1. Introduction
Atmospheric rivers (ARs) are long and narrow pathways of enhanced moisture transport, predominantly originating in the tropics (American Meteorological Society 2017). They carry the majority of the poleward moisture flux across midlatitude regions, despite only occupying approximately 10% of the total longitudinal length at these latitudes (Zhu and Newell 1998). When making landfall, ARs are typically associated with heavy precipitation (Lamjiri et al. 2017; Leung and Qian 2009; Ralph and Dettinger 2012) and/or strong surface winds (Waliser and Guan 2017). Flood/drought conditions and water resources in the U.S. West Coast are highly dependent on the frequency and intensity of incoming ARs from the Pacific Ocean (Dettinger 2011, 2013; Guan et al. 2010, 2012, 2013; Neiman et al. 2011; Paltan et al. 2017; Ralph et al. 2006). Regional studies outside the United States continue to establish the hydrological importance of ARs in midlatitude regions such as in Europe (Lavers and Villarini 2013; Lavers et al. 2012; Ramos et al. 2015), South America (Viale and Nuñez 2011; Viale et al. 2018), and New Zealand (Kingston et al. 2016). The importance of AR-related integrated vapor transport (IVT) is highlighted in (Ralph et al. 2019), where an AR scale is determined to predict the beneficial and harmful impacts of ARs based on maximum IVT and the event duration. While the impact of ARs on drought, flooding, and water resources has been widely studied, especially in the western United States, extreme winds associated with ARs have been relatively less studied. These extreme wind speed AR events may come with coastal storm surges (Khouakhi and Villarini 2016) and wind hazards inland. The study by Waliser and Guan (2017) examined AR winds and showed that, from 1997 to 2013, 14 of the 19 large European wind storms causing more than 1 billion U.S. dollars in insurance losses were associated with ARs. In addition, for many coastal locations around the globe with frequent landfalling ARs, such as the U.S. West Coast, southwestern South America (SA), western Europe (EU), and New Zealand (NZ), the highest or second-highest rank-order surface wind extreme during the study period was associated with an AR [Fig. 1a, reproduced from Waliser and Guan (2017)].
(a) Speed (size of circles) and rank (shading) of AR-related wind extremes from Waliser and Guan (2017). The U.S. West Coast (USA), southwestern South America (SA), western Europe (EU), and New Zealand (NZ) experience either the highest rank or second highest rank of wind extremes during AR events as based on the 6-hourly ECMWF ERA-Interim analysis from 1997 to 2010. (b) A visual interpretation of our hypothesis. For a given IVT of a landfalling AR (thick blue arrow), extreme surface winds are more likely to occur when vertical stability is weakened near the surface. Under stable conditions (left), the wind profile and speed are represented by the curved black line and black arrow, respectively. Near-surface stability is shown in temperature, with a vertical gradient line for stable conditions (from orange to blue going downward) indicating a relatively warmer atmosphere over a colder skin surface temperature and unstable conditions (from blue to orange) indicating a colder atmosphere over a warmer skin surface temperature. Under conditions of high vertical instability (right) the surface winds are enhanced (red arrow) as a result of increased vertical mixing from higher levels (red curved line).
Citation: Journal of Hydrometeorology 22, 6; 10.1175/JHM-D-20-0165.1
The connections between ARs, extreme IVT, and extreme surface winds, as discussed above, motivate a further study of their relationships. In particular, it remains to be understood why some ARs are more wind-driven than others, which has implications to the understanding and preparation for a full range of AR impacts beyond those directly associated with precipitation, including the potential occurrence of compound events (i.e., those involving more than one type of hazard) associated with ARs (Ridder et al. 2018; Rondanelli et al. 2019). One explanation is that near-surface atmospheric stability plays a role in the modulation of oceanic surface wind speed. Depending on the vertical structure of temperature and moisture, turbulence may be enhanced as a response to unstable conditions within a flow (Stull 2006). Cobb et al. (2021) used dropsondes to investigate the variability of atmospheric stability within ARs over the open ocean; however, to date the influence of stability on surface wind speed in landfalling ARs has yet to be quantified. Although not related to ARs, studies have indicated that atmospheric stability plays a role in modulating surface wind speed (Archer et al. 2016; Barthelmie 1999).
In this study, we investigate if near-surface vertical thermodynamic stability over the ocean near the landfall location plays a role in the magnitude of surface winds during an AR event. Specifically, we test the hypothesis that extreme surface winds associated with ARs are more likely to occur under conditions of weaker vertical stability in the atmospheric layer between the surface and the typical altitude of the AR low-level jet core (~1 km) (Ralph et al. 2004). Under weaker vertical stability conditions, the temperature difference between the surface and the low-level jet core will increase leading to air heating from lower levels upward, which will result in turbulence and mixing. Figure 1b is a visual representation of the hypothesis. The blue arrows represent a given IVT value, and the curved black lines represent the two different associated wind speed profiles. The near to vertical gradient-colored line represents the temperature gradient between the low-level wind speed jet at approximately 1 km and the surface (warmer and colder temperatures in orange and blue, respectively), while the color of the parallelogram represents the relative skin surface temperature of the coastal ocean. It is hypothesized that under conditions of weaker vertical stability, mixing of high momentum air from above would more readily occur and lead to stronger surface winds as shown in Fig. 1b.
To test the above hypothesis, the examination and quantification of the relationships between ARs, surface wind speed, and near-surface vertical thermodynamic static stability are examined. We will focus on the four west coast regions of the United States, SA, EU, and NZ that have strong and frequent landfalling ARs that can directly impact the wellbeing of their communities as well as having high-ranking orders of surface wind extremes associated with ARs. This paper is organized as follows. In section 2, we describe the data and methods used in this study, including a brief description of the AR detection procedure, relevant reanalysis and satellite fields, and the linear regression models used to quantify the relationship between AR IVT, surface wind speed, and near-surface stability. In section 3, we present the associated results, including an examination of the veracity of the results that is based on comparisons between reanalysis and satellite observations. Last AR cases within the top and bottom 25% of near-surface stability are examined in the context of understanding the associated frequency of occurrence of extreme wind speeds in ARs. We conclude in section 4 with a summary and discussion.
2. Data and methods
a. AR database
In this study, the Guan and Waliser (2015) global AR database is used. The database has been comprehensively evaluated (Guan and Waliser 2015; Guan et al. 2018), has been compared with other detection algorithms (Ralph et al. 2019; Shields et al. 2018), and was publicly available (https://ucla.box.com/ARcatalog) at the time of this writing. The temporal range of the AR dataset used in this analysis is 15 years (1999–2014) for the reanalysis portion of this study and from 2002 to 2009 for the satellite portion. The reason for the temporal range of the satellite portion will be explained in section 2b. For AR detection, 6-hourly IVT intensity at 0.5 × 0.625 resolution is calculated using specific humidity and wind fields at 17 pressure levels between 1000 and 300 hPa provided by the Modern-Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2) (Gelaro et al. 2017). The first step in the detection of ARs is identifying the grid cells where the IVT intensity exceeds the 85th percentile specific to each season and grid cell but is not lower than a fixed limit of 100 kg m−1 s−1, resulting in a set of “objects,” that is, contiguous areas, where the IVT intensity threshold is met. Then, geometric and IVT direction requirements on the IVT objects are implemented such as length > 2000 km, length-to-width ratio > 2, and mean poleward IVT component > 50 kg m−1 s−1, resulting in a set of detected ARs. Last, when an AR object intersects the coastline and the mean IVT is directed from the ocean toward the coast, the coastal grid cell with the largest onshore IVT is designated as the landfall location of the AR object. An example of AR objects from the global AR detection algorithm for 0000 UTC 18 January 2012 is shown in Fig. 2a. Variables provided by the AR database and used in the current study include the AR’s shape, zonal and meridional IVT, and landfall location.
(a) Atmospheric river objects extracted from the Guan and Waliser (2015) detection algorithm at 0000 UTC on 18 Jan 2012. The red-outlined box corresponds to the area shown in (b). (b) A visual representation of a landfalling AR and the extraction process for the surface and meteorological parameters used in our study. The red dot is the landfall location, and the red grid cells experience AR conditions offshore and are averaged to represent the mean value of the parameters of interest for a given landfalling AR.
Citation: Journal of Hydrometeorology 22, 6; 10.1175/JHM-D-20-0165.1
b. Surface winds and near-surface static stability indicators
The primary data used for surface winds and the low-level vertical thermodynamic stability over the ocean under AR conditions are provided by instantaneous MERRA-2 fields and subsampled temporally when needed to match the 6-hourly IVT AR data. For surface winds, the 10 m fields are used. The main metric of near-surface thermodynamic static stability used in our study is the delta T method, which is similar in calculation to a two-point lapse rate and known here after as ΔT (Mohan and Siddiqui 1998; Wallace and Hobbs 2006). We calculate ΔT as the 1-km temperature minus the skin surface temperature with the 1-km temperature derived from temperature and geopotential height using linear interpolation. Note that 1 km is approximately the average altitude of the low-level jet in an AR, as mentioned earlier. We use the skin surface temperature rather than the near-surface air temperature because it has been found to be more accurate over the ocean (Luo et al. 2020). The bulk Richardson number (hereinafter RB), an alternate measure of near-surface stability, is employed to supplement the results using ΔT. Using the above products provided from MERRA-2, we examine how AR intensity (in terms of IVT) and near-surface static stability (represented here as low-level ΔT) influence the strength of the ocean surface winds of landfalling ARs.
Because the MERRA-2 reanalysis fields discussed above are partly dependent on the underlying model and its physical parameterizations, we examine the fidelity of the analysis and findings with satellite observations. To do so, we utilize QuikSCAT ocean surface wind speed and AIRS near-surface static stability to quantify whether the reanalysis-based results are unduly sensitive to observational data sources. It should be known that the overlap in AIRS and QuikSCAT data record is from 2002 to 2009, shorter than the temporal period analyzed with the MERRA-2 data. Despite this difference, we believe that the length of the satellite record is sufficient to ensure that our findings are not solely replicating a model-based relationship in the reanalysis fields.
The QuikSCAT spacecraft carried the Sea Winds scatterometer, a dual-beam microwave radar utilizing ocean backscatter to determine wind speed and direction over the ocean (Tsai et al. 2000). Due to its wide swath, it covers 90% of the world’s ocean in 24 h. The spatial resolution of its ellipsoid footprint is 25 × 37 km2 and its data record extends from July 1999 to November 2009. Although the accuracy of QuikSCAT’s wind measurement decreases nearshore, it is still sufficient for coastal studies of the type considered here (i.e., approximately hundreds of kilometers offshore) and has been validated against buoy data off the U.S. West Coast (Pickett et al. 2003). As a scatterometer, the instrument can obtain a measurement in any cloud condition over the ocean (Callahan and Lungu 2006). The level 3 daily gridded product with a spatial resolution of 0.25° × 0.25° from October 2002 to November 2009 is used. Unusable data resulting from rain or contamination on one or more of the four beams needed to obtain a measurement are discarded in accordance with the rain flag matrix (Callahan and Lungu 2006). The remaining usable data are spatially binned to match the AR reanalysis grid by using the centroid of each QuikSCAT pixel. In addition, the data are temporally matched up to the AR grid by consulting the timestamp in the level-3 product and finding the closest time field within the 6-hourly AR data.
The Atmospheric Infrared Sounder (AIRS) and Advanced Microwave Sounding Unit (AMSU) are infrared and microwave instruments aboard the Aqua spacecraft that provide temperature profile measurements with a data record starting in October 2002 and are both still operating at the time of writing this paper (Chahine et al. 2006). For more information on how these two instruments work together to obtain a temperature profile please refer to (Susskind et al. 2010). The AIRS instrument has a spatial resolution of 13.5 km at nadir (scan angle = 0) and has a swath of 1650 km, allowing for near-global daily coverage. AIRS temperature profiles have been compared with radiosonde profiles (Divakarla et al. 2006), and the uncertainties and validation associated with the current level-2, version-6 product’s cloud-induced retrievals have been documented (Wong et al. 2015). AIRS temperature profiles are used in this study as an alternative source of information on near-surface stability, namely, using the AIRS 1-km temperature in combination with reanalysis surface skin temperature in the construction of ΔT. Note that AIRS surface air temperatures are the most impacted by cloud induced uncertainties (Wong et al. 2015). We use the MERRA-2 skin surface temperature as it utilizes a near-surface ocean diurnal layer model in combination with in situ buoys, surface-sensitive radiances, and other observations in the atmospheric assimilation system (Akella et al. 2017; Gelaro et al. 2017). To obtain the 1-km temperature, the level-2, version-6 (AIRS + AMSU) standard product with a spatial resolution of approximately 45 km is used to extract a temperature profile and geopotential height, which in turn, through interpolation, yields an observed estimate of the 1-km temperature. For the current analysis, AIRS measurements were selected in accordance with the quality flags in the level-2 product, keeping values of “1” (good) and “0” (best). After all adequate retrievals were selected, the centroid of each footprint is spatially binned and temporally matched up to the AR spatiotemporal grid.
c. Extraction of AR-related surface meteorological parameters
Winter ARs making landfall during 1999–2014 within the months of November–March for the Northern Hemisphere regions (North America and Europe) and May–September for the Southern Hemisphere (South America and New Zealand) are examined. We first need to extract the meteorological parameters of interest for each landfalling AR within the region of interest. These include the 1-km temperature, surface skin temperature, as well as the zonal and meridional components of the surface wind and IVT. To do this, we isolate the landfall location identified by the Guan and Waliser (2015) AR catalog and apply a land mask to set meteorological parameters outside the oceanic fraction of the AR shape to be missing. To dampen out variability and decrease noise, adjacent grid cells are spatially averaged to reduce the uncertainty in the meteorological parameters used as inputs into the regression. For each meteorological parameter, the nonmissing (oceanic) values over a block of five grid cells by five grid cells within the AR and centered at the landfall location are averaged to represent the value of the parameter near (i.e., immediately offshore from) the landfall location. Figure 2b serves as a visual representation of the above process. The red dot is marked by the AR detection algorithm as the landfall location, while the cyan dots indicate AR conditions. The red outlined grid cells indicate AR conditions offshore and are averaged together to represent the mean value of the parameter of interest. The white grid cells either do not experience AR conditions or are over land and are not used.
When performing sensitivity analysis based on satellite data, the same approach is used but one additional grid cell zonally and meridionally is considered. The purpose of the larger grid is to increase the number of matchups between the landfalling AR, AIRS, and QuikSCAT. In addition, for a grid cell to be included in the averaging, it must have good measurements from both QuikSCAT and AIRS. Fewer ARs are analyzed using satellite data because of the more limited sampling relative to the reanalysis data, resulting from a smaller temporal range of the data due to satellite lifetime, a field of view limitation due to instrument orbit, and quality control discarding unusable measurements.
d. Linear regression models and sensitivity tests
3. Results
a. Basic relationship between ARs, surface wind, and near-surface stability
As mentioned in the introduction, the regions of the western United States, southwestern SA, western EU, and western NZ are chosen for further investigation because these locations have frequent and strong landfalling ARs. Figure 3 shows the AR counts (i.e., the number of landfalling ARs using the landfall location and 6-h time steps) for the averaged offshore grid cells of landfalling ARs for all four study regions. It can be observed that each region to be analyzed has over 1000 landfalling AR events summed across all grid points. The variability in the landfall count is partially related to the coastal geometry as well as potential topographic features inland that may impact particular grid cells to have stronger IVT flows within a given coastal geometry. An alternative version of Fig. 3 using the U.S. West Coast as an example (Fig. S1 in the online supplemental material) shows the impact of coastal grid cell geometry where samplings at each latitude are summed with coastal grid cells overlaid.
Counts of landfalling ARs for the four areas investigated [(a) USA, (b) EU, (c) SA, and (d) NZ], utilizing the full MERRA-2 AR record analyzed in this study: 1999–2014 (see section 2).
Citation: Journal of Hydrometeorology 22, 6; 10.1175/JHM-D-20-0165.1
The extracted average meteorological and surface parameters from section 2c are further examined in Fig. 4 (U.S. West Coast) and Figs. S2–S4 in the online supplemental material (SA, EU, and NZ, respectively). In the U.S. West Coast, western SA, and western EU there is a trend of higher regional wind speeds at higher latitudes, as well as colder 1-km and skin surface temperatures. The relationship coincides with the assessment in Ralph et al. (2017) that ARs in the midlatitudes have larger winds but smaller integrated water vapor (IWV) relative to ARs that are closer to the tropics, because the approximate temperature of the AR-conducive low-level jet is slightly colder at higher latitudes. IVT has the largest regional variability of all considered parameters; however, western boundary regions with locations that are fairly close to the tropics tend to have lower IVT values.
The spatial distribution of offshore grid cells and the associated average of the geophysical variables used in the analysis for the U.S. West Coast region: (a) wind speed, (b) IVT, (c) 1-km temperature, and (d) oceanic skin surface temperature.
Citation: Journal of Hydrometeorology 22, 6; 10.1175/JHM-D-20-0165.1
Using the U.S. West Coast as a demonstration, a fundamental result from this study is shown in Fig. 5a. Here, AR surface wind speed is plotted against AR IVT, with the near-surface stability (ΔT) measure for each landfalling AR plotted as the third/colored dimension. As can be expected, there is a linear relationship between surface winds and IVT (R2 = 0.38). A further breakdown of the correlation and contribution of wind speed and specific humidity per layer of atmosphere to the overall IVT can be found in Fig. S5 in the online supplemental material. The majority of the contribution and the largest correlation do not come from near-surface wind, but rather from the upper (100–400 hPa) and lower to midlevels (700–850 hPa), respectively. Based on the results, it is suggested that although physically related, the surface wind speed may not be representative of the wind speed profile (300–1000 hPa) contributing to the IVT calculation. Apart from the linear surface wind–IVT relationship, it is also evident in Fig. 5a that there is variation following ΔT, with low near-surface stability conditions (red dots) having higher surface wind speeds for a given IVT than the high near-surface stability conditions (blue dots). It is also apparent along the extremities of ΔT that high IVT ARs tend to have a more stable lower atmosphere than low IVT ARs. Our hypothesis is that extreme surface winds associated with ARs are more likely to occur in unstable near-surface stability conditions in the region between the AR jet (~1 km) and the surface, which can result in turbulent mixing and higher surface wind speeds. At least qualitatively, the relationship in Fig. 5a supports the study’s hypothesis.
Wind speed magnitude and IVT vs stability metric, where reddish tones represent an unstable lower atmosphere and bluish tones represent a stable lower atmosphere as reported by (a) ΔT, (b) RB, (c) Δθυ, and (d) wind shear magnitude. All panels are for the U.S. West Coast region using MERRA-2 data.
Citation: Journal of Hydrometeorology 22, 6; 10.1175/JHM-D-20-0165.1
To further quantify this relationship, the approach outlined in section 2d is taken. Namely, we first use a simple linear regression model to quantify the variance in surface wind speed by IVT [outlined in Eq. (1)] and then use a bivariate linear regression model [outlined in Eq. (2)] to determine the explained variance due to the combination of IVT and near-surface stability. From the difference in the two R2 values we determine the value of incorporating near-surface stability into the regression and help to infer its physical importance to AR-related surface wind extremes. The procedure is followed separately for each of the four regions (U.S. West Coast, EU, NZ, and SA) and is displayed in Fig. 6a. For the regions considered, IVT alone accounts for about 22%–38% (depending on the region) of the explained variance in surface wind speed. The combination of IVT and ΔT together account for about 36%–52% (depending on the region) of the surface wind speed variation. Of the four regions examined, the regression model fits the observations the best for the U.S. West Coast (52%) due to the large initial correlation with IVT (38%). In terms of total quantifiable variance in surface wind speed, the incorporation of near-surface stability improves the regression model by about 13%–16% and has a relative increase of about 35%–70% from IVT alone depending on the region.
Coefficient of determination from linear regression models: (a) both IVT and ∆T together and IVT alone for the U.S. West Coast, southwestern South America, western Europe, and western New Zealand and (b) sensitivity tests for the U.S. West Coast relationship (see Fig. 4a), when satellite observations from AIRS are used for 1-km temperature and those from QuikSCAT are used for ocean surface wind speed. The number of landfalling ARs, and thus points used in the regression, are also shown.
Citation: Journal of Hydrometeorology 22, 6; 10.1175/JHM-D-20-0165.1
b. Sensitivity analysis using an alternate stability measure
The bulk Richardson number as defined in section 2d is used as an additional stability indicator to compare with the results obtained using ΔT. Figure 5b shows the resulting relationship between the ocean surface wind, IVT, and RB for the western U.S. region. Low and negative RB values represent dynamic instability, while larger values represent a more stable flow. RB tends to decrease as surface wind speed increases, with the lowest RB associated with high wind speeds and lower IVT, similarly to ΔT in those cases. Components of RB, namely, the virtual potential temperature difference Δθυ and the wind shear magnitude were plotted against IVT and wind speed in Figs. 5c and 5d, respectively. The distribution of surface wind speeds corresponding to IVT and virtual potential temperature difference has similar behavior to that of the original, simple measurement of stability, ΔT. Wind shear generally increases with increasing IVT and surface wind speed, and is lowest at low wind speeds. The linear regression model shown in Eq. (3) is constructed to determine the explained variance quantified by the RB. For the RB case, the combined linear regression model produces a coefficient of determination of 0.47, just a bit less than that using ΔT for vertical stability (i.e., 0.52). In addition, Δθυ was also accessed as a stability parameter and produces an explained variance of 0.51. All three stability metrics support the robustness of the underlying relationship.
c. Sensitivity analysis using satellite data
To ensure that the “observed” relationship identified in the analysis in section 3a is not solely replicating a model-based relationship in the reanalysis fields, the results are reconstructed using a combination of AIRS near-surface stability and QuikSCAT surface wind observations in place of the corresponding reanalysis fields, as described in section 2. The near-surface stability indicator ΔT and the surface wind are substituted individually, then also together, in Eq. (1) while the AR IVT values remain from MERRA-2. Using the temporal range corresponding to the overlap in data record (2002–09), QuikSCAT and AIRS observed and obtained valid measurements in accordance with quality control parameters described in section 2b 455 and 384 times, respectively, with 258 overlapping cases. Figure 6b shows the coefficient of determination from these sensitivity tests for the U.S. region as well as the number of ARs observed per data source(s). Substituting both satellite variables into the linear regression gives an R2 value of 0.44 for IVT and ΔT combined, and 0.33 for IVT alone (i.e., cf. IVT: 0.38 and combined: 0.52 using all MERRA-2). As can be expected, the relationship deteriorates a bit, due to the smaller sample size as well as the combination of respective uncertainties in both satellite variables and the reanalysis, all of which can hinder the performance of the regression. However, overall the qualitative nature of the relationship is still evident from this sensitivity test, namely, that IVT and near-surface stability each represent an important source of information for surface winds associated with landfalling ARs. It should also be stated for clarity that select spectral channels of AIRS clear sky observations are radiance assimilated into MERRA-2 (McCarty et al. 2016) but use a different algorithm for producing temperature and water vapor profiles than used in the AIRS retrievals. In our analysis, AIRS 1-km temperature, however, had a mean cold bias of 1.57 K during AR observations when compared with MERRA-2, although by definition the mean bias does not affect the linear model’s explained variance.
d. Considerations of ARs with the most stable and unstable near surface conditions
The results of the linear regression can be further resolved to examine the expected wind speeds and the impacts associated with the most unstable ARs. Using the MERRA-2 data, we first start with the U.S. West Coast region and compare AR cases within the top and bottom 25% of near-surface stability (simply referred to hereinafter as stable and unstable ARs) over a weak (150–600 kg m−1 s−1) and a strong (600–1250 kg m−1 s−1) IVT range. Breaking down the sampling into two IVT ranges helps further to isolate the impact of near-surface instability because the wind speed profile is expected to affect the column-integrated IVT. Over the weak IVT range, we have 1372 ARs with 343 in each percentile range, while over the strong IVT range there are 537 ARs with 134 in each percentile range. The mean IVT profile with the resulting mean wind speed over this region is shown in Fig. 7. The result is consistent with our original hypothesis (see Fig. 1b), showing that although stable ARs tend to have generally higher IVT values, unstable ARs have higher surface wind speeds regardless of IVT values. Specifically, unstable ARs have 26.5% higher wind speeds for the lower IVT range and 12.5% for the higher IVT range relative to their stable counterparts. It can also be seen that the role instability plays in modifying surface wind speed in this region may be greater in lower IVT ARs; however, it is still very much present in high IVT ARs as well.
IVT profile and surface wind speed as based on the top 25th percentile of stable and unstable ARs for weak and strong IVT ARs. The red tones represent unstable ARs, and the blue tones represent stable ARs. The x axis is the IVT at each pressure-level range, and the red and blue numbers are the mean surface wind speeds for each subset of ARs.
Citation: Journal of Hydrometeorology 22, 6; 10.1175/JHM-D-20-0165.1
To better understand the distribution of surface wind speeds in stable and unstable ARs, the 25% most stable and 25% most unstable of the cases over all 4 regions are extracted and binned according to surface wind speed. This is again done for the same two IVT ranges. To be more easily associated with the impacts of different wind speed ranges, the BWS is chosen as the binning range. The frequency distribution plots for both IVT ranges can be shown in Figs. 8a and 8b. It can be observed that over these four regions, the peak of the distribution in wind speed is shifted by approximately 1 BWS number when going from high to low stability. In addition, there is an additional shift of 1 BWS increment between the low and high IVT ranges. Although there are slight geographic variations among the regional distributions (Fig. S6 in the online supplemental material), extreme winds resulting from ARs such as those of gale force or above when all 4 regions considered together are more likely to be present in unstable ARs in the higher IVT range (14.7% probability of occurrence) but are also present in the lower IVT range (6.15%). They are also present in the high IVT range stable ARs due to the correlation with IVT (5.3%), but rarely occur in stable ARs with an IVT under 600 kg m−1 s−1 (0.58%).
Histogram of surface wind speed for the top 25th percentile of all regions for unstable and stable ARs binned by BWS number. The impact of stability can be seen in the distribution of wind speed. It is clear that for each IVT range the unstable ARs score higher on the BWS than do stable ARs. In addition, the higher IVT range tends to score higher on the BWS than does the lower IVT range.
Citation: Journal of Hydrometeorology 22, 6; 10.1175/JHM-D-20-0165.1
4. Summary and discussion
In this study, the roles of the offshore lower-level vertical thermodynamic stability and IVT on surface wind speed for landfalling ARs were analyzed for four regions previously shown to experience frequent and strong AR landfalls: the U.S. West Coast, southwestern South America, western Europe, and New Zealand. Our objective was to test the hypothesis that weaker low-level vertical stability conditions under the AR are associated with stronger surface winds for a given AR IVT strength. Using all MERRA-2 data and a linear regression model, it was shown that IVT is an important variable associated with surface winds in conjunction with landfalling ARs (R2 = 0.38; U.S. West Coast). However, the combined influence of IVT and lower-level vertical thermodynamic stability (ΔT) accounts for more variance than IVT alone (R2 = 0.52; U.S. West Coast).
Using ΔT as the measure of low-level vertical thermodynamic stability, it is shown that for a given IVT strength a negative ΔT (i.e., unstable) is associated with higher surface winds, while positive ΔT (i.e., stable) is associated with lower surface winds for a given IVT strength (Fig. 5a). The relationship is shown to hold in several regions where extreme surface winds associated with ARs are found (Fig. 6a), with the U.S. West Coast having the most pronounced relationship among the four regions considered (Fig. 5) due to having the highest initial correlation with IVT. It is interesting to note that all the other regions have strikingly similar R2 values for both IVT alone (R2 = 0.22–0.23) and IVT with ΔT (R2 = 0.36–0.38). Overall, IVT was shown to be a decent indicator of surface wind speed within a landfalling AR. However, ARs with weak to moderate IVT and a relatively unstable lower atmosphere can still exhibit strong surface winds. The incorporation of near-surface stability is most influential in low IVT ARs, where together with IVT they provide a more accurate understanding of the near-surface atmospheric state, enabling the ability to better resolve the variation in surface wind speed per a given IVT.
The reanalysis results are reproduced using satellite data to ensure the relationship holds when using observations independent of the reanalysis model for the variables of interest. For example, Fig. 6b shows the relationship when substituting one satellite observed variable (AIRS R2 = 0.53; QuikSCAT R2 = 0.48), as well as the combination of both AIRS and QuikSCAT variables (R2 = 0.44). The coupling effect by IVT and near-surface stability is still observed with the introduction of alternate metrics of stability, such as the bulk Richardson number RB (R2 = 0.47).
In addition to quantifying the overall relationship, the 25% most stable and unstable ARs over the lowest 1 km were isolated over two IVT ranges to better understand the distribution of extreme winds in terms of their potential impacts, using the BWS gale force and above as an indicator. It was shown that instability in ARs is more likely to be accompanied with extreme wind speeds (lower range of IVT: 5.3%, higher range: 14.7%) than in a stable regime (lower range: 0.58%, higher range: 6.15%), with IVT also playing a role as quantified previously.
Waliser and Guan (2017) highlighted the close association between ARs and extreme coastal surface winds across the globe. The current study sheds more light on what near-surface atmospheric conditions, in addition to the presence of enhanced IVT in a landfalling AR, are strongly associated with extreme surface winds. Understanding these factors, such as near-surface stability, can potentially lead to more guidance and situational awareness when it comes to understanding the hazardous impacts of ARs. Ralph et al. (2019) created an impact-oriented scale for AR events based on maximum IVT and the event duration. As with the scale for hurricanes, the AR scale aims to be used as a tool to efficiently convey potential hazards and/or benefits associated with ARs based on the expected hydrometeorological impacts largely from the AR precipitation. The current study offers an indication of what might be additionally considered when predicting potential hazards associated with ARs, i.e., those more closely related to AR winds.
Acknowledgments
This research was funded in part by National Aeronautics and Space Administration (NASA) MIRO NNX15AQ06A. Author Waliser’s contributions to this study were carried out on behalf of the Jet Propulsion Laboratory, California Institute of Technology, under a contract with NASA. Author Pagano thanks Shakeel Asharaf, Peter Gibson, Mike Deflorio, Elias Massoud, and Alex Goodman from the AR research group at JPL for their comments, suggestions, and support. He also thanks Dr. Jingjing Li for serving on his thesis committee at California State University, Los Angeles, because this work was also part of his master’s thesis.
Data availability statement
The AR detection algorithm used for this research is available online (https://ucla.app.box.com/v/ARcatalog/). MERRA-2 collections and the AIRS data used in this research are publicly available (https://disc.gsfc.nasa.gov/). The QuikSCAT data are online (https://podaac-tools.jpl.nasa.gov/).
REFERENCES
Akella, S., R. Todling, and M. Suarez, 2017: Assimilation for skin SST in the NASA GEOS atmospheric data assimilation system. Quart. J. Roy. Meteor. Soc., 143, 1032–1046, https://doi.org/10.1002/qj.2988.
American Meteorological Society, 2017: Atmospheric river. Glossary of Meteorology, http://glossary.ametsoc.org/wiki/Atmospheric_river.
Archer, C. L., B. A. Colle, D. L. Veron, F. Veron, and M. J. Sienkiewicz, 2016: On the predominance of unstable atmospheric conditions in the marine boundary layer offshore of the U.S. northeastern coast. J. Geophys. Res. Atmos., 121, 8869–8885, https://doi.org/10.1002/2016JD024896.
Barthelmie, R. J., 1999: The effects of atmospheric stability on coastal wind climates. Meteor. Appl., 6, 39–47, https://doi.org/10.1017/S1350482799000961.
Callahan, P. S., and T. Lungu, Eds., 2006: QuikSCAT science data product user’s manual, version 3.0. JPL/NASA Doc. D-18053–Rev A, 85 pp., https://podaac-tools.jpl.nasa.gov/drive/files/allData/quikscat/L2B12/docs/QSUG_v3.pdf.
Chahine, M. T., and Coauthors, 2006: AIRS: Improving weather forecasting and providing new data on greenhouse gases. Bull. Amer. Meteor. Soc., 87, 911–926, https://doi.org/10.1175/BAMS-87-7-911.
Cobb, A., A. Michaelis, S. Iacobellis, F. M. Ralph, and L. Delle Monache, 2021: Atmospheric river sectors: Definition and characteristics observed using dropsondes from 2014–20 CalWater and AR Recon. Mon. Wea. Rev., 149, 623–644, https://doi.org/10.1175/MWR-D-20-0177.1.
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, 514–523, https://doi.org/10.1111/J.1752-1688.2011.00546.X.
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.
Divakarla, M., and Coauthors, 2006: Validation of Atmospheric Infrared Sounder temperature and water vapor retrievals with matched radiosonde measurements and forecasts. J. Geophys. Res., 111, D09S15, https://doi.org/10.1029/2005JD006116.
Gelaro, R., and Coauthors, 2017: The Modern-Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2). J. Climate, 30, 5419–5454, https://doi.org/10.1175/JCLI-D-16-0758.1.
Guan, B., and D. E. Waliser, 2015: Detection of atmospheric rivers: Evaluation and application of an algorithm for global studies. J. Geophys. Res. Atmos., 120, 12 514–12 535, https://doi.org/10.1002/2015JD024257.
Guan, B., N. Molotch, D. Waliser, E. Fetzer, and P. 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.
Guan, B., D. E. Waliser, N. P. Molotch, E. J. Fetzer, and P. J. Neiman, 2012: Does the Madden–Julian oscillation influence wintertime atmospheric rivers and snowpack in the Sierra Nevada? Mon. Wea. Rev., 140, 325–342, https://doi.org/10.1175/MWR-D-11-00087.1.
Guan, B., N. P. Molotch, D. E. Waliser, E. J. Fetzer, and P. J. Neiman, 2013: The 2010/2011 snow season in California’s Sierra Nevada: Role of atmospheric rivers and modes of large-scale variability. Water Resour. Res., 49, 6731–6743, https://doi.org/10.1002/wrcr.20537.
Guan, B., D. E. Waliser, and F. M. Ralph, 2018: An intercomparison between reanalysis and dropsonde observations of the total water vapor transport in individual atmospheric rivers. J. Hydrometeor., 19, 321–337, https://doi.org/10.1175/JHM-D-17-0114.1.
Khouakhi, A., and G. Villarini, 2016: On the relationship between atmospheric rivers and high sea water levels along the U.S. West Coast. Geophys. Res. Lett., 43, 8815–8822, https://doi.org/10.1002/2016GL070086.
Kingston, D. G., D. A. Lavers, and D. M. Hannah, 2016: Floods in the southern Alps of New Zealand: The importance of atmospheric rivers. Hydrol. Processes, 30, 5063–5070, https://doi.org/10.1002/hyp.10982.
Lamjiri, M. A., M. D. Dettinger, F. M. Ralph, and B. Guan, 2017: Hourly storm characteristics along the U.S. West Coast: Role of atmospheric rivers in extreme precipitation. Geophys. Res. Lett., 44, 7020–7028, https://doi.org/10.1002/2017GL074193.
Lavers, D. A., and G. Villarini, 2013: The nexus between atmospheric rivers and extreme precipitation across Europe. Geophys. Res. Lett., 40, 3259–3264, https://doi.org/10.1002/grl.50636.
Lavers, D. A., G. Villarini, R. P. Allan, E. F. Wood, and A. J. Wade, 2012: The detection of atmospheric rivers in atmospheric reanalyses and their links to British winter floods and the large-scale climatic circulation. J. Geophys. Res., 117, D20106, https://doi.org/10.1029/2012JD018027.
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, https://doi.org/10.1029/2008GL036445.
Luo, B., P. J. Minnett, M. Szczodrak, N. R. Nalli, and V. R. Morris, 2020: Accuracy assessment of MERRA-2 and ERA-Interim sea surface temperature, air temperature, and humidity profiles over the Atlantic Ocean using AEROSE measurements. J. Climate, 33, 6889–6909, https://doi.org/10.1175/JCLI-D-19-0955.1.
McCarty, W., and Coauthors, 2016: MERRA-2 input observations: Summary and assessment. NASA Tech. Rep. Series on Global Modeling and Data Assimilation Rep. NASA/TM-2016-104606, Vol. 46, 64 pp., https://gmao.gsfc.nasa.gov/pubs/docs/McCarty885.pdf.
Mohan, M., and T. A. Siddiqui, 1998: Analysis of various schemes for the estimation of atmospheric stability classification. Atmos. Environ., 32, 3775–3781, https://doi.org/10.1016/S1352-2310(98)00109-5.
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.
Paltan, H., D. Waliser, W. H. Lim, B. Guan, D. Yamazaki, R. Pant, and S. Dadson, 2017: Global floods and water availability driven by atmospheric rivers. Geophys. Res. Lett., 44, 10 387–10 395, https://doi.org/10.1002/2017GL074882.
Pickett, M. H., W. Tang, L. K. Rosenfeld, and C. H. Wash, 2003: QuikSCAT satellite comparisons with nearshore buoy wind data off the U.S. west coast. J. Atmos. Oceanic Technol., 20, 1869–1879, https://doi.org/10.1175/1520-0426(2003)020<1869:QSCWNB>2.0.CO;2.
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, 783–790, https://doi.org/10.1175/BAMS-D-11-00188.1.
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 1997/98. Mon. Wea. Rev., 132, 1721–1745, https://doi.org/10.1175/1520-0493(2004)132<1721:SACAOO>2.0.CO;2.
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., and Coauthors, 2017: Dropsonde observations of total integrated water vapor transport within North Pacific atmospheric rivers. J. Hydrometeor., 18, 2577–2596, https://doi.org/10.1175/JHM-D-17-0036.1.
Ralph, F. M., and Coauthors, 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.
Ramos, A. M., R. M. Trigo, M. L. R. Liberato, and R. Tomé, 2015: Daily precipitation extreme events in the Iberian Peninsula and its association with atmospheric rivers. J. Hydrometeor., 16, 579–597, https://doi.org/10.1175/JHM-D-14-0103.1.
Richardson, L. F., and W. N. Shaw, 1920: The supply of energy from and to atmospheric eddies. Proc. Roy. Soc. London, 97A, 354–373, https://doi.org/10.1098/RSPA.1920.0039.
Ridder, N., H. de Vries, and S. Drijfhout, 2018: The role of atmospheric rivers in compound events consisting of heavy precipitation and high storm surges along the Dutch coast. Proc. Roy. Soc., A18, 3311–3326, https://doi.org/10.1098/RSPA.1920.0039.
Rondanelli, R., B. Hatchett, J. Rutllant, D. Bozkurt, and R. Garreaud, 2019: Strongest MJO on record triggers extreme Atacama rainfall and warmth in Antarctica. Geophys. Res. Lett., 46, 3482–3491, https://doi.org/10.1029/2018GL081475.
Shields, C. A., and Coauthors, 2018: Atmospheric River Tracking Method Intercomparison Project (ARTMIP): Project goals and experimental design. Geosci. Model Dev., 11, 2455–2474, https://doi.org/10.5194/GMD-11-2455-2018.
Stull, R., 2006: The atmospheric boundary layer. Atmospheric Science: An Introductory Survey, J. M. Wallace and P. V. Hobbs, Eds., Academic Press, 375–417, https://doi.org/10.1016/B978-0-12-732951-2.50014-4.
Susskind, J., J. M. Blaisdell, L. Iredell, and F. Keita, 2010: Improved temperature sounding and quality control methodology using AIRS/AMSU data: The AIRS science team version 5 retrieval algorithm. IEEE Trans. Geosci. Remote Sens., 49, 883–907, https://doi.org/10.1109/TGRS.2010.2070508.
Tsai, W., M. Spencer, C. Wu, C. Winn, and K. Kellogg, 2000: SeaWinds on QuikSCAT: Sensor description and mission overview. IEEE 2000 Int. Geoscience and Remote Sensing Symp., Honolulu, HI, IEEE, 1021–1023, https://doi.org/10.1109/IGARSS.2000.858008.
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, 481–507, https://doi.org/10.1175/2010JHM1284.1.
Viale, M., R. Valenzuela, R. D. Garreaud, and F. M. Ralph, 2018: Impacts of atmospheric rivers on precipitation in southern South America. J. Hydrometeor., 19, 1671–1687, https://doi.org/10.1175/JHM-D-18-0006.1.
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.
Wallace, J. M., and P. V. Hobbs, 2006: Introduction and overview. Atmospheric Science: An Introductory Survey, J. M. Wallace and P. V. Hobbs, Eds., Academic Press, 1–23, https://doi.org/10.1016/B978-0-12-732951-2.50006-5.
Wong, S., E. J. Fetzer, M. Schreier, G. Manipon, E. F. Fishbein, B. H. Kahn, Q. Yue, and F. W. Irion, 2015: Cloud-induced uncertainties in AIRS and ECMWF temperature and specific humidity. J. Geophys. Res. Atmos., 120, 1880–1901, https://doi.org/10.1002/2014JD022440.
Zhu, Y., and R. E. Newell, 1998: A proposed algorithm for moisture fluxes from atmospheric rivers. Mon. Wea. Rev., 126, 725–735, https://doi.org/10.1175/1520-0493(1998)126<0725:APAFMF>2.0.CO;2.
Zoumakis, N. M., and A. G. Kelessis, 1991: The dependence of the bulk Richardson number on stability in the surface layer. Bound.-Layer Meteor., 57, 407–414, https://doi.org/10.1007/BF00120057.
