Illustrating Ensemble Predictability across Scales Associated with the 13–15 February 2019 Atmospheric River Event

Chad W. Hecht Center for Western Weather and Water Extremes, Scripps Institution of Oceanography, University of California, San Diego, La Jolla, California;

Search for other papers by Chad W. Hecht in
Current site
Google Scholar
PubMed
Close
,
Allison C. Michaelis Department of Earth, Atmosphere, and Environment, Northern Illinois University, DeKalb, Illinois;

Search for other papers by Allison C. Michaelis in
Current site
Google Scholar
PubMed
Close
,
Andrew C. Martin Department of Geography, Portland State University, Portland, Oregon;

Search for other papers by Andrew C. Martin in
Current site
Google Scholar
PubMed
Close
,
Jason M. Cordeira Meteorology Program, Plymouth State University, Plymouth, New Hampshire

Search for other papers by Jason M. Cordeira in
Current site
Google Scholar
PubMed
Close
,
Forest Cannon Center for Western Weather and Water Extremes, Scripps Institution of Oceanography, University of California, San Diego, La Jolla, California;

Search for other papers by Forest Cannon in
Current site
Google Scholar
PubMed
Close
, and
F. Martin Ralph Center for Western Weather and Water Extremes, Scripps Institution of Oceanography, University of California, San Diego, La Jolla, California;

Search for other papers by F. Martin Ralph in
Current site
Google Scholar
PubMed
Close
Full access

© 2022 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Allison Michaelis, amichaelis@niu.edu

© 2022 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Allison Michaelis, amichaelis@niu.edu

The “Valentine’s Day” atmospheric river (AR) event that affected a majority of California during 13–15 February 2019 ranked as an AR3 (Ralph et al. 2019) along most of the California coast and reached AR4 intensity in Southern California. The strong onshore flow and dynamically favorable characteristics of the Valentine’s Day AR produced both an intense and long-duration precipitation event resulting in widespread hydrometeorological impacts across California. Palomar Observatory in northern San Diego County observed >10 in. (>254 mm) of precipitation in 24 h, the highest 24-h accumulation since record keeping began in 1943 (Hatchet et al. 2020). Avalanches occurred near Mount Shasta in Northern California and San Gorgonio in Southern California, evacuations near burn areas were ordered due to the risk of debris flows and flash flooding, and heavy snow in the Sierra Nevada and northern Central Valley resulted in disruptions to transportation and commerce. The dynamically favorable characteristics of this AR included an amplifying upper-level midlatitude trough, the presence of multiple surface low pressure systems over the North Pacific basin, and the intensification of a mesoscale frontal wave (MFW) into a secondary low pressure system that all contributed to the evolution of this AR. We examine the 20-member National Centers for Environmental Prediction (NCEP) Global Ensemble Forecast System (GEFS) and the 50-member European Centre for Medium-Range Weather Forecasts (ECMWF) Ensemble Prediction System (EPS) to illustrate the forecast uncertainty surrounding two primary interactions during the evolution of the AR prior to landfall: 1) the phasing between an upper-level midlatitude trough and subtropical cyclone (i.e., kona low; hereafter referred to as the primary cyclone) that led to the initial landfalling AR, and 2) the subsequent intensification of a mesoscale frontal wave into a secondary cyclone that prolonged AR conditions and subsequent precipitation over California. By focusing on the forecast uncertainty associated with the Valentine’s Day AR in the days preceding landfall, we highlight the combination of predictability challenges that can arise within multiple modeling frameworks at varying temporal and spatial scales throughout the offshore evolution of an AR.

Synoptic overview

There were two key synoptic-scale features present over the eastern North Pacific at 0000 UTC 12 February that interacted and ultimately led to the formation and landfall of the AR: 1) a quasi-stationary primary cyclone and tropical moisture export to the northeast of Hawaii and 2) an amplifying 500-hPa trough off the coast of the Pacific Northwest and British Columbia (Fig. 1a). The primary cyclone interacted with the amplifying trough to produce a narrowing corridor of poleward integrated water vapor transport (IVT; calculated from 1,000 to 300 hPa) along a developing AR directed toward the U.S. West Coast (Fig. 1b). As the primary cyclone and intensifying IVT continued to migrate toward the coast, a secondary cyclone developed from an intensifying MFW along the warm front of the primary cyclone in a region favorable for synoptic-scale forcing for ascent located between two 250-hPa jet streaks and immediately downstream of the amplifying trough along the West Coast (Fig. 1c). MFWs are small-scale features that typically develop along fronts of mature extratropical cyclones, which, when they occur along an AR, often affect the position, orientation, and/or intensity of the system (Martin et al. 2019; Michaelis et al. 2021; Ralph et al. 2011). Here, the MFW and subsequent formation of this secondary cyclone led to a secondary lobe of enhanced IVT along the poleward-expanding AR that made landfall at ∼0000 UTC 13 February (Fig. 1c). The secondary lobe of enhanced IVT made landfall prior to the primary corridor later at ∼1200 UTC leading to two maxima in IVT along north-coastal California on 13 February. Another factor that contributed to long durations of AR conditions over Northern California was the relatively slow zonal propagation speeds of the amplifying 500-hPa trough (Fig. 1), which allowed for the onshore IVT to remain persistent for an extended period.

Fig. 1.
Fig. 1.

GFS analysis (F-0) of 500-hPa geopotential height (red; 550-dam contour), SLP (black; 1,000- and 996-hPa contours), 250-hPa wind speed (green; 130-kt contour), IVT magnitude (color coded according to scale), and IVT vectors (kg m−1 s−1; plotted according to the reference vector in the top right) valid at (a) 0000 UTC 12 Feb, (b) 1200 UTC 12 Feb, (c) 0000 UTC 13 Feb, and (d) 1200 UTC 13 Feb 2019. (e) Stage-IV precipitation (color coded according to scale) valid from 1200 UTC 12 Feb through 1200 UTC 15 Feb 2019 and gridpoint AR scale (circles, according to scale and corresponding to lateral coastal grid point). The watersheds used for the analysis in Fig. 6 are outlined and labeled in (e).

Citation: Bulletin of the American Meteorological Society 103, 3; 10.1175/BAMS-D-20-0292.1

The aforementioned synoptic-scale features combined to produce a long-duration, high-intensity AR that brought AR3 or greater conditions on the Ralph et al. (2019) AR scale to a large portion of California (Fig. 1e; Ralph et al. 2019). The AR scale ranks AR events on a scale of 1–5 based on intensity and duration with events lower on the scale (e.g., AR1 and AR2) being considered largely beneficial and events higher on the scale (e.g., AR4 and AR5) mostly hazardous (Ralph et al. 2019). For example, the duration of IVT magnitudes ≥ 250 kg m−1 s−1exceeded 36 h and maximum IVT magnitudes exceeded 500 kg m−1 s−1 associated with the AR over Northern California (37°–42°N) where the secondary cyclone and lobe of enhanced IVT produced an earlier onset to AR conditions than would have been produced by the primary cyclone and corridor of enhanced IVT alone (Fig. 1e; AR2–AR3 conditions). Farther south, the duration of IVT magnitudes ≥ 250 k m−1 s−1 exceeded 40 h and IVT magnitudes exceeded 1,000 kg m−1 s−1 that produced AR4 conditions on the AR scale (Fig. 1e). The largest 72-h precipitation accumulations (>7 in.; >178 mm) occurred over the higher elevations of the Coastal Mountains and Sierra Nevada in Northern California as well as the Peninsular and far eastern Transverse Range (San Gorgonio Mountain) of Southern California (Fig. 1e).

Ensemble forecast analysis

Key synoptic-scale features.

The ensemble-mean GEFS and ECMWF EPS forecasts both contained the key synoptic-scale features associated with a primary cyclone located to the north-northeast of Hawaii with SLP < 1,004 hPa, enhanced IVT magnitudes > 500 kg m−1 s−1 extending poleward and eastward into the subtropical northeast Pacific, and a 500-hPa trough over the northeast Pacific with geopotential heights < 550 dam (Figs. 2a–d). Run-to-run and model-to-model variabilities within the two ensemble forecast systems in forecasts initialized between 0000 UTC 5 February and 0000 UTC 11 February valid at 0000 UTC 12 February principally included 1) a northeast displacement in the location of the primary cyclone converging toward the location of the analyzed cyclone centered at ∼25°N, ∼149°W on 12 February (Figs. 1a, 2a), 2) a northeast (GEFS) and east-northeast (EPS) displacement of the leading edge of enhanced IVT magnitudes > 500 kg m−1 s−1 (Figs. 2b,c), and 3) amplification and location of the 500-hPa trough over the northeast Pacific (Figs. 2a,b). Note that the latter amplification was better forecast by the EPS as compared to the GEFS with differences primarily related to the propagation speed and southeastward location of the trough (cf. Figs. 2a,b). Analysis of the ensemble spread via domain minimum geopotential height demonstrates that while both the GEFS and EPS trended toward lower 500-hPa heights (i.e., a more amplified trough) over the northeast Pacific, both ensemble forecast systems demonstrated characteristically large ensemble spread prior to forecasts initialized at 0000 UTC 8 February (Fig. 2e).

Fig. 2.
Fig. 2.

(a),(b) Ensemble-mean 500-hPa geopotential height (solid; 550-dam contour) and SLP (dashed; 1,004-hPa contour) for the (a) GEFS and (b) ECMWF EPS forecasts initialized every 24 h from 0000 UTC 5 Feb 2019 (F-168) through the valid time of 0000 UTC 12 Feb 2019 (F-0). (c),(d) As in (a) and (b), but for IVT (solid; 500 kg m−1 s−1 contour). (e) Box-and-whisker plots of GEFS (blue) and ECMWF (red) ensemble forecasts of minimum geopotential height (dam) within a domain expanding from 35° to 50°N and from 150° to 135°W (inset in top right with GFS analysis) for forecasts initialized every 24 h from 0000 UTC 5 Feb 2019 through the valid time of 0000 UTC 12 Feb. Ensemble outliers (circles) and extreme outliers (asterisks) are values that are more than 1.5 and 3.0 times the interquartile range, respectively.

Citation: Bulletin of the American Meteorological Society 103, 3; 10.1175/BAMS-D-20-0292.1

Alternatively, ensemble-mean forecasts verifying at 0000 UTC 13 February 2019 did not forecast SLP values < 996 hPa associated with the secondary cyclone and 250-hPa winds > 130 kt associated with the upper-tropospheric jet streak over the Pacific Northwest until initializations at 0000 UTC 10 and 11 February, respectively (Figs. 3a,b). At these short lead times, <72 h, variability in both the spatial extent and locations of SLP < 996 hPa and winds > 130 kt were misrepresented in both ensemble forecast systems (Figs. 3a,b). Similarly, while both ensemble forecast systems contained a large increase in ensemble-mean 250-hPa wind speeds over the Pacific Northwest and decrease in SLP over the northeast Pacific from one forecast initialization to another, with decreasing ensemble spread, very few of the ensemble members and a majority (i.e., the interquartile range) failed to accurately forecast the intensity of either feature at lead times > 48 h (Fig. 3c).

Fig. 3.
Fig. 3.

(a),(b) Ensemble-mean 250-hPa wind speed (solid; 130-kt contour) and SLP (dashed; 996 hPa) for the (a) GEFS and (b) ECMWF EPS forecasts initialized every 24 h from 0000 UTC 5 Feb 2019 (F-192) through the valid time of 0000 UTC 13 Feb 2019 (F-0). (c) Box-and-whisker plots of GEFS (blue) and ECMWF (red) ensemble forecasts of maximum wind speed (kt) within a domain from 47.5° to 57.5°N and from 135° to 110°W (inset in bottom right with GFS analysis) for forecasts initialized every 24 h from 0000 UTC 5 Feb 2019 through the valid time of 0000 UTC 13 Feb. (d) As in (c), but for minimum SLP (hPa) within a domain spanning from 30° to 40°N and from 135° to 125°W.

Citation: Bulletin of the American Meteorological Society 103, 3; 10.1175/BAMS-D-20-0292.1

The formation of the secondary cyclone in the GEFS and EPS forecasts at ∼0000 UTC 10 February following the amplification of the midlevel trough over the northeast Pacific (Fig. 2) and subsequent strengthening of the Pacific Northwest jet streak at ∼0000 UTC 11 February suggests that uncertainty in the dynamics related to secondary cyclogenesis along the MFW may have been related to phasing of the trough with the primary cyclone and latent heat release within the system, which would promote upper-tropospheric jet streak intensification aloft and downstream. Similarly, the northeastward elongation of ensemble-mean SLP < 996 hPa in the forecasts coincided with the northeastward extension of enhanced IVT magnitudes > 500 kg m−1 s−1 toward the Northern California coast, a typical characteristic observed during the development of mesoscale frontal waves and secondary lows (Figs. 4a,b). While both ensemble forecast systems progressively extended the 500 kg m−1 s−1 IVT contour toward the northeast as lead time progressed closer to valid time (0000 UTC 13 February), the ECMWF EPS 500 kg m−1 s−1 IVT contour approached landfall at F-48, ∼24 h earlier than the GEFS. The 24-h forecast in the GEFS aligns well with the AR structure in its analysis whereas the analysis in the ECMWF EPS was the only scenario that suggested that IVT magnitudes > 500 kg m−1 s−1would extend over land.

Fig. 4.
Fig. 4.

Ensemble-mean IVT (solid; 500 kg m−1 s−1 contour) for the (a) GEFS and (b) ECMWF EPS forecasts initialized every 24 h from 0000 UTC 5 Feb through the valid time of 0000 UTC 13 Feb 2019.

Citation: Bulletin of the American Meteorological Society 103, 3; 10.1175/BAMS-D-20-0292.1

Probabilities of atmospheric river conditions.

The GEFS and ECMWF EPS displayed varying degrees of uncertainty in association with the synoptic-scale features that contributed to the formation, evolution, and landfall of the Valentine’s Day AR. These uncertainties were also evident within the landfalling AR via time–latitude-over-threshold forecasts (“AR Landfall Tool”; Cordeira et al. 2017; Cordeira and Ralph 2021) initialized every 24 h from 0000 UTC 5 February through 0000 UTC 12 February (Fig. 5). The AR Landfall Tool plots the percentage of ensemble members that forecast IVT > 250 kg m−1 s−1 at points along the U.S. West Coast where forecast lead time increases from right to left, as though the features are approaching the coast. The GEFS and EPS AR Landfall Tool highlighted probabilities of IVT ≥ 250 kg m−1 s−1(i.e., AR conditions) over 50% as early as ∼9 days in advance over the Southern California coast, while the EPS suggested higher ensemble probabilities (>20% higher) over a majority of coastal California (Figs. 5a,b and 5i,j). The EPS continuously predicted higher ensemble probabilities of AR conditions over the next 2 days and saw a considerable convergence of ensemble scenarios during the forecast initialized at 0000 UTC 8 February, extending probabilities of AR conditions > 70% to 46°N and confining the higher probabilities of AR conditions to within a 24-h period along a majority of the California coast (Fig. 5l). This EPS forecast coincided with a reduction in the large ensemble spread and a shift toward a deeper 500-hPa ensemble-mean trough (Figs. 2b,e). The GEFS also experienced a trend toward higher probabilities of AR conditions north of 40°N, though these probabilities were ∼20% lower than the EPS and exhibited weaker agreement in landfall timing (Figs. 5d–h and 5l–p). Both ensemble systems began predicting probabilities of AR conditions > 75% between 35° and 40°N during the forecast initialized at 0000 UTC 9 February, with the EPS exhibiting probabilities > 90% (10%–50% higher than the GEFS), the time at which the EPS also suggested stronger 250-hPa wind speeds over the Pacific Northwest (Fig. 3; Figs. 5e,m). In the forecasts initialized 24-h later (0000 UTC 10 February), both ensemble systems began displaying probabilities of AR conditions > 95% over two coastal regions, one region spanning from Oregon to central California and a second from Southern California to the central Baja Peninsula, Mexico (Figs. 5f,n). During this forecast, the ECMWF EPS exhibited a higher probability of a longer-duration AR over a larger coastal extent within the northern region when compared to the GEFS coinciding with the potential development of a secondary low pressure system based on ensemble spread of domain-minimum sea level pressure (Fig. 3; Figs. 5g,h and 5o,p). During the forecast initialized at 0000 UTC 12 February, ∼24–36 h before landfall, both ensemble forecast systems displayed 100% ensemble probability of AR conditions lasting more than 24 h over the entirety of California, though the GEFS exhibited lower ensemble agreement toward the latter half of the event resulting in larger uncertainty pertaining to the overall duration of the event (Figs. 5h,p).

Fig. 5.
Fig. 5.

Time–latitude ensemble probabilities (color shaded according to scale) of IVT > 250 kg m−1 s−1at coastal points for the (a)–(h) GEFS and (i)–(p) ECMWF EPS forecasts initialized every 24 h from 0000 UTC 5 Feb through 0000 UTC 12 Feb 2019.

Citation: Bulletin of the American Meteorological Society 103, 3; 10.1175/BAMS-D-20-0292.1

Summary and discussion

In the present study, we evaluated the ensemble forecast performance of several key dynamical features that played important roles in the formation and evolution prior to landfall of the Valentine’s Day AR that impacted California from 13 to 15 February 2019. We identified uncertainty associated with multiple synoptic-to-mesoscale features in the GEFS and ECMWF EPS forecast progression leading up to the landfall of the AR over California. The first key development within each ensemble forecast system was the formation of an amplifying 500-hPa trough near the Gulf of Alaska, a feature that consistently appeared in EPS, but was only first suggested by GEFS ∼96 h before the analysis period. Run-to-run fluctuations, along with ensemble uncertainty, in the phase and amplitude of the forecast 500-hPa trough overlapped with changes in probabilities of IVT ≥ 250 kg m−1 s−1 along the California coast, further highlighting the connection between the trough and AR. In both ensemble forecast systems, the poleward extension of higher AR condition probabilities along coastal California is consistent with the forecast evolution of the 500-hPa trough. In the GEFS, higher probabilities of IVT ≥ 250 kg m−1 s−1 extended farther northward during the forecast initialized at 0000 UTC 6 February compared to 0000 UTC 5 February and 0000 UTC 7 February before converging on the poleward probabilities at 0000 UTC 8 February. Similarly, the ECMWF EPS consistently predicted high probabilities of IVT ≥ 250 kg m−1 s−1 north of 29°N, before converging toward larger ensemble agreement during the forecast initialized at 0000 UTC 8 February, the forecast run that also exhibited a large shift in the 500-hPa trough toward the analysis.

While the 500-hPa trough was the first feature identified within this analysis of the forecast evolution that contributed to higher probabilities of AR landfall, there were other synoptic-to-mesoscale features that ultimately influenced the Valentine’s Day AR and its impacts over California. For example, a short, compacted jet streak over the Pacific Northwest provided quasigeostrophic support for cyclogenesis over the northern edge of the AR, playing a role in the formation of an MFW and subsequent secondary cyclone. The formation of the secondary cyclone resulted in a northward extension and intensification of the AR, leading to an earlier onset and prolongation of enhanced AR conditions over Northern California. The EPS identified the formation of the Pacific Northwest jet streak and secondary cyclone development ∼24 h earlier than GEFS which is consistent with the higher probabilities of AR conditions over Northern California forecast by the EPS. The approximate ∼24 h of additional lead time for skillful prediction provided by the ECMWF EPS across each of the meteorological features and AR condition probabilities discussed is consistent with findings from Stewart (2021).

Trends in precipitation forecasts across California are generally consistent with the trends in meteorological features (e.g., trough, secondary low, AR) in the sense that as GEFS and ECMWF EPS forecasts trended closer toward a landfalling AR, precipitation forecasts trended toward higher amounts throughout the state (Fig. 6). Similar to the 24 h of additional lead time the ECMWF EPS provided in relation to the important meteorological features of this case, the ECMWF EPS consistently predicted higher watershed average precipitation accumulations compared to the GEFS in the days leading up to AR landfall. Any forecast adjustments in the synoptic-to-mesoscale features created large fluctuations in the precipitation forecast on a watershed scale, as demonstrated by the sharp increase in precipitation forecasts over California, specifically, the Russian, Upper Yuba, and Santa Ana River watersheds. For example, the ECMWF EPS experienced a large increase in 72-h precipitation forecasts over the Russian and Santa Ana River watersheds during the forecast initialized at 0000 UTC 9 February, which coincided with the shift toward a stronger Pacific Northwest jet and secondary low intensification (Figs. 2c,d). An increasing trend in watershed average precipitation was also seen in each California–Nevada River Forecast Center forecast leading up to the event, demonstrating how changes in ensemble forecasts can link closely to that of operational centers.

Fig. 6.
Fig. 6.

Watershed-averaged ensemble 72-h precipitation forecasts by the GEFS (blue) and ECMWF EPS (red) initialized every 24 h from 0000 UTC 5 Feb (F-180 to F-252) to 0000 UTC 12 Feb 2019 (F-12 to F-84) valid from 1200 UTC 12 Feb through 1200 UTC 15 Feb 2019 for the (a) Russian, (b) Upper Yuba, and (c) Santa Ana River watersheds. The California–Nevada River Forecast Center forecast initialized every 24 h from 1200 UTC 9 Feb (F-72 to F-144) through 1200 UTC 12 Feb 2019 (F-00 to F-72) for the same valid time period and each watershed is plotted in the gold triangles. The Stage-IV analysis is shown with the green line. The outline for each watershed is plotted in Fig. 1e.

Citation: Bulletin of the American Meteorological Society 103, 3; 10.1175/BAMS-D-20-0292.1

By examining the forecast uncertainty surrounding several dynamical features present throughout the evolution of the impactful Valentine’s Day AR using two ensemble forecast systems, we highlight the complexities associated with forecasting AR landfall position and intensity as well as extreme precipitation over the U.S. West Coast (e.g., Ralph et al. 2020). Although these results are drawn from a single case, it is clear that mesoscale interactions between synoptic-scale features leading to AR conditions over the entire western United States can have dramatic influence over watershed-scale impacts, with implications for skillful hydrometeorological prediction.

Acknowledgments.

This research was funded by the AR Program Phase II, Grant 4600013361, sponsored by the California Department of Water Resources and Forecast Informed Reservoir Operations (FIRO), Grant W912HZ-19-SOI-0027, sponsored by the U.S. Army Corps of Engineers Research and Development Center. The authors thank Benjamin Moore and one anonymous reviewer whose constructive feedback improved the quality of this manuscript.

References

  • Cordeira, J. M. , and F. M. Ralph , 2021: A summary of GFS ensemble integrated water vapor transport forecasts and skill along the U.S. West Coast during water years 2017–20. Wea. Forecasting, 36, 361377, https://doi.org/10.1175/WAF-D-20-0121.1.

    • Search Google Scholar
    • Export Citation
  • Cordeira, J. M. , F. M. Ralph , A. Martin , N. Gaggini , J. R. Spackman , P. J. Neiman , J. J. Rutz , and R. Pierce , 2017: Forecasting atmospheric rivers during CalWater 2015. Bull. Amer. Meteor. Soc., 98, 449459, https://doi.org/10.1175/BAMS-D-15-00245.1.

    • Search Google Scholar
    • Export Citation
  • Hatchett, B. , and Coauthors, 2020. Observations of an extreme atmospheric river storm with a diverse sensor network. Earth Space Sci., 7, e2020EA001129, https://doi.org/10.1029/2020EA001129.

    • Search Google Scholar
    • Export Citation
  • Martin, A. C. , F. M. Ralph , A. Wilson , L. DeHaan , and B. Kawzenuk , 2019: Rapid cyclogenesis from a mesoscale frontal wave on an atmospheric river: Impacts on forecast skill and predictability during atmospheric river landfall. J. Hydrometeor., 20, 17791794, https://doi.org/10.1175/JHM-D-18-0239.1.

    • Search Google Scholar
    • Export Citation
  • Michaelis, A. C. , A. C. Martin , M. A. Fish , C. W. Hecht , and F. M. Ralph , 2021: Modulation of atmospheric rivers by mesoscale frontal waves and latent heating: Comparison of two U.S. West Coast events. Mon. Wea. Rev., 149, 27552776, https://doi.org/10.1175/MWR-D-20-0364.1.

    • Search Google Scholar
    • Export Citation
  • Ralph, F. M. , P. J. Neiman , G. N. Kiladis , K. Weickmann , and D. W. Reynolds , 2011: A multiscale observational case study of a Pacific atmospheric river exhibiting tropical–extratropical connections and a mesoscale frontal wave. Mon. Wea. Rev., 139, 11691189, https://doi.org/10.1175/2010MWR3596.1.

    • Search Google Scholar
    • Export Citation
  • Ralph, F. M. , J. J. Rutz , J. M. Cordeira , M. Dettinger , M. Anderson , D. Reynolds , L. J. Schick , and C. Smallcomb , 2019: A scale to characterize the strength and impacts of atmospheric rivers. Bull. Amer. Meteor. Soc., 100, 269289, https://doi.org/10.1175/BAMS-D-18-0023.1.

    • Search Google Scholar
    • Export Citation
  • Ralph, F. M. , and Coauthors, 2020: West Coast forecast challenges and development of atmospheric river reconnaissance. Bull. Amer. Meteor. Soc., 101, E1357E1377, https://doi.org/10.1175/BAMS-D-19-0183.1.

    • Search Google Scholar
    • Export Citation
  • Stewart, B. E. , 2021: Evaluating GFS and ECMWF ensemble forecasts of integrated water vapor transport along the U.S. West Coast. M.S. thesis, Dept. of Atmospheric Science and Chemistry, Plymouth State University, 68 pp., www.proquest.com/openview/c6c90b995a320d7ecb09c4d6009d5ee0/1?pq-origsite=gscholar&cbl=18750&diss=y.

Save
  • Cordeira, J. M. , and F. M. Ralph , 2021: A summary of GFS ensemble integrated water vapor transport forecasts and skill along the U.S. West Coast during water years 2017–20. Wea. Forecasting, 36, 361377, https://doi.org/10.1175/WAF-D-20-0121.1.

    • Search Google Scholar
    • Export Citation
  • Cordeira, J. M. , F. M. Ralph , A. Martin , N. Gaggini , J. R. Spackman , P. J. Neiman , J. J. Rutz , and R. Pierce , 2017: Forecasting atmospheric rivers during CalWater 2015. Bull. Amer. Meteor. Soc., 98, 449459, https://doi.org/10.1175/BAMS-D-15-00245.1.

    • Search Google Scholar
    • Export Citation
  • Hatchett, B. , and Coauthors, 2020. Observations of an extreme atmospheric river storm with a diverse sensor network. Earth Space Sci., 7, e2020EA001129, https://doi.org/10.1029/2020EA001129.

    • Search Google Scholar
    • Export Citation
  • Martin, A. C. , F. M. Ralph , A. Wilson , L. DeHaan , and B. Kawzenuk , 2019: Rapid cyclogenesis from a mesoscale frontal wave on an atmospheric river: Impacts on forecast skill and predictability during atmospheric river landfall. J. Hydrometeor., 20, 17791794, https://doi.org/10.1175/JHM-D-18-0239.1.

    • Search Google Scholar
    • Export Citation
  • Michaelis, A. C. , A. C. Martin , M. A. Fish , C. W. Hecht , and F. M. Ralph , 2021: Modulation of atmospheric rivers by mesoscale frontal waves and latent heating: Comparison of two U.S. West Coast events. Mon. Wea. Rev., 149, 27552776, https://doi.org/10.1175/MWR-D-20-0364.1.

    • Search Google Scholar
    • Export Citation
  • Ralph, F. M. , P. J. Neiman , G. N. Kiladis , K. Weickmann , and D. W. Reynolds , 2011: A multiscale observational case study of a Pacific atmospheric river exhibiting tropical–extratropical connections and a mesoscale frontal wave. Mon. Wea. Rev., 139, 11691189, https://doi.org/10.1175/2010MWR3596.1.

    • Search Google Scholar
    • Export Citation
  • Ralph, F. M. , J. J. Rutz , J. M. Cordeira , M. Dettinger , M. Anderson , D. Reynolds , L. J. Schick , and C. Smallcomb , 2019: A scale to characterize the strength and impacts of atmospheric rivers. Bull. Amer. Meteor. Soc., 100, 269289, https://doi.org/10.1175/BAMS-D-18-0023.1.

    • Search Google Scholar
    • Export Citation
  • Ralph, F. M. , and Coauthors, 2020: West Coast forecast challenges and development of atmospheric river reconnaissance. Bull. Amer. Meteor. Soc., 101, E1357E1377, https://doi.org/10.1175/BAMS-D-19-0183.1.

    • Search Google Scholar
    • Export Citation
  • Stewart, B. E. , 2021: Evaluating GFS and ECMWF ensemble forecasts of integrated water vapor transport along the U.S. West Coast. M.S. thesis, Dept. of Atmospheric Science and Chemistry, Plymouth State University, 68 pp., www.proquest.com/openview/c6c90b995a320d7ecb09c4d6009d5ee0/1?pq-origsite=gscholar&cbl=18750&diss=y.

  • Fig. 1.

    GFS analysis (F-0) of 500-hPa geopotential height (red; 550-dam contour), SLP (black; 1,000- and 996-hPa contours), 250-hPa wind speed (green; 130-kt contour), IVT magnitude (color coded according to scale), and IVT vectors (kg m−1 s−1; plotted according to the reference vector in the top right) valid at (a) 0000 UTC 12 Feb, (b) 1200 UTC 12 Feb, (c) 0000 UTC 13 Feb, and (d) 1200 UTC 13 Feb 2019. (e) Stage-IV precipitation (color coded according to scale) valid from 1200 UTC 12 Feb through 1200 UTC 15 Feb 2019 and gridpoint AR scale (circles, according to scale and corresponding to lateral coastal grid point). The watersheds used for the analysis in Fig. 6 are outlined and labeled in (e).

  • Fig. 2.

    (a),(b) Ensemble-mean 500-hPa geopotential height (solid; 550-dam contour) and SLP (dashed; 1,004-hPa contour) for the (a) GEFS and (b) ECMWF EPS forecasts initialized every 24 h from 0000 UTC 5 Feb 2019 (F-168) through the valid time of 0000 UTC 12 Feb 2019 (F-0). (c),(d) As in (a) and (b), but for IVT (solid; 500 kg m−1 s−1 contour). (e) Box-and-whisker plots of GEFS (blue) and ECMWF (red) ensemble forecasts of minimum geopotential height (dam) within a domain expanding from 35° to 50°N and from 150° to 135°W (inset in top right with GFS analysis) for forecasts initialized every 24 h from 0000 UTC 5 Feb 2019 through the valid time of 0000 UTC 12 Feb. Ensemble outliers (circles) and extreme outliers (asterisks) are values that are more than 1.5 and 3.0 times the interquartile range, respectively.

  • Fig. 3.

    (a),(b) Ensemble-mean 250-hPa wind speed (solid; 130-kt contour) and SLP (dashed; 996 hPa) for the (a) GEFS and (b) ECMWF EPS forecasts initialized every 24 h from 0000 UTC 5 Feb 2019 (F-192) through the valid time of 0000 UTC 13 Feb 2019 (F-0). (c) Box-and-whisker plots of GEFS (blue) and ECMWF (red) ensemble forecasts of maximum wind speed (kt) within a domain from 47.5° to 57.5°N and from 135° to 110°W (inset in bottom right with GFS analysis) for forecasts initialized every 24 h from 0000 UTC 5 Feb 2019 through the valid time of 0000 UTC 13 Feb. (d) As in (c), but for minimum SLP (hPa) within a domain spanning from 30° to 40°N and from 135° to 125°W.

  • Fig. 4.

    Ensemble-mean IVT (solid; 500 kg m−1 s−1 contour) for the (a) GEFS and (b) ECMWF EPS forecasts initialized every 24 h from 0000 UTC 5 Feb through the valid time of 0000 UTC 13 Feb 2019.

  • Fig. 5.

    Time–latitude ensemble probabilities (color shaded according to scale) of IVT > 250 kg m−1 s−1at coastal points for the (a)–(h) GEFS and (i)–(p) ECMWF EPS forecasts initialized every 24 h from 0000 UTC 5 Feb through 0000 UTC 12 Feb 2019.

  • Fig. 6.

    Watershed-averaged ensemble 72-h precipitation forecasts by the GEFS (blue) and ECMWF EPS (red) initialized every 24 h from 0000 UTC 5 Feb (F-180 to F-252) to 0000 UTC 12 Feb 2019 (F-12 to F-84) valid from 1200 UTC 12 Feb through 1200 UTC 15 Feb 2019 for the (a) Russian, (b) Upper Yuba, and (c) Santa Ana River watersheds. The California–Nevada River Forecast Center forecast initialized every 24 h from 1200 UTC 9 Feb (F-72 to F-144) through 1200 UTC 12 Feb 2019 (F-00 to F-72) for the same valid time period and each watershed is plotted in the gold triangles. The Stage-IV analysis is shown with the green line. The outline for each watershed is plotted in Fig. 1e.

All Time Past Year Past 30 Days
Abstract Views 12 0 0
Full Text Views 1184 268 17
PDF Downloads 679 158 10