1. Introduction
Atmospheric rivers (ARs) are synoptic-scale phenomena which are associated with long, narrow corridors of enhanced low-level water vapor transport (American Meteorological Society 2018). Landfalling ARs often produce a wide variety of coastal and inland impacts worldwide, that may range from beneficial precipitation for drought amelioration (Dettinger 2013) to hazardous precipitation leading to flooding (Ralph et al. 2006). These beneficial precipitation events associated with ARs can benefit watersheds across the western United States by contributing up to 30%–50% of annual precipitation totals (Dettinger et al. 2011), increasing the winter snowpack in high elevations to later recharge watersheds in spring and summer (Neiman et al. 2008), and acting as “drought busters” by breaking prolonged periods of dry conditions (Dettinger 2013). Alternatively, the hazardous precipitation events associated with ARs may lead to flooding along rivers in California (Ralph et al. 2006) and heavy snow events in the Sierra Nevada (Guan et al. 2010); hazards may also include high wind events along the coast and inland mountain ranges (Waliser and Guan 2017). In anticipation of these potentially hazardous events, the National Weather Service (NWS) weather forecast offices will often issue watches, warnings, and advisories (WWAs) for their county warning areas and public forecast zones (PFZs). For example, 50%–70% of days with flood-related and 60%–80% of days with winter weather-related WWAs in California occurred on days with landfalling ARs (Cordeira et al. 2018). The goal of the present study is to further investigate this spatial and temporal relationship between landfalling ARs and the potential for hazardous weather via issuance of WWAs across the entire western United States for the period 2006–18.
The aforementioned benefit–hazard distribution associated with ARs may be summarized using a scale developed to assess the intensity and duration of landfalling ARs (Ralph et al. 2019). The AR scale subjectively ranges from primarily beneficial ARs (AR1) to primarily hazardous ARs (AR5) with the possibility of some storms providing impacts that may be both beneficial and hazardous (AR2–4; see Table 2 and Fig. 4 within Ralph et al. 2019). The AR scale combines the intensity of an AR [defined as the maximum integrated vapor transport (IVT) magnitude during an event] and its duration (defined as the total number of hours with IVT magnitudes ≥ 250 kg m−1 s−1) to describe the potential for hazardous weather associated with ARs. The hazards associated with landfalling ARs may also be summarized, for example, by their financial impact and propensity to cause flood-related damages. An analysis of landfalling ARs across the western United States combined with data from the National Flood Insurance Program demonstrates that the cost from flood damages increased exponentially as landfalling ARs increased in intensity and duration from an AR1 to an AR5 (Corringham et al. 2019). Although the potential for hazardous weather in the Ralph et al. (2019) AR scale increases on average as AR intensity and duration increases, it is important to recognize that the actual hazards may depend heavily on antecedent conditions such as saturated soils prior to enhanced streamflow (e.g., Ralph et al. 2013) or wildfire burn scars prior to debris flows (e.g., Oakley et al. 2017).
Examples of ARs and their AR scale ranking that occurred in association with both numerous WWAs and observations of hazardous weather across the western United States are illustrated for an AR3 event in the Pacific Northwest (Figs. 1a–c), an AR3 event in Northern California (Figs. 1d–f), and an AR4 event in Southern California (Figs. 1g–i):
The AR3 event in the Pacific Northwest occurred during 26 January–2 February 2020 and contained a maximum IVT magnitude > 810 kg m−1 s−1 and a 45-h duration of IVT magnitudes ≥ 250 kg m−1 s−1 in Washington and Oregon (Fig. 1a) that produced ~125–375 mm (~5–15 in.) of precipitation across the Olympic peninsula (Fig. 1b), and 30–40 individual WWA actions (e.g., actions herein refer to WWA status classifications when a WWA is issued, upgraded, extended in area, etc., see section 2) in PFZs primarily located in the northern Cascades (Fig. 1c).
The AR3 event in Northern California occurred during 25–27 February 2019 and contained a maximum IVT magnitude > 630 kg m−1 s−1 and a 51-h duration of IVT magnitudes ≥ 250 kg m−1 s−1 in North-Coastal California (Fig. 1d) that produced ~500–600 mm (~20–25 in.) of precipitation in Coastal California, ~225–375 mm (9–15 in.) of precipitation in the southern Klamath Mountain range (Fig. 1e), and 50–70 WWA actions in PFZs primarily located around the California San Francisco Bay Area (Fig. 1f).
The AR4 event in Southern California occurred during 13–15 February 2019 and contained a maximum IVT magnitude > 1000 kg m−1 s−1 and a 42-h duration of IVT magnitudes ≥ 250 kg m−1 s−1 (Fig. 1g) that produced ~125–225 mm (5–9 in.) of precipitation in the inland mountains of Southern California between Los Angeles and San Diego (Fig. 1h), and 50–80 WWA actions for PFZs primarily north of San Diego (Fig. 1i; note that this AR initially made landfall on 13 February in the San Francisco Bay Area and thus the widespread hazards and WWAs also in Northern California).
(a),(d),(g) National Centers for Environmental Prediction Global Forecast System 0-h IVT magnitude (kg m−1 s−1; shaded according to scale), IVT vectors plotted for magnitudes > 250 kg m−1 s−1 according to the reference vector and sea level pressure (hPa; contours); (b),(e),(h) PRISM observed precipitation (shaded in in.); and (c),(f),(i) WWA actions within each NWS public forecast zone for duration of three landfalling ARs in the (top) Pacific Northwest, (middle) Pacific Central, and (bottom) Pacific Southwest Regions. Dates of events are included in each panel. Panels (a), (d), and (g) are provided by the Center for Western Weather and Water Extremes at the UCSD Scripps Institution of Oceanography.
Citation: Weather and Forecasting 36, 3; 10.1175/WAF-D-20-0212.1
The aforementioned goal of this study seeks to investigate the relationship between landfalling ARs and the potential for hazardous weather via WWAs across the western United States, and to also investigate whether or not more intense and longer-duration ARs defined by the Ralph et al. (2019) scale may contribute to a higher potential for hazardous weather as quantified by the frequency and spatial extent of WWAs. It is hypothesized that, in a manner similar to Cordeira et al. (2018) and Corringham et al. (2019), that 1) a majority of WWAs issued in advance of potentially hazardous weather across the western United States occur in association with landfalling ARs, 2) WWAs issued in advance of potentially hazardous weather are increasingly more likely to occur in association with more intense and longer-duration ARs increases, 3) not all landfalling ARs necessarily require WWAs (i.e., not all ARs are hazardous), but 4) more intense and longer-duration ARs are more likely to require WWAs as compared to less intense and shorter-duration ARs (i.e., more intense and longer-duration ARs are more hazardous). Section 2 of this study provides the data and methods, section 3 illustrates the results of evaluating hypotheses 1 and 2 related to the “WWA perspective,” and section 4 illustrates the results of evaluating hypotheses 3 and 4 related to the landfalling “AR perspective.” Section 5 consists of our conclusions and offers concepts for future work.
2. Data and methodology
The analysis methodology focuses on two perspectives: the WWA perspective and the AR perspective. The WWA perspective used a catalog of WWA days (i.e., a day with at least one WWA of a given hazard type in a PFZ) to investigate how frequently WWA days occurred concurrently with an AR day (i.e., a day with a landfalling AR at the coast). Alternatively, the AR perspective used a catalog of AR days to investigate how frequently AR days occurred concurrently with WWA days. The WWA perspective may be thought of as answering the question “What fraction of hazardous weather days are related to landfalling ARs?” whereas the AR perspective may be thought of as answering the question “How likely is a landfalling AR to produce hazardous weather?” Additional details of the associated data and methods are included below.
The relationship between ARs and potentially hazardous weather is studied for the cool-season months of October through March during a 13-yr period between 1 January 2006 and 31 December 2018. This study defines the western United States using the state boundaries inclusive of Washington, Idaho, Oregon, California, Nevada, Utah, and Arizona. This region includes 22 NWS weather forecast offices (Fig. 2a) that contain 392 PFZs (Fig. 2b). In this study, the western United States is subset into Pacific Northwest, Pacific Central, and Pacific Southwest regions that were subjectively defined using the boundaries of the NWS weather forecast office county warning areas (Fig. 2c). These three regional subsets are primarily used to analyze locations of landfalling ARs along the coast and their potential for hazardous weather within PFZs in the same region. The PFZs across the western United States often encompass unique orographic features and may be different from county boundaries, such as the long, rectangular PFZs that run parallel to the Sierra Nevada and Cascade mountains (Fig. 2d).
Map of the Western Region with (a) county warning areas (CWAs; thick black line) and state boundaries (thin gray lines), (b) CWAs (thick black lines) and PFZs (thin gray lines), (c) terrain (m; shaded) and state lines (black lines), and (d) PFZs (thin black lines) and state lines (thick blank lines) with Pacific Northwest (green), Pacific Central (blue), and Pacific Southwest (orange) regions used to assess relationships with landfalling ARs in this study at coastal locations denoted by circle, triangle, and square symbols, respectively. The numbers next to the symbols in (d) give the latitude. Topography data in (c) were provided by the NOAA National Centers for Environmental Information ETOPO1 Global Relief Model (https://www.ngdc.noaa.gov/mgg/global/).
Citation: Weather and Forecasting 36, 3; 10.1175/WAF-D-20-0212.1
This study quantifies the potential for hazardous weather using WWAs issued by the NWS obtained as ESRI shapefiles from the Iowa State University Iowa Environmental Mesonet online archive (Iowa State University 2019). WWAs of all status classifications (e.g., action types that include “new event,” “event continued,” “event extended in time,” “event extended in area,” “event extended in both time and area,” “event upgraded,” “event canceled,” “event expired,” “correction,” or “routine”) and all significances (e.g., “warning,” “watch,” “advisory,” and “statement”) are retained for analysis, subset to the western United States, and matched with PFZs for which they overlap if necessary (i.e., a flash flood warning that may be issued for a certain area that is not a PFZ). The valid times associated with the WWAs are then filtered by day, PFZ, and by phenomenon (i.e., hazard) type to create the WWA catalog.
The WWA catalog, which at this point may be spatially and temporally redundant (e.g., may contain multiple records on the same day in the same PFZ for the same hazard if a given WWA is upgraded or canceled), is next spatially and temporally uniquely filtered to create the “WWA day” catalog. A WWA day is defined for each PFZ and includes whether or not a given hazard was at some point anticipated by the local NWS weather forecast office, even if the WWA was later modified following the other status classifications. This study does not remove a WWA if it was canceled, does not consider whether or not WWAs verified, and does not consider whether or not a hazard or report of hazardous weather occurred without a WWA. Therefore, the WWA catalog and the results herein rely on the short-term forecast skill of forecasters and guidance from numerical weather prediction that led to issuance of these WWAs, and variability in best practices for use of modifications (e.g., letting a WWA expire, canceling a WWA, or extending it in area and/or time). The representativeness of this methodology that focuses on the potential for hazardous weather via WWA to the actual occurrence of hazardous weather (e.g., via storm reports) is therefore a function of that skill and likely varies from region to region and from weather forecast office to weather forecast office. Using WWAs is a spatially and temporally continuous record of potentially hazardous weather across the western United States that a dataset containing local storm reports may not be able to provide in population sparse areas.
The WWA phenomenon types (e.g., flash flood, areal flood, winter storm, high wind, etc.) are then grouped into subset categories for further analysis (Fig. 3). The WWA subset categories are grouped based on common hazards frequently associated with landfalling ARs such as: hydrologic (e.g., flood-related), cold-precipitation (e.g., winter weather), and wind-related WWAs. Regionally varying criteria for issuing WWAs also provide a relative baseline from which to illustrate the localized potential for hazardous weather. For example, the 24-h snowfall thresholds for a winter storm warning in the California Central Valley near Sacramento at elevations < 3000 ft is 7 in. and nearby in the Sierra Nevada at elevations > 7000 ft is 18 in. (National Weather Service 2020, p. D-5). Note that some of the WWA types included in this analysis are not climatically applicable to all regions and also may no longer be in use. For example, PFZs in the southern Central Valley of California and deserts in Southern California and western Arizona had very few or zero cold-precipitation WWAs during the period of study. There may also be differences from region to region, from weather forecast office to weather forecast office, and from forecaster to forecaster, on best practices for marginal forecast parameters that may warrant a WWA for a given hazard or not. Grouping hazards by type does help alleviate some of these concerns, but it is worth emphasizing that the WWA catalog contains regionally varying inputs that are influenced by human forecasters, numerical weather prediction model skill, and climate/frequency of hazards.
Total number of (a) hydrologic (green), (b) cold-precipitation (purple), and (c) wind-related (brown) WWAs, plotted on a logarithmic scale for the Western Region between 1 Jan 2006 and 31 Dec 2018.
Citation: Weather and Forecasting 36, 3; 10.1175/WAF-D-20-0212.1
Landfalling ARs are identified from a catalog of AR data provided by Rutz et al. (2014) using the NASA MERRA-2 reanalysis (Rienecker et al. 2011) and 3-h values of IVT magnitude at 19 latitude and longitude locations along the U.S. West Coast (Fig. 2c). This AR catalog is grouped into whether or not a landfalling AR was present during any of the eight 3-h times on a UTC calendar day (i.e., an AR day) and then grouped into the three Pacific Northwest (41°–48°N), Pacific Central (34°–42°N), and Pacific Southwest (30°–35°N) regions (Fig. 2c). Note that Rutz et al. (2014) defines an AR using a combination of both IVT magnitudes ≥ 250 kg m−1 s−1 and a length criterion of at least 2000 km. In each region, the location with the maximum IVT magnitude and duration of IVT magnitudes ≥ 250 kg m−1 s−1 associated with landfalling AR events are used to calculate the AR scale following the methodology of Ralph et al. (2019) for each event, which is then assigned to each day of the landfalling AR event.
There were 1045 AR days in the Pacific Northwest, 812 AR days in the Pacific Central, and 305 AR days in the Pacific Southwest regions during the cool-season months studied in this analysis. These values correspond to an average cool-season number of AR days of 80.4 in the Pacific Northwest, 62.5 in the Pacific Central, and 23.5 in the Pacific Southwest regions (Fig. 4a). A breakdown of the average cool-season number of AR days for varying intensities and durations according to the Ralph et al. (2019) scale (Fig. 4b) demonstrates that AR1–2 days occur more frequently in the Pacific Central (34.0), and Pacific Southwest (17.9) regions as compared to AR3–5 days (28.5 and 5.6, respectively). The distribution of average cool-season number of AR days was approximately the same for AR1–2 days (38.5) as compared to AR3–5 days (42.0) in the Pacific Northwest region. There were zero AR5 days in the Pacific Southwest region during the period of study.
(a) The average annual number of AR days of varying scale in each region and (b) the AR scale as developed by Ralph et al. (2019).
Citation: Weather and Forecasting 36, 3; 10.1175/WAF-D-20-0212.1
3. WWA perspective
a. Given a WWA day, did it occur concurrently with an AR day?
The frequency of an AR at the coast on cool-season days with any type of WWA was 50%–70% for PFZs in the Pacific Northwest and Pacific Central regions, with maximum frequencies in PFZs across the northern Cascades and Olympic Peninsula (Fig. 5a). Hydrologic WWAs were most frequently associated with ARs in coastal PFZs in the Pacific Northwest and Pacific Central region (70%–90%), followed by PFZs in the inland mountainous areas of the Pacific Southwest region (50%–70%) (Fig. 5b). Cold-precipitation WWAs were most frequently associated with ARs in PFZs in the Pacific Northwest and Pacific Central regions, with regional maximum frequencies in the many mountainous PFZs of the northern Cascades (50%–70%) and the Sierra Nevada (40%–60%) mountains (Fig. 5c). Similarly, wind-related WWAs were most frequently associated with ARs in the Pacific Northwest and Pacific Central regions, with maxima extending inland across PFZs in Washington, Oregon, and Idaho (70%–90%) and from Northern California into the Great Basin (40%–60%) (Fig. 5d). These frequency maxima across all hazard types appear to strongly correlate spatially with topography. For example, the highest values are >70% in PFZs that contain the high elevation regions of the western United States, including the Cascades, northern Sierra Nevada, and Coastal Ranges. This variability is consistent with previous studies that illustrate many hazards associated with ARs are the result of upslope water vapor flux (e.g., Ralph et al. 2013) and strong winds in regions of complex terrain (e.g., Waliser and Guan 2017), with similar observed fractions of precipitation and snowfall associated with ARs (see Fig. 8 of Rutz et al. 2014).
Frequency of WWA days during the cool season with (a) any, (b) hydrologic (HL), (c) cold-precipitation (CP), and (d) wind-related (WR) WWAs (shaded) and an AR at the coast (shaded) in each PFZ (thin black lines). NWS county warning areas (thin white lines) are overlaid on the PFZs in each region (thick black lines), which are analyzed for landfalling ARs in the same region at coastal locations given by the symbols: Pacific Northwest (circles), Pacific Central (triangles), and Pacific Southwest (squares) regions.
Citation: Weather and Forecasting 36, 3; 10.1175/WAF-D-20-0212.1
b. Given a WWA day, did it occur concurrently with an AR1–2 or AR3–5 event?
Most cool-season WWA days (60%–80%) of any type in Pacific Southwest PFZs occurred in association with an AR1–2 event at the coast (Fig. 6a) with fewer (<30%) that occurred in association with AR3–5 events (Fig. 6b). This result is illustrative of the large number of observed AR1–2 events across the Pacific Southwest region relative to the number of AR3–5 events (Fig. 4). Alternatively, the frequency of an AR at the coast with cool-season WWA days of any type increased from 40%–50% to 50%–70% for PFZs in the Pacific Northwest and Pacific Central regions for AR1–2 versus AR3–5 events, respectively (Figs. 6a,b). This result suggests that WWA days are more likely to be associated with more intense and longer-duration landfalling ARs for PFZs in the Pacific Northwest and Pacific Central regions than less intense and shorter-duration landfalling ARs. When partitioned by phenomenon type, hydrologic and wind-related WWA days in the Pacific Northwest and Pacific Central regions are also more likely to be associated with AR3–5 events at the coast (~60%–100%) as compared to AR1–2 events (<40%) (Figs. 7a,b,e,f). Alternatively, cold-precipitation WWA days in PFZs across all three regions were more or equally likely to be associated with AR1–2 events at the coast (~50%–70%) as compared to AR3–5 events (<50%) (Figs. 7c,d). This latter result suggests that cold-precipitation WWA days are more likely to occur in association with less intense ARs, which may be related to storms with lower integrated water vapor, cooler temperatures, and lower freezing levels. In other words, more intense ARs may be too warm to produce widespread cold-precipitation-related hazards.
As in Fig. 5a, but for ARs defined at (a) AR1 or AR2 events and (b) AR3, AR4, or AR5 events.
Citation: Weather and Forecasting 36, 3; 10.1175/WAF-D-20-0212.1
As in Fig. 6, but partitioned by (a),(b) HL; (c),(d) CP; and (e),(f) WR WWA types.
Citation: Weather and Forecasting 36, 3; 10.1175/WAF-D-20-0212.1
4. AR perspective
a. Given an AR day, did it occur concurrently with a WWA day?
A majority of cool-season days with landfalling ARs occurred in association with at least one WWA of any type in coastal and mountainous PFZs (Fig. 8a). For example, regional landfalling ARs occurred in association with at least one WWA of any type with maximum frequencies of 40%–70% in PFZs in the Pacific Northwest region, 50%–80% in the in PFZs in the Pacific Central region, and 50%–70% in PFZs in the Pacific Southwest region. The maximum frequencies occurred predominantly in the mountainous PFZs, including those along the Coastal Ranges of Washington and Oregon, the Sierra Nevada throughout California, and the Transverse Mountains and Coastal Ranges of Southern California. Days with landfalling ARs occurred in association with hydrologic WWA days with a frequency of 30%–40% and 20%–30% in the PFZs of the Olympic Peninsula and along the Transverse and Coastal Ranges of Southern California, respectively (Fig. 8b), whereas days with a landfalling AR occurred in association with cold-precipitation WWA days with a frequency of 30%–50% in PFZs in the Cascades and Sierra Nevada and 40%–70% in the PFZs of the Upper Colorado River basin in Utah (Fig. 8c). Finally, days with a landfalling AR occurred in association with wind-related WWAs days with a frequency of 40%–60% in PFZs along the coastal ranges of Washington and Oregon, southern Cascades, Sierra Nevada, and the coastal ranges of Southern California (Fig. 8d). In summary, this analysis demonstrates that while a majority of days with landfalling ARs occur simultaneously with at least some type of WWA, ~50% or less consistently produce hydrologic, cold-precipitation, and wind-related WWAs. In other words, not all ARs are associated with WWAs and an increase in the potential for hazardous weather. Those ARs that do produce hazardous weather are likely to produce different hazards regionally and from event to event.
Frequency of AR days of any intensity at the coast with (a) any, (b) HL, (c) CP, and (d) WR WWA days for each PFZ (thin black lines) the CWAs (thin white lines) of the three defined regions (thick black lines).
Citation: Weather and Forecasting 36, 3; 10.1175/WAF-D-20-0212.1
b. Given an AR1–2 or AR3–5 day, did it occur concurrently with a WWA day?
Cool-season days with a landfalling AR meeting the criteria of an AR1–2 event occurred in association with at least one WWA of any type in PFZs with a frequency (40%–60%) that was similar to all ARs shown above (cf. Figs. 8a and 9a). These values are similarly highest in PFZs on the Olympic Peninsula and along the coastal mountains of Washington and Oregon, in the Sierra Nevada, across the Upper Colorado River basin and the California coastal ranges (Fig. 9a). Alternatively, days with a landfalling AR meeting the criteria of an AR3–5 event occurred in association with at least one WWA of any type in PFZs with a noticeably higher frequency of 60%–80% for Pacific Northwest PFZs on the Olympic Peninsula and in the coastal mountains of Washington and Oregon, 50%–80% for Pacific Central PFZs in the northern Sierra Nevada, and 70%–90% for Pacific Southwest PFZs in the southern Sierra Nevada, Southern California coastal ranges, and the Upper Colorado River basin (Fig. 9b).
As in Fig. 8a, but for ARs defined at (a) AR1 or AR2 events and (b) AR3, AR4, or AR5 events.
Citation: Weather and Forecasting 36, 3; 10.1175/WAF-D-20-0212.1
Cool-season AR1–2 days were infrequently (10%–20%) associated with hydrologic WWAs (Fig. 10a); however, cool-season AR3–5 days were frequently (50%–70%) associated with hydrologic WWAs in PFZs across the coastal Pacific Northwest and Pacific Southwest regions (Fig. 10b). The frequency of cold-precipitation WWAs was 30%–50% on AR1–2 days in PFZs in the northern Cascades and Sierra Nevada and 40%–70% in the Upper Colorado River basin (Fig. 10c); these values increased on AR3–5 days to 40%–60% in PFZs in the Sierra Nevada and 50%–70% in the Upper Colorado River basin (Fig. 10d). AR1–2 days were associated with 30%–50% of wind-related WWAs in PFZs along coastal Washington and Oregon, southern Cascades, Sierra Nevada, and the Southern California coastal ranges (Fig. 10e), while on AR3–5 days the frequencies in the PFZs of each of these regions increased to 50%–80% (Fig. 10f). In summary, the potential for hazardous weather (i.e., a higher likelihood of a WWA) increases across the western United States for more intense and longer-duration ARs as compared to less intense and shorter-duration ARs.
As in Fig. 9, but partitioned by (a),(b) HL; (c),(d) CP; and (e),(f) WR WWA types.
Citation: Weather and Forecasting 36, 3; 10.1175/WAF-D-20-0212.1
c. Given an AR day, what was the spatial footprint of the WWAs?
The sum of the average cumulative area over the entire western United States associated with WWAs on days with landfalling ARs in each region was analyzed in order to assess the potential “hazard footprint” of landfalling ARs (Fig. 11). In this analysis, the hazard footprint associated with WWAs of any type increased from ~160 000 km2 on AR1 days to ~240 000 km2 on AR5 days for landfalling ARs in the Pacific Northwest region and from ~80 000 km2 for AR1 days to ~300 000 km2 on AR5 days for landfalling ARs in the Pacific Central region (Fig. 11a). The hazard footprint increases from ~170 000 km2 on AR1 days to ~375 000 km2 on AR3 days, but decreases to ~250 000 km2 on AR4 days for landfalling ARs in the Pacific Southwest region (Fig. 11a; see additional details in next paragraph). For reference, the areas of the states of Washington, Oregon, and California are ~184 000 km2, ~255 000 km2, and ~424 000 km2, respectively.
(a) Sum of average cumulative area headlined by HL, CP, and WR WWAs on cool-season AR days with landfalling ARs in the Pacific Northwest (red), Pacific Central (black), and Pacific Southwest (blue) regions and the average cumulative area (i.e., hazard footprint) headlined by HL (green), CP (purple), and WR (brown) WWAs on cool-season days with landfalling ARs in the (b) Pacific Northwest, (c) Pacific Central, and (d) Pacific Southwest regions, classified by their intensity according to the Ralph et al. (2019) scale.
Citation: Weather and Forecasting 36, 3; 10.1175/WAF-D-20-0212.1
The hazard footprints associated with hydrologic, cold-precipitation, and wind-related WWAs also vary on days with more intense and longer-duration landfalling ARs in each region (Figs. 11b–d). The hazard footprints of hydrologic and wind-related WWAs increased exponentially and linearly, respectively, as AR intensity and duration increased following the AR scale, with the exception of AR4 days in the Pacific Southwest region (Figs. 11b–d). For example, the hazard footprint for hydrologic WWAs increased exponentially from ~15 000 km2 on AR1 days to ~160 000 km2 on AR5 days in the Pacific Central region. The hazard footprints of cold-precipitation WWAs decreased for landfalling ARs in the Pacific Northwest region, remained relatively constant or decreased in the Pacific Central region, and increased in the Pacific Southwest region as AR intensity and duration increased according to the AR scale (Figs. 11b–d). The decrease in the hazard footprint for cold-precipitation WWAs for landfalling ARs in the Pacific Northwest and Pacific Central regions for more intense and longer-duration ARs may be related to the propensity for more intense ARs to occur in association with higher freezing levels, more rain, and less snow. Alternatively, the increase in the hazard footprint for cold-precipitation WWAs for landfalling ARs in the Pacific Southwest for more intense and longer-duration ARs may be related to locations farther north having already experienced an AR (i.e., for a landfalling AR propagating south along the coast) and those PFZs now residing within the post–cold frontal air mass that may be more likely to produce winter weather–related hazards.
The hazard footprint for hydrologic WWAs also increased as landfalling ARs made landfall in the Pacific Central and Southwest regions where frequencies of landfalling ARs are much lower than in the Pacific Northwest region (Fig. 4). These results are likely related to the climatological southwest–northeast orientation of landfalling ARs (Neiman et al. 2011) that may influence a larger hazard footprint across the western United States as an AR propagates south along the coast and “sweeps” a larger area of the western United States. The hazard footprint for landfalling ARs in the Pacific Central and Pacific Southwest regions may also be larger due to a higher likelihood of or predisposition for flooding due to the drier hydroclimate of California and the southwest United States as compared to the Pacific Northwest, which may be compounded by an increased likelihood of issuance of a flood- or flash flood–related WWA owing to the presence of features such as wildfire burn scars. Additional analyses of individual events, beyond the scope of the current investigation, are needed in order to verify relationships between landfalling ARs of different intensities and durations and hazard footprints for different types of WWAs.
5. Conclusions
The WWA perspective of this study illustrated that a majority of cool-season days with the potential for hazardous weather via WWAs related to hydrologic, wind-related, or cold-precipitation hazards occurred in association with landfalling ARs. The majority of hazardous weather linked to AR days agrees with prior research on AR-related hazards in the western United States. For example, the 60%–90% of hydrologic WWA days that occurred in association with cool-season AR days in the western United States (Fig. 5b) are similar to the 50%–70% of flood-related WWA days associated with ARs in California found by Cordeira et al. (2018). The 60%–90% frequency values also agree with similar research that link ~64% of floods, flash floods, and debris flows over Northern California to ARs (Young et al. 2017), 82% of California Bay Area landslides (Cordeira et al. 2019) to ARs, and the 60%–90% of extreme precipitation events (Oakley et al. 2018) that may produce shallow landslides. Similarly, the 40%–60% frequency of cold-precipitation WWA days that occurred in association with AR days in the high elevation regions of the Cascades, Olympic Peninsula, and Sierra Nevada (Fig. 5c) are similar to the 40%–60% of snowfall in the western United States attributed to ARs by Rutz et al. (2014; their Fig. 8b) and the 60%–80% frequencies of winter weather–related WWAs associated with ARs in California by Cordeira et al. (2018). The 50%–90% frequency of wind-related WWAs days that occurred in association with cool-season AR days in the Pacific Northwest and Pacific Central regions (Fig. 5d) also agrees with the 40%–75% of extreme wind events along 40% of the worlds coastlines that occur with landfalling ARs (Waliser and Guan 2017). These results confirm our first hypothesis that a majority of days with the potential for hazardous weather via WWAs across the western United States occur in association with landfalling ARs.
The potential for hazardous weather via WWAs in this study also increased for more intense and longer-duration ARs, confirming the second hypothesis of this study. Cool-season days with WWAs of any type were more likely to be associated with AR3–5 days in the Pacific Northwest and Pacific Central regions (50%–70%), while WWAs were more likely to be associated with less intense AR1–2 days in the Pacific Southwest region (70%–80%) (Figs. 6a,b). These results were similar for hydrologic and wind-related WWA days in the Pacific Northwest and Pacific Central regions with higher frequencies for AR3–5 days (~60%–100%) as compared to AR1–2 days (Figs. 7a,b,e,f). Alternatively, cold-precipitation WWA days were more frequently associated with less intense AR1–2 days across a majority of all three regions (Fig. 7c). This result is most likely related to warmer temperatures and higher freezing levels associated with more intense ARs (see Fig. 9 of Ralph et al. 2019). The warmer temperatures could lead to less cold-precipitation related hazards on AR3–5 days in a majority of lower-elevation PFZs (Fig. 7d), but additional analyses of freezing levels and precipitation type among landfalling ARs with different intensities and durations is needed.
The AR perspective in this study demonstrated that not all ARs result in a large potential for hazardous weather requiring or necessitating a WWA. This result is similar to past studies that illustrate, for example, that not all ARs produce flooding (e.g., Ralph et al. 2019) and not all ARs lead to landslides (e.g., Cordeira et al. 2019). In this study, ~40%–70% of cool-season AR days occurred in association with hydrologic, wind-related, or cold-precipitation WWAs in regions of complex topography, while <10%–30% of cool-season AR days occurred in association with the aforementioned WWAs in PFZs in regions without complex topography (Fig. 8). This result confirms our third hypothesis that not all landfalling ARs necessarily produce a potential for hazardous weather illustrated by WWAs.
Analysis of the relationships among AR intensity, duration, and WWA frequency demonstrated that more intense and longer-duration ARs are more frequently associated with an increase in the potential for hazardous weather via WWAs. For example, landfalling ARs characterized as an AR3–5 event according to the Ralph et al. (2019) scale occurred in association with WWAs of any type with a frequency of ~60%–90% whereas AR1–2 events contained a frequency of ~40%–60% (Figs. 9a,b). Landfalling ARs produced the highest frequencies of WWAs on AR3–5 days in PFZs located in the high elevation regions of the Olympic Peninsula, northern Cascades, Sierra Nevada, the Coastal Ranges of Southern California, and Upper Colorado River basin. These results confirmed the fourth hypothesis that more intense and longer-duration ARs are potentially more hazardous.
The results of this study motivated an examination of the spatial extent of WWAs on AR days, which was herein referred to as the potential AR hazard footprint. The hazard footprints of landfalling ARs related to hydrologic and wind-related WWAs both increased as AR intensity and duration increased from an AR1 to AR5 (Fig. 11); the former occurred in association with an exponential increase in spatial area from ~15 000 km2 on AR1 days to ~160 000 km2 on AR5 days for Pacific Central region ARs. Alternatively, the hazard footprint of landfalling ARs related to cold-precipitation WWAs decreased as AR intensity and duration increased, likely related to higher freezing levels in these warmer ARs. It should be noted in this analysis that the hazard footprint of WWAs 1) contains a fixed upper limit in spatial area (i.e., ultimately the size of the western United States), 2) is limited to or influenced by the spatial area physically susceptible to a given hazard (e.g., spatially varying climates as a function of orography, latitude, proximity to the coast, etc.), and 3) is limited to predefined regionally varying hazard thresholds (e.g., a 2-in. snowfall may still be hazardous, yet not result in a WWA). Nonetheless, there were large and at least quasi-exponential increases in the hazard footprints associated with increasingly more intense and longer-duration ARs. The exponential increase in the hazard footprint of landfalling ARs related to hydrologic WWAs is similar to the exponential increase in dollar-based flood damages associated with increasingly more intense and longer-duration ARs using the AR scale by Corringham et al. (2019). Together, this study and the Corringham et al. (2019) study provide quantitative evidence to the subjective benefit–hazard distribution offered by Ralph et al. (2019) that is summarized by Fig. 12.
A generalized schematic illustration of the exponential increase in the potential for hazardous weather in the western United States, in this study defined by the “hazard footprint” of WWAs and in Corringham et al. (2019) defined by flood damages, associated with the Ralph et al. (2019) AR scale. Text annotation embedded within the schematic is intended to support interpretation, including the influence of antecedent conditions in modulating the potential for hazardous weather.
Citation: Weather and Forecasting 36, 3; 10.1175/WAF-D-20-0212.1
Future work aims to further investigate the spatial and temporal relationships between hazards described by WWAs and ARs, including how the IVT magnitude and direction of a landfalling AR, landfall location proximal to terrain gaps, freezing level, and aforementioned antecedent conditions may influence the spatial frequencies of hazards across the western United States. This study also grouped AR landfalls by region to simplify the illustration of the linkages among potentially hazardous weather, WWAs, and ARs. Prior work by Bartlett (2020) demonstrated that the WWA perspective could identify the likelihood of an AR at the PFZ location in lieu of the likelihood of an AR at the coast. Bartlett (2020) also demonstrated that the AR perspective could identify the likelihood of a WWA in any PFZ for any given combination of AR location/intensity/duration, whether at the coast or collocated with the PFZ. Additional work aims to 1) develop an analog tool from the AR perspective that allows forecasters to see the historical likelihood and distribution of WWAs at PFZs across the western United States for given combinations of AR intensity, duration, and landfall location, and 2) incorporate antecedent conditions and numerical weather prediction forecast uncertainty to better inform the analog-derived real-time potential for hazardous weather and WWAs.
Acknowledgments
Support for this project was primarily provided by awards from the State of California, Department of Water Resources (4600013361) and the U.S. Army Corps of Engineers (W912HZ-15-2-0019, W912HZ-19-2-0023) as part of broader projects led by the Center for Western Weather and Water Extremes (CW3E) at the University of California, San Diego Scripps Institution of Oceanography. A portion of this project was also completed while the first author held a graduate assistantship sponsored by Plymouth State University. We greatly acknowledge feedback by Dr. Jonathan Rutz (National Weather Service Western Region Headquarters), Dr. Eric Hoffman (Plymouth State University), and two anonymous reviewers that improved the quality of this research.
Data availability statement
Data analyzed in this study were a reanalysis and derivation of existing data, which are openly available at locations cited in the data and methods section.
REFERENCES
American Meteorological Society, 2018: Atmospheric river. Glossary of Meteorology, http://glossary.ametsoc.org/wiki/Atmospheric_river.
Bartlett, S., 2020: A climatological study of National Weather Service watches, warnings, and advisories in association with atmospheric rivers in the western U.S. 2006–2018. M.S. thesis, Meteorology Program, Plymouth State University, 41 pp.
Cordeira, J. M., M. M. Neureuter, and L. D. Kelleher, 2018: Atmospheric rivers and National Weather Service watches, warnings, and advisories issued over California 2007–2016. J. Oper. Meteor., 6, 87–94, https://doi.org/10.15191/nwajom.2018.0608.
Cordeira, J. M., J. Stock, M. D. Dettinger, A. M. Young, J. F. Kalansky, and F. M. Ralph, 2019: A 142-year climatology of Northern California landslides and atmospheric rivers. Bull. Amer. Meteor. Soc., 100, 1499–1509, https://doi.org/10.1175/BAMS-D-18-0158.1.
Corringham, T. W., F. M. Ralph, A. Gershunov, D. R. Cayan, and C. A. Talbot, 2019: Atmospheric rivers drive flood damages in the western United States. Sci. Adv., 5, eaax4631, https://doi.org/10.1126/sciadv.aax4631.
Dettinger, M. D., 2013: Atmospheric rivers as drought busters on the U.S. West Coast. J. Hydrometeor., 14, 1721–1732, https://doi.org/10.1175/JHM-D-13-02.1.
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, 445–478, https://doi.org/10.3390/w3020445.
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, https://doi.org/10.1029/2010GL044696.
Hatchett, B. J., and Coauthors, 2020: Observations of an extreme atmospheric river storm with a diverse sensor network. Earth Space Sci., 6, e2020EA001129, https://doi.org/10.1029/2020EA001129.
Iowa State University, 2019: Archived NWS watch/warnings. Iowa Environmental Mesonet, accessed 18 July 2019, https://mesonet.agron.iastate.edu/request/gis/watchwarn.phtml.
National Weather Service, 2020: National Weather Service Instruction 10-513: WFO winter weather products specification. NOAA/NWS, 51 pp., https://www.nws.noaa.gov/directives/sym/pd01005013curr.pdf.
Neiman, P. J., F. M. Ralph, G. A. Wick, J. D. Lundquist, and M. D. Dettinger, 2008: Meteorological characteristics and overland precipitation impacts of atmospheric rivers affecting the West Coast of North America based on eight years of SSM/I satellite observations. J. Hydrometeor., 9, 22–47, https://doi.org/10.1175/2007JHM855.1.
Neiman, P. J., L. J. Schick, F. M. Ralph, M. Hughes, and G. A. Wick, 2011: Flooding in western Washington: The connection to atmospheric rivers. J. Hydrometeor., 12, 1337–1358, https://doi.org/10.1175/2011JHM1358.1.
Oakley, N. S., J. T. Lancaster, M. L. Kaplan, and F. M. Ralph, 2017: Synoptic conditions associated with cool season post-fire debris flows in the Transverse Ranges of southern California. Nat. Hazards, 88, 327–354, https://doi.org/10.1007/s11069-017-2867-6.
Oakley, N. S., J. T. Lancaster, B. J. Hatchett, J. Stock, F. M. Ralph, S. Roj, and S. Lukashov, 2018: A 22-year climatology of cool-season hourly precipitation thresholds conducive to shallow landslides in California. Earth Interact., 22, https://doi.org/10.1175/EI-D-17-0029.1.
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, https://doi.org/10.1029/2006GL026689.
Ralph, F. M., T. Coleman, P. J. Neiman, R. 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, 443–459, https://doi.org/10.1175/JHM-D-12-076.1.
Ralph, F. M., J. J. Rutz, J. M. Cordeira, M. D. 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, 269–289, https://doi.org/10.1175/BAMS-D-18-0023.1.
Rienecker, M. M., and Coauthors, 2011: MERRA: NASA’s Modern-Era Retrospective Analysis for Research and Applications. J. Climate, 24, 3624–3648, https://doi.org/10.1175/JCLI-D-11-00015.1.
Rutz, J. J., W. J. Steenburgh, and F. M. Ralph, 2014: Climatological characteristics of atmospheric rivers and their inland penetration over the western United States. Mon. Wea. Rev., 142, 905–921, https://doi.org/10.1175/MWR-D-13-00168.1.
Waliser, D., and B. Guan, 2017: Extreme winds and precipitation during landfall of atmospheric rivers. Nat. Geosci., 10, 179–183, https://doi.org/10.1038/ngeo2894.
Young, A. M., K. T. Skelly, and J. M. Cordeira, 2017: High-impact hydrologic events and atmospheric rivers in California: An investigation using the NCEI Storm Events Database. Geophys. Res. Lett., 44, 3393–3401, https://doi.org/10.1002/2017GL073077.
