• Balsamo, G., , A. Beljaars, , K. Scipal, , P. Viterbo, , B. van den Hurk, , M. Hirschi, , and A. K. Betts, 2009: A revised hydrology for the ECMWF model: Verification from field site to terrestrial water storage and impact in the Integrated Forecast System. J. Hydrometeor., 10, 623643, doi:10.1175/2008JHM1068.1.

    • Search Google Scholar
    • Export Citation
  • Banacos, P. C., , and D. M. Schultz, 2005: The use of moisture flux convergence in forecasting convective initiation: Historical and operational perspectives. Wea. Forecasting, 20, 351366, doi:10.1175/WAF858.1.

    • Search Google Scholar
    • Export Citation
  • Becker, A., , P. Finger, , A. Meyer-Christoffer, , B. Rudolf, , K. Schamm, , U. Schneider, , and M. Ziese, 2013: A description of the global land-surface precipitation data products of the Global Precipitation Climatology Centre with sample applications including centennial (trend) analysis from 1901–present. Earth Syst. Sci. Data, 5, 7199, doi:10.5194/essd-5-71-2013.

    • Search Google Scholar
    • Export Citation
  • Berckmans, J., , T. Woollings, , M.-E. Demory, , P.-L. Vidale, , and M. Roberts, 2013: Atmospheric blocking in a high resolution climate model: Influences of mean state, orography and eddy forcing. Atmos. Sci. Lett., 14, 3440, doi:10.1002/asl2.412.

    • Search Google Scholar
    • Export Citation
  • Blackmon, M. L., 1976: A climatological spectral study of the 500 mb geopotential height of the Northern Hemisphere. J. Atmos. Sci., 33, 16071623, doi:10.1175/1520-0469(1976)033<1607:ACSSOT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Bony, S., , J.-L. Dufresne, , H. Le Treut, , J.-J. Morcrette, , and C. Senior, 2004: On dynamic and thermodynamic components of cloud changes. Climate Dyn., 22 (2–3), 7186, doi:10.1007/s00382-003-0369-6.

    • Search Google Scholar
    • Export Citation
  • Champion, A. J., , K. I. Hodges, , L. O. Bengtsson, , N. S. Keenlyside, , and M. Esch, 2011: Impact of increasing resolution and a warmer climate on extreme weather from Northern Hemisphere extratropical cyclones. Tellus, 63A, 893906, doi:10.1111/j.1600-0870.2011.00538.x.

    • Search Google Scholar
    • Export Citation
  • Chang, E. K. M., , Y. Guo, , and X. Xia, 2012: CMIP5 multimodel ensemble projection of storm track change under global warming. J. Geophys. Res., 117, D23118, doi:10.1029/2012JD018578.

    • Search Google Scholar
    • Export Citation
  • Colle, B. A., , Z. Zhang, , K. A. Lombardo, , E. Chang, , P. Liu, , and M. Zhang, 2013: Historical evaluation and future prediction of eastern North American and western Atlantic extratropical cyclones in the CMIP5 models during the cool season. J. Climate, 26, 68826903, doi:10.1175/JCLI-D-12-00498.1.

    • Search Google Scholar
    • Export Citation
  • Dee, D. P., and et al. , 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553597, doi:10.1002/qj.828.

    • Search Google Scholar
    • Export Citation
  • Demory, M.-E., , P. Vidale, , M. Roberts, , P. Berrisford, , J. Strachan, , R. Schiemann, , and M. Mizielinski, 2013: The role of horizontal resolution in simulating drivers of the global hydrological cycle. Climate Dyn., 42, 2201–2225, doi:10.1007/s00382-013-1924-4.

    • Search Google Scholar
    • Export Citation
  • Efron, B., , and R. J. Tibshirani, 1993: An Introduction to the Bootstrap.Chapman & Hall, 456 pp.

  • Emori, S., , and S. J. Brown, 2005: Dynamic and thermodynamic changes in mean and extreme precipitation under changed climate. Geophys. Res. Lett., 32, L17706, doi:10.1029/2005GL023272.

    • Search Google Scholar
    • Export Citation
  • Flato, G., and et al. , 2014: Evaluation of climate models. Climate Change 2013: The Physical Science Basis, T. F. Stocker et al., Eds., Cambridge University Press, 741–866.

  • Graff, L. S., , and J. H. LaCasce, 2012: Changes in the extratropical storm tracks in response to changes in SST in an AGCM. J. Climate, 25, 18541870, doi:10.1175/JCLI-D-11-00174.1.

    • Search Google Scholar
    • Export Citation
  • Haarsma, R. J., , W. Hazeleger, , C. Severijns, , H. de Vries, , A. Sterl, , R. Bintanja, , G. J. van Oldenborgh, , and H. W. van den Brink, 2013: More hurricanes to hit Western Europe due to global warming. Geophys. Res. Lett., 40, 17831788, doi:10.1002/grl.50360.

    • Search Google Scholar
    • Export Citation
  • Hack, J. J., , J. M. Caron, , G. Danabasoglu, , K. W. Oleson, , C. Bitz, , and J. E. Truesdale, 2006: CCSMCAM3 climate simulation sensitivity to changes in horizontal resolution. J. Climate, 19, 22672289, doi:10.1175/JCLI3764.1.

    • Search Google Scholar
    • Export Citation
  • Hawcroft, M. K., , L. C. Shaffrey, , K. I. Hodges, , and H. F. Dacre, 2012: How much Northern Hemisphere precipitation is associated with extratropical cyclones? Geophys. Res. Lett., 39, doi:10.1029/2012GL053866.

    • Search Google Scholar
    • Export Citation
  • Haylock, M. R., , N. Hofstra, , A. M. G. K. Tank, , E. J. Klok, , P. D. Jones, , and M. New, 2008: A European daily high-resolution gridded data set of surface temperature and precipitation for 1950–2006. J. Geophys. Res.,113, D20119, doi:10.1029/2008JD010201.

  • Hazeleger, W., and et al. , 2010: EC-Earth: A seamless Earth-system prediction approach in action. Bull. Amer. Meteor. Soc., 91, 13571363, doi:10.1175/2010BAMS2877.1.

    • Search Google Scholar
    • Export Citation
  • Hazeleger, W., and et al. , 2012: EC-Earth v2.2: Description and validation of a new seamless Earth system prediction model. Climate Dyn., 39, 26112629, doi:10.1007/s00382-011-1228-5.

    • Search Google Scholar
    • Export Citation
  • Jones, P. D., , and I. Harris, 2014: CRU TS3.22: Climatic Research Unit (CRU) Time-Series (TS) version 3.22 of high resolution gridded data of month-by-month variation in climate (Jan. 1901-Dec. 2013). Centre for Environmental Data Archival Dataset, doi:10.5285/18BE23F8-D252-482D-8AF9-5D6A2D40990C.

  • Jones, P. D., , E. B. Horton, , C. K. Folland, , M. Hulme, , D. E. Parker, , and T. A. Basnett, 1999: The use of indices to identify changes in climatic extremes. Climatic Change, 42, 131149, doi:10.1023/A:1005468316392.

    • Search Google Scholar
    • Export Citation
  • Jung, T., and et al. , 2012: High-resolution global climate simulations with the ECMWF model in Project Athena: Experimental design, model climate, and seasonal forecast skill. J. Climate, 25, 31553172, doi:10.1175/JCLI-D-11-00265.1.

    • Search Google Scholar
    • Export Citation
  • Katz, R. W., , M. B. Parlange, , and P. Naveau, 2002: Statistics of extremes in hydrology. Adv. Water Resour., 25 (8–12), 12871304, doi:10.1016/S0309-1708(02)00056-8.

    • Search Google Scholar
    • Export Citation
  • Pfahl, S., , and H. Wernli, 2012: Quantifying the relevance of cyclones for precipitation extremes. J. Climate, 25, 67706780, doi:10.1175/JCLI-D-11-00705.1.

    • Search Google Scholar
    • Export Citation
  • Pope, V., , and R. Stratton, 2002: The processes governing horizontal resolution sensitivity in a climate model. Climate Dyn., 19 (3–4), 211236, doi:10.1007/s00382-001-0222-8.

    • Search Google Scholar
    • Export Citation
  • Reynolds, R. W., , N. A. Rayner, , T. M. Smith, , D. C. Stokes, , and W. Wang, 2002: An improved in situ and satellite SST analysis for climate. J. Climate, 15, 16091625, doi:10.1175/1520-0442(2002)015<1609:AIISAS>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Seager, R., , and N. Henderson, 2013: Diagnostic computation of moisture budgets in the ERA-Interim reanalysis with reference to analysis of CMIP-archived atmospheric model data. J. Climate, 26, 7876–7901, doi:10.1175/JCLI-D-13-00018.1.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., , J. T. Fasullo, , and J. Mackaro, 2011: Atmospheric moisture transports from ocean to land and global energy flows in reanalyses. J. Climate, 24, 4907–4924, doi:10.1175/2011JCLI4171.1.

    • Search Google Scholar
    • Export Citation
  • van den Hurk, B. J., , P. Viterbo, , A. Beljaars, , and A. Betts, 2000: Offline validation of the ERA40 surface scheme. ECMWF Tech. Rep., 43 pp.

    • Search Google Scholar
    • Export Citation
  • van Haren, R., , G. J. van Oldenborgh, , G. Lenderink, , M. Collins, , and W. Hazeleger, 2013a: SST and circulation trend biases cause an underestimation of European precipitation trends. Climate Dyn., 40, 120, doi:10.1007/s00382-012-1401-5.

    • Search Google Scholar
    • Export Citation
  • van Haren, R., , G. J. van Oldenborgh, , G. Lenderink, , and W. Hazeleger, 2013b: Evaluation of modeled changes in extreme precipitation in Europe and the Rhine basin. Environ. Res. Lett.,8, 014053, doi:10.1088/1748-9326/8/1/014053.

  • Willison, J., , W. A. Robinson, , and G. M. Lackmann, 2013: The importance of resolving mesoscale latent heating in the North Atlantic storm track. J. Atmos. Sci., 70, 2234–2250, doi:10.1175/JAS-D-12-0226.1.

    • Search Google Scholar
    • Export Citation
  • Zahn, M., , and R. P. Allan, 2011: Changes in water vapor transports of the ascending branch of the tropical circulation. J. Geophys. Res.,116, D18111, doi:10.1029/2011JD016206.

  • Zahn, M., , and R. P. Allan, 2013: Quantifying present and projected future atmospheric moisture transports onto land. Water Resour. Res., 49, 72667277, doi:10.1002/2012WR013209.

    • Search Google Scholar
    • Export Citation
  • Zappa, G., , L. C. Shaffrey, , and K. I. Hodges, 2013: The ability of CMIP5 models to simulate North Atlantic extratropical cyclones. J. Climate, 26, 53795396, doi:10.1175/JCLI-D-12-00501.1.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Comparison precipitation datasets for average winter (NDJFMA) precipitation for 1982–2011 (mm day−1): (a) ERA-Interim; (b) E-OBS; (c) CRU; (d) GPCC; (e) standard deviation between (a)–(d).

  • View in gallery

    Outline of the study area.

  • View in gallery

    (a) Average precipitation in the study area and (b) 2-yr return value of maximum daily precipitation in the study area. The shaded area indicates the 95% confidence interval as determined from bootstrapping.

  • View in gallery

    Average NDJFMA (a),(b) precipitation and (d),(e) evaporation for the medium- and high-resolution models and (c),(f) the differences between the model resolutions (mm day−1). Differences with p > 0.05 (estimated with a two-sided Student’s t test) have been masked white.

  • View in gallery

    Moisture convergence for our study area (mm day−1) computed using (a) PE and (b) Gauss’s theorem. The shaded areas in (a) indicate the 95% confidence interval determined by bootstrapping.

  • View in gallery

    Difference average NDJFMA moisture convergence (mm day−1). Differences with p > 0.05 (estimated with a two-sided Student’s t test) have been masked white.

  • View in gallery

    Average NDJFMA precipitation difference with E-OBS (mm day−1). Differences with p > 0.05 (estimated with a two-sided Student’s t test) have been masked white.

  • View in gallery

    (left) Specific humidity at 850 hPa (kg kg−1), (center) wind speed at 850 hPa (m s−1), and (right) integrated water vapor transport (kg m−1 s−1). Differences with p > 0.05 (estimated with a two-sided Student’s t test) have been masked white.

  • View in gallery

    Storm track calculated as the variance of 2–8-day bandpass filtered Z500 (m4 s−4).

  • View in gallery

    Relative frequency distribution of daily NDJFMA precipitation averaged over the study area. The inset panel shows the tail of the distribution.

  • View in gallery

    (left) Relative frequency distribution of w500 and (right) probability distribution of precipitation per w500 bin averaged over the study area: (top) T159, (middle) T799, and (bottom) T799 − T159. The error bars are standard errors. On (left), the error bars are represented by the standard error of Poisson counting (n1/2/k) and the error bars on (right) are calculated assuming a normal distribution within each bin (σ/n1/2) (n is the number of elements in a bin and k is the total number of elements).

  • View in gallery

    Decomposition of precipitation difference between the medium- and high-resolution models (%) according to the method by (Bony et al. 2004). Negative values indicate less precipitation in the high-resolution model.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 42 42 10
PDF Downloads 26 26 10

Resolution Dependence of European Precipitation in a State-of-the-Art Atmospheric General Circulation Model

View More View Less
  • 1 Royal Netherlands Meteorological Institute, De Bilt, and Netherlands eScience Center, Amsterdam, Netherlands
  • | 2 Royal Netherlands Meteorological Institute, De Bilt, Netherlands
  • | 3 Royal Netherlands Meteorological Institute, De Bilt, and Meteorology and Air Quality Section, Wageningen University, Wageningen, and Netherlands eScience Center, Amsterdam, Netherlands
© Get Permissions
Full access

Abstract

In this study, the authors investigate the effect of GCM spatial resolution on modeled precipitation over Europe. The objectives of the analysis are to determine whether climate models have sufficient spatial resolution to have an accurate representation of the storm tracks that affect precipitation. They investigate if there is a significant statistical difference in modeled precipitation between a medium-resolution (~112-km horizontal resolution) and a high-resolution (~25-km horizontal resolution) version of a state-of-the-art AGCM (EC-EARTH), if either model resolution gives a better representation of precipitation in the current climate, and what processes are responsible for the differences in modeled precipitation. The authors find that the high-resolution model gives a more accurate representation of northern and central European winter precipitation. The medium-resolution model has a larger positive bias in precipitation in most of the northern half of Europe. Storm tracks are better simulated in the high-resolution model, providing for a more accurate horizontal moisture transport and moisture convergence. Using a decomposition of the precipitation difference between the medium- and high-resolution model in a part related and a part unrelated to a difference in the distribution of vertical atmospheric velocity, the authors find that the smaller precipitation bias in central and northern Europe is largely unrelated to a difference in vertical velocity distribution. The smaller precipitation amount in these areas is in agreement with less moisture transport over this area in the high-resolution model. In areas with orography the change in vertical velocity distribution is found to be more important.

Corresponding author address: Ronald van Haren, Netherlands eScience Center, Science Park 140, 1098 XG Amsterdam, The Netherlands. E-mail: r.vanharen@esciencecenter.nl

Abstract

In this study, the authors investigate the effect of GCM spatial resolution on modeled precipitation over Europe. The objectives of the analysis are to determine whether climate models have sufficient spatial resolution to have an accurate representation of the storm tracks that affect precipitation. They investigate if there is a significant statistical difference in modeled precipitation between a medium-resolution (~112-km horizontal resolution) and a high-resolution (~25-km horizontal resolution) version of a state-of-the-art AGCM (EC-EARTH), if either model resolution gives a better representation of precipitation in the current climate, and what processes are responsible for the differences in modeled precipitation. The authors find that the high-resolution model gives a more accurate representation of northern and central European winter precipitation. The medium-resolution model has a larger positive bias in precipitation in most of the northern half of Europe. Storm tracks are better simulated in the high-resolution model, providing for a more accurate horizontal moisture transport and moisture convergence. Using a decomposition of the precipitation difference between the medium- and high-resolution model in a part related and a part unrelated to a difference in the distribution of vertical atmospheric velocity, the authors find that the smaller precipitation bias in central and northern Europe is largely unrelated to a difference in vertical velocity distribution. The smaller precipitation amount in these areas is in agreement with less moisture transport over this area in the high-resolution model. In areas with orography the change in vertical velocity distribution is found to be more important.

Corresponding author address: Ronald van Haren, Netherlands eScience Center, Science Park 140, 1098 XG Amsterdam, The Netherlands. E-mail: r.vanharen@esciencecenter.nl

1. Introduction

General circulation models (GCMs) attempt to simulate Earth’s climate. Often these models are used to isolate the drivers of climate change in response to natural and/or anthropogenic forcings. While some features are well represented in GCMs (e.g., global temperature), other aspects remain uncertain (Flato et al. 2014). One of these aspects is (regional) precipitation in Europe (van Haren et al. 2013a,b). A correct representation of precipitation in climate models is, among others, relevant for hydrological applications, such as flood risk management, navigation and energy production.

It remains to be seen whether the current generation of GCMs have sufficient spatial resolution to resolve the physical processes affecting climate (Pope and Stratton 2002). However, running high-resolution models is often expensive in terms of computing and data storage. Previous studies have shown an improvement in the representation of atmospheric circulation with increased horizontal model resolution, as well as an improvement of the representation of precipitation with improved atmospheric circulation. These studies have focused on one of the following points: 1) changes in the representation of the storm track or large-scale circulation with resolution (Jung et al. 2012; Willison et al. 2013; Zappa et al. 2013; Colle et al. 2013; Hack et al. 2006); 2) changes of precipitation within the storm track with resolution (Champion et al. 2011); 3) changes in blocking frequency with resolution (Berckmans et al. 2013; Jung et al. 2012); 4) circulation dependence of precipitation (not resolution dependent; van Haren et al. 2013a,b); and 5) effect of resolution on global average land–ocean precipitation partitioning (Demory et al. 2013). For different regions the impact of any of these caused by a change in model resolution is different. Here we combine the results of these studies in an analysis of regional precipitation over Europe. The objectives of the paper are to determine whether climate models have sufficient spatial resolution to have an accurate representation of storm tracks affecting precipitation over Europe. We investigate if there is a significant statistical difference in modeled precipitation between a medium-resolution and a high-resolution ensemble of a state-of-the-art global AGCM, if either of the model resolutions gives a better representation of precipitation in the actual climate system, and what processes are responsible for the differences in modeled precipitation. AGCMs simplify the climate system by constraining it by observed boundary conditions (sea surface temperatures and sea ice cover) that 1) make their results more comparable to observations and reanalyses; 2) allow for a better comparison between models; and 3) make it easier to isolate atmospheric processes responsible for affecting the hydrological cycle in climate models with various resolutions (Demory et al. 2013).

This paper is outlined as follows: In section 2, we define our study area, introduce the datasets used in this study, and discuss the methodology. In section 3, we calculate the difference in precipitation and moisture convergence between the two model resolutions. In search for causes of the differences in modeled precipitation, we investigate differences in moisture transport and storm tracks in section 4. In section 5, we discuss the distribution of daily precipitation over our study area and try to link differences with differences in strength/frequency of atmospheric disturbances. Finally, some conclusions are drawn in section 6.

2. Data and methods

a. Data

The model data are from EC-EARTH, version 2.3, a state-of-the-art GCM developed by a consortium of European research institutions. The atmospheric component of the model is derived from the weather forecast model [Integrated Forecast System (IFS) cycle 31r1] of the European Centre for Medium-Range Weather Forecasts (ECMWF; Hazeleger et al. 2010). EC-EARTH differs from the weather forecast model by a small number of changes in the physics parameterizations, applied to optimize the model for climate simulations (Hazeleger et al. 2012): An improved description of the entrainment of environmental air in deep convecting plumes from IFS cycle 32r3 was used. Also the inhomogeneity scaling factor for shortwave cloud optical thickness has been reduced from 0.7 to 0.57 and an improved mass conservation correction scheme has been applied (the scheme from IFS cycle 33r2). The land surface component in IFS cycle 31r1, Tiled ECMWF Scheme for Surface Exchanges over Land (TESSEL; van den Hurk et al. 2000), was replaced by H-TESSEL (Balsamo et al. 2009), which uses an improved representation of hydrology, as in more recent cycles of IFS.

The experiment (Haarsma et al. 2013) consists of two sets of 5-yr six-member ensemble simulations from 2002 to 2006, resulting in 30 yr of data for each set. The sets differ both in horizontal and vertical resolution. The model resolution of the medium-resolution ensemble is T159L62 (spectral triangular truncation at wavenumber 159 equivalent to ~112-km horizontal resolution, 62 vertical levels). The high-resolution ensemble is at T799L91 (spectral triangular truncation at wavenumber 799 equivalent to ~25-km horizontal resolution, 91 vertical levels). The parameterizations packages of the high- and medium-resolution model runs are the same. Observed greenhouse gases and aerosol concentrations were used in the simulations. Also, observed sea surface temperatures (SSTs) and sea ice coverage were prescribed. The daily SSTs and sea ice data were taken from the daily optimum interpolation (OI) SST analysis (Reynolds et al. 2002) at 0.25° resolution and interpolated on the model grid. A 10-yr spinup run at medium resolution (T159) was made, followed by a 9-month (from January to October) spinup run at the desired resolution. The initial spinup was done at medium resolution for both model ensembles in order to save computing resources. The six-member ensemble was made by taking the atmospheric state of one of the first six days of October as initial state for each member. Thereafter, the model was run for another three months until 1 January before the data were used for the analysis. After this spinup, the spread in the atmospheric states was sufficient to treat the six runs as independent members. This computational very expensive experiment was done for multiple research questions. One of those was the impact of climate change on teleconnection responses to specific tropical SST patterns (Haarsma et al. 2013). This motivated the larger ensemble approach of shorter runs. The research questions discussed in this paper could also be studied with a fewer longer runs.

The model data are verified against ERA-Interim (Dee et al. 2011), a global atmospheric reanalysis produced by the ECMWF, extending back to 1979. We used the period 1982–2011 for the ERA-Interim data. ERA-Interim has a T255L60 resolution (spectral triangular truncation at wavenumber 255 equivalent to ~80-km horizontal resolution, 60 vertical levels). Reanalysis data provide a multivariate, spatially homogeneous, and coherent record of the global atmospheric circulation. This reanalysis uses a single data assimilation system and is therefore not affected by changes in method. A sufficiently realistic model is able to extrapolate information from locally observed parameters to unobserved parameters at nearby locations, and it can also propagate this information forward in time. ERA-Interim uses a four-dimensional data assimilation system that allows the analyzed fields to evolve smoothly in time instead of with jumps at times of analyses.

In this way it is possible, for example, to obtain meaningful precipitation estimates from a reanalysis of temperature, humidity, and wind observations (Dee et al. 2011). The forecast parameters (precipitation and evaporation) were calculated from averaging the accumulated values from the beginning of the forecast, initialized at 12 h from 0000 and 1200 UTC, over the range of 12 h.

Because precipitation in ERA-Interim is in fact calculated using short-range model forecast, a second evaluation of the simulated precipitation is performed using the state-of-the-art gridded high-resolution (0.5° horizontal resolution) daily precipitation fields of the European ENSEMBLES project version 9.0 (E-OBS; Haylock et al. 2008). The dataset is based on meteorological station measurements and is designed to provide the best estimate of gridbox averages to enable direct comparison with climate models. The same period is selected as for the ERA-Interim data (1982–2011). Figure 1 compares daily average November–April (NDJFMA) winter precipitation for ERA-Interim, E-OBS, and two additional observational datasets: the Climatic Research Unit (CRU) time series (TS), version 3.22 (Jones and Harris 2014), and the dataset from the Global Precipitation Climatology Centre (GPCC), version 6 (Becker et al. 2013). All four datasets agree well over most of the European area for this period (Fig. 1e; exceptions are areas with orography), providing confidence in the representation of precipitation in ERA-Interim and E-OBS.

Fig. 1.
Fig. 1.

Comparison precipitation datasets for average winter (NDJFMA) precipitation for 1982–2011 (mm day−1): (a) ERA-Interim; (b) E-OBS; (c) CRU; (d) GPCC; (e) standard deviation between (a)–(d).

Citation: Journal of Climate 28, 13; 10.1175/JCLI-D-14-00279.1

Because of storage limitations, the data were first averaged to daily means where applicable. To allow a fair comparison the data were then regridded onto the grid of the medium-resolution (T159) ensemble. Regridding was done by means of second-order conservative remapping (Jones et al. 1999). In conservative remapping, the flux on the new (destination) grid results in the same energy or water exchange as the flux on the old (source) grid. Second-order conservative remapping is more accurate compared to first-order conservative remapping, at the expense of computational demands. Note that we use a 30-yr continuous period for the reanalysis and observations to verify the two sets of 5-yr six-member ensemble simulations. While this makes the influence of natural variability on the estimated quantities much better comparable, some differences between model results and observations may be due to different decadal variability because of different SSTs.

b. Study area

We focus in this study on European winter precipitation. In earlier studies, we found that, for this area and season, circulation-driven precipitation trends are not well represented in climate models (van Haren et al. 2013a,b). Increased spatial resolution in climate models may give a more accurate circulation and the associated precipitation. We discuss some results in more detail for a smaller area which is outlined in Fig. 2. This area consists of the Netherlands, Belgium, Luxembourg, Germany, and part of France. It covers two major river basins (Meuse and Rhine) and no major orography. Winter precipitation in this area is relevant for hydrological applications such as flood risk management, navigation, and energy production. When we refer to the study area in later parts of this paper, we refer to this smaller area.

Fig. 2.
Fig. 2.

Outline of the study area.

Citation: Journal of Climate 28, 13; 10.1175/JCLI-D-14-00279.1

c. Methods

1) Moisture convergence

Moisture convergence follows from the conservation of water vapor in the atmosphere (e.g., Banacos and Schultz 2005). Under the assumption that condensed water immediately precipitates out, moisture convergence is equal to the difference between evaporation e and precipitation p,
e1
where the vector V represents the horizontal wind components, w is the vertical velocity, and q is the specific humidity. Vertical integration between the surface and the top of the atmosphere yields
e2
where g is the gravitational constant and P and E are surface precipitation and evaporation, which follow from vertically integrating p e, respectively. The subscript s refers to surface quantities. The lowest and highest available model output levels are 850 and 200 hPa, respectively. Under assumption that the majority of the horizontal moisture convergence takes place within this layer, Eq. (2) can be simplified as (Seager and Henderson 2013)
e3
The 30-yr-averaged NDJFMA moisture convergence is related to P and E via
e4
where the angle brackets indicate a seasonal average and the overline indicates a 30-yr average. We find that the last term on the right-hand side of Eq. (4) is small compared to the other terms [O(10−4)] and is ignored. The difference in moisture convergence between medium- and high-resolution models is finally written as
e5
While the (approximate) equality in Eq. (5) may be technically true, in reality this may not be the case due to (Zahn and Allan 2013): 1) numerical issues, 2) the use of instantaneous velocity and humidity fields whereas P and E are fluxes and noninstantaneous, and 3) limited vertical resolution of saved output fields (only atmospheric data at 850, 700, 500, 300, and 200 hPa were saved). Moreover, for reanalysis the moisture budget is generally not closed, which effectively adds another term to Eq. (5) (Trenberth et al. 2011). Because quantitative agreement may not be possible for these reasons, we will only look for qualitative agreement between the left- and right-hand side of Eq. (5).

Note that the integrations in the above equation and further integrations in this study are discretized before they are calculated.

2) Integrated vapor transport

Integrated vapor transport (IVT) is defined as
e6
where g is the gravitational constant, q is the specific humidity, p is the atmospheric pressure, and u and υ are the horizontal wind speed components. The integration takes place between 850 and 200 hPa (the lowest and highest vertical output level available).

3) Precipitation decomposition

The precipitation difference between the two models is decomposed using the ascending motion of the atmosphere into three different parts: a vertical velocity component, a nonvertical velocity component, and a covariation term. The vertical velocity term indicates the precipitation response is due to differences in strength/frequency of dynamic disturbances (change in distribution of w500). The nonvertical velocity term indicates the difference in precipitation for a given w500, which includes every influence that is not captured by changes of w500 (e.g., horizontal moisture transport). The last term in the equation arises from the correlation of the two effects. The decomposition is given by (Bony et al. 2004)
e7
where pw is the probability distribution of w500 from the medium-resolution model, the precipitation for each w500 bin from the medium-resolution model, and δ indicates the difference between the high- and medium-resolution models (T799 − T159) for the accompanying term. The integration takes place over the whole range of vertical motions and is approximated by summations where the probability distributions are approximated by frequency distributions. We tried different bin sizes for the discretized integration, all giving similar results.

3. Comparison between T159 and T799

a. Precipitation difference between high resolution and medium resolution

The monthly averaged daily precipitation for our study area is shown in Fig. 3a for the two model ensembles, as well as ERA-Interim and E-OBS. We quantify the robustness of our results by bootstrapping [Efron and Tibshirani 1993; bias-corrected accelerated (BCa) method] the 30 yr of data, assuming all years are independent. The error band in the figure represents the 95% confidence interval. The months May–October show in general better agreement between both the ensembles and ERA-Interim. During most of the winter months the average monthly precipitation in the ensembles deviates significantly from the quasi-observed reanalysis amounts. Although both model ensembles significantly overestimate the amount of winter precipitation compared to ERA-Interim, the bias in the high-resolution model is much smaller. The difference in NDJFMA winter precipitation between the medium- and high-resolution models is approximately 20%. E-OBS observed precipitation data agree well with the quasi-observed ERA-Interim data, confirming the overestimation of the modeled precipitation. Similar results were obtained for daily extreme precipitation [Fig. 3b; 2-yr return value estimated by fitting a generalized extreme value (GEV) distribution; Katz et al. 2002], although the signal is much more noisy.

Fig. 3.
Fig. 3.

(a) Average precipitation in the study area and (b) 2-yr return value of maximum daily precipitation in the study area. The shaded area indicates the 95% confidence interval as determined from bootstrapping.

Citation: Journal of Climate 28, 13; 10.1175/JCLI-D-14-00279.1

b. P − E and moisture convergence

To explore the physical processes that can cause these differences, we consider the average NDJFMA winter precipitation and evaporation in Europe in Fig. 4. The amount of winter precipitation is less in the high-resolution model in much of northern Europe (exceptions are areas with orographic precipitation), with differences as large as 0.5 mm day−1 found in France, the Netherlands, Belgium, Luxembourg, Germany, the Czech Republic, and Poland (roughly equal to our smaller study area). Larger amounts of winter precipitation in the higher-resolution model are found in southern Europe, where also differences as large as 0.5 mm day−1 are found for much of Portugal, Spain, and Italy. Areas with orographic precipitation have in general more precipitation (>1 mm day−1) in the high-resolution model.

Fig. 4.
Fig. 4.

Average NDJFMA (a),(b) precipitation and (d),(e) evaporation for the medium- and high-resolution models and (c),(f) the differences between the model resolutions (mm day−1). Differences with p > 0.05 (estimated with a two-sided Student’s t test) have been masked white.

Citation: Journal of Climate 28, 13; 10.1175/JCLI-D-14-00279.1

Over land, we find only small differences in average NDJFMA evaporation between the two model resolutions. The most noteworthy difference is found in our study area, with less evaporation (0.2 mm day−1) in the high-resolution model. This could be related to the decrease in precipitation in the same area, as described in the previous paragraph. In most of continental Europe, the difference in precipitation (Fig. 4c) is larger than the difference in evaporation between the two model resolutions (Fig. 4f).

To include the effect of circulation in our analysis of the differences, we study the moisture convergence [see section 2c(1)]. Calculations of the right-hand side of Eq. (5) resulted in very noisy patterns, likely related to numerical issues and the limited number of vertical model output levels available. For an area average, this term may also be computed as a line integral around a boundary using Gauss’s theorem (e.g., Zahn and Allan 2011). Moisture convergence found from applying Gauss’s theorem to our study area is given in Fig. 5b. The left-hand side of the Eq. (5) is shown in Fig. 5a. Both methods agree qualitatively that higher-resolution model has less moisture convergence in the winter months for our smaller study area.

Fig. 5.
Fig. 5.

Moisture convergence for our study area (mm day−1) computed using (a) PE and (b) Gauss’s theorem. The shaded areas in (a) indicate the 95% confidence interval determined by bootstrapping.

Citation: Journal of Climate 28, 13; 10.1175/JCLI-D-14-00279.1

Figure 6 considers the moisture convergence as estimated from the left-hand side of Eq. (5) for the larger European area. Figure 6c shows lower moisture convergence in much of the northern half of Europe in the high-resolution model compared to the medium-resolution model. Exceptions are areas with high orography in Scotland and Norway, where the high-resolution model has higher and steeper mountains. In the southern half of Europe there is an increase in moisture convergence, between the medium to the high-resolution model. Comparing Fig. 6c with the difference in average precipitation (Fig. 4c) and evaporation (Fig. 4f) in the two model resolutions in Fig. 4, we conclude that, over land, precipitation is the dominant term in the left-hand side of Eq. (5). Therefore, the results suggest that the difference in precipitation over Europe between the two models (i.e., the dipole in Fig. 4c) is at least partly related to the convergence and advection of moisture. A comparison of the moisture convergence of the two models with ERA-Interim (Figs. 6a,b), shows a more accurate representation of moisture convergence in most of the central and northern part of continental Europe in the high-resolution model. There is no clear improvement in the southern part of Europe, in fact, in highland areas the agreement with ERA-Interim is in general worse for the high-resolution model. In other areas the magnitude of the bias is roughly the same, although the sign changes in some areas. The worse agreement in highland areas is likely related to the orography in ERA-Interim, which is more comparable with the orography in the medium-resolution model than it is with the orography in the high-resolution model. Figure 7 indeed shows that the precipitation bias in these areas is not larger in the high-resolution model when compared to actual observations (E-OBS).

Fig. 6.
Fig. 6.

Difference average NDJFMA moisture convergence (mm day−1). Differences with p > 0.05 (estimated with a two-sided Student’s t test) have been masked white.

Citation: Journal of Climate 28, 13; 10.1175/JCLI-D-14-00279.1

Fig. 7.
Fig. 7.

Average NDJFMA precipitation difference with E-OBS (mm day−1). Differences with p > 0.05 (estimated with a two-sided Student’s t test) have been masked white.

Citation: Journal of Climate 28, 13; 10.1175/JCLI-D-14-00279.1

4. Circulation

a. Moisture transport

To better understand the difference in moisture convergence between the medium- and high-resolution models, we consider the individual quantities from the moisture convergence definition in Fig. 8: specific humidity in the left column, wind speed in the center column, and IVT (see section 2) in the right column of the figure.

Fig. 8.
Fig. 8.

(left) Specific humidity at 850 hPa (kg kg−1), (center) wind speed at 850 hPa (m s−1), and (right) integrated water vapor transport (kg m−1 s−1). Differences with p > 0.05 (estimated with a two-sided Student’s t test) have been masked white.

Citation: Journal of Climate 28, 13; 10.1175/JCLI-D-14-00279.1

Specific humidity at 850 hPa is in the high-resolution model on average lower in the northern half of continental Europe and higher in the southern half of continental Europe, compared to the medium-resolution model (Fig. 8j). A similar difference pattern is found at 700 hPa but, as moisture decreases with height, with somewhat lower values (not shown). Upon closer inspection of the vertical profile of the atmospheric moisture content in the models, we find that the difference in atmospheric moisture content between the high- and medium-resolution models in our study area is in the order of 5% (not shown), about 4 times smaller than the difference in average precipitation. This suggests that circulation differences play a large role.

The center column of Fig. 8 shows that the average wind speed of the low level flow is too large in the medium-resolution model compared to reanalysis data. The high-resolution model is much closer to the reanalysis data. The combined effect of specific humidity and wind speed is less transport of moisture (IVT: Fig. 8, right) from the ocean to the western part of Europe at higher resolution. Figure 6 suggests that the extra moisture transported into our study area in the medium-resolution model (partly) converges and rains out, thereby increasing precipitation. This is confirmed by considering only the zonal component of the moisture transport.

b. Storm track

The results shown in Fig. 8 suggest that the main differences are found over the Atlantic. This is the storm-track region where extratropical cyclones form (Blackmon 1976). These storm-track regions are associated with increased precipitation and winds and are subject to extreme weather events (e.g., Graff and LaCasce 2012). The weather in Europe is heavily influenced by the storm track over the North Atlantic: Hawcroft et al. (2012) estimate that over 70% of total winter precipitation in large parts of Europe is associated with the passage of an extratropical cyclone, and Pfahl and Wernli (2012) found a high percentage of precipitation extremes within the storm-track region to be directly related to cyclones.

We calculate the storm track as the seasonal variance of 2–8-day bandpass filtered geopotential height at 500 hPa (Z500) (Blackmon 1976), which is shown in Fig. 9. Note that the storm-track definition does not discriminate between cyclones, anticyclones, and variability not related to geopotential minima. The storm track in the medium-resolution model appears to be too zonal, which is a common problem in coarse-resolution GCMs (Chang et al. 2012; Zappa et al. 2013). The location of the storm track in the high-resolution model is more realistic: fewer and/or less intense storms pass over central Europe. For our study area, the mean area-averaged absolute bias of the storm track as shown in Fig. 9 improves with respect to ERA-Interim from 12.3 m4 s−4 in the medium-resolution model (Fig. 9e) to 9.8 m4 s−4 in the high-resolution model (Fig. 9f). In areas north and south of our study area in general larger improvements are found for the high-resolution model.

Fig. 9.
Fig. 9.

Storm track calculated as the variance of 2–8-day bandpass filtered Z500 (m4 s−4).

Citation: Journal of Climate 28, 13; 10.1175/JCLI-D-14-00279.1

The better representation of the storm track seems to be in agreement with Zappa et al. (2013); Colle et al. (2013) found that the performance of GCMs in representing North Atlantic cyclones was strongly dependent on model resolution: models from phase 5 of CMIP (CMIP5) performed better than models from phase 3 of CMIP (CMIP3) and higher-resolution CMIP5 models performed better than lower-resolution CMIP5 models. They found that higher-resolution models had a more realistic representation of the North Atlantic storm track in terms of location, track density, and intensity. The resolution used here for our high-resolution model is not found in CMIP5. In fact, our medium-resolution model is included in CMIP5 and is one of the highest-resolution models. Colle et al. (2013) showed that our medium-resolution model was the best performing CMIP5 model in simulating western Atlantic extratropical cyclones in both track density and intensity, and Zappa et al. (2013) rank it among the best performing CMIP5 models in simulating the storm-track position, tilt, their number, and their intensity. Figure 9 shows that increasing the resolution even further still improves the modeled storm track.

5. Precipitation decomposition

The results of the previous sections suggest that the high-resolution model provides more accurate horizontal moisture transport and moisture convergence, caused partly by a lower humidity but mainly by a more realistic representation of the North Atlantic storm track. To look in more detail to the area-averaged precipitation in our study area, we consider the precipitation distribution in both models as well as ERA-Interim in Fig. 10. We find that both models underestimate the number of “dry” days (0–1 mm): 12% for the medium-resolution model and 5% for the high-resolution model. Both models overestimate the frequency of area-averaged precipitation across the range of intensities larger than 2 mm day−1. Similar to the frequency of dry days, the bias with respect to ERA-Interim is smaller for the high-resolution model.

Fig. 10.
Fig. 10.

Relative frequency distribution of daily NDJFMA precipitation averaged over the study area. The inset panel shows the tail of the distribution.

Citation: Journal of Climate 28, 13; 10.1175/JCLI-D-14-00279.1

To understand these differences, we consider the change in the strength and/or frequency of dynamic disturbances and the associated change in precipitation, averaged over the study area, in Fig. 11. Similar to Emori and Brown (2005), we use daily mean 500-hPa vertical velocity (w500) as a proxy of the strength of dynamic disturbance. The change in distribution of w500 between the medium- and high-resolution models is shown in the left panels of Fig. 11. Note that positive values of w500 represent upward motion here, unlike the usual definition of pressure velocity (ω500 scaled by −1). The figure shows an increase in extremes of w500 (both positive and negative) and a decrease of days with moderate vertical velocities when increasing the resolution from T159 to T799. The widening of the distribution is a general effect of the increase in horizontal resolution in the model and is also found in other areas of Europe (not shown).

Fig. 11.
Fig. 11.

(left) Relative frequency distribution of w500 and (right) probability distribution of precipitation per w500 bin averaged over the study area: (top) T159, (middle) T799, and (bottom) T799 − T159. The error bars are standard errors. On (left), the error bars are represented by the standard error of Poisson counting (n1/2/k) and the error bars on (right) are calculated assuming a normal distribution within each bin (σ/n1/2) (n is the number of elements in a bin and k is the total number of elements).

Citation: Journal of Climate 28, 13; 10.1175/JCLI-D-14-00279.1

The increase in frequency of days with subsidence in the high-resolution model is in agreement with the increase in number of dry days, as was shown in Fig. 10. An increase in days with extreme precipitation associated with the increase in strong positive upward motion of the atmosphere is however not found. The latter is related to a decrease in precipitation in the high-resolution model for the same upward motion (Fig. 11, bottom right). Reduced horizontal moisture transport likely reduces the amount of available moisture to precipitate.

To confirm this, we decompose the precipitation difference between the two models using the ascending motion of the atmosphere into three different parts: a vertical velocity component, a nonvertical velocity component, and a covariation term (see section 3). The distributions of the individual quantities are shown in Fig. 11 for the study area.

Figure 12 shows the spatial distribution of the three components in the precipitation decomposition. The nonvertical velocity component is responsible for much of the lower precipitation in the high-resolution model in central and northern Europe, indicating that it is caused by less moisture available to precipitate. The decrease in precipitation in these areas is in agreement with reduced moisture transport over this area in the high-resolution model (Fig. 8l).

Fig. 12.
Fig. 12.

Decomposition of precipitation difference between the medium- and high-resolution models (%) according to the method by (Bony et al. 2004). Negative values indicate less precipitation in the high-resolution model.

Citation: Journal of Climate 28, 13; 10.1175/JCLI-D-14-00279.1

The vertical velocity component is mainly positive along areas with high orography. Orography is much more pronounced in the high-resolution model, resulting in increased precipitation and a change in w500 distribution. Increased storm-track activity in the high-resolution model (Fig. 9d) could also be a factor in the positive vertical velocity component in southern Spain and Scandinavia, where the difference in moisture transport between the two model resolutions is relatively small (Fig. 8l). The covariation term is small everywhere: that is, there is low correlation between the change in precipitation resulting from changes in vertical velocity distribution and the change in precipitation unrelated to changes in vertical velocity distribution.

6. Conclusions

In this study, we investigated if there is a significant statistical difference in modeled precipitation between a medium-resolution and a high-resolution state-of-the-art AGCM model and if either of the model resolutions gives a better representation of precipitation in the actual climate system. The same AGCM was used for both model resolutions.

We found that the high-resolution model gives a more accurate representation of northern and central European winter precipitation than the medium-resolution model, both in the mean state and in the extremes. The medium-resolution model has a larger positive bias in precipitation in most of the northern half of Europe. In the southern half of Europe, the magnitude of the precipitation bias is approximately the same, but the sign of the bias changes at some locations. We found a large difference in moisture transport and moisture convergence between the two model resolutions. In agreement with recent studies using multimodel ensembles, we found that the performance of the model in representing the North Atlantic storm track was strongly dependent on model resolution: the high-resolution model gives a more realistic representation.

A closer inspection of precipitation in our study area (the western part of continental mid-Europe) reveals a higher frequency of dry days in the high-resolution model, closer to the observed frequency. We found that this is related to a thickening of the tail in the downward motion regime (subsidence) of the w500 distribution in the high-resolution model. A thickening of the tail in the upward motion regime is not found to increase precipitation extremes. The latter is related to a decrease in precipitation for the same upward motion in the high-resolution model.

Using a decomposition of the precipitation difference between the medium- and high-resolution models in a part related (vertical velocity component) and unrelated (nonvertical velocity component) to a difference in the distribution of w500, we found that the nonvertical velocity component is responsible for much of the smaller precipitation bias in central and northern Europe. The decrease in precipitation in these areas is in agreement with reduced moisture transport over this area in the high-resolution model. The difference in w500 distribution is only a minor factor in the difference in total precipitation over large parts of central and northern Europe. We found that the vertical velocity component is much more important along areas with high orography. Major orography is better represented in the high-resolution model, resulting in increased precipitation and a change in w500 distribution.

These results are relevant for climate studies that assess present and future precipitation changes. To get climate information on the fine spatial scale that is required by decision makers, statistical downscaling or dynamical downscaling is often applied. Dynamical downscaling is done by embedding a high-resolution regional climate model within a coarse-resolution global model, allowing for a better representation of orographic and coastal effects, as well as more resolved model physics. Regional climate models however, are often largely dependent on the storm tracks in the driving GCM because of their relatively small spatial domain. Our findings show, assuming that sufficient temporal- and vertical-resolution data are saved to do a detailed moisture budget analysis, that our high-resolution AGCM has a better representation of the North Atlantic storm track and therefore precipitation. This may be valid for other GCMs as well, showing the necessity to analyze other GCMs that may become available in the future with such high horizontal resolutions.

Acknowledgments

We thank the reviewers for their valuable comments which helped to considerably improve the quality of the paper. The research was supported by the Dutch research program Knowledge for Climate.

REFERENCES

  • Balsamo, G., , A. Beljaars, , K. Scipal, , P. Viterbo, , B. van den Hurk, , M. Hirschi, , and A. K. Betts, 2009: A revised hydrology for the ECMWF model: Verification from field site to terrestrial water storage and impact in the Integrated Forecast System. J. Hydrometeor., 10, 623643, doi:10.1175/2008JHM1068.1.

    • Search Google Scholar
    • Export Citation
  • Banacos, P. C., , and D. M. Schultz, 2005: The use of moisture flux convergence in forecasting convective initiation: Historical and operational perspectives. Wea. Forecasting, 20, 351366, doi:10.1175/WAF858.1.

    • Search Google Scholar
    • Export Citation
  • Becker, A., , P. Finger, , A. Meyer-Christoffer, , B. Rudolf, , K. Schamm, , U. Schneider, , and M. Ziese, 2013: A description of the global land-surface precipitation data products of the Global Precipitation Climatology Centre with sample applications including centennial (trend) analysis from 1901–present. Earth Syst. Sci. Data, 5, 7199, doi:10.5194/essd-5-71-2013.

    • Search Google Scholar
    • Export Citation
  • Berckmans, J., , T. Woollings, , M.-E. Demory, , P.-L. Vidale, , and M. Roberts, 2013: Atmospheric blocking in a high resolution climate model: Influences of mean state, orography and eddy forcing. Atmos. Sci. Lett., 14, 3440, doi:10.1002/asl2.412.

    • Search Google Scholar
    • Export Citation
  • Blackmon, M. L., 1976: A climatological spectral study of the 500 mb geopotential height of the Northern Hemisphere. J. Atmos. Sci., 33, 16071623, doi:10.1175/1520-0469(1976)033<1607:ACSSOT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Bony, S., , J.-L. Dufresne, , H. Le Treut, , J.-J. Morcrette, , and C. Senior, 2004: On dynamic and thermodynamic components of cloud changes. Climate Dyn., 22 (2–3), 7186, doi:10.1007/s00382-003-0369-6.

    • Search Google Scholar
    • Export Citation
  • Champion, A. J., , K. I. Hodges, , L. O. Bengtsson, , N. S. Keenlyside, , and M. Esch, 2011: Impact of increasing resolution and a warmer climate on extreme weather from Northern Hemisphere extratropical cyclones. Tellus, 63A, 893906, doi:10.1111/j.1600-0870.2011.00538.x.

    • Search Google Scholar
    • Export Citation
  • Chang, E. K. M., , Y. Guo, , and X. Xia, 2012: CMIP5 multimodel ensemble projection of storm track change under global warming. J. Geophys. Res., 117, D23118, doi:10.1029/2012JD018578.

    • Search Google Scholar
    • Export Citation
  • Colle, B. A., , Z. Zhang, , K. A. Lombardo, , E. Chang, , P. Liu, , and M. Zhang, 2013: Historical evaluation and future prediction of eastern North American and western Atlantic extratropical cyclones in the CMIP5 models during the cool season. J. Climate, 26, 68826903, doi:10.1175/JCLI-D-12-00498.1.

    • Search Google Scholar
    • Export Citation
  • Dee, D. P., and et al. , 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553597, doi:10.1002/qj.828.

    • Search Google Scholar
    • Export Citation
  • Demory, M.-E., , P. Vidale, , M. Roberts, , P. Berrisford, , J. Strachan, , R. Schiemann, , and M. Mizielinski, 2013: The role of horizontal resolution in simulating drivers of the global hydrological cycle. Climate Dyn., 42, 2201–2225, doi:10.1007/s00382-013-1924-4.

    • Search Google Scholar
    • Export Citation
  • Efron, B., , and R. J. Tibshirani, 1993: An Introduction to the Bootstrap.Chapman & Hall, 456 pp.

  • Emori, S., , and S. J. Brown, 2005: Dynamic and thermodynamic changes in mean and extreme precipitation under changed climate. Geophys. Res. Lett., 32, L17706, doi:10.1029/2005GL023272.

    • Search Google Scholar
    • Export Citation
  • Flato, G., and et al. , 2014: Evaluation of climate models. Climate Change 2013: The Physical Science Basis, T. F. Stocker et al., Eds., Cambridge University Press, 741–866.

  • Graff, L. S., , and J. H. LaCasce, 2012: Changes in the extratropical storm tracks in response to changes in SST in an AGCM. J. Climate, 25, 18541870, doi:10.1175/JCLI-D-11-00174.1.

    • Search Google Scholar
    • Export Citation
  • Haarsma, R. J., , W. Hazeleger, , C. Severijns, , H. de Vries, , A. Sterl, , R. Bintanja, , G. J. van Oldenborgh, , and H. W. van den Brink, 2013: More hurricanes to hit Western Europe due to global warming. Geophys. Res. Lett., 40, 17831788, doi:10.1002/grl.50360.

    • Search Google Scholar
    • Export Citation
  • Hack, J. J., , J. M. Caron, , G. Danabasoglu, , K. W. Oleson, , C. Bitz, , and J. E. Truesdale, 2006: CCSMCAM3 climate simulation sensitivity to changes in horizontal resolution. J. Climate, 19, 22672289, doi:10.1175/JCLI3764.1.

    • Search Google Scholar
    • Export Citation
  • Hawcroft, M. K., , L. C. Shaffrey, , K. I. Hodges, , and H. F. Dacre, 2012: How much Northern Hemisphere precipitation is associated with extratropical cyclones? Geophys. Res. Lett., 39, doi:10.1029/2012GL053866.

    • Search Google Scholar
    • Export Citation
  • Haylock, M. R., , N. Hofstra, , A. M. G. K. Tank, , E. J. Klok, , P. D. Jones, , and M. New, 2008: A European daily high-resolution gridded data set of surface temperature and precipitation for 1950–2006. J. Geophys. Res.,113, D20119, doi:10.1029/2008JD010201.

  • Hazeleger, W., and et al. , 2010: EC-Earth: A seamless Earth-system prediction approach in action. Bull. Amer. Meteor. Soc., 91, 13571363, doi:10.1175/2010BAMS2877.1.

    • Search Google Scholar
    • Export Citation
  • Hazeleger, W., and et al. , 2012: EC-Earth v2.2: Description and validation of a new seamless Earth system prediction model. Climate Dyn., 39, 26112629, doi:10.1007/s00382-011-1228-5.

    • Search Google Scholar
    • Export Citation
  • Jones, P. D., , and I. Harris, 2014: CRU TS3.22: Climatic Research Unit (CRU) Time-Series (TS) version 3.22 of high resolution gridded data of month-by-month variation in climate (Jan. 1901-Dec. 2013). Centre for Environmental Data Archival Dataset, doi:10.5285/18BE23F8-D252-482D-8AF9-5D6A2D40990C.

  • Jones, P. D., , E. B. Horton, , C. K. Folland, , M. Hulme, , D. E. Parker, , and T. A. Basnett, 1999: The use of indices to identify changes in climatic extremes. Climatic Change, 42, 131149, doi:10.1023/A:1005468316392.

    • Search Google Scholar
    • Export Citation
  • Jung, T., and et al. , 2012: High-resolution global climate simulations with the ECMWF model in Project Athena: Experimental design, model climate, and seasonal forecast skill. J. Climate, 25, 31553172, doi:10.1175/JCLI-D-11-00265.1.

    • Search Google Scholar
    • Export Citation
  • Katz, R. W., , M. B. Parlange, , and P. Naveau, 2002: Statistics of extremes in hydrology. Adv. Water Resour., 25 (8–12), 12871304, doi:10.1016/S0309-1708(02)00056-8.

    • Search Google Scholar
    • Export Citation
  • Pfahl, S., , and H. Wernli, 2012: Quantifying the relevance of cyclones for precipitation extremes. J. Climate, 25, 67706780, doi:10.1175/JCLI-D-11-00705.1.

    • Search Google Scholar
    • Export Citation
  • Pope, V., , and R. Stratton, 2002: The processes governing horizontal resolution sensitivity in a climate model. Climate Dyn., 19 (3–4), 211236, doi:10.1007/s00382-001-0222-8.

    • Search Google Scholar
    • Export Citation
  • Reynolds, R. W., , N. A. Rayner, , T. M. Smith, , D. C. Stokes, , and W. Wang, 2002: An improved in situ and satellite SST analysis for climate. J. Climate, 15, 16091625, doi:10.1175/1520-0442(2002)015<1609:AIISAS>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Seager, R., , and N. Henderson, 2013: Diagnostic computation of moisture budgets in the ERA-Interim reanalysis with reference to analysis of CMIP-archived atmospheric model data. J. Climate, 26, 7876–7901, doi:10.1175/JCLI-D-13-00018.1.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., , J. T. Fasullo, , and J. Mackaro, 2011: Atmospheric moisture transports from ocean to land and global energy flows in reanalyses. J. Climate, 24, 4907–4924, doi:10.1175/2011JCLI4171.1.

    • Search Google Scholar
    • Export Citation
  • van den Hurk, B. J., , P. Viterbo, , A. Beljaars, , and A. Betts, 2000: Offline validation of the ERA40 surface scheme. ECMWF Tech. Rep., 43 pp.

    • Search Google Scholar
    • Export Citation
  • van Haren, R., , G. J. van Oldenborgh, , G. Lenderink, , M. Collins, , and W. Hazeleger, 2013a: SST and circulation trend biases cause an underestimation of European precipitation trends. Climate Dyn., 40, 120, doi:10.1007/s00382-012-1401-5.

    • Search Google Scholar
    • Export Citation
  • van Haren, R., , G. J. van Oldenborgh, , G. Lenderink, , and W. Hazeleger, 2013b: Evaluation of modeled changes in extreme precipitation in Europe and the Rhine basin. Environ. Res. Lett.,8, 014053, doi:10.1088/1748-9326/8/1/014053.

  • Willison, J., , W. A. Robinson, , and G. M. Lackmann, 2013: The importance of resolving mesoscale latent heating in the North Atlantic storm track. J. Atmos. Sci., 70, 2234–2250, doi:10.1175/JAS-D-12-0226.1.

    • Search Google Scholar
    • Export Citation
  • Zahn, M., , and R. P. Allan, 2011: Changes in water vapor transports of the ascending branch of the tropical circulation. J. Geophys. Res.,116, D18111, doi:10.1029/2011JD016206.

  • Zahn, M., , and R. P. Allan, 2013: Quantifying present and projected future atmospheric moisture transports onto land. Water Resour. Res., 49, 72667277, doi:10.1002/2012WR013209.

    • Search Google Scholar
    • Export Citation
  • Zappa, G., , L. C. Shaffrey, , and K. I. Hodges, 2013: The ability of CMIP5 models to simulate North Atlantic extratropical cyclones. J. Climate, 26, 53795396, doi:10.1175/JCLI-D-12-00501.1.

    • Search Google Scholar
    • Export Citation
Save