• Dettinger, M. D., 2011: Climate change, atmospheric rivers, and floods in California—A multimodel analysis of storm frequency and magnitude changes. J. Amer. Water Resour. Assoc., 47, 514523, doi:10.1111/j.1752-1688.2011.00546.x.

    • Search Google Scholar
    • Export Citation
  • Dettinger, M. D., , F. M. Ralph, , T. Das, , P. J. Neiman, , and D. R. Cayan, 2011: Atmospheric rivers, floods, and the water resources of California. Water, 3, 445478, doi:10.3390/w3020445.

    • Search Google Scholar
    • Export Citation
  • Du, Q., , V. Faber, , and M. Gunzburger, 1999: Centroidal Voroni tessellations: Applications and algorithms. SIAM Rev., 41, 637676, doi:10.1137/S0036144599352836.

    • Search Google Scholar
    • Export Citation
  • Evans, K. J., , P. Lauritzen, , S. K. Mishra, , R. Neale, , M. A. Taylor, , and J. J. Tribbia, 2013: AMIP simulations with the CAM4 spectral element dynamical core. J. Climate, 26, 689709, doi:10.1175/JCLI-D-11-00448.1.

    • Search Google Scholar
    • Export Citation
  • Gates, W. L., and et al. , 1999: An overview of the results of the Atmospheric Model Intercomparison Project (AMIP I). Bull. Amer. Meteor. Soc., 80, 2955, doi:10.1175/1520-0477(1999)080<0029:AOOTRO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Gent, P. R., and et al. , 2011: The Community Climate System Model version 4. J. Climate, 24, 49734991, doi:10.1175/2011JCLI4083.1.

    • Search Google Scholar
    • Export Citation
  • Guan, B., , N. P. Molotch, , D. E. Waliser, , E. J. Fetzer, , and P. J. Neiman, 2010: Extreme snowfall events linked to atmospheric rivers and surface air temperature via satellite measurements. Geophys. Res. Lett., 37, L20401, doi:10.1029/2010GL044696.

    • Search Google Scholar
    • Export Citation
  • Hagos, S., , R. Leung, , S. A. Rauscher, , and T. Ringler, 2013: Errors characteristics of two grid refinement approaches in aquaplanet simulations: MPAS-A and WRF. Mon. Wea. Rev.,141, 3022–3036, doi:10.1175/MWR-D-12-00338.1.

  • Ju, L. T., , T. Ringler, , and M. Gunzburger, 2011: Voroni tessellations and their application to climate and global modeling. Numerical Techniques for Global Atmospheric Models, P. H. Lauritzen et al., Eds., Lecture Notes in Computational Science and Engineering, Vol. 80, Springer, 313–342.

  • Kanamitsu, M., , W. Ebisuzaki, , J. Woollen, , S.-K. Yang, , J. J. Hnilo, , M. Fiorino, , and G. L. Potter, 2002: NCEP–DOE AMIP-II Reanalysis (R-2). Bull. Amer. Meteor. Soc., 83, 16311643, doi:10.1175/BAMS-83-11-1631.

    • Search Google Scholar
    • Export Citation
  • Landu, K., , L. R. Leung, , S. Hagos, , V. Vinoj, , S. Rauscher, , T. Ringler, , and M. Taylor, 2014: The dependence of ITCZ structure on model resolution and dynamical core in aquaplanet simulations. J. Climate, 27, 23752385, doi:10.1175/JCLI-D-13-00269.1.

    • Search Google Scholar
    • Export Citation
  • Lavers, D. A., , R. P. Allan, , G. Villarini, , B. Lloyd-Hughes, , D. J. Brayshaw, , and A. J. Wade, 2013: 2013: Future changes in atmospheric rivers and their implications for winter flooding in Britain. Environ. Res. Lett., 8, 034010, doi:10.1088/1748-9326/8/3/034010.

    • Search Google Scholar
    • Export Citation
  • Leung, L. R., , and Y. Qian, 2009: Atmospheric rivers induced heavy precipitation and flooding in the western U.S. simulated by the WRF regional climate model. Geophys. Res. Lett., 36, L03820, doi:10.1029/2008GL036445.

    • Search Google Scholar
    • Export Citation
  • Leung, L. R., , T. Ringler, , W. D. Collins, , and M. Ashfaq, 2013: A hierarchical evaluation of regional climate simulations. Eos, Trans. Amer. Geophys. Union, 94, 297298.

    • Search Google Scholar
    • Export Citation
  • Luce, C. H., , J. T. Abatzoglou, , and Z. A. Holden, 2013: The missing mountain water: Slower westerlies decrease orographic enhancement in the Pacific Northwest USA. Science, 342, 13601364, doi:10.1126/science.1242335.

    • Search Google Scholar
    • Export Citation
  • Neale, R. B., , and B. J. Hoskins, 2000: A standard test for AGCMs including their physical parametrizations: I: The proposal. Atmos. Sci. Lett., 1, 101107, doi:10.1006/asle.2000.0022.

    • Search Google Scholar
    • Export Citation
  • Neelin, J. D., , B. Langenbrunner, , J. E. Meyerson, , A. Hall, , and N. Berg, 2013: California winter precipitation change under global warming in the Coupled Model Intercomparison Project phase 5 ensemble. J. Climate, 26, 62386256, doi:10.1175/JCLI-D-12-00514.1.

    • Search Google Scholar
    • Export Citation
  • Neiman, P. J., , F. M. Ralph, , G. A. Wick, , J. D. Lundquist, , and M. D. Dettinger, 2008: Meteorological characteristics and overland precipitation impacts of atmospheric rivers affecting the west coast of North American based on eight years of SSM/I satellite observations. J. Hydrometeor., 9, 2247, doi:10.1175/2007JHM855.1.

    • Search Google Scholar
    • Export Citation
  • Neiman, P. J., , L. J. Schick, , F. M. Ralph, , M. Hughes, , and G. A. Wick, 2011: Flooding in western Washington: The connection to atmospheric rivers. J. Hydrometeor., 12, 13371358, doi:10.1175/2011JHM1358.1.

    • Search Google Scholar
    • Export Citation
  • Payne, A. E., , and G. Magnusdottir, 2014: Dynamics of landfalling atmospheric rivers over the North Pacific in 30 years of MERRA reanalysis. J. Climate, 27, 71337150, doi:10.1175/JCLI-D-14-00034.1.

    • Search Google Scholar
    • Export Citation
  • Pierce, D. W., and et al. , 2013: The key role of heavy precipitation events in climate model disagreements of future annual precipitation changes in California. J. Climate, 26, 58795896, doi:10.1175/JCLI-D-12-00766.1.

    • Search Google Scholar
    • Export Citation
  • Ralph, F. M., , and M. D. Dettinger, 2011: Storms, floods, and the science of atmospheric rivers. Eos, Trans. Amer. Geophys. Union, 92, 265, doi:10.1029/2011EO320001.

    • Search Google Scholar
    • Export Citation
  • Ralph, F. M., , and M. D. Dettinger, 2012: Historical and national perspectives on extreme West Coast precipitation associated with atmospheric rivers during December 2010. Bull. Amer. Meteor. Soc., 93, 783790, doi:10.1175/BAMS-D-11-00188.1.

    • Search Google Scholar
    • Export Citation
  • Ralph, F. M., , P. J. Neiman, , and G. A. Wick, 2004: Satellite and CALJET aircraft observations of atmospheric rivers over the eastern North Pacific Ocean during the winter of 1997. Mon. Wea. Rev., 132, 17211745, doi:10.1175/1520-0493(2004)132<1721:SACAOO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Ralph, F. M., , P. J. Neiman, , G. A. Wick, , S. I. Gutman, , M. D. Dettinger, , D. R. Cayan, , and A. B. White, 2006: Flooding on California's Russian River: Role of atmospheric rivers. Geophys. Res. Lett., 33, L13801, doi:10.1029/2006GL026689.

    • Search Google Scholar
    • Export Citation
  • Ralph, F. M., , T. Coleman, , P. J. Neiman, , R. J. Zamora, , and M. D. Dettinger, 2013: Observed impacts of duration and seasonality of atmospheric-river landfalls on soil moisture and runoff in coastal Northern California. J. Hydrometeor., 14, 443459, doi:10.1175/JHM-D-12-076.1.

    • Search Google Scholar
    • Export Citation
  • Rauscher, S., , T. Ringler, , W. C. Skamarock, , and A. A. Mirin, 2013: Exploring a global multiresolution modeling approach using aquaplanet simulations. J. Climate, 26, 24322452, doi:10.1175/JCLI-D-12-00154.1.

    • Search Google Scholar
    • Export Citation
  • Ringler, T., , L. Ju, , and M. Gunzburger, 2008: A multi resolution method for climate system modeling: Application of spectral centroidal Voroni tessellations. Ocean Dyn., 58, 475498, doi:10.1007/s10236-008-0157-2.

    • Search Google Scholar
    • Export Citation
  • Ringler, T., , J. Thuburn, , J. B. Klemp, , and W. C. Skamarock, 2010: A unified approach to energy conservation and potential vorticity dynamics for arbitrarily structured C-grids. J. Comput. Phys., 229, 30653090, doi:10.1016/j.jcp.2009.12.007.

    • Search Google Scholar
    • Export Citation
  • Ryoo, J., , Y. Kaspi, , D. Waugh, , G. Kiladis, , D. Waliser, , E. Fetzer, , and J. Kim, 2013: Impact of Rossby wave breaking on U.S. West Coast winter precipitation during ENSO events. J. Climate, 26, 63606382, doi:10.1175/JCLI-D-12-00297.1.

    • Search Google Scholar
    • Export Citation
  • Skamarock, W. C., , J. B. Klemp, , M. G. Duda, , L. Fowler, , S.-H. Park, , and T. D. Ringler, 2012: A multiscale nonhydrostatic atmospheric model using centroidal Voronoi tessellations and C-grid staggering. Mon. Wea. Rev., 140, 30903105, doi:10.1175/MWR-D-11-00215.1.

    • Search Google Scholar
    • Export Citation
  • Stohl, A., , C. Forster, , and H. Sodemann, 2008: Remote sources of water vapor forming precipitation on the Norwegian west coast at 60°N—A tale of hurricanes and an atmospheric river. J. Geophys. Res., 113, D05102, doi:10.1029/2007JD00900.

    • Search Google Scholar
    • Export Citation
  • Taylor, M. A., , J. Edwards, , and A. St. Cyr, 2008: Petascale atmospheric models for the Community Climate System Model: New developments and evaluation of scalable dynamical cores. J. Phys., 125, 012023, doi:10.1088/1742-6596/125/1/012023.

    • Search Google Scholar
    • Export Citation
  • Thuburn, J., , T. D. Ringler, , W. C. Skamarock, , and J. B. Klemp, 2009: Numerical representation of geostrophic modes on arbitrarily structured C-grids. J. Comput. Phys., 228, 83218335, doi:10.1016/j.jcp.2009.08.006.

    • Search Google Scholar
    • Export Citation
  • Viale, M., , and M. N. Nuñez, 2011: Climatology of winter orographic precipitation over the subtropical central Andes and associated synoptic and regional characteristics. J. Hydrometeor., 12, 481507, doi:10.1175/2010JHM1284.1.

    • Search Google Scholar
    • Export Citation
  • Welch, W. T., , and K. K. Tung, 1998: On the equilibrium spectrum of transient waves in the atmosphere. J. Atmos. Sci., 55, 28332851, doi:10.1175/1520-0469(1998)055<2833:OTESOT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Wick, G. A., , P. J. Neiman, , F. M. Ralph, , and T. M. Hamill, 2013: Evaluation of forecasts of the water vapor signature of atmospheric rivers in operational numerical weather prediction models. Wea. Forecasting, 28, 13371352, doi:10.1175/WAF-D-13-00025.1.

    • Search Google Scholar
    • Export Citation
  • Williamson, D. L., 2008: Convergence of aquaplanet simulations with increasing resolution in the Community Atmospheric Model, version 3. Tellus, 60A, 848862, doi:10.1111/j.1600-0870.2008.00339.x.

    • Search Google Scholar
    • Export Citation
  • Zhu, Y., , and R. E. Newell, 1998: A proposed algorithm for moisture fluxes from atmospheric rivers. Mon. Wea. Rev., 126, 725735, doi:10.1175/1520-0493(1998)126<0725:APAFMF>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    A typical atmospheric river event detected in a real-world MPAS simulation at 30-km grid spacing.

  • View in gallery

    Number of atmospheric rivers per month in the eight aquaplanet simulations.

  • View in gallery

    (a) Zonal-mean CPW (cm) in the eight aquaplanet simulations using MPAS and HOMME dynamical cores and (b) CDF of CPW of the longitudinal band between 20° and 40°N.

  • View in gallery

    (a) Thresholds of CPW (cm) defined in such a way that the threshold remains the same for MPAS30 while it increases for others according to the CDFs (Fig. 3). (b) Frequencies of ARs after differences in CPW are accounted for using the new thresholds.

  • View in gallery

    (a) Zonal mean of the fraction of CPW below 800 hPa (LLPW/CPW) and (b) the CDF of LLPW/CPW, which is used to define thresholds that account for differences in precipitable water profiles.

  • View in gallery

    The thresholds of LLPW/CPW derived from the CDFs (Fig. 5): note that the threshold for MPAS30 remains 0.8, while the thresholds of the others increase. (b) The frequency of AR events after the differences in CPW and LLPW/CPW are accounted for.

  • View in gallery

    (a) Zonal-mean zonal wind at 800 hPa (m s−1). (b) The CDF of the minimum latitude at which the zonal wind at 800 hPa is positive and the latitude at which 77.76% of the grid points have positive (westerly): note that this latitude is 20°N for MPAS30 and lower for others. (c) The latitude at which the zonal wind changes direction.

  • View in gallery

    (a) The new threshold latitude defined using the CDFs in Fig. 7 and (b) the frequency of AR events after accounting for the differences in PW, LLPW/PW, and latitudinal structure of the zonal winds.

  • View in gallery

    Power spectrum eddy kinetic energy at 800 hPa and 20°N for the eight simulations.

  • View in gallery

    The frequency of atmospheric river events over the (a) northeast and (b) southeast Pacific Ocean from real-world MPAS simulations at 120- and 30-km resolutions and NCEP-2 reanalysis.

  • View in gallery

    (a) CPW (cm) and (b) zonal wind at 800 hPa (m s−1) from real-world MPAS simulations at two resolutions and NCEP-2 reanalysis. All are averaged between 180° and 122.5°W.

  • View in gallery

    Difference between MPAS30R and MPAS120R mean precipitation (mm day−1; shaded), wind vectors (max = 3 m s−1), and streamfunction (contours) at 800 hPa. The blue and red dots mark the mean locations of the ARs in the MPAS30RW and MPAS120RW simulations, respectively.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 187 187 19
PDF Downloads 135 135 9

Resolution and Dynamical Core Dependence of Atmospheric River Frequency in Global Model Simulations

View More View Less
  • 1 Pacific Northwest National Laboratory, Richland, Washington
© Get Permissions
Full access

Abstract

This study examines the sensitivity of atmospheric river (AR) frequency simulated by a global model with different grid resolutions and dynamical cores. Analysis is performed on aquaplanet simulations using version 4 of the Community Atmosphere Model (CAM4) at 240-, 120-, 60-, and 30-km model resolutions, each with the Model for Prediction Across Scales (MPAS) and High-Order Methods Modeling Environment (HOMME) dynamical cores. The frequency of AR events decreases with model resolution and the HOMME dynamical core produces more AR events than MPAS. Comparing the frequencies determined using absolute and percentile thresholds of large-scale conditions used to define an AR, model sensitivity is found to be related to the overall sensitivity of subtropical westerlies, atmospheric precipitable water content and profile, and to a lesser extent extratropical Rossby wave activity to model resolution and dynamical core. Real-world simulations using MPAS at 120- and 30-km grid resolutions also exhibit a decrease of AR frequency with increasing resolution over the southern east Pacific, but the difference is smaller over the northern east Pacific. This interhemispheric difference is related to the enhancement of convection in the tropics with increased resolution. This anomalous convection sets off Rossby wave patterns that weaken the subtropical westerlies over the southern east Pacific but has relatively little effect on those over the northern east Pacific. In comparison to the NCEP-2 reanalysis, MPAS real-world simulations are found to underestimate AR frequencies at both resolutions likely because of their climatologically drier subtropics and poleward-shifted jets. This study highlights the important links between model climatology of large-scale conditions and extremes.

Corresponding author address: Samson Hagos, Pacific Northwest National Laboratory, 902 Battelle Boulevard, Richland, WA 99352. E-mail: samson.hagos@pnnl.gov

Abstract

This study examines the sensitivity of atmospheric river (AR) frequency simulated by a global model with different grid resolutions and dynamical cores. Analysis is performed on aquaplanet simulations using version 4 of the Community Atmosphere Model (CAM4) at 240-, 120-, 60-, and 30-km model resolutions, each with the Model for Prediction Across Scales (MPAS) and High-Order Methods Modeling Environment (HOMME) dynamical cores. The frequency of AR events decreases with model resolution and the HOMME dynamical core produces more AR events than MPAS. Comparing the frequencies determined using absolute and percentile thresholds of large-scale conditions used to define an AR, model sensitivity is found to be related to the overall sensitivity of subtropical westerlies, atmospheric precipitable water content and profile, and to a lesser extent extratropical Rossby wave activity to model resolution and dynamical core. Real-world simulations using MPAS at 120- and 30-km grid resolutions also exhibit a decrease of AR frequency with increasing resolution over the southern east Pacific, but the difference is smaller over the northern east Pacific. This interhemispheric difference is related to the enhancement of convection in the tropics with increased resolution. This anomalous convection sets off Rossby wave patterns that weaken the subtropical westerlies over the southern east Pacific but has relatively little effect on those over the northern east Pacific. In comparison to the NCEP-2 reanalysis, MPAS real-world simulations are found to underestimate AR frequencies at both resolutions likely because of their climatologically drier subtropics and poleward-shifted jets. This study highlights the important links between model climatology of large-scale conditions and extremes.

Corresponding author address: Samson Hagos, Pacific Northwest National Laboratory, 902 Battelle Boulevard, Richland, WA 99352. E-mail: samson.hagos@pnnl.gov

1. Introduction

Atmospheric rivers (ARs) are prominent features of the global water cycle. These relatively narrow filaments of warm and moist air are responsible for up to 90% of the global poleward moisture transport (Zhu and Newell 1998) from the tropics. On long-term average, 20%–50% of annual precipitation over California is attributed to six to seven AR events in the cold season that produce heavy precipitation (Dettinger et al. 2011; Guan et al. 2010). Flooding is more likely to occur (Neiman et al. 2008; Leung and Qian 2009; Ralph et al. 2013), when an AR makes a landfall on preexisting snowpack and high antecedent soil moisture conditions. Over the last 10 yr, every major flooding of the Russian River in California has been associated with an AR, and streamflow in the Merced River increases by an order of magnitude with the arrival of ARs (Ralph et al. 2006; Dettinger et al. 2011; Ralph and Dettinger 2012). Of the 48 annual peak daily flows on four watersheds in western Washington, 46 were associated with landfalling AR conditions (Neiman et al. 2011). ARs are also important cause of extreme precipitation in the west coasts of Europe (Stohl et al. 2008; Lavers et al. 2013) and South America (Viale and Nuñez 2011).

Multimodel analysis of the response of ARs to global warming by Dettinger (2011) suggests that, while their average number may not change, the number of the extreme events and years with a greater number of events are projected to increase. Furthermore, he showed that the peak season for AR occurrence may lengthen in the warming climate. Uncertainties in the frequency of extreme precipitation events associated with ARs are strongly reflected in uncertainty of projected annual-mean precipitation over California. Pierce et al. (2013) showed that intermodel divergence of the frequency of rare high precipitation (>60 mm day−1) events explains much of the disagreement in the annual-mean precipitation over California. The response of the frequency and intensity of AR events to warming appears to be very sensitive to the latitude and topography of the region of interest. For example, Luce et al. (2013) find that slower westerlies reduce the topographic enhancement of precipitation over the mountains of the Pacific Northwest; farther south, over Northern California, the eastward extension of the westerlies favor increased precipitation (Neelin et al. 2013).

In any case, the response of AR frequency to global warming is likely to depend on the comparative response of the subtropical westerlies and atmospheric water vapor content and its vertical structure to global warming, which may vary from model to model depending on the physics, dynamical core, and resolution of the model. From a 20-yr regional climate simulation, Leung and Qian (2009) found that their model captured realistic spatial patterns of temperature and precipitation anomalies associated with ARs in the western United States. Wick et al. (2013) evaluated AR predictions from five operational ensemble forecast systems for three winter seasons with lead time out to 10 days in the northeastern Pacific Ocean and the west coast of North America. They found reasonable skill in the overall forecast of the occurrence of ARs but also significant errors in forecasting the position of landfall. They identified spatial resolution as a key factor in model biases associated with the forecast of AR width and that the models they examined also tend to have a slight moist bias averaged over the entire AR.

This study examines the dependence of AR frequency on model dynamical core and resolution and explores the dynamic and thermodynamic roots of these dependencies. A suite of aquaplanet and real-world simulations from a community model with two dynamical cores, applied at model resolutions ranging from 30 to 240 km, are used in our investigation. Section 2 describes the model and simulation setup. Section 3 discusses the definition of ARs derived from observed conditions. Sections 4 and 5 describe our analysis approach and findings from the aquaplanet and real-world simulations, respectively. Section 6 discusses and concludes this research.

2. Model and simulation description

We use output from CAM4 (Gent et al. 2011) aquaplanet simulations, generated by the Model for Prediction Across Scales (MPAS; Skamarock et al. 2012) and High-Order Methods Modeling Environment (HOMME; Taylor et al. 2008) dynamical cores as part of a project to examine the veracity of different approaches to modeling regional climate by the application of a hierarchical evaluation framework (Leung et al. 2013).

MPAS uses unstructured Voronoi meshes and a C-grid discretization as the basis of the model formulation. The unstructured Voronoi meshes allow for both quasi-uniform discretization of the sphere and local refinement (Du et al. 1999; Ringler et al. 2008; Ju et al. 2011). The underlying numerical method used in MPAS is described in Thuburn et al. (2009) and Ringler et al. (2010). In this study, MPAS is run for 5 yr with quasi-uniform grid spacings of 240, 120, 60, and 30 km. Further details about model configurations for these experiments are discussed in Rauscher et al. (2013) and Hagos et al. (2013). The aquaplanet MPAS simulations are referred to as MPAS240, MPAS120, MPAS60 and MPAS30, respectively. HOMME uses spectral finite-element discretization on a relatively isotropic cubed sphere grid (Taylor et al. 2008; Evans et al. 2013). The HOMME spectral element discretization uses a compatible formulation that conserves dry mass and total energy and it also enables simulations with local refinement. HOMME simulations are run at resolutions comparable to MPAS, at 220, 120, 55, and 28 km. These simulations are referred to as HOMME220, HOMME110, HOMME55, and HOMME28, respectively.

Following the aquaplanet experiment protocol of Neale and Hoskins (2000), all simulations are run with zonally symmetric fixed sea surface temperatures, which are symmetrical about the equator and decrease poleward. In addition, solar insolation is fixed at the March equinoctial condition. To examine the impact of zonal nonuniformity on the response of ARs to increased resolution, two additional Atmospheric Model Intercomparison Program (AMIP)-style (Gates et al. 1999; Evans et al. 2013) real-world simulations are performed at 120- and 30-km grid resolutions using the MPAS dynamical core. These two simulations are referred to as MPAS120RW and MPAS30RW. (Corresponding HOMME simulations were unavailable.) Six-hourly outputs of three-dimensional water vapor mixing ratio and horizontal winds are obtained from all the simulations and regridded using an area conservative remapping to 240-km grid resolution for AR-related calculations. They are compared with AR counts calculated from NCEP–DOE AMIP-II reanalysis (NCEP-2; Kanamitsu et al. 2002) in the same manner.

3. Definition of atmospheric river events

In the first part of this study, various thresholds are set to define AR events. Based on satellite observations, Ralph et al. (2004) define ARs as contiguous narrow bands of moisture with column-integrated precipitable water (CPW) greater than 2 cm that are more than 2000 km in length and less than 1000 km in width. Recent definitions by Ralph and Dettinger (2011, 2012) include additional criteria that 80% of the water vapor transport occurs below 2 km, the associated near surface generally exceeds 10 m s−1, and Northern Hemisphere ARs generally occur poleward of 20° latitude. We adopt this definition and apply the following criteria to define ARs as contiguous objects with:

  1. CPW > 2 cm;
  2. 80% of the vertically integrated precipitable water is below 800-hPa level;
  3. wind speed at 800 hPa > 10 m s−1, with both zonal wind and meridional winds at 800 hPa > 0 m s−1;
  4. length > 2000 km and width < 1000 km; and
  5. the mode latitude of the grid points constituting the AR poleward of ±20°.
Figure 1 shows an example of an AR detected from the MPAS120RW simulation using the above criteria. The vertically integrated moisture extends from the Hawaiian Islands to California. The frequency of an AR event is defined as the sum of all AR occurrences detected in each 6-hourly sample of model outputs from the 5-yr-long simulation divided by the 5 yr. For the aquaplanet simulations, the AR counts are the sums of occurrences in both hemispheres.
Fig. 1.
Fig. 1.

A typical atmospheric river event detected in a real-world MPAS simulation at 30-km grid spacing.

Citation: Journal of Climate 28, 7; 10.1175/JCLI-D-14-00567.1

4. Resolution and dynamical core dependence in the aquaplanet simulations

In the aquaplanet simulations, a large number of AR events are detected because of zonal symmetry. On examination of the frequency of AR events from the 6-hourly snapshots of the aquaplanet simulations, the AR frequency is consistently very sensitive to the dynamical core and resolution. Figure 2 shows that the frequency of AR events decreases as the resolution increases. The HOMME dynamical core produces between 2 and 5 times as many AR events as in the MPAS dynamical core. The decrease in AR frequency with increasing resolution is monotonic in the MPAS aquaplanet simulations but not in HOMME, which produces slightly more frequent AR events at 110-km grid resolution than at 220-km grid resolution.

Fig. 2.
Fig. 2.

Number of atmospheric rivers per month in the eight aquaplanet simulations.

Citation: Journal of Climate 28, 7; 10.1175/JCLI-D-14-00567.1

Since ARs are defined based on absolute thresholds of moisture, winds, etc., we hypothesize that the basic climate state simulated by the models contributes to the dynamical core and resolution dependence of the simulations. To test this hypothesis, we examine how variables related to three of the thresholds used in defining ARs are related to the overall climate state of the model simulations based on their cumulative distribution function (CDF). The variables analyzed include the vertically integrated and vertical profile of precipitable water content and meridional structure of zonal-mean zonal winds. We hypothesize that the resolution and dynamical core dependences of the frequency distribution of these variables explain the overall sensitivity of AR frequency.

a. Precipitable water content

To quantify the extent to which variations in vertically integrated precipitable water explain the resolution and dynamical core dependence of AR frequency, first we examine how precipitable water content varies with resolution and how the threshold associated with it can be adjusted to account for the sensitivity of AR frequency. Figure 3a shows the zonal- and time-mean CPW. In general, simulations with the HOMME dynamical core have higher CPW than those with MPAS. Furthermore, the lower-resolution simulations are generally moister than those at higher resolution, with the exception of HOMME55, which has the highest CPW.

Fig. 3.
Fig. 3.

(a) Zonal-mean CPW (cm) in the eight aquaplanet simulations using MPAS and HOMME dynamical cores and (b) CDF of CPW of the longitudinal band between 20° and 40°N.

Citation: Journal of Climate 28, 7; 10.1175/JCLI-D-14-00567.1

Zonal- and time-mean CPW, however, may not reflect the CPW during AR events, which are associated with above normal CPW. Hence, the CDF of CPW is more informative of differences that are relevant to ARs. To construct the CDF, the 6-hourly CPW of each grid point between 20° and 40°N is sorted into one of 1000 bins between 0 and 10 cm of CPW, with 0.01-cm increments. A grid point is assigned to a specific bin if its CPW is less than the upper end of the bin. The number of grid points assigned to each bin is then divided by the total number of grid points.

Figure 3b shows the CDFs for the eight simulations. The rightward displacement of the CDFs for HOMME compared to MPAS indicates that both the mean and extreme values of CPW in the HOMME simulations are higher than in the MPAS simulations. Similarly, the CDFs of the low-resolution simulations from both dynamical cores are displaced toward higher values, indicating larger mean and extreme values compared to the high-resolution simulations. In MPAS30, the 2-cm CPW threshold used in defining ARs corresponds to 95.27% in the CDF. In other words, 2 cm of CPW is higher than 95.27% of all the CPW values in the MPAS30 simulation. To account for the effect of differences in precipitable water content in AR frequency, we identify the CPW content that corresponds to this 95.27% in each CDF as the threshold in the other seven simulations. In other words, we revise the CPW component of our definition of ARs to one based on the percentile threshold of CPW in each model simulation instead of using a common threshold based on absolute CPW value. A similar approach is used by Lavers et al. (2013) to account for intermodel differences on the response of number of ARs making landfall over the United Kingdom. Figure 4 shows the new CPW values corresponding to 95.27% of each CDF. These values are, in order of increasing resolution, 2.27, 2.15, 2.11, and 2.0 cm for MPAS and 2.4, 2.32, 2.49, and 2.26 cm for HOMME. The frequency of AR events is recalculated after adjusting the original 2-cm threshold to the new percentile based simulation dependent values that account for the differences in CPW. The resolution and dynamical core dependence of the new AR frequencies shown in Fig. 4b is dramatically reduced, indicating that the CPW differences can largely explain the sensitivity of the AR frequency in Fig. 2. With the new thresholds, all simulations produce less than 11 AR events per year, an order of magnitude smaller than the 4–160 events with the original definition. Interestingly, no AR event is detected for HOMME55.

Fig. 4.
Fig. 4.

(a) Thresholds of CPW (cm) defined in such a way that the threshold remains the same for MPAS30 while it increases for others according to the CDFs (Fig. 3). (b) Frequencies of ARs after differences in CPW are accounted for using the new thresholds.

Citation: Journal of Climate 28, 7; 10.1175/JCLI-D-14-00567.1

b. Precipitable water profile

In defining AR events, we also required that the low-level precipitable water (LLPW) integrated between the surface and 800 hPa constitute 80% of the total CPW. Differences in the simulated specific humidity profiles in the global simulations may also be reflected in the frequency of AR events. Figure 5a shows the meridional structure of the zonally averaged LLPW/CPW for the various simulations. In general, the specific humidity profile of the MPAS simulations is more bottom heavy (greater fraction of the water vapor in the column is near surface) in comparison to that of HOMME. The profile of specific humidity is also found to be sensitive to resolution. With increasing resolution, the specific humidity becomes more bottom heavy. HOMME55 again is an outlier in that it has the smallest LLPW/CPW or the most top-heavy specific humidity profile.

Fig. 5.
Fig. 5.

(a) Zonal mean of the fraction of CPW below 800 hPa (LLPW/CPW) and (b) the CDF of LLPW/CPW, which is used to define thresholds that account for differences in precipitable water profiles.

Citation: Journal of Climate 28, 7; 10.1175/JCLI-D-14-00567.1

To assess the impact of this dependence of specific humidity profiles on dynamical core and resolution and impact on the AR frequency, we construct the CDF of LLPW/CPW using the method applied in the CPW analysis (Fig. 3b) for all grid points between 20° and 40°N using 1000 bins with LLPW/CPW values between 0.0 and 1.0 (Fig. 5b). The number of grid points with LLPW/CPW less than a certain value (e.g., the 80% in our definition of ARs) is sensitive to the model resolution and dynamical core. Comparing the different simulations, the mean and extreme values of LLPW/CPW are higher (the bottom-heavy specific humidity profiles are more dominant) in MPAS than in HOMME and generally increase with increasing resolution.

Examination of the CDF shows that, for MPAS30, the percentage of grid points with LLPW/CPW less than 0.8 is 50.89%. This 50.89% is used as a new percentile threshold for LLPW/CPW (Fig. 6a), such that the AR count for MPAS30 remains the same while those of the other simulations will vary depending on their respective thresholds corresponding to this percentile. These new threshold values are 0.763, 0.784, 0.796, and 0.80 for MPAS240, MPAS120, MPAS60, and MPAS30, respectively, and 0.756, 0.774, 0.748, and 0.775 for HOMME220, HOMME110, HOMME55, and HOMME28, respectively. Recalculating the frequency of AR events with the new LLPW/CPW thresholds, in addition to the new CPW thresholds set used in Fig. 4, Fig. 6b shows the new dependence of AR frequency on dynamical core and resolution. Because differences in PW and LLPW/CPW among the simulations have opposite effects on AR frequency, the resolution and dynamical core dependence shown in Fig. 2 is partly recovered when the percentile thresholds for CPW and LLPW/CPW are applied simultaneously to define ARs. Figure 6b shows more AR events in the low-resolution simulations, although the range of model differences is smaller compared to Fig. 2. With the absolute thresholds, for example, MPAS240 produced about 20 times more AR events than MPAS30. After using percentile thresholds, MPAS240 only produced 5 times more AR events than MPAS30. The dynamical core dependence is also reduced significantly. For example, in Fig. 2, HOMME110 produced 6 times as many AR events as MPAS120; however, after accounting for differences in CPW and LLPW/CPW, it is only about 3 times that of MPAS120.

Fig. 6.
Fig. 6.

The thresholds of LLPW/CPW derived from the CDFs (Fig. 5): note that the threshold for MPAS30 remains 0.8, while the thresholds of the others increase. (b) The frequency of AR events after the differences in CPW and LLPW/CPW are accounted for.

Citation: Journal of Climate 28, 7; 10.1175/JCLI-D-14-00567.1

c. Meridional structure of zonal wind

In the last two subsections, the roles of total moisture and moisture profiles in the sensitivities of AR frequency to resolution and dynamical core are examined. In this subsection, we investigate dynamical aspects. For large-scale conditions to qualify as AR events, the zonal and meridional winds within the contiguous moist air must be eastward and poleward, respectively, and their center latitudes must be poleward of 20° latitude. Figure 7a shows the meridional structure of the zonally averaged zonal wind at 800 hPa from the eight aquaplanet simulations. As resolution increases, the mean westerly winds shift poleward in both MPAS and HOMME simulations. This poleward shift is related to increased poleward transport of angular momentum with increasing resolution (Williamson 2008), which is a noted feature in aquaplanet simulations using multiple CAM dynamical cores, including MPAS (Rauscher et al. 2013).

Fig. 7.
Fig. 7.

(a) Zonal-mean zonal wind at 800 hPa (m s−1). (b) The CDF of the minimum latitude at which the zonal wind at 800 hPa is positive and the latitude at which 77.76% of the grid points have positive (westerly): note that this latitude is 20°N for MPAS30 and lower for others. (c) The latitude at which the zonal wind changes direction.

Citation: Journal of Climate 28, 7; 10.1175/JCLI-D-14-00567.1

To assess the contribution of this poleward shift of westerlies to the resolution and dynamical core dependence of AR frequency, the minimum threshold set for the latitude of the of the AR (20°N/S) is allowed to vary, depending on the latitude at which the subtropical winds transition from easterlies to westerlies. The new latitudinal threshold is again determined based on percentile information from the CDF of the latitude of the zero crossing of the zonal-mean zonal wind. For each 6-hourly zonal-mean wind sample, the latitude between 0° and 30°N at which the zonal wind is equal to zero is calculated by linear interpolation. Then, a CDF is constructed with 150 bins that cover that range. If the latitude of a point is less than the upper bound of the bin, the point is assigned to the bin.

The resulting CDFs are shown in Fig. 7b. In the low-resolution simulations, the zonal wind frequently switches sign at significantly lower latitudes than higher resolutions: for example, 51.91% of the latitude points at which the zonal wind is zero south of 20°N for MPAS30. That percentage point corresponds to 16.2°, 17.6°, and 19.4°N in MPAS220, MPAS120, and MPAS60, respectively. For the HOMME simulations, the 51.91% values are at 16°, 18.8°, 20.6°, and 19.8°N for HOMME220, HOMME110, HOMME55, and HOMME28, respectively (Fig. 7c).

To account for the impact of the meridional structure of zonal wind on AR frequency, the minimum latitude is redefined using based on the 51.91% in the CDF discussed above. For each simulation, the minimum latitude is now defined as , such that it remains at 20°N for MPAS30 but is shifted north for all other simulations to compensate for the equatorward shift of the westerlies with decreasing resolution. Likewise, for the Southern Hemisphere the minimum latitude is defined as . Figure 8a shows the new minimum latitudes used to recalculate the frequency of AR events, and Fig. 8b shows the frequency of AR events after accounting for the differences in precipitable water content, moisture profile, and the meridional structure of zonal winds. The resolution dependence is greatly reduced. For MPAS, all the simulations produce between 4 and 6 AR events per year; for HOMME, all but HOMME110 produce between 7 and 12 events per year. The HOMME110 simulation produces much larger number: 44.

Fig. 8.
Fig. 8.

(a) The new threshold latitude defined using the CDFs in Fig. 7 and (b) the frequency of AR events after accounting for the differences in PW, LLPW/PW, and latitudinal structure of the zonal winds.

Citation: Journal of Climate 28, 7; 10.1175/JCLI-D-14-00567.1

The analysis discussed in this and the previous sections shows much of the resolution and dynamical core (dycore) dependence of the AR frequency can be explained by the combined differences in precipitable water content, its profile, and the position of the subtropical jet. However, as Fig. 8b shows, these differences do not completely explain all the dycore dependences. HOMME simulations still consistently produce more ARs than MPAS simulations both before and after accounting for all the above-discussed factors. Thus, finally we hypothesize that the dependence of wave activities relevant to the formation of AR events might play a role. This hypothesis is examined in the next subsection.

d. Wave amplitude

To examine the potential for differences in Rossby wave amplitudes to explain the remaining resolution and dynamical core dependence of ARs, the resolution and dynamical core dependence of the eddy kinetic energy associated with the wavenumber 4–7 Rossby waves are analyzed. These waves constitute the prominent circulation patterns in the midlatitudes (Welch and Tung 1998), whose equatorward and downward propagation and breaking are often associated with AR events (Ryoo et al. 2013; Payne and Magnusdottir 2014). A fast Fourier transform (FFT) is performed on the zonal wind and meridional winds at 800 hPa at 20°. These are chosen because Rossby waves have to propagate far enough equatorward to reverse the direction of the otherwise moist tropical easterlies (Figs. 3a and 7a) and cause ARs; as we show in Fig. 7a, this happens at about 20°. Then eddy kinetic energy (EKE) at each wavenumber (k) is calculated as , where and are the amplitudes of contributions of zonal wavenumber waves to the zonal and meridional winds at 20°. Figure 9 shows the spectrum of EKE with respect to k. For both MPAS and HOMME simulations, EKE peaks at wavenumbers 5 and 6 as expected. However, the dynamical core dependence of the amplitudes of wavenumbers 5 and 6 is more interesting. In general, HOMME simulations have larger EKE than the corresponding MPAS simulations. This might explain the consistently larger number of ARs in HOMME even after accounting for the factors examined in the last section. EKE associated with wavenumber 4–7 Rossby waves decreases with increased resolution. This is consistent with the fact that the high-resolution models shift the subtropical westerlies and the waves embedded in them. Hence, EKE at 20°N would decrease with resolution. Since we have accounted for the effects of the above-discussed processes, the resolution dependence of AR frequency is relatively small and not monotonic with respect to increasing resolution. This suggests what remains might at least partially be related to internal variability.

Fig. 9.
Fig. 9.

Power spectrum eddy kinetic energy at 800 hPa and 20°N for the eight simulations.

Citation: Journal of Climate 28, 7; 10.1175/JCLI-D-14-00567.1

5. Resolution dependence in real-world simulations

While the main focus of this study is to examine the impacts of the global model dynamics and resolution on the frequency of AR events in idealized aquaplanet simulations, in real-world simulations these effects may be obscured by spatial heterogeneity related to regional processes that are neither steady nor zonally uniform. For example, unlike in aquaplanet simulations, in real-world simulations (and in observations) ARs often originate from the moisture-rich tropical central and western Pacific and transport moisture poleward toward the west coast of North America and South America. As such, their frequency is likely modulated by processes in those regions. In this section, we examine the 10-yr-long real-world AMIP-style simulations using the MPAS dynamical core and from NCEP-2 reanalysis data to evaluate the impact of zonal nonuniformity on the resolution dependence of AR frequency over eastern Pacific Ocean and to compare the AR frequencies obtained from the simulations at the two resolutions with the reanalysis. Since we are only considering ARs over the east Pacific, the representativeness of the sample size is reduced compared to the aquaplanet simulations, but this is partly compensated by the increase of the length of the simulations from 5 to 10 yr. The definition of ARs is the same as the absolute thresholds used in the analysis of aquaplanet simulations, except in this section we focus only on the ARs between the date line and longitude of the California coast (180°–122.5°W) for both hemispheres for consistency and to capture as many eastern Pacific AR events as possible to make the comparison representative. For the NCEP reanalysis, the 2.5° resolution is regridded to the 240-km horizontal grid and the 20 levels before the AR frequencies are calculated in the same manner as for the models. Figure 10a shows a comparison among AR frequencies from the 10-yr-long simulations at 120- and 30-km grid resolutions as well as the NCEP-2 reanalysis.

Fig. 10.
Fig. 10.

The frequency of atmospheric river events over the (a) northeast and (b) southeast Pacific Ocean from real-world MPAS simulations at 120- and 30-km resolutions and NCEP-2 reanalysis.

Citation: Journal of Climate 28, 7; 10.1175/JCLI-D-14-00567.1

The frequency of northeast Pacific AR events in the real-world simulations does not decrease with resolution as dramatically as in the aquaplanet simulations. Over the southeast Pacific, on the other hand, the frequency decreases with resolution in agreement with what was found in aquaplanet simulations. At both resolutions, MPAS underestimates the frequency of ARs relative to NCEP-2 and increased resolution moves the frequency further from those of the reanalysis in both hemispheres. That is expected from the analysis of aquaplanet simulations and the fact that the NCEP-2 resolution is considerably lower. To determine the reason for the contrast between the two resolutions and their deviation from reanalysis, two hypotheses are examined based on the findings of the last section. One possible explanation is that the amount of precipitable water over the eastern Pacific is somehow increased with resolution. Figure 11a shows that is not the case. The zonal averaged (only over eastern Pacific) vertically integrated precipitable water decreases with resolution as in the aquaplanet simulations. In comparison to the reanalysis, both MPAS simulations are drier over the east Pacific north (south) of 30°N (30°S), which may partly explain the fact that they both have fewer AR events than the reanalysis. The westerly winds, however, behave quite differently over the northern east Pacific from our findings from the aquaplanet simulations (Fig. 11b). Over the southern east Pacific, the subtropical winds shift poleward even more dramatically than was shown in the aquaplanet simulations, and the low-resolution (MPAS120RW) simulation is in better agreement with the reanalysis. Over the northern east Pacific, there is little difference in subtropical zonally averaged zonal wind between the two simulations. They are both shifted poleward compared to the reanalysis, which likely contributed to the relatively fewer ARs in the model simulations in comparison to the reanalysis. However, the model simulations and the reanalysis are consistent in that there are more AR events over the southern east Pacific than over the northern east Pacific.

Fig. 11.
Fig. 11.

(a) CPW (cm) and (b) zonal wind at 800 hPa (m s−1) from real-world MPAS simulations at two resolutions and NCEP-2 reanalysis. All are averaged between 180° and 122.5°W.

Citation: Journal of Climate 28, 7; 10.1175/JCLI-D-14-00567.1

To understand the root of the differences between the two hemispheres in the response of the zonal-mean zonal wind to increased resolution, the resolution dependence of precipitation and circulation patterns over Pacific Ocean and the surrounding land are considered. Figure 12 shows the difference between the MPAS30RW and MPAS120RW mean precipitation (mm day−1), wind vectors, and streamfunction at 800 hPa, with blue and red dots marking the mean locations of the ARs detected. Increased resolution increases precipitation over the western Pacific Ocean, Central America, and the Amazon region. The cyclonic circulation pairs associated with the diabatic heating of the enhanced precipitation are fairly symmetrical about the equator, except over the central western Pacific (between 180° and 120°E). Over this longitudinal band, the responses of the Southern Hemisphere circulations are more or less zonally uniform because the anomalous circulations associated with the South Pacific convergence zone extend far enough eastward and merge with the cyclonic circulation associated with the anomalous heating over South America to form a basinwide easterly anomaly. Over the Northern Hemisphere, on the other hand, the anomalous cyclonic circulations associated with the anomalous heating over the western Pacific and those off the coast of Mexico are separated by anomalous high pressure, where the anomalous winds (i.e., the response to increased resolution) are very weak.

Fig. 12.
Fig. 12.

Difference between MPAS30R and MPAS120R mean precipitation (mm day−1; shaded), wind vectors (max = 3 m s−1), and streamfunction (contours) at 800 hPa. The blue and red dots mark the mean locations of the ARs in the MPAS30RW and MPAS120RW simulations, respectively.

Citation: Journal of Climate 28, 7; 10.1175/JCLI-D-14-00567.1

Given the above response of the circulation patterns to increased resolution, the differences in the response of AR frequency to increased resolution become clearer. Over the Southern Ocean, similar to the results from the aquaplanet simulations, the poleward shifts of the subtropical westerlies play a primary role in the resolution dependence. Over the Northern Hemisphere, on the other hand, this effect is absent because of the wave patterns forced by the heating anomalies from the enhanced precipitation with increasing resolution. As a result, the resolution dependence of AR frequency is not discernible. The spatial distribution of the AR events over southern east Pacific indicate yet another effect of the weakened subtropical westerlies on the AR events in the high-resolution simulations (Fig. 12, red dots), which shifts the AR locations westward compared to those in the low-resolution simulations. No such effect is apparent over the northern east Pacific.

6. Discussion

This study examines the sensitivity of the frequency of model simulated atmospheric river events to resolution and dynamical core using aquaplanet and real-world (AMIP) simulations. The AR frequency is calculated based on several criteria defined by absolute thresholds on atmospheric conditions simulated by CAM4, using two different dynamical cores with the same physics parameterizations across a range of model resolutions. It is found that in the aquaplanet simulations the sensitivity of AR frequency is related to the overall resolution dependence of the column-integrated precipitable water and its vertical profile in the atmosphere as well as the meridional structure of the westerlies. In general, with increasing resolution in the aquaplanet simulations, the subtropical westerly wind is shifted poleward. This poleward shift displaces the westerlies toward colder waters and hence evaporation over the subtropical band also decreases with increased resolution (not shown). This reduced evaporation from the surface is a likely factor that contributes to the decline in precipitable water in the subtropical atmosphere as the resolution increases (Fig. 4). On the other hand, the fraction of the atmospheric water vapor in the lower troposphere increases with increasing resolution.

Comparing the AR frequency obtained using percentile and absolute thresholds, our analysis shows that the two thermodynamical effects related to precipitable water (its content and profile) in combination explain much of the dynamical core and resolution dependence of the AR frequency (Fig. 6b). Adding the poleward shift of the westerlies, the three factors together explain almost all the dynamical core and resolution dependence of AR frequency (Fig. 8b) in the aquaplanet simulations. We note that the dynamical core dependence of atmospheric moisture has been demonstrated by Landu et al. (2014) to be an important factor, leading to a single versus double intertropical convergence zone (ITCZ) structure in the HOMME versus MPAS simulations at high resolution. While the resolution dependence of atmospheric moisture may be partly explained by the poleward shift of the subtropical westerlies (Williamson 2008), the moisture dependence on the dynamical core remains unclear and should be a topic of future investigation.

In real-world simulations, the resolution dependence is complicated by a zonal response of convection and associated wave dynamics to regional forcing (warm SST in the western Pacific). With increasing resolution, convective activity over the western Pacific increases and the anomalous diabatic heating induces Rossby wave patterns that enhance the westerlies over the northern east Pacific. The latter response compensates for the other factors considered in the analysis of the aquaplanet simulation. Therefore, the likelihood of ARs events that transport tropical moisture to the west coast of North America shows little dependence on resolution. Over the southern east Pacific, on the other hand, the frequency of AR events decreases with resolution in a manner similar to that in the aquaplanet simulations. In comparison to the NCEP-2 reanalysis, MPAS real-world simulations are found to underestimate AR frequencies at both resolutions, likely because of their climatologically drier subtropics and poleward-shifted jets.

In summary, this study links the frequency of AR events to various aspects of model climatology. Specifically, the impacts of CPW and its vertical profile, the latitudinal position of the subtropical winds, and tropical convection in the western Pacific and their respective sensitivities to model resolution and dynamical core on the frequency of AR events are examined. This work highlights the importance of understanding biases and uncertainties in the simulated climate state in reducing uncertainties in model simulations of extreme events. Idealized simulations such as aquaplanet experiments provide a useful framework for testing hypotheses of processes that contribute to model sensitivity. In translating the findings from aquaplanet simulations to real-world simulations, we caution that model sensitivity may be amplified regionally that complicates interpretation of the model response, but insights from aquaplanet simulations still provide useful guidance for analysis. Such analysis combined with model evaluation using observations should provide further insights on model uncertainty and biases to help future developments. Although the evaluating simulations of high-impact, low-probability events such as ARs are hampered by lack of data in some cases, model climatology of large-scale conditions can be routinely evaluated using a wide range of satellite and in situ observations. Therefore, understanding the links between model climatology of large-scale conditions and extremes will likely contribute to improving our understanding of model limitations in simulating both.

Simulation of ARs is likely to be also sensitive to physics parameterizations. While this study only examined the sensitivity of AR frequency to model resolution and dynamical cores with the CAM4 physics parameterizations, future studies should extend the scope to better understand how large-scale conditions such as precipitable water and the subtropical winds may vary with model resolution, dynamical cores, and physics parameterizations and the interactions among these modeling factors. In addition to AR frequency, it would be important to determine how modeling uncertainties influence other aspects of AR such as the position and precipitation associated with AR landfall. The latter, in particular, may have a stronger dependence on model resolution for ARs that make landfall in the complex terrain of the western United States.

Acknowledgments

This research is based on work supported by the U.S. Department of Energy Office of Science Biological and Environmental Research as part of the Regional and Global Climate Modeling Program. Computing resources for the model simulations are provided by the National Energy Research Scientific Computing Center (NERSC). Pacific Northwest National Laboratory is operated by Battelle for the U.S. Department of Energy under Contract DE-AC05-76RLO1830.

REFERENCES

  • Dettinger, M. D., 2011: Climate change, atmospheric rivers, and floods in California—A multimodel analysis of storm frequency and magnitude changes. J. Amer. Water Resour. Assoc., 47, 514523, doi:10.1111/j.1752-1688.2011.00546.x.

    • Search Google Scholar
    • Export Citation
  • Dettinger, M. D., , F. M. Ralph, , T. Das, , P. J. Neiman, , and D. R. Cayan, 2011: Atmospheric rivers, floods, and the water resources of California. Water, 3, 445478, doi:10.3390/w3020445.

    • Search Google Scholar
    • Export Citation
  • Du, Q., , V. Faber, , and M. Gunzburger, 1999: Centroidal Voroni tessellations: Applications and algorithms. SIAM Rev., 41, 637676, doi:10.1137/S0036144599352836.

    • Search Google Scholar
    • Export Citation
  • Evans, K. J., , P. Lauritzen, , S. K. Mishra, , R. Neale, , M. A. Taylor, , and J. J. Tribbia, 2013: AMIP simulations with the CAM4 spectral element dynamical core. J. Climate, 26, 689709, doi:10.1175/JCLI-D-11-00448.1.

    • Search Google Scholar
    • Export Citation
  • Gates, W. L., and et al. , 1999: An overview of the results of the Atmospheric Model Intercomparison Project (AMIP I). Bull. Amer. Meteor. Soc., 80, 2955, doi:10.1175/1520-0477(1999)080<0029:AOOTRO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Gent, P. R., and et al. , 2011: The Community Climate System Model version 4. J. Climate, 24, 49734991, doi:10.1175/2011JCLI4083.1.

    • Search Google Scholar
    • Export Citation
  • Guan, B., , N. P. Molotch, , D. E. Waliser, , E. J. Fetzer, , and P. J. Neiman, 2010: Extreme snowfall events linked to atmospheric rivers and surface air temperature via satellite measurements. Geophys. Res. Lett., 37, L20401, doi:10.1029/2010GL044696.

    • Search Google Scholar
    • Export Citation
  • Hagos, S., , R. Leung, , S. A. Rauscher, , and T. Ringler, 2013: Errors characteristics of two grid refinement approaches in aquaplanet simulations: MPAS-A and WRF. Mon. Wea. Rev.,141, 3022–3036, doi:10.1175/MWR-D-12-00338.1.

  • Ju, L. T., , T. Ringler, , and M. Gunzburger, 2011: Voroni tessellations and their application to climate and global modeling. Numerical Techniques for Global Atmospheric Models, P. H. Lauritzen et al., Eds., Lecture Notes in Computational Science and Engineering, Vol. 80, Springer, 313–342.

  • Kanamitsu, M., , W. Ebisuzaki, , J. Woollen, , S.-K. Yang, , J. J. Hnilo, , M. Fiorino, , and G. L. Potter, 2002: NCEP–DOE AMIP-II Reanalysis (R-2). Bull. Amer. Meteor. Soc., 83, 16311643, doi:10.1175/BAMS-83-11-1631.

    • Search Google Scholar
    • Export Citation
  • Landu, K., , L. R. Leung, , S. Hagos, , V. Vinoj, , S. Rauscher, , T. Ringler, , and M. Taylor, 2014: The dependence of ITCZ structure on model resolution and dynamical core in aquaplanet simulations. J. Climate, 27, 23752385, doi:10.1175/JCLI-D-13-00269.1.

    • Search Google Scholar
    • Export Citation
  • Lavers, D. A., , R. P. Allan, , G. Villarini, , B. Lloyd-Hughes, , D. J. Brayshaw, , and A. J. Wade, 2013: 2013: Future changes in atmospheric rivers and their implications for winter flooding in Britain. Environ. Res. Lett., 8, 034010, doi:10.1088/1748-9326/8/3/034010.

    • Search Google Scholar
    • Export Citation
  • Leung, L. R., , and Y. Qian, 2009: Atmospheric rivers induced heavy precipitation and flooding in the western U.S. simulated by the WRF regional climate model. Geophys. Res. Lett., 36, L03820, doi:10.1029/2008GL036445.

    • Search Google Scholar
    • Export Citation
  • Leung, L. R., , T. Ringler, , W. D. Collins, , and M. Ashfaq, 2013: A hierarchical evaluation of regional climate simulations. Eos, Trans. Amer. Geophys. Union, 94, 297298.

    • Search Google Scholar
    • Export Citation
  • Luce, C. H., , J. T. Abatzoglou, , and Z. A. Holden, 2013: The missing mountain water: Slower westerlies decrease orographic enhancement in the Pacific Northwest USA. Science, 342, 13601364, doi:10.1126/science.1242335.

    • Search Google Scholar
    • Export Citation
  • Neale, R. B., , and B. J. Hoskins, 2000: A standard test for AGCMs including their physical parametrizations: I: The proposal. Atmos. Sci. Lett., 1, 101107, doi:10.1006/asle.2000.0022.

    • Search Google Scholar
    • Export Citation
  • Neelin, J. D., , B. Langenbrunner, , J. E. Meyerson, , A. Hall, , and N. Berg, 2013: California winter precipitation change under global warming in the Coupled Model Intercomparison Project phase 5 ensemble. J. Climate, 26, 62386256, doi:10.1175/JCLI-D-12-00514.1.

    • Search Google Scholar
    • Export Citation
  • Neiman, P. J., , F. M. Ralph, , G. A. Wick, , J. D. Lundquist, , and M. D. Dettinger, 2008: Meteorological characteristics and overland precipitation impacts of atmospheric rivers affecting the west coast of North American based on eight years of SSM/I satellite observations. J. Hydrometeor., 9, 2247, doi:10.1175/2007JHM855.1.

    • Search Google Scholar
    • Export Citation
  • Neiman, P. J., , L. J. Schick, , F. M. Ralph, , M. Hughes, , and G. A. Wick, 2011: Flooding in western Washington: The connection to atmospheric rivers. J. Hydrometeor., 12, 13371358, doi:10.1175/2011JHM1358.1.

    • Search Google Scholar
    • Export Citation
  • Payne, A. E., , and G. Magnusdottir, 2014: Dynamics of landfalling atmospheric rivers over the North Pacific in 30 years of MERRA reanalysis. J. Climate, 27, 71337150, doi:10.1175/JCLI-D-14-00034.1.

    • Search Google Scholar
    • Export Citation
  • Pierce, D. W., and et al. , 2013: The key role of heavy precipitation events in climate model disagreements of future annual precipitation changes in California. J. Climate, 26, 58795896, doi:10.1175/JCLI-D-12-00766.1.

    • Search Google Scholar
    • Export Citation
  • Ralph, F. M., , and M. D. Dettinger, 2011: Storms, floods, and the science of atmospheric rivers. Eos, Trans. Amer. Geophys. Union, 92, 265, doi:10.1029/2011EO320001.

    • Search Google Scholar
    • Export Citation
  • Ralph, F. M., , and M. D. Dettinger, 2012: Historical and national perspectives on extreme West Coast precipitation associated with atmospheric rivers during December 2010. Bull. Amer. Meteor. Soc., 93, 783790, doi:10.1175/BAMS-D-11-00188.1.

    • Search Google Scholar
    • Export Citation
  • Ralph, F. M., , P. J. Neiman, , and G. A. Wick, 2004: Satellite and CALJET aircraft observations of atmospheric rivers over the eastern North Pacific Ocean during the winter of 1997. Mon. Wea. Rev., 132, 17211745, doi:10.1175/1520-0493(2004)132<1721:SACAOO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Ralph, F. M., , P. J. Neiman, , G. A. Wick, , S. I. Gutman, , M. D. Dettinger, , D. R. Cayan, , and A. B. White, 2006: Flooding on California's Russian River: Role of atmospheric rivers. Geophys. Res. Lett., 33, L13801, doi:10.1029/2006GL026689.

    • Search Google Scholar
    • Export Citation
  • Ralph, F. M., , T. Coleman, , P. J. Neiman, , R. J. Zamora, , and M. D. Dettinger, 2013: Observed impacts of duration and seasonality of atmospheric-river landfalls on soil moisture and runoff in coastal Northern California. J. Hydrometeor., 14, 443459, doi:10.1175/JHM-D-12-076.1.

    • Search Google Scholar
    • Export Citation
  • Rauscher, S., , T. Ringler, , W. C. Skamarock, , and A. A. Mirin, 2013: Exploring a global multiresolution modeling approach using aquaplanet simulations. J. Climate, 26, 24322452, doi:10.1175/JCLI-D-12-00154.1.

    • Search Google Scholar
    • Export Citation
  • Ringler, T., , L. Ju, , and M. Gunzburger, 2008: A multi resolution method for climate system modeling: Application of spectral centroidal Voroni tessellations. Ocean Dyn., 58, 475498, doi:10.1007/s10236-008-0157-2.

    • Search Google Scholar
    • Export Citation
  • Ringler, T., , J. Thuburn, , J. B. Klemp, , and W. C. Skamarock, 2010: A unified approach to energy conservation and potential vorticity dynamics for arbitrarily structured C-grids. J. Comput. Phys., 229, 30653090, doi:10.1016/j.jcp.2009.12.007.

    • Search Google Scholar
    • Export Citation
  • Ryoo, J., , Y. Kaspi, , D. Waugh, , G. Kiladis, , D. Waliser, , E. Fetzer, , and J. Kim, 2013: Impact of Rossby wave breaking on U.S. West Coast winter precipitation during ENSO events. J. Climate, 26, 63606382, doi:10.1175/JCLI-D-12-00297.1.

    • Search Google Scholar
    • Export Citation
  • Skamarock, W. C., , J. B. Klemp, , M. G. Duda, , L. Fowler, , S.-H. Park, , and T. D. Ringler, 2012: A multiscale nonhydrostatic atmospheric model using centroidal Voronoi tessellations and C-grid staggering. Mon. Wea. Rev., 140, 30903105, doi:10.1175/MWR-D-11-00215.1.

    • Search Google Scholar
    • Export Citation
  • Stohl, A., , C. Forster, , and H. Sodemann, 2008: Remote sources of water vapor forming precipitation on the Norwegian west coast at 60°N—A tale of hurricanes and an atmospheric river. J. Geophys. Res., 113, D05102, doi:10.1029/2007JD00900.

    • Search Google Scholar
    • Export Citation
  • Taylor, M. A., , J. Edwards, , and A. St. Cyr, 2008: Petascale atmospheric models for the Community Climate System Model: New developments and evaluation of scalable dynamical cores. J. Phys., 125, 012023, doi:10.1088/1742-6596/125/1/012023.

    • Search Google Scholar
    • Export Citation
  • Thuburn, J., , T. D. Ringler, , W. C. Skamarock, , and J. B. Klemp, 2009: Numerical representation of geostrophic modes on arbitrarily structured C-grids. J. Comput. Phys., 228, 83218335, doi:10.1016/j.jcp.2009.08.006.

    • Search Google Scholar
    • Export Citation
  • Viale, M., , and M. N. Nuñez, 2011: Climatology of winter orographic precipitation over the subtropical central Andes and associated synoptic and regional characteristics. J. Hydrometeor., 12, 481507, doi:10.1175/2010JHM1284.1.

    • Search Google Scholar
    • Export Citation
  • Welch, W. T., , and K. K. Tung, 1998: On the equilibrium spectrum of transient waves in the atmosphere. J. Atmos. Sci., 55, 28332851, doi:10.1175/1520-0469(1998)055<2833:OTESOT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Wick, G. A., , P. J. Neiman, , F. M. Ralph, , and T. M. Hamill, 2013: Evaluation of forecasts of the water vapor signature of atmospheric rivers in operational numerical weather prediction models. Wea. Forecasting, 28, 13371352, doi:10.1175/WAF-D-13-00025.1.

    • Search Google Scholar
    • Export Citation
  • Williamson, D. L., 2008: Convergence of aquaplanet simulations with increasing resolution in the Community Atmospheric Model, version 3. Tellus, 60A, 848862, doi:10.1111/j.1600-0870.2008.00339.x.

    • Search Google Scholar
    • Export Citation
  • Zhu, Y., , and R. E. Newell, 1998: A proposed algorithm for moisture fluxes from atmospheric rivers. Mon. Wea. Rev., 126, 725735, doi:10.1175/1520-0493(1998)126<0725:APAFMF>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
Save