• Bates, B. C., Kundzewicz Z. W. , Wu S. , and Palutikof J. P. , Eds., 2008: Climate Change and Water. Intergovernmental Panel on Climate Change, 214 pp.

  • Bower, D., Hannah D. M. , and McGregor G. R. , 2004: Techniques for assessing the climatic sensitivity of river flow regimes. Hydrol. Processes, 18, 25152543.

    • Search Google Scholar
    • Export Citation
  • Cassou, C., Deser C. , Terray L. , Hurrell J. W. , and Drevillon M. , 2004: Summer sea surface temperature conditions in the North Atlantic and their impact upon the atmospheric circulation in early winter. J. Climate, 17, 33493363.

    • Search Google Scholar
    • Export Citation
  • Colman, A., 1997: Prediction of summer central England temperature from preceding North Atlantic winter sea surface temperature. Int. J. Climatol., 17, 12851300.

    • Search Google Scholar
    • Export Citation
  • Colman, A., and Davey M. , 1999: Prediction of summer temperature, rainfall and pressure in Europe from preceding winter North Atlantic Ocean temperature. Int. J. Climatol., 19, 513536.

    • Search Google Scholar
    • Export Citation
  • Czaja, A., and Frankignoul C. , 2002: Observed impact of Atlantic SST anomalies on the North Atlantic Oscillation. J. Climate, 15, 606623.

    • Search Google Scholar
    • Export Citation
  • Dixon, H., Lawler D. M. , and Shamseldin A. , 2006: Streamflow trends in western Britain. Geophys. Res. Lett., 33, L19406, doi:10.1029/2006GL027325.

    • Search Google Scholar
    • Export Citation
  • Fleig, A. K., Tallaksen L. M. , Hisdal H. , and Demuth S. , 2006: A global evaluation of discharge drought characteristics. Hydrol. Earth Syst. Sci., 10, 535552.

    • Search Google Scholar
    • Export Citation
  • Fleig, A. K., Tallaksen L. M. , Hisdal H. , Stahl K. , and Hannah D. M. , 2010: Inter-comparison of weather and circulation type classifications for hydrological drought development. Phys. Chem. Earth, 35, 507515.

    • Search Google Scholar
    • Export Citation
  • Fleig, A. K., Tallaksen L. M. , Hisdal H. , and Hannah D. M. , 2011: Regional hydrological drought in north-western Europe: Linking a new regional drought area index with weather types. Hydrol. Processes, 25, 11631179.

    • Search Google Scholar
    • Export Citation
  • Folland, C. K., Knight J. , Linderholm H. W. , Fereday D. , Ineson S. , and Hurrell J. W. , 2009: The summer North Atlantic Oscillation: Past, present, and future. J. Climate, 22, 10821103.

    • Search Google Scholar
    • Export Citation
  • Fowler, H. J., and Kilsby C. G. , 2002: Precipitation and the North Atlantic Oscillation: A study of climatic variation in northern England. Int. J. Climatol., 22, 843866.

    • Search Google Scholar
    • Export Citation
  • Gámiz-Fortis, S., Pozo-Vázquez D. , Trigo R. M. , and Castro-Díez Y. , 2008: Quantifying the predictability of winter river flow in Iberia. Part 2: Seasonal predictability. J. Climate, 21, 25032518.

    • Search Google Scholar
    • Export Citation
  • Hannah, D. M., Demuth S. , van Lanen H. A. J. , Looser U. , Prudhomme C. , Rees G. , Stahl K. , and Tallaksen L. M. , 2011: Large-scale river flow archives: Importance, current status and future needs. Hydrol. Processes, 25, 11911200.

    • Search Google Scholar
    • Export Citation
  • Hertig, E., and Jacobeit J. , 2011: Predictability of Mediterranean climate variables from oceanic variability. Part 1: Sea surface temperature regimes. Climate Dyn., 36, 811823.

    • Search Google Scholar
    • Export Citation
  • Hu, Z., and Huang B. , 2006: Air–sea coupling in the North Atlantic during summer. Climate Dyn., 26, 441457.

  • Hurrell, J. W., and van Loon H. , 1997: Decadal variations in climate associated with the North Atlantic Oscillation. Climatic Change, 36, 301326.

    • Search Google Scholar
    • Export Citation
  • James, P. M., 2007: An objective classification method for Hess and Brezowsky Grosswetterlagen over Europe. Theor. Appl. Climatol., 88, 1742.

    • Search Google Scholar
    • Export Citation
  • Jones, P. D., Jónsson T. , and Wheeler D. , 1997: Extension to the North Atlantic Oscillation using early instrumental pressure observations from Gibraltar and south-west Iceland. Int. J. Climatol., 17, 14331450.

    • Search Google Scholar
    • Export Citation
  • Kapala, A., Mächel H. , and Flohn H. , 1998: Behaviour of the centres of action above the Atlantic since 1881. Part 2: Associations with regional climate anomalies. Int. J. Climatol., 18, 2326.

    • Search Google Scholar
    • Export Citation
  • Kingston, D. G., Lawler D. M. , and McGregor G. R. , 2006a: Linkages between atmospheric circulation, climate and streamflow in the northern North Atlantic: Research prospects. Prog. Phys. Geogr., 30, 143174.

    • Search Google Scholar
    • Export Citation
  • Kingston, D. G., McGregor G. R. , Hannah D. M. , and Lawler D. M. , 2006b: River flow teleconnections across the northern North Atlantic region. Geophys. Res. Lett., 33, L14705, doi:10.1029/2006GL026574.

    • Search Google Scholar
    • Export Citation
  • Kingston, D. G., McGregor G. R. , Hannah D. M. , and Lawler D. M. , 2007: Climatic controls on New England streamflow. J. Hydrometeor., 8, 367379.

    • Search Google Scholar
    • Export Citation
  • Kingston, D. G., Hannah D. M. , Lawler D. M. , and McGregor G. R. , 2009: Climate–river flow relationships across montane and lowland environments in northern Europe. Hydrol. Processes, 23, 985996.

    • Search Google Scholar
    • Export Citation
  • Kingston, D. G., Fleig A. K. , Tallaksen L. M. , and Hannah D. M. , 2010: North Atlantic sea surface temperature, atmospheric circulation and summer drought in Great Britain. Global Change: Facing Risks and Threats to Water Resources, E. Servat et al., Eds., IAHS Publ. 340, 598–604.

  • Kingston, D. G., Hannah D. M. , Lawler D. M. , and McGregor G. R. , 2011: Regional classification, variability and trends of northern North Atlantic river flow. Hydrol. Processes, 25, 10211033.

    • Search Google Scholar
    • Export Citation
  • Kushnir, Y., Robinson W. A. , Chang P. , and Robertson A. W. , 2006: The physical basis for predicting Atlantic sector seasonal-to-interannual climate variability. J. Climate, 19, 59495969.

    • Search Google Scholar
    • Export Citation
  • Laizé, C. L. R., and Hannah D. M. , 2010: Modification of climate-river flow associations by basin properties. J. Hydrol., 389, 186204.

    • Search Google Scholar
    • Export Citation
  • Lavers, D., Prudhomme C. , and Hannah D. M. , 2010a: Large-scale climatic influences on precipitation and discharge for a British river basin. Hydrol. Processes, 24, 25552563.

    • Search Google Scholar
    • Export Citation
  • Lavers, D., Prudhomme C. , and Hannah D. M. , 2010b: Large-scale climate, precipitation and British river flows: Identifying hydroclimatological connections and dynamics. J. Hydrol., 395, 242255.

    • Search Google Scholar
    • Export Citation
  • Mitchell, T. D., and Jones P. I. , 2005: An improved method of constructing a database of monthly climate observations and associated high-resolution grids. Int. J. Climatol., 25, 693712.

    • Search Google Scholar
    • Export Citation
  • Monk, W. A., Wood P. J. , Hannah D. M. , and Wilson D. A. , 2008: Macroinvertebrate community response to inter-annual and regional river flow regime dynamics. River Res. Appl., 24, 9881001.

    • Search Google Scholar
    • Export Citation
  • Phillips, I. D., and McGregor G. R. , 2001: Western European water vapor flux—Southeast England rainfall associations. J. Hydrometeor., 2, 505524.

    • Search Google Scholar
    • Export Citation
  • Portis, D. H., Walsh J. E. , El Hamly M. , and Lamb P. J. , 2001: Seasonality of the North Atlantic Oscillation. J. Climate, 14, 20692078.

    • Search Google Scholar
    • Export Citation
  • Rayner, N. A., Parker D. E. , Horton E. B. , Folland C. K. , Alexander L. V. , Rowell D. P. , Kent E. C. , and Kaplan A. , 2003: Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J. Geophys. Res., 108, 4407, doi:10.1029/2002JD002670.

    • Search Google Scholar
    • Export Citation
  • Rodwell, M. J., Rowell D. P. , and Folland C. K. , 1999: Oceanic forcing of the wintertime North Atlantic Oscillation and European climate. Nature, 398, 320323.

    • Search Google Scholar
    • Export Citation
  • Svensson, C., and Prudhomme C. , 2005: Prediction of British summer river flows using winter predictors. Theor. Appl. Climatol., 82, 115.

    • Search Google Scholar
    • Export Citation
  • Tallaksen, L. M., and van Lanen H. A. J. , Eds., 2004: Hydrological Drought: Processes and Estimation Methods for Streamflow and Groundwater. Developments in Water Sciences, Vol. 48, Elsevier, 580 pp.

  • Tallaksen, L. M., Madsen H. , and Clausen B. , 1997: On the definition and modelling of streamflow drought duration and deficit volume. Hydrol. Sci. J., 42, 1533.

    • Search Google Scholar
    • Export Citation
  • Todd, M. C., Taylor R. G. , Osborn T. J. , Kingston D. G. , Arnell N. W. , and Gosling S. N. , 2011: Uncertainty in climate change impacts on basin-scale freshwater resources–Preface to the special issue: The QUEST-GSI methodology and synthesis of results. Hydrol. Earth Syst. Sci., 15, 10351046.

    • Search Google Scholar
    • Export Citation
  • Uppala, S. M., and Coauthors, 2005: The ERA-40 Re-Analysis. Quart. J. Roy. Meteor. Soc., 131, 29613012.

  • Wedgbrow, C. S., Wilby R. L. , Fox H. R. , and O’Hare G. , 2002: Prospects for seasonal forecasting of summer drought and low river flow anomalies in England and Wales. Int. J. Climatol., 22, 219236.

    • Search Google Scholar
    • Export Citation
  • Wilby, R. L., O’Hare G. , and Barnsley N. , 1997: The North Atlantic Oscillation and British Isles climate variability, 1865–1996. Weather, 52, 266276.

    • Search Google Scholar
    • Export Citation
  • Wilby, R. L., Wedgbrow C. S. , and Fox H. R. , 2004: Seasonal predictability of the summer hydrometeorology of the River Thames, UK. J. Hydrol., 295, 116.

    • Search Google Scholar
    • Export Citation
  • View in gallery
    Fig. 1.

    Location of streamflow gauging stations and drought regions.

  • View in gallery
    Fig. 2.

    Monthly drought duration for each of the four regions (light gray: 0 < RDAI < 0.5; medium gray: 0.5 < RDAI < 0.7; dark gray: RDAI > 0.7).

  • View in gallery
    Fig. 3.

    SST composite anomalies (°C) for (a) GB1 May, (b) GB2 May, (c) GB3 April, and (d) GB4 March. Shading indicates statistical significance at p = 0.1 (light), 0.05 (medium), and 0.01 (dark).

  • View in gallery
    Fig. 4.

    The 500-hPa geopotential height composite anomalies (m) for (a) GB1 May, (b) GB2 May, (c) GB3 March, and (d) GB4 February. Shading as in Fig. 3.

  • View in gallery
    Fig. 5.

    May mean 1000-hPa water vapor flux (kg m−1 Pa−1 s−1) (left) during GB1 drought years and (right) the 1964–2001 mean. Water vapor flux indicated by arrow length (see scale).

  • View in gallery
    Fig. 6.

    Precipitation composite anomalies (mm) for (a) GB1 May–June, (b) GB2 May–June, (c) GB3 March–April, and (d) GB4 February–March 2-month means. Shading as in Fig. 3.

  • View in gallery
    Fig. 7.

    March mean 1000-hPa water vapor flux (kg m−1 Pa−1 s−1) (left) during GB3 drought years and (right) the 1964–2001 mean. Water vapor flux indicated by arrow length (see scale).

  • View in gallery
    Fig. 8.

    February mean 1000-hPa water vapor flux (kg m−1 Pa−1 s−1) (left) during GB4 drought years and (right) the 1964–2001 mean. Water vapor flux indicated by arrow length (see scale).

  • View in gallery
    Fig. 9.

    Schematic model of the relationship between SST, climate, and December–May drought development.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 176 92 7
PDF Downloads 88 39 4

Ocean–Atmosphere Forcing of Summer Streamflow Drought in Great Britain

Daniel G. KingstonDepartment of Geography, University of Otago, Dunedin, New Zealand, and Department of Geosciences, University of Oslo, Oslo, Norway

Search for other papers by Daniel G. Kingston in
Current site
Google Scholar
PubMed
Close
,
Anne K. FleigDepartment of Geosciences, University of Oslo, and Norwegian Water Resources and Energy Directorate, Oslo, Norway

Search for other papers by Anne K. Fleig in
Current site
Google Scholar
PubMed
Close
,
Lena M. TallaksenDepartment of Geosciences, University of Oslo, Oslo, Norway

Search for other papers by Lena M. Tallaksen in
Current site
Google Scholar
PubMed
Close
, and
David M. HannahSchool of Geography, Earth and Environmental Sciences, University of Birmingham, Birmingham, United Kingdom

Search for other papers by David M. Hannah in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

Droughts are high-impact events that have substantial implications for both human and natural systems. As such, improved understanding of the hydroclimatological processes involved in drought development is a major scientific imperative of direct practical relevance. To address this research need, this paper investigates the chain of processes linking antecedent ocean–atmosphere variation to summer streamflow drought in Great Britain. Analyses are structured around four distinct drought regions (defined using hierarchical cluster analysis) for the period 1964–2001. Droughts were identified using a novel regional drought area index. Composite analysis of monthly sea surface temperature (SST) prior to drought onset reveals a horseshoe- or tripole-shaped pattern of North Atlantic SST anomalies that is similar to patterns of SST anomalies associated with the North Atlantic Oscillation (NAO). Patterns in geopotential height, wind, moisture vapor flux, and precipitation prior to drought onset support the influence of the NAO but also demonstrate that the atmospheric bridge linking North Atlantic SST to drought development is too complex to be described solely by indices of the NAO. In revealing new information on the chain of processes leading to the development of hydrological drought in Great Britain, this paper has the potential to inform drought-forecasting research and so improve drought preparedness and management.

Corresponding author address: Daniel Kingston, Department of Geography, University of Otago, P.O. Box 56, Dunedin 9054, New Zealand. E-mail: daniel.kingston@geography.otago.ac.nz

Abstract

Droughts are high-impact events that have substantial implications for both human and natural systems. As such, improved understanding of the hydroclimatological processes involved in drought development is a major scientific imperative of direct practical relevance. To address this research need, this paper investigates the chain of processes linking antecedent ocean–atmosphere variation to summer streamflow drought in Great Britain. Analyses are structured around four distinct drought regions (defined using hierarchical cluster analysis) for the period 1964–2001. Droughts were identified using a novel regional drought area index. Composite analysis of monthly sea surface temperature (SST) prior to drought onset reveals a horseshoe- or tripole-shaped pattern of North Atlantic SST anomalies that is similar to patterns of SST anomalies associated with the North Atlantic Oscillation (NAO). Patterns in geopotential height, wind, moisture vapor flux, and precipitation prior to drought onset support the influence of the NAO but also demonstrate that the atmospheric bridge linking North Atlantic SST to drought development is too complex to be described solely by indices of the NAO. In revealing new information on the chain of processes leading to the development of hydrological drought in Great Britain, this paper has the potential to inform drought-forecasting research and so improve drought preparedness and management.

Corresponding author address: Daniel Kingston, Department of Geography, University of Otago, P.O. Box 56, Dunedin 9054, New Zealand. E-mail: daniel.kingston@geography.otago.ac.nz

1. Introduction

Although the climate of Great Britain is influenced heavily by moist maritime air masses originating from the North Atlantic Ocean, droughts are an important natural hazard that may adversely affect society and ecosystems for prolonged periods across large geographical domains (e.g., Fleig et al. 2006; Tallaksen and van Lanen 2004; Monk et al. 2008). The 2003 European drought demonstrated the adverse effects of drought, with forest fires, agricultural losses, disruption to power supply, and transport systems among the many impacts attributed to this event (Tallaksen and van Lanen 2004). As such, improving understanding of the physical causes of drought development is of great importance because it has potential to improve drought-forecasting skills and, thus, drought preparedness and management. The need for such hydroclimatological research is made even more pressing given projected changes in twenty-first-century climate, including hypothesized future acceleration of the hydrological cycle and the probability that a warmer climate will alter the frequency–magnitude distribution of hydrological extremes (Bates et al. 2008; Todd et al. 2011).

Streamflow variation is driven primarily by climate (Bower et al. 2004; Kingston et al. 2009), with catchment properties as a second-order control (Laizé and Hannah 2010). Streamflow at the catchment outlet can be considered both a spatial integration and a temporal average of the net water balance across the catchment. Streamflow drought is defined as a sustained period of streamflow below a certain threshold (Tallaksen et al. 1997) and so, at least in part, may be considered a function of climate variation over longer time periods (from months to years, depending on the underlying climatic and physiographic setting).

In terms of understanding the large-scale climate system processes driving the development of drought, sea surface temperature (SST) offers a potential source of drought predictability given its slowly varying nature and greater persistence in comparison to more rapid atmospheric variability (Kingston et al. 2006a). Although the extratropical North Atlantic atmospheric circulation is considered generally to drive SST, examples have been documented of the opposite situation, with SST leading atmospheric variation (as discussed below). Thus, there is ongoing research into whether (or not) analysis of patterns of SST variation can be used to improve understanding of the underlying climate conditions leading to drought occurrence. In this article, we address this important but unresolved research problem.

Previous research has demonstrated the existence of persistent patterns of SST variation associated with atmospheric circulation in the North Atlantic region, such as the North Atlantic horseshoe (NAH) and tripole patterns (Czaja and Frankignoul 2002; Kushnir et al. 2006). The NAH pattern consists of an SST anomaly neighboring the East Coast of the United States, surrounded by a semicircle-shaped anomaly of the opposite sign, extending westward from the areas south of Greenland and the Caribbean into the eastern North Atlantic (Czaja and Frankignoul 2002). The tripole pattern is similar to the NAH pattern, but with a more distinct separation between subpolar and subtropical regions and with stronger SST features in the western Atlantic compared to the NAH pattern (Kushnir et al. 2006). Both the NAH and tripole patterns have been identified by a number of studies as being linked to the North Atlantic Oscillation (NAO) mode of atmospheric variability at a variety of lagged, leading, and concurrent time steps (e.g., Czaja and Frankignoul 2002; Cassou et al. 2004; Hu and Huang 2006).

The NAO is a large-scale variation in atmospheric mass between the two semipermanent centers of action in the North Atlantic region, the Icelandic low and the Azores high, and is the dominant mode of atmospheric variation in the North Atlantic region (Hurrell and van Loon 1997). The strength and phase of the NAO is measured typically by the difference in standardized air pressure between Iceland and the Azores or an equivalent southern location [i.e., an NAO index (NAOI)] (e.g., Jones et al. 1997). The NAOI is in a positive state when the Azores–Iceland difference in pressure anomalies is positive, with the index negative when the anomaly difference is reversed. A positive winter NAOI is associated with a stronger maritime (Atlantic) influence on the climate of Britain, typically leading to milder and wetter conditions in the west and north and drier conditions in the south and east (Wilby et al. 1997).

Previous studies have identified two distinct relationships between SSTs and the NAO. First, the occurrence of the NAH pattern of SSTs in summer and autumn has been used as a predictor of wintertime NAO variation (e.g., Cassou et al. 2004; Czaja and Frankignoul 2002). For convenience, this SST-to-atmosphere relationship is henceforth referred to as chain 1. Second, and conversely, the tripole pattern has been identified as the primary imprint of strong winter NAO phases on concurrent North Atlantic SSTs (Cassou et al. 2004), with this SST pattern persisting into spring and summer following strong wintertime atmospheric forcing (Hu and Huang 2006). This atmosphere-to-SST relationship is referred to herein as chain 2. Both patterns appear to influence European climate through changes in the meridional temperature gradient associated with variation in tropical SSTs, which then project onto the extratropical atmospheric Rossby waves (Cassou et al. 2004; Hu and Huang 2006). However, the extent, direction, and mechanisms by which North Atlantic SST and atmospheric variation are coupled remain a matter of ongoing research (e.g., Cassou et al. 2004; Hu and Huang 2006; Rodwell et al. 1999).

A number of earlier studies have explored the use of the NAH and/or tripole pattern and NAO as predictors for a range of hydroclimatological variables. Colman (1997) and Colman and Davey (1999) identified a tripole-like pattern as the leading mode of January–February SST variation in the North Atlantic, which they demonstrated could be used as a predictor for the monthly central England temperature time series, European temperature, precipitation, and sea level pressure during the following summer. Other researchers, such as Wedgbrow et al. (2002) and Wilby et al. (2004), have taken this work further to examine the potential for the NAH–tripole pattern and the NAO as a predictor of British streamflow. Although both Wedgbrow et al. (2002) and Wilby et al. (2004) found statistically significant correlation between the winter SST index defined by Colman and Davey (1999), the NAO, and summer streamflow, correlation coefficients were modest (mostly between 0.2 and 0.3). The exploratory study of Kingston et al. (2010) also suggested that SST variation consistent with the NAH–tripole generally preceded British summer streamflow droughts, although the possible nature of the atmospheric bridge linking SST to drought development was discussed only briefly. Predictive relationships between a pattern resembling the NAH in autumn and winter river flow have been identified on the Iberian Peninsula (Gámiz-Fortis et al. 2008), and Kingston et al. (2007) linked winter streamflow in eastern North America to a pattern of surface air temperature anomalies resembling the NAH pattern.

Although connections between the NAH–tripole pattern and hydroclimatological variables in Britain have been found in previous works (as described in the preceding paragraph), these studies have focused typically on identifying maximum statistical predictive power on the basis of indices of atmospheric circulation and SST variation and/or have analyzed explicitly only part of the ocean–atmosphere–river flow process cascade. While such indices can be useful general descriptors of large-scale variation, they inevitably neither capture the full range of variability of the field in question nor always provide a clear insight into the processes involved, as discussed by Kingston et al. (2006a, 2009) and Lavers et al. (2010a,b). Therefore, there is a clear need for more explicit investigation of the processes linking SST to climate and land-surface hydrology. This important research gap is addressed in this paper, which aims to investigate links between monthly North Atlantic SST, atmospheric variables, precipitation, and summer (i.e., June–August) streamflow drought in Great Britain. Summer drought is the focus because this is typically when British water resource supply is least and demand is greatest. This paper advances considerably the exploratory analyses of Kingston et al. (2010) and makes use of British drought regions and a novel regional drought area index (RDAI) developed by Fleig et al. (2010, 2011). The paper starts by overviewing data sources and methodology, after which SST–RDAI relationships are presented. Section 3 provides interpretation of the SST–RDAI relationships, with supporting analyses of atmospheric variables and precipitation used to underpin hydroclimatological process inference. As part of this discussion, the issue of whether atmospheric or SST variation initiates drought development is addressed. The conclusions highlight the wider scientific and practical relevance of this research.

2. Data and methods

a. Data

Streamflow data comprised more than 30 yr of records for gauges with good quality and natural or naturalized daily streamflow data from the U.K. National River Flow Archive. Selection of these records avoids undue human influence on streamflows (e.g., reservoirs or flow augmentation or abstraction), which may cloud analysis of hydrology–climate system links (Hannah et al. 2011). The stations were chosen to cover the whole of Great Britain and included catchments of varying area, streamflow characteristics, and catchment properties (Fleig et al. 2011).

Monthly SST data were obtained from the gridded 1° latitude–longitude resolution Hadley Centre Sea Ice and Sea Surface Temperature (HadISST) dataset (Rayner et al. 2003). The climate dataset includes gridded monthly geopotential height (1000 and 500 hPa), wind vector (1000 and 500 hPa), specific humidity (1000 hPa), relative vorticity (500 hPa), and air temperature (1000 hPa) data from the 2.5° resolution 40-yr European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis (ERA-40) (Uppala et al. 2005). The 1000-hPa fields are taken as indicators of surface climate, and 500-hPa fields are taken as indicators of upper-atmosphere conditions. A visual comparison (not shown) of ERA-40 1000-hPa wind and geopotential fields with 10-m wind vector and mean sea level pressure fields (respectively) determined that the 1000-hPa fields were representative of surface climate conditions, despite occasional instances where the 1000-hPa level was below the land surface. The 1000-hPa specific humidity and wind vector data were used to calculate water vapor flux [following Phillips and McGregor (2001)]; this was used as an indicator of overall moisture transport. Precipitation data were obtained from the 0.5° resolution Climatic Research Unit (CRU) time series 3.0 dataset (Mitchell and Jones 2005). These data were chosen to provide a comprehensive description of hydroclimatically relevant atmospheric variation [following Kingston et al. (2007)]. Streamflow data availability and the use of ERA-40 data restricted the common period for analysis to 1964–2001.

b. Derivation of regional drought area index and drought time series

This paper focuses on Great Britain (GB), for which drought time series were derived for the same regions and novel RDAI method as described in Fleig et al. (2011) and as summarized below. Homogeneous drought regions (GB1–4; Fig. 1) were identified using hierarchical cluster analysis with the complete linkage algorithm. These regions were shown to differ little across a range of clustering algorithms and were relatively insensitive to the time periods analyzed (1986–2000, 1971–85, and 1971–2000). Similar separation of streamflow variation within Britain has been found in previous regional classifications (e.g., Svensson and Prudhomme 2005; Laizé and Hannah 2010; Kingston et al. 2011), adding further credence to the regions identified herein.

Fig. 1.
Fig. 1.

Location of streamflow gauging stations and drought regions.

Citation: Journal of Hydrometeorology 14, 1; 10.1175/JHM-D-11-0100.1

The monthly RDAI was calculated as follows. Monthly-flow time series were classified binarily (i.e., drought or nondrought) for each river gauging station using the Q80 exceedence threshold (i.e., the flow exceeded more than or equal to 80% of the time). From these binary series, monthly regional drought series were derived for each of the four regions of homogenous drought response. The RDAI was calculated as the proportion of the total catchment area within each region affected by drought for each month of the year, resulting in a fractional index ranging between 0 and 1. A region is considered as being in drought when the RDAI > 0.7.

c. Characteristics of drought regions

Fleig et al. (2011) provide a detailed description of the four drought regions. In the context of this paper, perhaps the most relevant characteristics of the regions are those related to climatology, topography, and geology. There is a general (high-to-low) precipitation gradient from northwestern to southeastern Britain as a consequence of both proximity to moisture-bearing Atlantic weather systems and their interception by topographic features. As such, GB2 is relatively wet and GB4 is relatively dry. GB1–2 are located mostly in upland areas on impermeable substrate, but with GB1 catchments principally oriented to the north and east and experiencing greater winter snowfall than GB2 catchments. GB3–4 are primarily lowland regions, with the key difference between these two regions being that GB3 catchments are influenced less by groundwater, whereas GB4 catchments are located over highly productive chalk aquifers.

There are a number of common drought summers across the four regions, but the typical characteristics of drought events vary widely according to regional climatology, topography, and geology. A key indicator of drought characteristics is the drought response time of each region. The response time was derived separately for each region by Fleig et al. (2011) and was based on the highest correlation between the daily RDAI series and drought-supporting circulation types for different time windows (i.e., 10, 15, 30, 45, and 60 days and then in steps of 30 days until 360 days). The circulation types were defined on the basis of the objective version of the Hess–Brezowsky Grosswetterlagen over Europe by James (2007). Catchments in GB4 have the longest mean response time over which atmospheric circulation influences drought development (210 days). This is reflected in the high base flow index (BFI; i.e., the ratio of base flow to total flow volume) of approximately 0.9 for rivers in GB4 (Fleig et al. 2011), emphasizing the importance of long-residence-time groundwater as a source of river flow in this region. In contrast, upland-dominated GB1 and GB2 rivers have low BFI values (0.4–0.5) and a drought response time of just 45 days. GB3 rivers also have low BFI values (~0.4), but with a larger spread in station BFI values and a drought response time of 90 days. The drought response and BFI statistics are reflected in the overall drought characteristics of the regions, with GB1–2 experiencing the most frequent but shortest-duration droughts and GB4 experiencing the least frequent but longest-duration droughts (Fig. 2). Although faster responding, GB3 is most similar to GB4 in terms of drought frequency and duration.

Fig. 2.
Fig. 2.

Monthly drought duration for each of the four regions (light gray: 0 < RDAI < 0.5; medium gray: 0.5 < RDAI < 0.7; dark gray: RDAI > 0.7).

Citation: Journal of Hydrometeorology 14, 1; 10.1175/JHM-D-11-0100.1

d. Identification of ocean–atmosphere–drought relationships

Relationships between the RDAI and SSTs were explored from 1964 to 2001 using composite analysis. One-, two-, and three-month running-mean composites were examined, to cover both monthly and seasonal variation. Composite analysis was used as a robust yet easy-to-interpret means by which to identify statistical linkages and associated physical mechanisms (Kingston et al. 2006b, 2007, 2009). Droughts of at least 3 months duration that occurred in at least one of the summer months (June–August) were selected for the analysis (Fig. 2). In practice, this means that for regions GB1–3 drought conditions were present in at least two of the summer months. However, a number of GB4 events do not pass the RDAI > 0.7 threshold until August, instead building up during the summer and then persisting into autumn and winter. This resulted in the identification of between 6 and 10 drought summers over the study period for each region, with the lowest number of 3-month drought events for GB2 and highest number for GB1 (Fig. 2).

As a check of the sensitivity of composite analysis results to composite size (i.e., number of drought summers), exploratory analyses were undertaken using composites ranging from a uniform 7-member composite across all regions to 11-member composites based on droughts lasting at least 2 months. These analyses revealed the composite anomalies to be relatively insensitive to composite size, with similar SST and atmospheric composite patterns occurring throughout. Additionally, the one-sample Student’s t test was used to compare the composite mean SST and atmospheric fields against the 1964–2001 mean and to determine if differences in SST and climate leading up to drought events were statistically significant. Field significance (i.e., the number of locally significant t tests that could be expected by chance) was taken into account following Kingston et al. (2007). Composite SST anomalies were examined for the 6 months preceding the June–August summer drought period (i.e., from December to May). This 6-month period was chosen partly because of the focus on the NAO (a pattern that is strongest from December to March) and partly because of exploratory analyses for longer time lags that indicated that the strongest and most consistent drought–SST relationships across all regions occur from December to May.

After investigation of SST–drought links, the next step in the analysis was to determine the existence of an “atmospheric bridge” between SST and land surface processes by investigating potential links between drought events and prior atmospheric variation and precipitation. The extent to which SSTs lead atmospheric variation (or vice versa) was also explored. Composite analysis of atmospheric variables and precipitation was employed using the same approach as for SST data.

3. Drought–SST relationships

Composite analysis shows that a number of statistically significant (p ≤ 0.05) SST anomalies occur prior to drought onset: these are the focus of this section. The month with the strongest composite SST anomalies for each drought region is illustrated in Fig. 3; however, SST patterns for all months leading up to drought onset are discussed in the text. Space constraints precluded inclusion of diagrams for all months across the four regions.

Fig. 3.
Fig. 3.

SST composite anomalies (°C) for (a) GB1 May, (b) GB2 May, (c) GB3 April, and (d) GB4 March. Shading indicates statistical significance at p = 0.1 (light), 0.05 (medium), and 0.01 (dark).

Citation: Journal of Hydrometeorology 14, 1; 10.1175/JHM-D-11-0100.1

For GB1, coherent patterns of SST anomalies are evident in summer drought years from December to June. At their strongest, these patterns consist of a relatively large zone of negative SST anomalies in the mid-Atlantic, approximately extending from 20° to 50°N and 60°W to 0°. A much smaller zone of positive SST anomalies occurs around Britain in January and February, moving northward from March to May and weakening in June. SST anomalies are strongest and most widespread from April to June, where the mid-Atlantic negative anomalies expand to assume a near-semicircular shape and are accompanied by positive SST anomalies to the east of Florida in some months (e.g., Fig. 3a).

SST anomalies associated with summer droughts in GB2 comprise primarily a zone of negative anomalies in the mid-Atlantic. These extend from 50°W to 0° at their peak extent in May but are smaller in spatial extent than those associated with GB1 (Fig. 3b). While negative SST anomalies also occur south of Greenland in some months, the pattern of anomalies does not merge into a semicircle to the same extent as for GB1. There is also no consistent zone of positive anomalies adjacent to the British Isles; instead, negative anomalies are present from March to June. Positive SST anomalies are present in the western Atlantic from January to April, but these are smaller, less spatially coherent, and more variable in position than those evident in GB1 composites.

There are similarities between SST anomalies associated with the development of summer drought in GB3 to those already described for GB1. Again, negative anomalies are present in the mid-Atlantic, although for GB3 they are strongest in March and April and are located and extend farther west than those for GB1 (e.g., Fig. 3c). The western Atlantic positive SST anomalies cover a larger area compared to GB1 composites, extending some way out from the eastern coast of North America. Statistically significant positive SST anomalies occur in the vicinity of Britain and north of Scandinavia from January onward, although these are relatively weak around Britain by May.

Patterns of SST anomalies associated with GB4 summer drought occurrence most closely resemble those of GB3 but are more confined to the western Atlantic. Anomalies are present from December onward and are strongest during February–April. The anomalies consist of a strong positive SST anomaly in the Gulf of Mexico, extending eastward from the eastern coast of North America to approximately 45°W, and negative anomalies to the south (Fig. 3d). Anomalies are displaced westward in comparison to GB1–3, extending into the Gulf of Mexico. Positive SST anomalies occur around Britain throughout the January–June period.

In summary, some consistency is evident between SST composite anomalies for all four regions in the 6 months preceding summer (i.e., June–August) drought occurrence. However, the regional SST anomalies are at peak strength during different months for different regions (GB1–2: May; GB3: March–April; GB4: January–March). Notably, these peaks are temporally consistent with the Fleig et al. (2011) drought response times of each region, as described in section 2c. Patterns and timings of SST anomalies are also very similar to the patterns of SST–RDAI correlation described by Kingston et al. (2010).

There are a number of shared drought summers between regions, and as a result, two-sample t tests show that the composite SST anomalies for each region are not always statistically significantly different from each other. Despite the absence of statistically significant differences, and the similarities noted in the previous paragraphs, there are some consistent differences between the SST anomalies for each region. In addition to the different timing of peak anomalies, there appears to be a westward displacement of North Atlantic SST anomalies for regions with longer hydrological response times (i.e., GB1–2 versus GB3–4). This displacement is consistent with the findings of Colman and Davey (1999), who found that North Atlantic SST anomalies preceding warm summers gradually moved from the western to the eastern North Atlantic from winter to summer. Furthermore, such a displacement suggests different forcing mechanisms between GB1–2 and GB3–4. Although not an exact match, GB1 (and, to a lesser extent, GB2) SST anomalies resemble the NAH SST pattern, whereas GB3 and GB4 anomalies display more similarity to the tripole SST pattern. The presence of SST anomalies resembling the NAH and tripole patterns in the months preceding summer drought suggests that there is not a simple SST–atmosphere–drought chain of causality in operation across the four British drought regions. Rather, it appears that there are subtle but important differences in forcing mechanisms between regions, following the distinct chain 1 and chain 2 relationships outlined in section 1. Furthermore, the presence of SST anomalies resembling the NAH and tripole patterns suggests that the NAO may be important for the development of streamflow drought. Both the role of the NAO and the extent to which SSTs lead or lag atmospheric variation are explored in the following section.

4. The atmospheric bridge

This section describes the results of the composite analysis of drought-year atmospheric fields as a basis for discussing the extent to which the NAH and tripole-like patterns of SST anomalies are associated with atmospheric anomalies (i.e., the atmospheric bridge between SST and land-surface processes). Precipitation composites are used to explore catchment water inputs associated with drought for each region and support inference of the process chain. A representative sample of composite climate anomalies associated with drought occurrence is used for graphical illustration (Figs. 48); however, the following text considers all months leading up to drought onset. To clarify explanation of the chain of physical processes, the connections between SST, atmosphere, and drought are outlined schematically for regions GB1, GB3, and GB4 in Fig. 9 (GB2 connections area are weaker and more difficult to generalize in this way). As noted in section 2, air temperature was investigated, but no statistically significant anomalies were detected for any region prior to drought onset (indicating that December–May drought development for all regions is driven primarily by precipitation deficits). As such, air temperature is not discussed further.

Fig. 4.
Fig. 4.

The 500-hPa geopotential height composite anomalies (m) for (a) GB1 May, (b) GB2 May, (c) GB3 March, and (d) GB4 February. Shading as in Fig. 3.

Citation: Journal of Hydrometeorology 14, 1; 10.1175/JHM-D-11-0100.1

Fig. 5.
Fig. 5.

May mean 1000-hPa water vapor flux (kg m−1 Pa−1 s−1) (left) during GB1 drought years and (right) the 1964–2001 mean. Water vapor flux indicated by arrow length (see scale).

Citation: Journal of Hydrometeorology 14, 1; 10.1175/JHM-D-11-0100.1

Fig. 6.
Fig. 6.

Precipitation composite anomalies (mm) for (a) GB1 May–June, (b) GB2 May–June, (c) GB3 March–April, and (d) GB4 February–March 2-month means. Shading as in Fig. 3.

Citation: Journal of Hydrometeorology 14, 1; 10.1175/JHM-D-11-0100.1

Fig. 7.
Fig. 7.

March mean 1000-hPa water vapor flux (kg m−1 Pa−1 s−1) (left) during GB3 drought years and (right) the 1964–2001 mean. Water vapor flux indicated by arrow length (see scale).

Citation: Journal of Hydrometeorology 14, 1; 10.1175/JHM-D-11-0100.1

Fig. 8.
Fig. 8.

February mean 1000-hPa water vapor flux (kg m−1 Pa−1 s−1) (left) during GB4 drought years and (right) the 1964–2001 mean. Water vapor flux indicated by arrow length (see scale).

Citation: Journal of Hydrometeorology 14, 1; 10.1175/JHM-D-11-0100.1

Fig. 9.
Fig. 9.

Schematic model of the relationship between SST, climate, and December–May drought development.

Citation: Journal of Hydrometeorology 14, 1; 10.1175/JHM-D-11-0100.1

Investigation of atmospheric anomalies associated with drought development for GB1 reveals that the SST anomalies for this region (Fig. 3) are accompanied by persistent patterns of upper-atmosphere circulation anomalies from December to May. Although varying between months, these circulation anomalies comprise typically a near-meridional dipole situated over the eastern Atlantic and northwestern Europe and are formed of negative anomalies in the southern center and positive anomalies in the northern center (e.g., Fig. 4a).

The presence of a north–south meridional geopotential height dipole in GB1 composites suggests a pivotal role for the NAO in the development of drought conditions, with the direction of the anomalies suggestive initially of a negative NAO. However, the situation is more complex than a straightforward NAO–drought correlation. For example, the NAO index (Jones et al. 1997) in drought years is not consistently strongly negative, with the mean May NAOI relatively neutral at −0.24 across the 10 GB1 drought years. Examination of climate fields in drought years shows that there are a number of departures from the typical NAO climate signal. December–May 500-hPa geopotential height dipoles for GB1 are located northeast of the typical Azores–Iceland pattern. In May, the maximum height anomalies are situated over and to the west of Britain (southern center), and midway between Iceland and Svalbard (northern center) (Fig. 4a). NAOIs such as the Jones et al. (1997) index are based on fixed locations. These fixed indices have difficulties in capturing more subtle NAO-related climate variation, as discussed previously (e.g., Kingston et al. 2006a; Lavers et al. 2010b). However, even seasonally varying characterizations of the NAO, such as the “mobile” NAO index (Portis et al. 2001) and summer NAO (Folland et al. 2009), vary somewhat from the pattern displayed for GB1 in May.

Irrespective of the link to the NAO, the climate composites are physically consistent with the development of drought conditions in GB1. From December to April the geopotential height anomaly patterns indicate decreased vigor of the westerly circulation over Britain and enhanced meridional airflow, which typically results in lower precipitation (e.g., Kapala et al. 1998; Lavers et al. 2010b). Additionally, in May (the key month for GB1 climate–drought sensitivity and the month of strongest SST anomalies), the north–south dipole is located far to the northeast of the typical Azores–Iceland axis, such that the negative (southern) center is located over Britain (Fig. 4a). As summarized in Fig. 9 and evidenced in Fig. 5, this is associated with stronger winds and increased moisture transport over northern Britain during drought years, but, consistent with Fig. 4a, these are from a southerly direction. Given the north-northeasterly aspect of GB1 catchments nd resultant orographic rain-shadow effects, these conditions may be expected to result in reduced delivery of precipitation to this region. In support of this explanation, precipitation composite anomalies show consistent deficits occurring over northeastern Scotland from December onward, with the largest deficits occurring in May and June (Fig. 6a).

Just as GB2 SST anomalies are somewhat different to (and weaker than) those of GB1, so are the geopotential height anomalies. A similar meridional geopotential height dipole to that of GB1 is present in February and March, with an associated weaker westerly circulation also occurring. However, this circulation pattern weakens in April, with a negative height anomaly present over the eastern Atlantic (including Britain) and a positive anomaly to the northeast in May (Fig. 4b). The cause of the weaker ocean–atmosphere–drought relationships for GB2 in comparison to GB1 is thought to be associated with the greater size and longitudinal extent of GB2, meaning that there is both a greater variety of climate fields that can lead to drought onset and potentially larger variability in hydrological response across this region. Similarly, Fig. 2 shows the starting month for GB2 droughts to be quite variable, especially in comparison with the other drought regions. Because of the seasonal variability in SST and atmospheric patterns and their linkages, such variation in the onset of drought may introduce noise and so reduce significance in the SST and atmospheric composite anomalies. Weaker geopotential height anomalies (in comparison with GB1) are reflected in precipitation anomalies, with deficits occurring in at least part of GB2 from February onward, but not always covering the whole region (Fig. 6b).

It is difficult to determine the precise relationship between the GB1 (and, to a lesser extent, GB2) NAH–like pattern and the concurrent composite climate anomalies. The summer/autumn NAH is associated primarily with the NAO in the following winter (i.e., chain 1; Kushnir et al. 2006), although the NAH has been speculated as originating from the winter tripole pattern (Czaja and Frankignoul 2002). Hertig and Jacobeit (2011) note that the summer NAH is linked concurrently to the Scandinavian atmospheric circulation pattern, although these findings are not replicated here.

SST–drought relationships for GB3 are at their strongest in March and April and for GB4 from February to April. Interestingly, geopotential height anomalies for GB3 and GB4 are at their strongest before this period, with the most-widespread anomalies occurring from February to March for GB3 and from January to February for GB4. The occurrence of strong atmospheric anomalies before strong SST anomalies suggests that atmospheric variation is leading that of the ocean (i.e., following chain 2). Although further analyses are needed before a more-definitive statement can be made, these results are consistent with the similarity of the GB3 and GB4 SST composites to the tripole pattern, which is considered to be generated primarily as an imprint of strong atmospheric variation (Kushnir et al. 2006).

Also consistent with the similarity of GB3 and GB4 SST anomalies to the tripole pattern is evidence of the influence of the NAO in geopotential height composites for these regions, with composite geopotential height anomalies displaying a near-meridional dipole pattern (Figs. 4c,d). These anomalies show a deeper Icelandic low and stronger Azores high prior to drought occurrence for GB4 and a similar (but west shifted) pattern for GB3, both of which indicate that the NAO is in a strong positive phase in these winter months. This is in contrast to the GB1 May anomalies, which make up the opposite meridional dipole pattern (albeit over the eastern, rather than western, North Atlantic). That opposing phases of the NAO are associated with drought development in different regions is not necessarily contradictory. The different locations, physiographic characteristics, and drought response times of the drought regions result in different regional responses to the same large-scale climate situation (Laizé and Hannah 2010).

Despite the apparent link between a strong positive NAO in winters preceding GB4 (and, to a lesser extent, GB3) drought, the mean drought-year (positive) NAOI for both regions varies in strength [mean drought-year NAOI values for GB3 are 1.5 (January), 1.7 (February), and 0.1 (March) and for GB4 are 0.7 (January), 1.83 (February), and 0.7 (March)]. Furthermore, the strongest NAOI values do not occur in the same months as the strongest geopotential height anomalies for either region. As with GB1, this apparent contradiction can be explained through further consideration of the geopotential height composites. The NAOI used here is based on data from fixed locations in Iceland and Gibraltar. The geopotential height composites for GB3 and GB4 display a clear meridional dipole, but with the southern anomaly weakening toward the Iberian Peninsula; hence, a conventional Iceland–Gibraltar NAOI cannot be expected to detect the strong anomalies patterns in Figs. 4c,d.

In terms of the link between NAO-type climate variation and drought occurrence in GB3–4, composite analysis of climate fields suggests that the results are physically consistent with drought propagation in southeastern Britain (i.e., GB3–4; summarized in Fig. 9). For example, geopotential height composites for GB3 (Fig. 4c) demonstrate that while North Atlantic atmospheric circulation is generally more vigorous during drought years, the configuration of the Icelandic low and Azores high results in Atlantic air masses (and so moisture transport) being directed away from southern and eastern Britain (Fig. 7). In addition, differences in relative vorticity (not shown) indicate a more anticyclonic circulation over Britain prior to drought onset. This suite of composite climate anomalies is consistent with reduced delivery of precipitation and the development of drought conditions in GB3 (Fig. 6c; summarized in Fig. 9).

More intense westerly winds and enhanced moisture transport occur over Britain during GB4 drought years (Fig. 8), which is nevertheless thought to be consistent with priming this region for drought. Previous work has demonstrated that an enhanced westerly circulation (as experienced typically under a strong positive NAO) is negatively correlated with precipitation in southern, eastern, and central England (Fowler and Kilsby 2002; Wilby et al. 1997; Dixon et al. 2006) and that a strong positive winter NAO often precedes summer low flow in southern and eastern England (Wedgbrow et al. 2002). This is attributed to rain-shadow effects in the lee of higher relief in western Britain, together with distance from moisture source (i.e., the Atlantic), and is reflected in the general northwest–southeast rainfall gradient across Britain. Although moisture transport from the Atlantic to Britain is relatively high during drought years (Fig. 8), relative vorticity immediately upwind of Britain is reduced (i.e., less cyclonic; not shown), suggesting fewer (rain bearing) weather systems are embedded in the westerly airflow. This is supported here by anomalously low precipitation totals over southeastern England during January–March in drought years (Fig. 6d).

5. Conclusions

This study has revealed the presence of coherent and consistent statistically significant SST and atmospheric anomalies prior to the occurrence of summer streamflow drought in Great Britain. More importantly, it has been demonstrated that these anomalies are physically consistent with the development of drought conditions. Ocean–atmosphere–drought connections are shown to vary for different regions of Great Britain and to be dependent on catchment hydrological response time to climatic inputs. Catchment response time is largely a function of catchment properties such as topography and hydrogeology. Despite catchment modification of ocean–atmosphere–drought relationships and resultant differences in relationships, there are also some similarities between drought regions. SST anomalies associated with drought occurrence for all regions resemble to some extent the NAH or tripole patterns of SST anomalies associated with the NAO. In turn, there is compelling evidence that atmospheric variation associated with drought occurrence is influenced by the NAO.

Differences in the spatial patterns of SST anomalies and their time of occurrence between regions provide insight into the drivers of the ocean–atmosphere system. GB3 and GB4 SST anomalies occur in late winter, resemble the tripole pattern, and follow strong NAO-like atmospheric anomalies. This is consistent with previous studies showing the tripole to be the result of atmospheric forcing (i.e., chain 2, as discussed in section 1). As such, the results suggest that GB3–4 drought development is initiated by atmospheric forcing, at least within the 6-month study period used here. In contrast, GB1 (and, to a lesser extent, GB2) SST anomalies are more similar to the NAH SST pattern, which is more commonly considered as a precursor rather than a result of atmospheric variation (i.e., chain 1). Further research is required to elaborate on the dynamics of ocean–atmosphere interaction leading to drought onset, particularly beyond the 6-month period analyzed here.

At the same time as acknowledging the role of the NAO, it is also important to emphasize that the NAO paradigm (and indices thereof) is unable to describe completely atmospheric variation preceding drought onset. This is, in part, due to shortcomings in the ability of NAOIs to capture fully atmospheric variation in the North Atlantic region and highlights the limitations of using circulation indices as sole descriptors of climate system variation (as discussed in section 1). Such shortcomings may provide some explanation as to why past studies have found only limited correlation between the winter NAO and summer streamflow, despite the importance of the NAO for North Atlantic climate. Basing the description of atmospheric variation solely in terms of the NAO, particularly through use of an index based on fixed locations, fails to capture fully the climatic implications of variation in atmospheric circulation. In turn, this emphasizes the value of composite analysis in developing understanding beyond straightforward linear correlations, as shown here by explanation of the initially contradictory links between the NAO and precipitation deficits in the different drought regions.

Acknowledgments

This work was facilitated by a Research Council of Norway Yggdrasil mobility grant to Daniel Kingston and forms a contribution to the UNESCO–IHP FRIEND water program. Comments and suggestions from the anonymous reviewers and editor are gratefully acknowledged.

REFERENCES

  • Bates, B. C., Kundzewicz Z. W. , Wu S. , and Palutikof J. P. , Eds., 2008: Climate Change and Water. Intergovernmental Panel on Climate Change, 214 pp.

  • Bower, D., Hannah D. M. , and McGregor G. R. , 2004: Techniques for assessing the climatic sensitivity of river flow regimes. Hydrol. Processes, 18, 25152543.

    • Search Google Scholar
    • Export Citation
  • Cassou, C., Deser C. , Terray L. , Hurrell J. W. , and Drevillon M. , 2004: Summer sea surface temperature conditions in the North Atlantic and their impact upon the atmospheric circulation in early winter. J. Climate, 17, 33493363.

    • Search Google Scholar
    • Export Citation
  • Colman, A., 1997: Prediction of summer central England temperature from preceding North Atlantic winter sea surface temperature. Int. J. Climatol., 17, 12851300.

    • Search Google Scholar
    • Export Citation
  • Colman, A., and Davey M. , 1999: Prediction of summer temperature, rainfall and pressure in Europe from preceding winter North Atlantic Ocean temperature. Int. J. Climatol., 19, 513536.

    • Search Google Scholar
    • Export Citation
  • Czaja, A., and Frankignoul C. , 2002: Observed impact of Atlantic SST anomalies on the North Atlantic Oscillation. J. Climate, 15, 606623.

    • Search Google Scholar
    • Export Citation
  • Dixon, H., Lawler D. M. , and Shamseldin A. , 2006: Streamflow trends in western Britain. Geophys. Res. Lett., 33, L19406, doi:10.1029/2006GL027325.

    • Search Google Scholar
    • Export Citation
  • Fleig, A. K., Tallaksen L. M. , Hisdal H. , and Demuth S. , 2006: A global evaluation of discharge drought characteristics. Hydrol. Earth Syst. Sci., 10, 535552.

    • Search Google Scholar
    • Export Citation
  • Fleig, A. K., Tallaksen L. M. , Hisdal H. , Stahl K. , and Hannah D. M. , 2010: Inter-comparison of weather and circulation type classifications for hydrological drought development. Phys. Chem. Earth, 35, 507515.

    • Search Google Scholar
    • Export Citation
  • Fleig, A. K., Tallaksen L. M. , Hisdal H. , and Hannah D. M. , 2011: Regional hydrological drought in north-western Europe: Linking a new regional drought area index with weather types. Hydrol. Processes, 25, 11631179.

    • Search Google Scholar
    • Export Citation
  • Folland, C. K., Knight J. , Linderholm H. W. , Fereday D. , Ineson S. , and Hurrell J. W. , 2009: The summer North Atlantic Oscillation: Past, present, and future. J. Climate, 22, 10821103.

    • Search Google Scholar
    • Export Citation
  • Fowler, H. J., and Kilsby C. G. , 2002: Precipitation and the North Atlantic Oscillation: A study of climatic variation in northern England. Int. J. Climatol., 22, 843866.

    • Search Google Scholar
    • Export Citation
  • Gámiz-Fortis, S., Pozo-Vázquez D. , Trigo R. M. , and Castro-Díez Y. , 2008: Quantifying the predictability of winter river flow in Iberia. Part 2: Seasonal predictability. J. Climate, 21, 25032518.

    • Search Google Scholar
    • Export Citation
  • Hannah, D. M., Demuth S. , van Lanen H. A. J. , Looser U. , Prudhomme C. , Rees G. , Stahl K. , and Tallaksen L. M. , 2011: Large-scale river flow archives: Importance, current status and future needs. Hydrol. Processes, 25, 11911200.

    • Search Google Scholar
    • Export Citation
  • Hertig, E., and Jacobeit J. , 2011: Predictability of Mediterranean climate variables from oceanic variability. Part 1: Sea surface temperature regimes. Climate Dyn., 36, 811823.

    • Search Google Scholar
    • Export Citation
  • Hu, Z., and Huang B. , 2006: Air–sea coupling in the North Atlantic during summer. Climate Dyn., 26, 441457.

  • Hurrell, J. W., and van Loon H. , 1997: Decadal variations in climate associated with the North Atlantic Oscillation. Climatic Change, 36, 301326.

    • Search Google Scholar
    • Export Citation
  • James, P. M., 2007: An objective classification method for Hess and Brezowsky Grosswetterlagen over Europe. Theor. Appl. Climatol., 88, 1742.

    • Search Google Scholar
    • Export Citation
  • Jones, P. D., Jónsson T. , and Wheeler D. , 1997: Extension to the North Atlantic Oscillation using early instrumental pressure observations from Gibraltar and south-west Iceland. Int. J. Climatol., 17, 14331450.

    • Search Google Scholar
    • Export Citation
  • Kapala, A., Mächel H. , and Flohn H. , 1998: Behaviour of the centres of action above the Atlantic since 1881. Part 2: Associations with regional climate anomalies. Int. J. Climatol., 18, 2326.

    • Search Google Scholar
    • Export Citation
  • Kingston, D. G., Lawler D. M. , and McGregor G. R. , 2006a: Linkages between atmospheric circulation, climate and streamflow in the northern North Atlantic: Research prospects. Prog. Phys. Geogr., 30, 143174.

    • Search Google Scholar
    • Export Citation
  • Kingston, D. G., McGregor G. R. , Hannah D. M. , and Lawler D. M. , 2006b: River flow teleconnections across the northern North Atlantic region. Geophys. Res. Lett., 33, L14705, doi:10.1029/2006GL026574.

    • Search Google Scholar
    • Export Citation
  • Kingston, D. G., McGregor G. R. , Hannah D. M. , and Lawler D. M. , 2007: Climatic controls on New England streamflow. J. Hydrometeor., 8, 367379.

    • Search Google Scholar
    • Export Citation
  • Kingston, D. G., Hannah D. M. , Lawler D. M. , and McGregor G. R. , 2009: Climate–river flow relationships across montane and lowland environments in northern Europe. Hydrol. Processes, 23, 985996.

    • Search Google Scholar
    • Export Citation
  • Kingston, D. G., Fleig A. K. , Tallaksen L. M. , and Hannah D. M. , 2010: North Atlantic sea surface temperature, atmospheric circulation and summer drought in Great Britain. Global Change: Facing Risks and Threats to Water Resources, E. Servat et al., Eds., IAHS Publ. 340, 598–604.

  • Kingston, D. G., Hannah D. M. , Lawler D. M. , and McGregor G. R. , 2011: Regional classification, variability and trends of northern North Atlantic river flow. Hydrol. Processes, 25, 10211033.

    • Search Google Scholar
    • Export Citation
  • Kushnir, Y., Robinson W. A. , Chang P. , and Robertson A. W. , 2006: The physical basis for predicting Atlantic sector seasonal-to-interannual climate variability. J. Climate, 19, 59495969.

    • Search Google Scholar
    • Export Citation
  • Laizé, C. L. R., and Hannah D. M. , 2010: Modification of climate-river flow associations by basin properties. J. Hydrol., 389, 186204.

    • Search Google Scholar
    • Export Citation
  • Lavers, D., Prudhomme C. , and Hannah D. M. , 2010a: Large-scale climatic influences on precipitation and discharge for a British river basin. Hydrol. Processes, 24, 25552563.

    • Search Google Scholar
    • Export Citation
  • Lavers, D., Prudhomme C. , and Hannah D. M. , 2010b: Large-scale climate, precipitation and British river flows: Identifying hydroclimatological connections and dynamics. J. Hydrol., 395, 242255.

    • Search Google Scholar
    • Export Citation
  • Mitchell, T. D., and Jones P. I. , 2005: An improved method of constructing a database of monthly climate observations and associated high-resolution grids. Int. J. Climatol., 25, 693712.

    • Search Google Scholar
    • Export Citation
  • Monk, W. A., Wood P. J. , Hannah D. M. , and Wilson D. A. , 2008: Macroinvertebrate community response to inter-annual and regional river flow regime dynamics. River Res. Appl., 24, 9881001.

    • Search Google Scholar
    • Export Citation
  • Phillips, I. D., and McGregor G. R. , 2001: Western European water vapor flux—Southeast England rainfall associations. J. Hydrometeor., 2, 505524.

    • Search Google Scholar
    • Export Citation
  • Portis, D. H., Walsh J. E. , El Hamly M. , and Lamb P. J. , 2001: Seasonality of the North Atlantic Oscillation. J. Climate, 14, 20692078.

    • Search Google Scholar
    • Export Citation
  • Rayner, N. A., Parker D. E. , Horton E. B. , Folland C. K. , Alexander L. V. , Rowell D. P. , Kent E. C. , and Kaplan A. , 2003: Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J. Geophys. Res., 108, 4407, doi:10.1029/2002JD002670.

    • Search Google Scholar
    • Export Citation
  • Rodwell, M. J., Rowell D. P. , and Folland C. K. , 1999: Oceanic forcing of the wintertime North Atlantic Oscillation and European climate. Nature, 398, 320323.

    • Search Google Scholar
    • Export Citation
  • Svensson, C., and Prudhomme C. , 2005: Prediction of British summer river flows using winter predictors. Theor. Appl. Climatol., 82, 115.

    • Search Google Scholar
    • Export Citation
  • Tallaksen, L. M., and van Lanen H. A. J. , Eds., 2004: Hydrological Drought: Processes and Estimation Methods for Streamflow and Groundwater. Developments in Water Sciences, Vol. 48, Elsevier, 580 pp.

  • Tallaksen, L. M., Madsen H. , and Clausen B. , 1997: On the definition and modelling of streamflow drought duration and deficit volume. Hydrol. Sci. J., 42, 1533.

    • Search Google Scholar
    • Export Citation
  • Todd, M. C., Taylor R. G. , Osborn T. J. , Kingston D. G. , Arnell N. W. , and Gosling S. N. , 2011: Uncertainty in climate change impacts on basin-scale freshwater resources–Preface to the special issue: The QUEST-GSI methodology and synthesis of results. Hydrol. Earth Syst. Sci., 15, 10351046.

    • Search Google Scholar
    • Export Citation
  • Uppala, S. M., and Coauthors, 2005: The ERA-40 Re-Analysis. Quart. J. Roy. Meteor. Soc., 131, 29613012.

  • Wedgbrow, C. S., Wilby R. L. , Fox H. R. , and O’Hare G. , 2002: Prospects for seasonal forecasting of summer drought and low river flow anomalies in England and Wales. Int. J. Climatol., 22, 219236.

    • Search Google Scholar
    • Export Citation
  • Wilby, R. L., O’Hare G. , and Barnsley N. , 1997: The North Atlantic Oscillation and British Isles climate variability, 1865–1996. Weather, 52, 266276.

    • Search Google Scholar
    • Export Citation
  • Wilby, R. L., Wedgbrow C. S. , and Fox H. R. , 2004: Seasonal predictability of the summer hydrometeorology of the River Thames, UK. J. Hydrol., 295, 116.

    • Search Google Scholar
    • Export Citation
Save