 are related to an integrated effects of ENSO, PDO, and NAO which explain nearly half (49%) of the spatial and temporal variance in R95 in this region. Additionally, the coupled effects of these three oceanic–atmospheric oscillations on winter R95 are evaluated by investigating the ENSO-R95 responses modulated by a combination of different PDO and NAO phases. Based on our analyses and predicted future states of these oceanic–atmospheric oscillations, we suggest possible heavy precipitation scenarios for upcoming decades which may be useful to forecasters and water managers. © 2013 Elsevier B.V. All rights reserved.

1. Introduction The western United States is undergoing rapidly changing social dynamics, pressure from an expanding population and a greater risk of water shortage and flooding (Piechota et al., 2004; Mote et al., 2005; Hamlet and Lettenmaier, 2007). As a result, the system becomes more vulnerable to the climatic extremes. A long-term view into the spatiotemporal pattern of precipitation extremes is expected to help plan for flood disasters (Yang et al., 2010). Several large-scale oceanic and atmospheric oscillations are thought to affect the delivery of moisture to the United States, thus influencing the spatiotemporal distribution of precipitation extremes. The predictability of future precipitation extreme scenarios is therefore possible if we can thoroughly understand the relationship between spatiotemporal variability of precipitation extremes and large-scale ocean oscillations, and will help us prepare for

⁎ Corresponding author at: Department of Geoscience, University of Nevada Las Vegas, 4505 S. Maryland Pkwy, Las Vegas, NV 89154-4010, USA. Tel.: +1 702 895 2447. E-mail address: [email protected] (Z. Yu). 0921-8181/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.gloplacha.2013.07.004

the future possible extreme precipitation scenarios in the west United States, which can be useful to water resources engineers and managers. Recent research regarding the influences of quasi-periodic variations in sea surface variables has shown growing promise for long-range probabilistic forecasts of precipitation extremes. Several oceanic– atmospheric indices are used to explain the variability of precipitation extremes in the United States. The best known of these is ENSO with two basic phases (warm phase: El Niño; cold phase: La Niña) of the tropical eastern Pacific Ocean. ENSO helps to explain the occurrence of heavy winter precipitations at inter-annual temporal scales in the western United States (Cayan et al., 1999; Meehl et al., 2007). The extreme precipitation is sensitive to the ENSO phase: Most of the southwest United States experiences more than double heavy precipitation events during El Niño years compared to La Niña years (Cayan et al., 1999). The PDO is another well-known oceanic index, which may help to depict the decadal variability of precipitation extremes. It is distinguished by warm and cold phases at decadal-scale periods of the North Pacific Ocean north of 20°N (Mantua et al., 1997). Correspondence between extreme events and PDO was investigated by Hidalgo (2004) suggesting that PDO phase is correlated with aboveand below-average precipitation in the Colorado River Basin. Other

J. Peng et al. / Global and Planetary Change 109 (2013) 38–45

available studies focus on the modulation of PDO cycle on the ENSO-precipitation signal (McCabe and Dettinger, 1999; Gutzler et al., 2002; Brown and Comrie, 2004; Kurtzman and Scanlon, 2007; Arriaga-Ramírez and Cavazos, 2010; Wise, 2010; Timm et al., 2011; Cai and van Rensch, 2012). Besides the two well-known indices, the recently developed index of the AMO is also found to be associated with the multi-decadal variability of boreal summer precipitation and extreme events (Enfield et al., 2001; Sutton and Hodson, 2005; Wang et al., 2006; Curtis, 2008; Mo et al., 2009). This long-term fluctuation of SST in the North Atlantic Ocean exhibits a multi-decadal shift between warm and cool periods with each lasting about 30 years (Kerr, 2000; Enfield et al., 2001). The warm phase (AMO+) is accompanied by a reduced rainfall over most of the United States (Enfield et al., 2001) including northwest United States and Great Plains (Wang et al., 2006) and an increase in precipitation intensity in southwest and coastal southeast United States (Curtis, 2008). The cool phase, on the contrary, has an opposite relationship with precipitation and extremes in these regions (Feng et al., 2011; Oglesby et al., 2012). Another long term oscillation associated with North Atlantic Ocean is NAO which is an atmospheric oscillation with two action centers located near Iceland and subtropical Atlantic. This pattern is identified in northern Europe, and the eastern United States. It is found to affect extreme precipitation regionally, such as on the eastern seaboard of the United States (Zhang et al., 2010). However, regions far away from the northern Atlantic Ocean also experience significant NAOrelated impact (Visbeck et al., 2001). Most current studies focus on the relationship between total precipitation and the large-scale oceanic–atmospheric oscillations. Changes in the total precipitation are responsible for disproportionate changes in precipitation extremes, but they don't always go in the same trend (Easterling et al., 2000). As a result, independent studies into the spatiotemporal variability of precipitation extremes shaped by the large-scale ocean oscillation are necessary. However, there have been relatively few studies regarding the spatiotemporal variability of precipitation extremes caused by multi-scale temporal fluctuations of SST in both Pacific Ocean and North Atlantic Ocean such as PDO and AMO. Available studies mostly investigate the impacts of large-scale oceanic–atmospheric oscillations especially the impacts of ENSO on precipitation extremes (Cayan et al., 1999; Meehl et al., 2007). However, it is highly possible that spatiotemporal variations in the occurrence of precipitation extremes in the western United States involve complex interactions between the Atlantic and Pacific Oceans (represented by ENSO, PDO, AMO, and NAO in this paper). We hypothesize that no single feature from them can explain all the spatiotemporal variations: it is the integrated effect of them or some other unidentified factors that control the multi-scale temporal variability and spatial distributions of precipitation extremes in this region. In this paper, we strive to offer a comprehensive analysis of impacts from Pacific and North Atlantic Ocean on the spatiotemporal variability of precipitation extremes in western United States. We seek here to, (1) determine how much variance of seasonal (summer period: April–September; winter period: October–March) precipitation extremes these ocean oscillations including ENSO, PDO, AMO, and NAO explain; (2) understand the integrated impacts from Pacific and North Atlantic Ocean on multi-scale temporal variability and spatial distributions of seasonal precipitation extremes; and (3) explore possible extreme precipitation scenarios for the upcoming decade based on the projected conditions of the three oceanic indices (Latif and Barnett, 1996; Yeh et al., 2009; Lapp et al., 2012). 2. Data and methods Analyses for this study are based on NOAA Climate Prediction Center (CPC) Daily US Unified Precipitation data (available from the

39

NOAA/OAR/ESRL PSD, Boulder, Colorado, USA, from their website at http://www.esrl.noaa.gov/psd/). This dataset is derived from 3 sources including NOAA's National Climate Data Center (NCDC) daily co-op stations from 1948, CPC dataset (River Forecast Centers data 1st order stations from 1992), and daily accumulations from hourly precipitation dataset from 1948. The daily data were gridded at a horizontal resolution of 0.25° × 0.25° using inverse-distance weighting interpolation algorithms of Cressman (1959). Quality control methods such as “duplicate station check”, “buddy check”, and standard deviation check were conducted against the dataset to exclude the keypunch errors and extreme values errors and to compare it with daily rain gauge data correspondingly (Higgins and Center, 2000). Although the station density used for this dataset is sparser in the west than the eastern two-thirds of the United States, the general coverage over this region could be considered (Chen et al., 2008; Higgins et al., 2008). Five key indices (Table 1) suggested by Frich et al. (2002) are usually chosen for the analysis of changing precipitation patterns. The five indices cover changes in intensity, frequency and duration of precipitation events and belong to 5 different categories (Alexander et al., 2006; Yang et al., 2011): (1) percentile-based indices (R95), (2) absolute indices representing maximum or minimum values within a season or year (R5D), (3) threshold indices defined as the number of days on which a temperature or precipitation value falls above or below a fixed threshold (R10), (4) duration indices representing periods of excessive warmth, cold, wetness or dryness or in the case of growing season length, period of mildness (CDD), and (5) other indices such as intensity index (SDII). R95 is used to measure heavy precipitation that exceeds 95 percentile thresholds which covers, but does not limit to most extreme precipitation events in a year (Alexander et al., 2006). R5D represents very heavy precipitation. R10 calculates the annual count of days when daily precipitation is larger than10 mm. Threshold indices such as R10 are region dependent. R10 is not necessarily meaningful in a global scope especially in the Intermountain West with a heterogeneous precipitation (Alexander et al., 2006). CDD represents the length of the longest dry period in a year and focuses more directly on the evaluation of droughts. SDII accounts for both the total amount of annual precipitation and the number of days when rainfall exceeds 1 mm. Rather than capturing the tail of the distribution such as R95 and R5D, SDII is more likely to show the middle of the distribution. Given the great spatial variability of precipitation in the West, R95 is selected for the calculation of the seasonal extreme precipitation based on the CPC daily precipitation data. EOF analysis (Lorenz, 1956; Preisendorfer, 1988; Roxy et al., 2013) is applied to the seasonal R95 for finding out the leading modes of variability. This method is a decomposition of dataset by orthogonal functions. It investigates the variability of a single field such as one scalar variable (Temperature, Precipitation, etc). The method finds both spatial and temporal patterns of variability, and measures the “importance” or “contribution” of each pattern. In this paper, the EOF analysis is used to understand the spatiotemporal structure of the long-term variations of the seasonal precipitation extremes over

Table 1 Five indices of precipitation extremes as described by Frich et al. (2002).a Index

Definitions

Units

R10 CDD R5D SDII R95

Total count of days when RR ≥ 10 mm Maximum number of consecutive dry days with RR b 1 mm Maximum 5-day precipitation total Total precipitation divided by the number of wet days Total precipitation when RR N 95th percentile

days days mm mm/day mm

a Abbreviations are as follows: RR, daily precipitation. A wet day is defined when RR ≥ 1 mm, and a dry day is defined when RR b 1 mm.

40

J. Peng et al. / Global and Planetary Change 109 (2013) 38–45

western United States. The EOF analysis is conducted based on the covariance matrix to produce a new dataset by its eigenvectors. These EOFs reveal the spatial patterns of its associated PCs, which provide the information of their temporal variability. During the analysis, the grid data is weighted by the square root of cosine of latitude to consider the latitudinal distortions; the rotated EOF analysis is conducted after the regular EOF analysis using the Varimax rotation (Richman, 1986) for a simpler and more physically explained structure. A rule of thumb introduced by North et al. (1982) is used to test the independence of rotated eigenvectors, which also provide criteria for deciding the numbers of EOFs retained. To investigate the dynamical context of the leading modes in R95 variability, we have used Kaplan SST V2 data (Kaplan et al., 1998) provided by the NOAA/OAR/ESRL PSD, Boulder, Colorado, USA, from their website at http://www.esrl.noaa.gov/psd/. The data is stored on a 5° × 5° grid and consists of monthly anomalies from 1856–present. In the present study, we calculate the annual average SST anomalies from the monthly values for 1948–2009. A correlation map is generated by calculating the correlation coefficient between the time series of each grid SST anomalies and PCs. Monthly values of Southern Oscillation Index (SOI), PDO, AMO, and NAO are obtained from the University of East Anglia's Climatic Research Unit, the University of Washington's Joint Institute for the Study of the Atmosphere and Ocean, the National Oceanic and Atmospheric Administration's Climate Prediction Center, and the National Center for Atmospheric Research's (NCAR's) Climate Analysis Section (Hurrell, 1995), respectively. The period of analysis includes three AMO phases (1948–1963 [AMO+], 1964–1994 [AMO −], and 1995– 2009 [AMO+]), three PDO phases, and five NAO phases (Table 2). To investigate the integrate effect of ENSO, PDO, and NAO on winter R95, the correlations between June and November SOI and winter R95 are calculated on each grid for the combinations of PDO and NAO phases listed in Table 2. Winter R95 anomalies are calculated for SOI/PDO/NAO phase combination subsets. Statistical significance of the anomaly patterns is tested using a permutation resampling method described by Wise (2010).

3. Results and discussions 3.1. Leading modes of the seasonal precipitation extremes The EOF modes of R95 are evaluated for both summer and winter period. We retain the first two EOF modes for our analysis, as the independent test (North et al., 1982) shows that the first two eigenvector are independent (Fig. 1). For the summer period, we limit our analysis to the first EOF mode, because the second mode explains only 7% of the total R95 variance.

The first mode of summer R95 explains 11% of the total variance (Fig. 2a). The correlation map (Fig. 2a) shows a northeastern– southwestern dipolar pattern with the strongest correlation over upper Mississippi River Basin and lower Colorado River Basin. It is correlated with SST anomalies over the northern Atlantic Ocean (Fig. 4a). Also, the corresponding PC is statistically significantly correlated to annual average AMO index (r = −0.38) (Fig. 3a). The SST anomalies correlation map and the significant correlation between the corresponding PC and AMO index imply that the first leading mode of the summer R95 may be associated with AMO. During AMO cold/warm phase (expressed as negative/positive AMO index), the upper Missouri River Basin and the Great Basin experience an increase/reduction of extreme precipitation events while the southwest, mainly centered at lower Colorado River Basin, exhibits a reduction/increase of extreme precipitation events. This is consistent with the results from Curtis (2008). For the winter period, the leading modes of R95 present more complex spatial and temporal patterns. The first EOF mode accounts for 35% of the total variance of winter R95. The respective spatial pattern (Fig. 2b) depicts a north-west dipolar pattern with one action center located on the northwest and another on the southwest. This mode is related to a complex pattern of SST anomalies over both Pacific Ocean and North Atlantic Ocean (Fig. 4b). It has a significant negative correlation with SST anomalies over North Pacific Ocean north of 20°N (considered as PDO SST field) and a significant positive correlation with SST anomalies over central Pacific Ocean which is similar to the ENSO SST pattern. Besides, it is also correlated with the North Atlantic tripole SST pattern described by Wu and Liu (2005) and Fan and Schneider (2012) (Fig. 4b). The tripole SST pattern may be associated with NAO through Air–Sea coupling (Peng et al., 2003; Wu and Liu, 2005). So, hereafter we use NAO to represent the North Atlantic tripole SST pattern. In other words, the first mode of winter R95 is associated with NAO. Correspondingly, its PC is statistically significantly correlated to June–November SOI, October–March PDO index, and annual NAO index (r = − 0.44/ 0.54/0.23) at 0.1 level (Fig. 3b). It further confirms that the northern–southern dipolar pattern of winter R95 is not caused by a single oceanic pattern, but an integrate effect of ENSO, PDO, and NAO. The second EOF mode of winter R95 explains 13.8% of the total variance of winter R95. The spatial pattern of this mode (Fig. 2c) is characterized by coherent R95 variations over Pacific Northwest Basin, northern and central California Basins which extends east to the Great Basin and Colorado River Basin headwater area and

Table 2 Positive (+) and negative (−) PDO and NAO periods and combinations. PDO phases

Period

NAO phases

Period

Combinations

Period

PDO−

1948–1976

PDO+

1977–1998

NAO+ NAO− NAO+ NAO− NAO+

1948–1952 1953–1972 1973–1976 1977–1979 1980–2009

PDO−

1999–2009

PDO−/NAO+ PDO−/NAO− PDO−/NAO+ PDO+/NAO− PDO+/NAO+ PDO−/NAO+

1948–1952 1953–1972 1973–1976 1977–1979 1980–1998 1999–2009

Four combinations

Period

PDO−/NAO+ PDO−/NAO− PDO+/NAO− PDO+/NAO+

1948–1952, 1973–1976, 1999–2009 1953–1972 1977–1979 1980–1998

Fig. 1. Independence test for the rotated eigenvectors of R95 during summer and winter.

J. Peng et al. / Global and Planetary Change 109 (2013) 38–45

41

Fig. 2. Spatial patterns (correlation maps) of (a) the first EOF mode of the summer, (b) the first, and (c) the second EOF mode of the winter R95 (1948–2009). Correlations significant at the 0.05 level are highlighted and contoured.

opposite variations over a minor center located on the southeastern corner of western United States. The teleconnection pattern between this mode and SST anomalies is similar with the one between the first mode of winter R95 and SST anomalies. However, the region in the central Pacific Ocean having significant relationship with the second mode is relatively small and is limited to the eastern central Pacific Ocean compared to the central Pacific Region of the first mode. The corresponding PC is statistically significantly correlated to October– March PDO index and annual NAO index (r = − 0.38/− 0.39), but not to the SOI. It implies that the second mode of winter R95 is controlled by both PDO and NAO.

Fig. 3. Principal components of (a) the first EOF mode of the summer, (b) the first, and (c) the second EOF mode of the winter R95 (1948–2009). Solid line depicts the inverse AMO index in Fig. 2a, dash line depicts the inverse SOI index in Fig. 2b, dash-dot line depicts the PDO index in Fig. 2b and inverse PDO index in Fig. 2c, and dot line depicts the NAO index in Fig. 2b and inverse NAO index in Fig. 2c.

3.2. Integrated impacts from Pacific Ocean and North Atlantic Ocean on winter precipitation extremes The EOF analysis reveals that the winter R95 is controlled by an integrate influence of ENSO, PDO, and NAO. To evaluate the coupling effects, four periods are indentified by considering all the possible

Fig. 4. Correlation field between annual SST anomalies and the time series of (a) the first rotated EOF mode of the summer R95, (b) the first, and (c) the second rotated EOF modes of the winter R95. Correlations significant at the 0.05 level are highlighted and contoured.

42

J. Peng et al. / Global and Planetary Change 109 (2013) 38–45

combinations of the PDO and NAO phases (Table 2). During each period, El Niño years (SOI b 0) and La Nina years (SOI N 0) are indentified by average June–November SOI. The combination of PDO+/NAO− is not reported as it covers a very short period of 3 years which may introduce random errors to the diagnostics. The correlation map between winter R95 and June–November SOI exhibits a north–south dipolar pattern (Fig. 5a). Winter R95 across the Pacific Northwest (except the low-lying region in the eastern Washington) and Southwest is significantly correlated with SOI. The region which separates the West into northern and southern centers of the dipole mainly covers the southern California Basin, the Great Basin, and most part of upper Colorado River Basin. This region is defined as “transition zone”. Within the transition zone, the correlations between R95 and SOI are near-zero which indicates that prediction of precipitation extremes based on the forecasting of ENSO states is difficult. The winter R95 dipole transition pattern is similar to the winter precipitation dipole transition pattern described by Wise (2010). However, obvious differences exist as the transition zone of R95 is much wider than that of total precipitation. For example, the total winter precipitation in most part of Great Basin is significantly related to SOI while the winter R95 is not. The phenomenon also highlights the fact that changes in extreme precipitation doesn't always go in the same direction with the changes in total precipitation. In the periods of different PDO/NAO combinations, the location and area of the R95-SOI action center changes: during PDO −/NAO+ period, the southern center is not obvious and is shrunk to a small area in the upper Colorado River Basin (Fig. 5b); during PDO−/NAO− period, the R95 southern action center mainly lies over lower Colorado River Basin (Fig. 5c); during PDO+/NAO+ period, the northern center

almost disappears, and the southern center is located at southern California Basin (Fig. 5d). The large-scale oceanic–atmospheric oscillations will introduce different pressure patterns which control the storm track position and impact the spatial patterns of precipitation. El Niño (SOI −) and PDO warm phase are accompanied with a deep Aleutian low which forces the storm track southwards and results in a higher probability of heavy precipitation (Gershunov and Barnett, 1998). In contrast, during La Nina (SOI+) years and PDO cold phase, the Pacific jet stream carves north and enters the North America through Pacific Northwest, bringing more heavy precipitation to this region. NAO pattern is not identified in winter precipitation patterns in the western United States by previous studies; however, it is associated with pressure difference between Azores and Iceland which changes the strength of westerlies across the mid-latitudes (Visbeck et al., 2001) and thus, can have a modulation effect on precipitation pattern far away from North Atlantic region. Their impacts on winter R95 could be seen by plotting the standard winter R95 anomalies for combinations of PDO/NAO/ENSO phases (Fig. 6). The winter R95 anomalies region looks sparsely distributed only when we draw contour lines to indicate the significant level. There is a dipolar pattern in Fig. 6b, d, and e where the PDO is in a constructive phase with ENSO. This pattern becomes more obvious when both PDO and NAO are in the opposite phase with ENSO (Fig. 6d, and e). This is consistent with the mechanisms for the precipitation responses to the oceanic–atmospheric oscillations discussed above. The dipolar pattern is weak and even disappears in other combinations (Fig. 6a, c, and f) as their effects on precipitation extremes may offset each other when PDO/NAO is not in the same phase with ENSO.

Fig. 5. Correlation coefficients between June and November SOI and winter R95 for (a) the whole period 1948–2009, (b) PDO−/NAO+ period, (c) PDO−/NAO− period, and (d) PDO+/NAO+ period. Correlations significant at the 0.05 level are highlighted and contoured.

J. Peng et al. / Global and Planetary Change 109 (2013) 38–45

43

Fig. 6. Standard winter R95 anomalies for PDO/NAO/ENSO phases. Anomalies significant at the 0.1 level are highlighted and contoured.

3.3. Implication for possible extreme precipitation scenarios for the upcoming decades Prediction of future oceanic–atmospheric oscillation states may help explain part of extreme precipitation variations for the upcoming decades. As found in this study and others (Sutton and Hodson, 2005; Wang et al., 2006; Curtis, 2008; Hu et al., 2011), summer extreme precipitation is associated with AMO. Recent studies suggest that the AMO returned to its positive phase in 1995 (Enfield et al., 2001) and the current positive AMO state may persist into one to three decades in the future (Schlesinger and Ramankutty, 1994; Sutton and Hodson, 2005; Enfield and Cid-Serrano, 2006). Based on the analysis of leading modes of the summer R95 (Figs. 2a, 3a), we would then suggest a forecast of decreased summer precipitation extremes in the upper Missouri River Basin and the Great Basin and increased summer precipitation extremes in the lower Colorado River Basin. Compared to summer precipitation extremes, winter precipitation extremes are impacted by an integrated effect of ENSO, PDO, and NAO which explains nearly half of its variance. The potential predictability of PDO and NAO is uncertain; however, recent studies including simulations from multiple Global Climate Models (GCMs) suggest that the current positive NAO and negative PDO phase might continue (Visbeck et al., 2001; Miller et al., 2006; Solomon, 2007; Mochizuki et al., 2010; Lapp et al., 2012). For future ENSO states, we are more confident at short term forecasting up to 8 months (Collins et al., 2002). Together with about 4 months lead time of SOI over precipitation; we are able to make up to one year ahead forecasting of winter precipitation extremes in the western United States. Long term prediction of ENSO is not reliable; however, Jin et al. (2003) and Timmermann et al. (1999) suggest an increased El Niño frequency in the future. In other words, we are facing an upcoming period with more El Niño years under the combination PDO negative phase and NAO positive phase (Fig. 6a). In

this situation, most Pacific Northwest area is expected to experience less than normal winter precipitation extremes while the Southwest will probably encounter insignificantly more than normal winter precipitation extremes (Fig. 6a).

4. Summary and conclusions We have presented a comprehensive analysis of relations between the spatiotemporal patterns of precipitation extremes in western United States and SST anomalies as well as oceanic–atmospheric oscillations including ENSO, PDO, AMO, and NAO. Our analysis indicates that the leading modes of R95 variability and the connections between local R95 and SST over western United States are seasonally dependent. Summer precipitation extremes are associated with AMO while winter precipitation extremes are impacted by an integrate effect of ENSO, PDO, and NAO. It is shown for the first time that the NAO is identified in winter extreme precipitation patterns over the western United States, which may impact the winter precipitation extremes by changing the strength of westerlies across the mid-latitudes. Although the influence of NAO on the spatiotemporal pattern of winter precipitation extremes is much smaller than PDO as suggested by PDO-PCs and NAO-PCs correlations, it is still non-negligible. In addition, our results have implications for predicting the seasonal precipitation extremes for next few decades over the western United States. The persistence of the current positive AMO state may cause decreased summer precipitation extremes in the upper Missouri River Basin and the Great Basin and increased summer precipitation extremes in the lower Colorado River Basin. Also, if the current positive NAO and negative PDO phase continues, as suggested by Lapp et al. (2012), Miller et al. (2006), Mochizuki et al. (2010) and Visbeck et al. (2001), a significant decrease and an insignificant increase in winter precipitation

44

J. Peng et al. / Global and Planetary Change 109 (2013) 38–45

extremes are expected over most Pacific Northwest area and Southwest, respectively. It is worth noting that besides the oceanic–atmospheric oscillation that affects the seasonal variations of precipitation extremes, other factors including land surface process and anthropogenic effects also appear to be important (Yang et al., 2008). Although the AMO explains for the first leading mode of summer R95, the land surface can also contribute to the summer precipitation variance, probably by affecting variation in moist convective precipitation (Sutton and Hodson, 2005; Zveryaev, 2006). Also, the anthropogenic effects may offset or enhance the oceanic–atmospheric impacts on precipitation extremes. The influence of anthropogenic effects such as human-induced warming on extreme precipitation events could be investigated by conducting analysis from multi-model ensemble projections (Yang et al., 2011, 2012). However, further studies on the nonlinear interactions between oceanic–atmospheric oscillations and the anthropogenic activities and their integrate effects on precipitation extremes is necessary. Finally, it should be noted that we discuss the precipitation extremes mainly from the aspect of heavy precipitation. Drought is another aspect of precipitation extremes. Several studies have examined the connections between drought and large-scale ocean oscillations including ENSO, PDO, and AMO (Barlow et al., 2001; Feng et al., 2011; Oglesby et al., 2012). An extension of the current research to investigate the connections between drought and the integrated effect which reflects the combined physical oceanic–atmospheric processes, is the subject of ongoing work. Acknowledgments This study was supported by the NSF EPSCoR Change Graduate Research Assistantship and project NSF EPS-0814372 awarded to the first author. The second author's work during sabbatical leave has been supported by University of Nevada Las Vegas. The authors acknowledge that the views herein are that of the authors only and not related the organizations they are affiliated to. We acknowledge with gratitude the helpful comments and suggestions by the two anonymous reviewers on the earlier version of the manuscript. We thank Michael Anderson at the California Department of Water Resources for his helpful suggestions on the earlier version of the manuscript. References Alexander, L., et al., 2006. Global observed changes in daily climate extremes of temperature and precipitation. Journal of Geophysical Research 111 (D05109), 22. Arriaga-Ramírez, S., Cavazos, T., 2010. Regional trends of daily precipitation indices in northwest Mexico and southwest United States. Journal of Geophysical Research 115 (D14), D14111. Barlow, M., Nigam, S., Berbery, E., 2001. ENSO, Pacific decadal variability, and US summertime precipitation, drought, and stream flow. Journal of Climate 14 (9), 2105–2128. Brown, D.P., Comrie, A.C., 2004. A winter precipitation ‘dipole’ in the western United States associated with multidecadal ENSO variability. Geophysical Research Letters 31 (9), L09203. Cai, W., van Rensch, P., 2012. The 2011 southeast Queensland extreme summer rainfall: a confirmation of a negative Pacific Decadal Oscillation phase? Geophysical Research Letters 39 (8), L08702. Cayan, D.R., Redmond, K.T., Riddle, L.G., 1999. ENSO and hydrologic extremes in the western United States. Journal of Climate 12, 2881–2893. Chen, M., et al., 2008. Assessing objective techniques for gauge-based analyses of global daily precipitation. Journal of Geophysical Research 113 (D4), D04110. Collins, M., Frame, D., Sinha, B., Wilson, C., 2002. How far ahead could we predict El Nino? Geophysical Research Letters 29 (10), 130–131. Cressman, G.P., 1959. An operational objective analysis system. Monthly Weather Review 87 (10), 367–374. Curtis, S., 2008. The Atlantic multidecadal oscillation and extreme daily precipitation over the US and Mexico during the hurricane season. Climate Dynamics 30 (4), 343–351. Easterling, D.R., et al., 2000. Climate extremes: observations, modeling, and impacts. Science 289 (5487), 2068–2074. Enfield, D.B., Cid-Serrano, L., 2006. Projecting the risk of future climate shifts. International Journal of Climatology 26 (7), 885–895. Enfield, D.B., Mestas-Nunez, A.M., Trimble, P.J., 2001. The Atlantic multidecadal oscillation and its relation to rainfall and river flows in the continental U. S. Geophysical Research Letters 28 (10), 2077–2080.

Fan, M., Schneider, E.K., 2012. Observed decadal North Atlantic tripole SST variability. Part I: weather noise and coupled response. Journal of the Atmospheric Sciences 69, 35–50. Feng, S., Hu, Q., Oglesby, R.J., 2011. Influence of Atlantic sea surface temperatures on persistent drought in North America. Climate Dynamics 37 (3), 569–586. Frich, P., et al., 2002. Observed coherent changes in climatic extremes during the second half of the twentieth century. Climate Research 19 (3), 193–212. Gershunov, A., Barnett, T.P., 1998. Interdecadal modulation of ENSO teleconnections. Bulletin of the American Meteorological Society 79, 2715–2726. Gutzler, D.S., Kann, D.M., Thornbrugh, C., 2002. Modulation of ENSO-based long-lead outlooks of southwestern US winter precipitation by the Pacific Decadal Oscillation. Weather and Forecasting 17 (6), 1163–1172. Hamlet, A.F., Lettenmaier, D.P., 2007. Effects of 20th century warming and climate variability on flood risk in the western US. Water Resources Research 43 (6), W06427. Hidalgo, H.G., 2004. Climate precursors of multidecadal drought variability in the western United States. Water Resources Research 40 (12), W12504. Higgins, R.W., Center, C.P., 2000. Improved United States Precipitation Quality Control System and Analysis. 40. Higgins, R., Silva, V., Kousky, V., Shi, W., 2008. Comparison of daily precipitation statistics for the United States in observations and in the NCEP Climate Forecast System. Journal of Climate 21 (22), 5993–6014. Hu, Q., Feng, S., Oglesby, R.J., 2011. Variations in North American summer precipitation driven by the Atlantic multidecadal oscillation. Journal of Climate 24 (21), 5555–5570. Hurrell, J.W., 1995. Decadal trends in the North Atlantic Oscillation: regional temperatures and precipitation. Science 269 (5224), 676–679. Jin, F.F., An, S.I., Timmermann, A., Zhao, J., 2003. Strong El Nino events and nonlinear dynamical heating. Geophysical Research Letters 30 (3), 1120. Kaplan, A., et al., 1998. Analyses of global sea surface temperature 1856–1991. Journal of Geophysical Research 103 (18), 18567–18589. Kerr, R.A., 2000. A North Atlantic climate pacemaker for the centuries. Science 288 (5473), 1984–1985. Kurtzman, D., Scanlon, B.R., 2007. El Nino-Southern Oscillation and Pacific Decadal Oscillation impacts on precipitation in the southern and central United States: evaluation of spatial distribution and predictions. Water Resources Research 43 (10), 10427. Lapp, S.L., St Jacques, J.M., Barrow, E.M., Sauchyn, D.J., 2012. GCM projections for the Pacific Decadal Oscillation under greenhouse forcing for the early 21st century. International Journal of Climatology 32 (9), 1423–1442. Latif, M., Barnett, T.P., 1996. Decadal climate variability over the North Pacific and North America: dynamics and predictability. Journal of Climate 9 (10), 2407–2423. Lorenz, E.N., 1956. Empirical Orthogonal Functions and Statistical Weather Prediction, 1. Department of Meteorology, Massachusetts Institute of Technology, Cambridge, MA. Mantua, N.J., Hare, S.R., Zhang, Y., Wallace, J.M., Francis, R.C., 1997. A Pacific interdecadal climate oscillation with impacts on salmon production. Bulletin of the American Meteorological Society 78 (6), 1069–1080. McCabe, G.J., Dettinger, M.D., 1999. Decadal variations in the strength of ENSO teleconnections with precipitation in the western United States. International Journal of Climatology 19 (13), 1399–1410. Meehl, G.A., Tebaldi, C., Teng, H., Peterson, T.C., 2007. Current and future US weather extremes and El Niño. Geophysical Research Letters 34 (20), L20704. Miller, R., Schmidt, G., Shindell, D., 2006. Forced variations of annular modes in the 20th century IPCC AR4 simulations. Journal of Geophysical Research 111, D18101. Mo, K.C., Schemm, J.K.E., Yoo, S.H., 2009. Influence of ENSO and the Atlantic multidecadal oscillation on drought over the United States. Journal of Climate 22 (22), 5962–5982. Mochizuki, T., et al., 2010. Pacific decadal oscillation hindcasts relevant to near-term climate prediction. Proceedings of the National Academy of Sciences 107 (5), 1833–1837. Mote, P.W., Hamlet, A.F., Clark, M.P., Lettenmaier, D.P., 2005. Declining mountain snowpack in western North America. Bulletin of the American Meteorological Society 86 (1), 39–49. North, G.R., Bell, T.L., Cahalan, R.F., Moeng, F.J., 1982. Sampling errors in the estimation of empirical orthogonal functions. Monthly Weather Review 110 (7), 699–706. Oglesby, R., Feng, S., Hu, Q., Rowe, C., 2012. The role of the Atlantic Multidecadal Oscillation on medieval drought in North America: synthesizing results from proxy data and climate models. Global and Planetary Change 84, 56–65. Peng, S., Robinson, W.A., Li, S., 2003. Mechanisms for the NAO responses to the North Atlantic SST tripole. Journal of Climate 16 (12), 1987–2004. Piechota, T., Timilsena, J., Tootle, G., Hidalgo, H., 2004. The western US drought: how bad is it. Eos 85 (32), 301–304. Preisendorfer, R.W., 1988. Principal Component Analysis in Meteorology and Oceanography. Elsevier. Richman, M.B., 1986. Rotation of principal components. Journal of Climatology 6 (3), 293–335. Roxy, M., Patil, N., Ashok, K., Aparna, K., 2013. Revisiting the Indian summer monsoonENSO links in the IPCC AR4 projections: a cautionary outlook. Global and Planetary Change 104, 51–60. Schlesinger, M.E., Ramankutty, N., 1994. An oscillation in the global climate system of period 65–70 years. Nature 367 (6465), 723–726. Solomon, S., 2007. Climate Change 2007: The Physical Science Basis: Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press. Sutton, R.T., Hodson, D.L.R., 2005. Atlantic Ocean forcing of North American and European summer climate. Science 309 (5731), 115–118.

J. Peng et al. / Global and Planetary Change 109 (2013) 38–45 Timm, O.E., Diaz, H., Giambelluca, T., Takahashi, M., 2011. Projection of changes in the frequency of heavy rain events over Hawaii based on leading Pacific climate modes. Journal of Geophysical Research 116 (D4), D04109. Timmermann, A., et al., 1999. Increased El Niño frequency in a climate model forced by future greenhouse warming. Nature 398, 694–697. Visbeck, M.H., Hurrell, J.W., Polvani, L., Cullen, H.M., 2001. The North Atlantic Oscillation: past, present, and future. Proceedings of the National Academy of Sciences of the United States of America 98 (23), 12876–12877. Wang, C., Enfield, D.B., Lee, S., Landsea, C.W., 2006. Influences of the Atlantic warm pool on Western Hemisphere summer rainfall and Atlantic hurricanes. Journal of Climate 19 (12), 3011–3028. Wise, E.K., 2010. Spatiotemporal variability of the precipitation dipole transition zone in the western United States. Geophysical Research Letters 37 (7), L07706. Wu, L., Liu, Z., 2005. North Atlantic decadal variability: air-sea coupling, oceanic memory, and potential northern hemisphere resonance. Journal of Climate 18 (2), 331–349.

45

Yang, T., et al., 2008. A spatial assessment of hydrologic alteration caused by dam construction in the middle and lower Yellow River, China. Hydrological Processes 22 (18), 3829–3843. Yang, T., et al., 2010. Regional frequency analysis and spatio-temporal pattern characterization of rainfall extremes in the Pearl River Basin, China. Journal of Hydrology 380 (3), 386–405. Yang, T., et al., 2011. Changes of climate extremes in a typical arid zone: observations and multimodel ensemble projections. Journal of Geophysical Research 116 (D19), D19106. Yang, T., et al., 2012. Multi-model ensemble projections in temperature and precipitation extremes of the Tibetan Plateau in the 21st century. Global and Planetary Change 80, 1–13. Yeh, S.W., et al., 2009. El Niño in a changing climate. Nature 461 (7263), 511–514. Zhang, X., Wang, J., Zwiers, F.W., Groisman, P.Y., 2010. The influence of large-scale climate variability on winter maximum daily precipitation over North America. Journal of Climate 23 (11), 2902–2915. Zveryaev, I.I., 2006. Seasonally varying modes in long-term variability of European precipitation during the 20th century. Journal of Geophysical Research 111 (D21), 21116.