The nexus between atmospheric rivers and extreme precipitation across Europe

Transcription

1 GEOPHYSICAL RESEARCH LETTERS, VOL. 40, , doi: /grl.50636, 2013 The nexus between atmospheric rivers and extreme precipitation across Europe David A. Lavers 1 and Gabriele Villarini 1 Received 6 May 2013; revised 4 June 2013; accepted 5 June 2013; published 28 June [1] Extreme precipitation and floods in Europe are a recurring natural hazard causing large socioeconomic damages. Here we investigate the connection between annual maxima (AM) daily precipitation at a pan-european scale and atmospheric rivers (ARs), narrow filaments that convey the majority of the poleward water vapor transport within extratropical cyclones. We show that ARs are responsible for many AM precipitation days in Western Europe. The relationship is especially strong along the western European seaboard, with some areas having eight of their top 10 AM related to ARs. The effects of ARs are also seen as far inland as Germany and Poland. Southern Europe was most affected by ARs under negative North Atlantic Oscillation (NAO) conditions, whereas northern Europe was more associated with a positive relationship between ARs and an NAO-type pattern. Our results suggest that ARs are critical in explaining the upper tail of the extreme precipitation distribution in Western Europe. Citation: Lavers, D. A., and G. Villarini (2013), The nexus between atmospheric rivers and extreme precipitation across Europe, Geophys. Res. Lett., 40, , doi: /grl Introduction [2] In the pre-cold frontal region (within the warm conveyor belt) of extratropical cyclones, there exists a narrow corridor of air with high water vapor content, which in combination with the low-level jet (LLJ) leads to a region of intense moisture transport [e.g., Ralph et al., 2004]. Due to the filament-like nature of this synoptic feature, this area has been labeled as an atmospheric river (AR) [Zhu and Newell, 1998]. It is especially when this fast-flowing moisture-rich air impinges on a mountain chain that large precipitation totals and flooding can take place. [3] Although the moisture transport in the LLJ region has been known for some time [Browning and Pardoe, 1973], accurate satellite retrievals of total column water vapor highlighted the area we now refer to as an AR. Subsequent research has uncovered strong relationships between ARs and flooding in the midlatitudes, with western North America [e.g., Ralph et al., 2006; Neiman et al., 2011; Ralph and Dettinger, 2012], South America [Viale and Nunez, 2011], and the British Isles [Lavers et al., 2011, 2012] having Additional supporting information may be found in the online version of this article. 1 IIHR-Hydroscience and Engineering, University of Iowa, Iowa City, Iowa, USA. Corresponding author: D. A. Lavers, IIHR-Hydrosciance and Engineering, University of Iowa, 100 C. Maxwell Stanley Hydraulics Laboratory, Iowa City, IA 52242, USA. (david-lavers@uiowa.edu) American Geophysical Union. All Rights Reserved /13/ /grl large floods caused by ARs. Given this link between ARs and extreme precipitation in certain midlatitude locations, it is timely and important to evaluate whether ARs are important agents in extreme precipitation at a pan-european scale (1) because of the socioeconomic damage associated with flooding and (2) as a changing climate is expected to lead to an increased risk of hydrological extremes [Held and Soden, 2006; Allan and Soden, 2008; O'Gorman and Schneider, 2009]. The identification of potential linkages across Europe may enable future process-based climate change assessments to be undertaken. [4] In this paper, we perform a pan-european study to investigate the strength of the link between ARs and annual maxima (AM) precipitation over the period In so doing, we extend an AR detection algorithm developed for use in Britain [Lavers et al., 2012] to capture AR occurrence across Europe, associate the detected ARs with AM precipitation, and analyze the large-scale atmospheric patterns associated with AR events. 2. Data and Methods [5] The specific humidity, mean sea level pressure (MSLP), and the zonal and meridional wind fields were retrieved from the European Centre for Medium-Range Weather Forecasts ERA-Interim (ERAIN) reanalysis at a resolution over [Dee et al., 2011]. The vertically integrated horizontal water vapor transport (hereafter, integrated vapor transport, IVT) was calculated from 1000 hpa to 300 hpa in an Eulerian framework [e.g., Neiman et al., 2008]: sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1 2 IVT ¼ g qudp þ 1 2 g qvdp where q is the layer-averaged specific humidity in kg/kg, u and v are the layer-averaged zonal and meridional winds in ms 1 respectively, g is the acceleration due to gravity, and dp is the pressure difference between two adjacent pressure levels. [6] The IVT fields were used in a modified AR detection algorithm [Lavers et al., 2012] to identify ARs striking the western European boundary between 35 N and 70 N near 10 W. A latitude-dependent IVT threshold was determined as follows. At 1200 UTC on each day from 1979 to 2011, we extracted the maximum IVT between 35 N and 70 N near 10 W and binned it into 5 latitude bins. The 85th percentile of the IVT in each latitude bin was used as the threshold value for ARs identified in that region because Lavers et al. [2012] showed that it corresponded to the IVT value of the most intense ARs. The IVT thresholds in each band were as follows: kg m 1 s 1 (35 N 40 N), kg m 1 s 1 (40 N 45 N), kg m 1 s 1 (45 N 50 N), kg m 1 s 1 (50 N 55 N), kg m 1 s 1 (55 N 60 N), kg m 1 s 1 (60 N 65 N), and kg m 1 s 1 (65 N 70 N). (1)

2 Figure 1. The IVT of ARs detected by the algorithm in the latitude bands (a) 35 N 40 N (at 1800 UTC 20th May 1994), (b) 45 N 50 N (at 0000 UTC 27th Oct 1980), (c) 55 N 60 N (at 0000 UTC 6th Mar 2002), and (d) 65 N 70 N (at 1800 UTC 1st Dec 1989). Units are kg m 1 s 1. [7] With an IVT threshold established, the following methodology was used at each 6 h time step in ERAIN from 1979 to We calculated the IVT at grid points spanning between 35 N and 70 N along 10 W and retained the maximum IVT value and location. If the maximum IVT exceeded the IVT threshold for that particular region, the grid point was recorded. We then performed a backward search from 10 W to 30 W to identify the maximum IVT at each longitude and tracked the location for the grid points where the IVT threshold (taken from 10 W) was exceeded. We also performed a forward search from 10 W to 25 E to identify the maximum IVT at each longitude, recording the locations where the IVT threshold was exceeded. Finally, we determined whether the extracted points satisfied the criterion of an appropriate length scale. If 30 continuous longitude points exceeded the threshold (on average across the domain this is roughly equal to 1500 km), we considered it an AR time step. Furthermore, only AR events that occurred for three or more time steps (18 h or more) were considered as potential extreme precipitation generating ARs (hereafter, the term AR refers to a persistent AR), as AR duration is fundamental in controlling precipitation totals [Ralph et al., 2013]. Moreover, we allowed only a 4.5 latitude movement to the north or south of the initial IVT maximum at 10 W in an 18 h period. Assuming that the midpoint of the AR (at 10 W) is given by the maximum IVT and that ARs are of the order of 1000 km wide [Neiman et al., 2008], a 4.5 latitude movement (approximately equal to 500 km) means that even if the central location of the AR moves by 4.5, the AR may still deliver heavy rainfall to a specific location. To have independent events, two ARs were considered distinct only if they were separated by more than 1 day. [8] We retrieved daily observed gauge-based precipitation produced by the European Union funded ENSEMBLES project [Hewitt and Griggs, 2004; Haylock et al., 2008] at a resolution across Europe (E-OBS version 7.0 data set). To identify extreme precipitation days, we used a block maxima approach over the period [Coles, 2001]. More specifically, the maximum daily precipitation was extracted for each calendar year, generating a time series of AM at each pixel. 3. Results and Discussion [9] Our algorithm detected a total of 432 ARs over , with the region of peak AR activity occurring between 45 N and 55 N. Figure 1 shows four examples of ARs (detected by our algorithm) that caused AM precipitation events in Europe. The narrow enhanced region of moisture transport highlights that ARs impact the whole of the western seaboard of Europe; the long AR signature noted in previous research [Neiman et al., 2008] is also clearly visible. [10] For each calendar year from 1979 to 2011 at each precipitation grid, we extracted the date of the AM daily precipitation total and placed it into separate seasons (Figure 2, left panels). Along the western seaboard of Europe from the Iberian Peninsula through France and the British Isles to Scandinavia, the majority of the AM occur in December, January, and February (DJF) and September, October, and November (SON). Conversely, in central and eastern Europe, the AM predominantly occur in June, July, and August (JJA). To associate an AM event with an AR, we assume that (1) an AM event occurred on the same day or a day after an AR was detected at 10 W allowing for AR 3260

3 Figure 2. (left) The number of AM precipitation events in each season over and (right) the percentage of the total AMs in each season that are related to ARs. 3261

4 Figure 3. The number of the top 10 AM precipitation events that are related to ARs. inland penetration into eastern Europe (similar to the approach by Rutz and Steenburgh [2012] for the western United States), and (2) the IVT during or up to 1 day after an AR (between 4.5 latitude to the north or south of the precipitation grid) exceeded the IVT threshold at the AR's point of entry into Europe at 10 W. The right panels of Figure 2 show the percentage of AM that are related to ARs for each season. The largest AR-AM links are found in SON and DJF, with many places having more than 40% of their AM caused by ARs (e.g., northern France in DJF). Furthermore, the plots highlight the inland penetration of ARs into Europe as far as Poland, indicating that ARs affect extreme precipitation at large distances from the North Atlantic Ocean. It is evident that the patterns of the strongest AR-AM relationship tend to follow the orography, as is seen along the Scandinavian mountains, western Britain, northern France, southern Benelux, the Pyrenees, and on the Iberian Peninsula. Therefore, it is the presence of mountains that provides the uplift of the moisture-rich air in the AR, in turn causing high precipitation totals [e.g., Ralph et al., 2005]. Note that in the lee of these mountains the linkage is weaker (e.g., Scandinavia) due to the rain shadow. [11] To further illustrate the nexus between ARs and extreme European precipitation, Figure 3 presents the number of the top 10 AM precipitation events that were related to ARs. Over large areas of the Iberian Peninsula, northern France, and southern Norway, six out of the top 10 AM were caused by ARs; moreover, some locations in Scotland, southwest England, northern France, and Norway had eight of the top 10 AM associated with ARs. These results indicate that ARs dominate the upper tail of the rainfall distributions over large regions of Europe. Figure 3 also confirms that the strongest AR-AM connection is found in mountainous areas. It is particularly striking how far inland the impacts of these catastrophic hazards are felt (see Rutz and Steenburgh [2012] for results on AR inland penetration for western North America). Among the AR events identified by our algorithm, we note that Cyclone Lothar (December 1999) and the Cockermouth flood in England (November 2009) were detected, which further illustrates the connection between ARs and well-known heavy precipitation/flood episodes. [12] The AR influence on AM precipitation events is mainly restricted to fall and winter rather than spring or summer (Figure 2, right panels). It is during the winter half-year across the North Atlantic Ocean that there is a stronger equator-to- North Pole temperature gradient resulting in a stronger baroclinic zone and storm track affecting Western Europe. The extratropical cyclones that grow in the baroclinic zone contain the ARs that strike the European land mass. During the summer, however, a weaker equator-to-north Pole temperature gradient and North Atlantic storm track means that extratropical cyclones are not as prevalent, and thus the AR effect on precipitation is weaker. Precipitation in summer tends to be more associated with convective storms [e.g., Berg et al., 2009], and the fact that a high proportion of central European AM precipitation events occur in JJA (Figure 2; left panels) suggests that convective storms are the key driver of extreme precipitation in this region. [13] For each AR event, the gridded daily precipitation was summed over the AR days to create AR storm total precipitation. These storm totals were then binned into 5 latitude bins (for the bins, see section 2) depending on the point of entry of the AR into Europe at 10 W. The composite mean of the AR storm total precipitation for each latitude band is shown in Figure S1 in the supporting information. It is evident that the areas with the highest rainfall accumulations depend on the location where the AR entered Europe (c.f. Figures S1a and S1e). [14] For each AR event, we calculated the average MSLP (at each grid point) over its lifetime. We then computed the MSLP anomaly pattern for that particular AR event with respect to the same time period over the years The MSLP anomaly fields were then placed into the aforementioned latitude bins, and a composite mean anomaly pattern was calculated for each latitude band (Figure 4). We also computed the composite standard deviation for each latitude band (Figure S2); the results support the signals shown in the composite mean anomaly patterns. For the southernmost occurring ARs (Figure 4a; 35 N 40 N), positive MSLP anomalies are located over Iceland and Greenland, and negative MSLP anomalies extend from Britain to the Iberian Peninsula. This setup relates to a negative North Atlantic Oscillation (NAO) pattern, with a blocked flow over northern Europe and the North Atlantic storm track (and their embedded ARs and heavy precipitation) impacting southern Europe; this is highlighted in the precipitation composite (Figure S1a) and corroborates earlier NAO findings [e.g., Hurrell, 1995; Hurrell et al., 2003; Pinto and Raible, 2012]. A negative, albeit weaker, NAO influence on ARs is also found between 40 N and 45 N (Figure 4b). In the latitude band 45 N 55 N (Figures 4c and 4d), the MSLP dipole pattern relates to a positive NAO phase, with the ARs within the extratropical cyclones delivering rainfall from northern France, through the western British Isles to Norway. Further north between 55 N and 70 N (Figures 4e, 4f, and 4g), ARs and their associated precipitation are related to an MSLP dipole of positive anomalies near the British Isles and negative anomalies over Greenland and Iceland. The storm track and the passage of extratropical cyclones can be thought of as passing through the region of negative 3262

5 LAVERS AND VILLARINI: ARS AND EXTREME EUROPEAN PRECIPITATION Figure 4. The composite mean MSLP anomaly patterns (for the lifetime of the ARs) in 5 latitude bins. Units are hpa. (a) In 35 N 40 N, there were 36 events, (b) in 40 N 45 N 48 events, (c) in 45 N 50 N 107 events, (d) in 50 N 55 N 89 events, (e) in 55 N 60 N 66 events, (f) in 60 N 65 N 53 events, and (g) in 65 N 70 N 33 events. 3263

6 MSLP anomalies with the ARs located to the south between the MSLP centers of action; hence, it is generally the land masses situated between the opposing MSLP anomalies that receive the highest precipitation totals. Although to a large extent the NAO is related to AR occurrence in Europe, we note that the MSLP patterns presented have a significant spatial displacement between different European regions and are thus not always colocated with the traditional locations used in calculating the NAO index (in particular for the northernmost regions). [15] The aim of this study was to undertake a pan-european analysis to assess the potential link between ARs and extreme precipitation over We modified an earlier algorithm based on IVT fields for AR detection in Europe and then linked the identified ARs with (1) extreme rainfall occurrence and (2) hemispheric MSLP patterns. We show that ARs do cause extreme precipitation particularly in fall and winter in Western Europe, with areas in Britain, France, and Norway having up to eight out of the 10 largest daily rainfall events caused by the identified ARs. These regions have hills and mountainous relief, which together with the vertical motion in some ARs provides the uplift necessary to cause significant rainfall totals. The MSLP patterns showed that the NAO affected AR activity in different parts of Europe; a negative NAO was shown to be concurrent with ARs in southern Europe, and a positive NAO pattern was found to be more associated with ARs in northern Europe, consistent with previous research. There is also a significant influence of ARs on extreme precipitation in the interior of Europe, with AR effects felt as far east as Poland. [16] Acknowledgments. The authors gratefully acknowledge financial support by IIHR-Hydroscience and Engineering and the Iowa Flood Center. We make use of the E-OBS data set from the EU-FP6 project ENSEMBLES ( and thank data providers in the ECA&D project ( The analyses were undertaken using Enthought Python Distribution (EPD) 7.3.1, Enthought, Inc., ( We also thank the two anonymous reviewers for their constructive comments. [17] The Editor thanks two anonymous reviewers for their assistance in evaluating this paper. References Allan, R. P., and B. J. Soden (2008), Atmospheric warming and the amplification of precipitation extremes, Science, 321(5895), Berg, P., J. O. Haerter, P. Thejll, C. Piani, S. Hagemann, and J. H. Christensen (2009), Seasonal characteristics of the relationship between daily precipitation intensity and surface temperature, J. Geophys. Res., 114, D18102, doi: /2009jd Browning, K. A., and C. W. Pardoe (1973), Structure of low-level jet streams ahead of mid-latitude cold fronts, Q. J. R. Meteorol. Soc., 99(422), , doi: /qj Coles, S. (2001), An Introduction to Statistical Modeling of Extreme Values, Springer, London. Dee, D. P., et al. (2011), The ERA-Interim reanalysis: Configuration and performance of the data assimilation system, Q. J. R. Meteorol. Soc., 137(656), Haylock, M. R., et al. (2008), European daily high-resolution gridded dataset of surface temperature and precipitation, J. Geophys. Res., 113, D20119, doi: /2008jd Held, I. M., and B. J. Soden (2006), Robust responses of the hydrological cycle to global warming, J. Clim., 19(21), Hewitt, C. D., and D. J. Griggs (2004), Ensembles-based predictions of climate changes and their impacts, Eos. Trans. AGU, 85, 566, doi: / 2004EO Hurrell, J. W. (1995), Decadal trends in the North Atlantic Oscillation: Regional temperatures and precipitation, Science, 269, Hurrell, J. W., Y. Kushnir, G. Ottersen, and M. Visbeck (2003), An overview of the North Atlantic Oscillation, in The North Atlantic Oscillation Climatic Significance and Environmental Impact, Geophys. Monogr. Ser., vol. 134, edited by J. W. Hurrell et al., pp. 1 36, AGU, Washington, D. C. Lavers, D. A., R. P. Allan, E. F. Wood, G. Villarini, D. J. Brayshaw, and A. J. Wade (2011), Winter floods in Britain are connected to atmospheric rivers, Geophys. Res. Lett., 38, L23803, doi: /2011gl Lavers, D. A., G. Villarini, R. P. Allan, E. F. Wood, and A. J. Wade (2012), The detection of atmospheric rivers in atmospheric reanalyses and their links to British winter floods and the large-scale climatic circulation, J. Geophys. Res., 117, D20106, doi: /2012jd Neiman, P. J., F. M. Ralph, G. A. Wick, J. D. Lundquist, and M. D. Dettinger (2008), Meteorological characteristics and overland precipitation impacts of atmospheric rivers affecting the West Coast of North America based on eight years of SSM/I satellite observations, J. Hydrometeorol., 9(1), Neiman, P. J., L. J. Schick, F. M. Ralph, M. Hughes, and G. A. Wick (2011), Flooding in Western Washington: The connection to atmospheric rivers, J. Hydrometeorol., 12(6), O'Gorman, P. A., and T. Schneider (2009), The physical basis for increases in precipitation extremes in simulations of 21st-century climate change, Proc. Natl. Acad. Sci. U. S. A., 106(35), 14,773 14,777. Pinto, J. G., and C. C. Raible (2012), Past and recent changes in the North Atlantic oscillation, WIREs Clim. Change, 3, 79 90, doi: / wcc.150. Ralph, F. M., and M. D. Dettinger (2012), Historical and national perspectives on extreme West Coast precipitation associated with atmospheric rivers during December 2010, Bull. Am. Meteorol. Soc., 93, , doi: /bams-d Ralph, F. M., P. J. Neiman, and G. A. Wick (2004), Satellite and CALJET aircraft observations of atmospheric rivers over the eastern North Pacific ocean during the winter of 1997/98, Mon. Weather Rev., 132(7), Ralph, F. M., P. J. Neiman, and R. Rotunno (2005), Dropsonde observations in low-level jets over the Northeastern Pacific Ocean from CALJET-1998 and PACJET-2001: Mean vertical-profile and atmospheric-river characteristics, Mon. Weather Rev., 133, Ralph, F. M., P. J. Neiman, G. A. Wick, S. I. Gutman, M. D. Dettinger, D. R. Cayan, and A. B. White (2006), Flooding on California's Russian River: Role of atmospheric rivers, Geophys. Res. Lett., 33, L13801, doi: /2006gl Ralph, F. M., T. Coleman, P. J. Neiman, R. Zamora, and M. D. Dettinger (2013), Observed impacts of duration and seasonality of atmosphericriver landfalls on soil moisture and runoff in coastal northern California, J. Hydrometeorol., 14, Rutz, J. J., and W. J. Steenburgh (2012), Quantifying the role of atmospheric rivers in the interior western United States, Atmos. Sci. Lett., 13, Viale, M., and M. N. Nunez (2011), Climatology of winter orographic precipitation over the subtropical central Andes and associated synoptic and regional characteristics, J. Hydrometeorol., 12(4), Zhu, Y., and R. E. Newell (1998), A proposed algorithm for moisture fluxes from atmospheric rivers, Mon. Weather Rev., 126(3), , doi: / (1998)126<0725:apafmf>2.0.co;