1. Introduction
Atmospheric rivers (ARs)—long, narrow filamentary features characterized by enhanced horizontal moisture transport and often associated with a pre-cold-frontal low-level jet—play a significant role in the global water cycle by contributing to the poleward transport of moisture (Browning and Pardoe 1973; Newell et al. 1992; Zhu and Newell 1998; Ralph et al. 2004, 2005; Knippertz and Wernli 2010; Newman et al. 2012; Knippertz et al. 2013; Ralph et al. 2017). While ARs are critical providers of water resources for certain regions, they can also be associated with high-impact, extreme precipitation events affecting areas around the world including the western, central, and eastern United States (Ralph et al. 2006; Neiman et al. 2008; Leung and Qian 2009; Neiman et al. 2011; Moore et al. 2012; Lavers and Villarini 2013a,b; Neiman et al. 2013; Ralph et al. 2013; Rutz et al. 2014; Mahoney et al. 2016; Demaria et al. 2017; Nayak and Villarini 2017; Moore et al. 2020), South America (Viale and Nuñez 2011; Viale et al. 2013, 2018), the United Kingdom and western Europe (Stohl et al. 2008; Lavers et al. 2011, 2012; Lavers and Villarini 2013b; Sodemann and Stohl 2013; Ramos et al. 2015; Eiras-Barca et al. 2016), New Zealand (Kingston et al. 2016; Little et al. 2019), East Asia (Hirota et al. 2016; Kamae et al. 2017), and high-latitude regions (Gorodetskaya et al. 2014; Nash et al. 2018; Adusumilli et al. 2021; Wille et al. 2021). Waliser and Guan (2017) recently showed that not only are ARs often accompanied by heavy precipitation, but ARs can also pose serious threats associated with extreme winds. Impacts from ARs are further exacerbated when they occur in succession with one another, creating an AR family (Fish et al. 2019).
The vast majority of heavy precipitation events for the U.S. West Coast are AR-related (e.g., Ralph et al. 2006; Smith et al. 2010; Dettinger et al. 2011; Yuter et al. 2011; Ralph et al. 2013; Lamjiri et al. 2017). In particular, ARs are responsible for almost half of California’s annual precipitation (Guan et al. 2010; Dettinger et al. 2011; Gershunov et al. 2017) and 33%–44% of drought breaks within the state (Dettinger 2013). The fractions of annual precipitation and floods produced by ARs are > 60% and almost 100%, respectively, for the Russian River watershed in Northern California (Ralph et al. 2006; Gershunov et al. 2017). The AR scale (Ralph et al. 2019) categorizes AR events on a scale of 1–5 based event intensity (i.e., maximum instantaneous IVT) and duration (i.e., duration of IVT ≥ 250 kg m−1 s−1). Events lower on the scale (e.g., AR1–AR3) are largely beneficial while more intense events (AR4–5) are considered mostly hazardous (Ralph et al. 2019). Flood damages related to ARs range from less than $1 million (U.S. dollars) for weak and moderate ARs (AR1 and AR2; Corringham et al. 2019; Ralph et al. 2019) to greater than $20 million for extreme and exceptional events (AR4 and AR5; Corringham et al. 2019; Ralph et al. 2019).
ARs form within complex environments in conjunction with other meso- to synoptic-scale features spanning ocean basins from the tropics to the midlatitudes (e.g., Lackmann and Gyakum 1999; Cordeira et al. 2013; Neiman et al. 2014; Zhang et al. 2019; Ralph et al. 2020). These additional features include, but are not limited to, extratropical cyclones, their associated fronts, and consequently, mesoscale frontal waves (MFWs). The presence of a MFW along an AR can modify the landfall position, orientation, duration, and intensity of the system, critical factors in determining where, and how much, precipitation will occur over land (Ralph et al. 2003; Neiman et al. 2004; Ralph et al. 2011; Neiman et al. 2016; Hecht and Cordeira 2017; Lamjiri et al. 2017; Martin et al. 2019), and consequently, the damages associated with the event (Corringham et al. 2019). Therefore, the prediction of MFWs is important for accurate forecasts of landfalling precipitation associated with ARs. MFWs, however, often form at short lead times and are typically poorly handled by global forecast models due to their small spatial scale, rapid growth, and sensitivity to the effects of latent heating, contributing to large errors in the landfalling AR and subsequent precipitation forecast (Parker 1998; Hewson 2009; Martin et al. 2019; Ralph et al. 2020), thus motivating the present study.
MFWs, also referred to as frontal waves and diminutive frontal waves, form due to dynamic instabilities along fronts associated with mature extratropical cyclones, and in some cases, intensify into secondary cyclones (e.g., Joly and Thorpe 1990; Schär and Davies 1990; Parker 1998). The Norwegian cyclone model (Bjerknes and Solberg 1922) assumes that frontal waves/secondary cyclones share the same dynamics as the parent cyclone, but a larger variety of mechanisms can cause secondary frontal waves to form, such as shear along the existing frontal zone (e.g., barotropic instability; Joly and Thorpe 1990) and latent heat release (Reed et al. 1993b; Dacre and Gray 2006; Ludwig et al. 2015; Schemm and Sprenger 2015). Previous work has highlighted the importance of both diabatically generated low-level PV anomalies along the frontal boundary and upper-level PV anomalies associated with upper-tropospheric waves in the development and growth of frontal disturbances into waves and secondary cyclones (Hoskins et al. 1985; Thorncroft and Hoskins 1990; Malardel et al. 1993; Appenzeller and Davies 1996; Parker 1998; Fehlmann and Davies 1999; Zhang et al. 2002; Dacre and Gray 2006; Ludwig et al. 2015; Schemm and Sprenger 2015). In other words, the dynamics associated with frontal wave development and potential secondary cyclogenesis should be considered largely independent of the parent cyclone’s dynamics (Parker 1998).
Likewise, ARs are typically maintained and propagated by a parent extratropical cyclone or series of cyclones (e.g., Zhang et al. 2019). The intensity of moisture transport, moisture content, and latent heat release within an AR, however, varies considerably and often in response to a multitude of other factors apart from the parent cyclone. Some examples include tropical moisture exports (e.g., Cordeira et al. 2013), low-level jets (e.g., Ralph et al. 2005; Martin et al. 2018; Demirdjian et al. 2020b), deep and slantwise convection (e.g., Cordeira et al. 2013; Lavers and Villarini 2013a; Mahoney et al. 2016), and cross-frontal circulations (e.g., Demirdjian et al. 2020a). This symmetry of interactions across scales, response to, and modification of diverse environmental features suggests that ARs and MFWs may form their own symbiotic relationship which, depending on a host of factors, may sever or reinforce the relationship to the parent cyclone. In particular, both AR mesoscale circulations and MFWs are sensitive to diabatic generation of PV through latent heat release (Joly and Thorpe 1990; Schär and Davies 1990; Dacre and Gray 2006; Ludwig et al. 2015; Schemm and Sprenger 2015; Cannon et al. 2020; Demirdjian et al. 2020a), a fundamental process linking several of the factors and studies mentioned above.
While a substantial amount of research has examined frontal waves, rapidly intensifying extratropical cyclones, and secondary cyclogenesis in the North Atlantic (e.g., Reed et al. 1993a,b, 1994; Blier and Wakimoto 1995; Chang et al. 1996; Joly et al. 1997; Lackmann et al. 1997; Baehr et al. 1999; Bouniol et al. 1999; Chaboureau and Thorpe 1999; Priestley et al. 2020), significantly less attention has been given to the relationship between mesoscale frontal waves and ARs in the North Pacific (cf. Neiman et al. 2004; Ralph et al. 2011; Neiman et al. 2016; Martin et al. 2019). To that end, we analyze two AR events that made landfall in Northern California and were associated with different MFW environments—one with significant upper-level support and one with little upper-level support, subjectively defined by the wave’s proximity to upper-level features. We conduct novel numerical model simulations using the Model for Prediction Across Scales-Atmosphere (MPAS-A) with the effects of latent heating removed to diminish the MFW and therefore weaken the AR–MFW interaction in each case. Doing so allows us to build on previous work and quantify the impact of MFWs and diabatic processes on the ARs and associated landfalling precipitation. Changes in the ARs are further analyzed in regard to dynamic and thermodynamic contributions to moisture transport, and orographic and dynamic forcings are considered in the analysis of precipitation differences. While removing latent heating has been a technique used in other models (e.g., Kuo et al. 1995; Posselt and Martin 2004; Tao and Zhang 2015; Steinfeld et al. 2020), to our knowledge, this is one of the first studies to implement this “no-latent-heating” capability into MPAS-A. Our paper continues with a description of the data used and methods for the numerical simulations in section 2. Section 3 provides an overview of the two selected cases. Sections 4 and 5 present results from the no-latent-heating experiments. Finally, section 6 summarizes the main findings and discusses avenues of future work.
2. Data, model experiments, and evaluation
a. Observations and model analyses
1) Stage IV quantitative precipitation estimates (QPE) and USGS stream gauge observations
We use the Stage IV QPE (accessed from https://data.eol.ucar.edu/dataset/21.093) and USGS stream gauge observations (accessed from https://waterdata.usgs.gov/nwis/sw) as ground truth for the December 2014 and January 2010 events. The Stage IV precipitation analyses are created using the multisensor analyses produced by the 12 River Forecast Centers across the United States and are provided on a 4-km horizontal grid with 1-, 6-, and 24-h temporal resolutions; we use the 6-h data for our purposes. Storm-total precipitation is evaluated over the 72-h period from 1200 UTC 10 December to 1200 UTC 13 December 2014 and over the 54-h period from 1200 UTC 24 January to 1800 UTC 26 January 2010. The USGS stream gauge data, made available by the USGS National Water Information System (NWIS), provides river stage observations every 15 min. Here, we use the observations recorded at the Johnson’s Beach location along the Russian River in Guerneville, California.
2) ERA5
We use ERA5 (Hersbach et al. 2020) for model initial conditions and sea surface temperature (SST) fields that are updated daily. We also evaluate the IVT forecasts and control simulations in the supplementary material using ERA5 for comparison. ERA5 is produced on a global 31-km horizontal grid with 137 vertical levels up to 0.01 hPa and 1-h temporal resolution; 6-h data with 37 pressure levels available from https://rda.ucar.edu/datasets/ds633.0/ were used here.
b. Model configuration
We utilize version 7.0 of the Model for Prediction Across Scales-Atmosphere (MPAS-A; Skamarock et al. 2012) for our simulations. MPAS-A is a global atmosphere-only nonhydrostatic numerical model with variable-resolution grids created using unstructured Voroni meshes (Du et al. 1999), which allow for the gradual degradation of resolution within a singular domain, effectively alleviating any issues arising from sharp discontinuities between nested boundary domains (Park et al. 2014). We used the 10–60-km variable-resolution mesh with the 10-km area centered just north of Hawaii and relaxing to 60 km elsewhere (Fig. 1).
MPAS-A model domain for the CNTL and noLH simulations. Variables are interpolated to a 0.1° × 0.1° latitude–longitude grid within the red box.
Citation: Monthly Weather Review 149, 8; 10.1175/MWR-D-20-0364.1
Initial conditions and daily updated SSTs are derived from ERA5. We initialize simulations 48 h prior to the initial MFW formation in each case and integrate through the event. The December 2014 and January 2010 cases are run for 96 and 78 h, respectively; output is recorded every 3 h. During postprocessing, we vertically interpolate three-dimensional variables onto 42 pressure levels, and spatially interpolate all variables to a 0.1° × 0.1° latitude–longitude grid within a limited portion of the domain (Fig. 1).
For all simulations, we use the Kain–Fritsch [as in the Weather Research and Forecast (WRF) Model v.3.2.1] convective parameterization scheme and the WRF single-moment 6-class (WSM6; as in WRF v.4.1) microphysics parameterization scheme. The Yonsei University (YSU; as in WRF model version 4.0.3) parameterization scheme represents the planetary boundary layer, the Rapid Radiative Transfer Model (RRTMG; as in WRF v.3.8.1) scheme parameterizes longwave and shortwave radiation, and the surface processes are handled by the Noah land surface model (as in WRF v.4.0.3). The model top is set to 30 km with 55 vertical levels. We tested other physics configurations using the scale-aware Grell–Freitas (as in WRF v.3.6.1) and the non-aerosol-aware Thompson (as in WRF v.3.8.1) convective and microphysics schemes, respectively. Due to the substantial spread in spatial precipitation patterns, we chose to retain only the model configuration that best represented the observed precipitation pattern over Northern California and the Russian River watershed in both cases for analysis (i.e., the Kain–Fritsch convective and WSM6 microphysics schemes). Evaluations of the control (CNTL) simulations for each case are presented in the supplementary material.
c. No-latent-heating experiments
For the no-latent-heating (noLH) experiments, we modify the MPAS-A microphysics module to remove the effects of latent heating from the microphysics parameterization scheme. This change is akin to activating the “no_mp_heating” flag in the WRF model namelist. When turning this flag on in WRF, however, one must also run without a convective parameterization scheme to effectively eliminate the effects of subgrid-scale convection and associated latent heat release on the simulated temperature profile. Because the grid spacing in our variable-resolution mesh (10–60 km) is not conducive for running with explicit convection, we made additional modifications to the MPAS-A physics module to remove the effects of latent heating from the convective and planetary boundary layer schemes. Therefore, in the noLH simulations presented here, the potential temperature tendencies due to planetary boundary layer processes, microphysics, and deep and shallow convection are set to zero; tendencies of other variables (e.g., water vapor, cloud water) are retained as normal.
d. Limitations of experimental design
No set of model experiments is without its limitations. First, removing the effects of latent heating throughout the entire model domain affects more than just the features of interest (e.g., MFW and AR). We aim to constrain the effects of removing diabatic heating to the MFW and AR by initializing the simulations relatively close to MFW development. In doing so, we allow for a credible control simulation (see the supplementary material) and allow the model sufficient time to adjust to the removal of latent heating while also limiting the influence of features remote from the MFW and AR. Despite the known limitations, this method has been shown to be a useful tool for investigating the impacts of latent heating on various meteorological phenomena (e.g., Hjelmfelt and Braham 1983; Kuo et al. 1995; Buzzi et al. 1998; Posselt and Martin 2004; Tao and Zhang 2015; Kolstad et al. 2016; Steinfeld et al. 2020). Additionally, while setting the potential tendencies in the planetary boundary layer and convective schemes to zero, it is possible that, in addition to removing tendencies due to diabatic processes, tendencies from subgrid-scale mass transport are also neglected.
Second, high-resolution is important for resolving the effects of latent heating (e.g., Willison et al. 2013), but MFWs are mesoscale phenomena, and therefore, may not be fully resolved with 10-km grid spacing. Nevertheless, our control runs simulate realistic MFW signatures in both cases (e.g., Figs. S4d, S8f).
Furthermore, we only tested four ensemble members by varying just the convective and microphysics parameterization schemes. We attempted to include two additional ensemble members using the Tiedtke convective parameterization scheme, but an incompatibility between the modifications to the code for the noLH simulations and Tiedtke led to unrealistically cold near-surface temperatures and widespread convection throughout the domain (not shown). This behavior is an interesting result in and of itself, especially considering the infancy of conducting noLH simulations with MPAS-A. While resolving this issue is beyond the scope of the current study, it remains an active area of future work. Because of the substantial differences between the four ensemble members in representing precipitation over our study area in Northern California, we opted to analyze the one simulation that best replicated the observed precipitation pattern. Future studies will explore incorporating additional ensemble members with perturbed initial conditions to better represent the full range of possibilities in the noLH experiments.
3. Event overviews, significance, and forecast challenges
a. 10–13 December 2014
An intense AR with maximum IVT > 1200 kg m−1 s−1 initially made landfall around 0000 UTC 10 December 2014 along the Pacific Northwest coast of the United States (Fig. 2a), propagating southward and penetrating inland to the Intermountain West over the next 24 h. As AR conditions began to weaken over the region, a MFW, noted by the depression in SLP and secondary branch, or “cusp,” in the IVT field (e.g., Ralph et al. 2011; Neiman et al. 2016; Martin et al. 2019), formed offshore at the intersection of two frontal zones ~37°N, 135°W (Fig. 2c). Martin et al. (2019) characterized this MFW as having strong upper-level support (e.g., Parker 1998; Ludwig et al. 2015), evident by its location relative to a 250-hPa jet and 500-hPa trough; the MFW is positioned near the left-exit region of the jet and downstream of a shallow trough, both favorable locations for quasigeostrophic forcing for ascent (Fig. 2d).
(left) SLP (hPa; contoured every 4 hPa) and IVT (kg m−1 s−1; shaded), (center) 500-hPa height (m; contoured every 60 m) and 250-hPa wind speed (m s−1; shaded), and (right) transect of potential vorticity [PV; PVU; 1 PVU = 10−6 K kg−1 m2 s−1) shaded] and potential temperature (θ; K; contoured every 1 K) from ERA5 at (a),(b) 0000 UTC 10 December; (c)–(e) 0000 UTC 11 Dec; (f),(g) 1200 UTC 11 Dec; and (h),(i) 0000 UTC 12 Dec 2014. The dashed black line in (c) represents the position of the cross section used in (e). Approximate locations of cold and warm fronts in the left column are shown in the dark blue and black lines, respectively. The location of the MFW and secondary low pressure centers are labeled “W” in (d) and “L” in (g) and (i), respectively. The Russian River watershed (~39°N, 123°W) is indicated with a white star in the left column.
Citation: Monthly Weather Review 149, 8; 10.1175/MWR-D-20-0364.1
As a result of sufficient upper-level forcing, the MFW underwent cyclogenesis over the next 12 h, intensifying into a secondary cyclone with a minimum central SLP ~976 hPa and bringing a resurgence of strengthened AR conditions to Northern California (Figs. 2f, 3a). AR conditions reached peak intensity over the Russian River Watershed ~1200 UTC 11 December 2014 with a maximum IVT > 1200 kg m−1 s−1 (Table 1; Fig. 3a) and subsequently began to weaken over the area as the cyclone propagated northward (Figs. 2h, 3a); the event dissipated throughout the Pacific Northwest region by ~1800 UTC 12 December (not shown).
Stage IV 6-hourly accumulated mean areal precipitation (mm) ending at the time indicated on the abscissa for the Russian River watershed (blue bars), ERA5 6-hourly mean areal IVT (kg m−1 s−1) for the Russian River watershed at the end of the precipitation accumulation period (black line), and USGS river stage (ft) at Guerneville, CA (magenta line), for (a) Dec 2014 and (b) Jan 2010 cases. Red asterisks represent the maximum IVT (kg m−1 s−1) over the Russian River watershed at each time as indicated by ERA5. Plus signs and open circles indicate times when the river height exceeded monitor and flood stages, respectively. Vertical black dashed lines indicate the onset and dissipation of AR conditions (IVT ≥ 250 kg m−1 s−1) defined by the maximum IVT.
Citation: Monthly Weather Review 149, 8; 10.1175/MWR-D-20-0364.1
Observed and simulated event summary statistics for Dec 2014 and Jan 2010 events. All values have been rounded to the nearest whole number and are reported for the Russian River watershed. Note that 1 ft ≈ 0.305 m.
PV within the 900–700-hPa layer has previously been used to indicate the influence of diabatic processes on surface circulations (e.g., Davis and Emanuel 1991; Reed et al. 1993b; Marciano et al. 2015; Michaelis et al. 2017). A PV transect through the MFW at the time of development (Fig. 2c) emphasizing the lower atmosphere up to 700 hPa, therefore helps elucidate the likely diabatic influences. Here, a small PV tower elevated from the surface and extending up to ~850 hPa with a maximum ~1.75 PVU (Fig. 2e) suggests that while diabatic processes are not irrelevant to MFW development in this case, upper-level forcing related to the upper-level trough and jet likely played a more dominant role. A PV transect of the full atmosphere further emphasizes the presence of a substantial upper-level trough upstream of the MFW with the tropopause extending down to ~500 hPa (not shown).
With an event maximum IVT > 1200 kg m−1 s−1 and AR conditions persisting for 36 h over the Russian River watershed, the December 2014 event ranked as an AR4 on the AR scale (Ralph et al. 2019). Most of the precipitation over the Russian River watershed [~129 mm (~5 in.) out of the storm total ~178 mm (~7 in.); Table 1] fell in conjunction with the second wave of AR conditions between 0600 UTC 11 December and 0000 UTC 12 December (Fig. 3a). Over this same time period, the river stage at Guerneville, California, rose from ~2 m (~6.4 ft) to ~7.3 m (~24 ft), approximately 1.5 m (~5 ft) below monitor stage, before cresting at a height slightly greater than 10 m (33 ft) around 1300 local time (LT) 12 December, thus exceeding flood stage (Fig. 3a).
Several forecast challenges were associated with the December 2014 AR event. Most notably, the “hiatus” in AR conditions between 0000 and 1200 UTC 11 December, which halted precipitation for ~6 h at the Atmospheric River Observatory (ARO) in Bodega Bay, California (Martin et al. 2019, their Fig. 3), was not well forecast (Fig. S1a). As a result, earlier forecasts misrepresented the timing and magnitude of precipitation over the Russian River watershed as well as the river stage at flood-vulnerable Guerneville, California (Fig. S1c; Martin et al. 2019). Interestingly, precipitation forecast errors were not uniform across the region; western portions of the watershed and areas to the northwest were continually underforecast while regions to the north of the watershed along the coast were overforecast (Fig. S2). The heterogenous pattern in spatial precipitation forecast error, error in timing of peak precipitation, and shift from under- to over-river-stage forecast illustrate the effect an MFW can have on the location, timing, and impact of landfalling precipitation, especially on the watershed scale.
b. 24–26 January 2010
Around 1200 UTC 24 January 2010, an AR associated with a strong extratropical cyclone (minimum SLP < 970 hPa) to the northwest made landfall along the Pacific Northwest coast (Fig. 4a). AR conditions (i.e., IVT ≥ 250 kg m−1 s−1) lingered along the coastlines of Washington, Oregon, and Northern California with slight inland penetration to the Cascade and Sierra Nevada mountain ranges as the offshore extratropical cyclone propagated northward and began to weaken (Fig. 4c). Just as AR conditions began to dissipate through much of the area around 1200 UTC 25 January 2010, a weak MFW formed on the periphery of the remnant AR along a cold-frontal zone left behind by the extratropical cyclone off the Pacific Northwest coast (Fig. 4e). This MFW concentrated the remaining IVT off the coast of Northern California and resulted in a second pulse of AR conditions inland over the Russian River watershed for an additional 18 h (Figs. 4h, 3b).
(left) SLP (hPa; contoured every 4 hPa) and IVT (kg m−1 s−1; shaded), (center) 500-hPa height (m; contoured every 60 m) and 250-hPa wind speed (m s−1; shaded), and (right) transect of potential vorticity (PV; PVU; shaded) and potential temperature (θ; K; contoured every 1 K) from ERA5 at (a),(b) 1200 UTC 24 Jan; (c),(d) 0000 UTC 25 Jan; (e)–(g) 1200 UTC 25 Jan; and (h),(i) 0000 UTC 26 Jan 2010. The dashed black line in (e) represents the position of the cross section used in (g). Approximate locations of cold and warm fronts in the left column are shown in the dark blue and black lines, respectively. The location of the MFW and secondary low pressure centers are labeled “W” in (f) and “L” in (i), respectively. The Russian River watershed (~39°N, 123°W) is indicated with a white star in the left column.
Citation: Monthly Weather Review 149, 8; 10.1175/MWR-D-20-0364.1
The upper-level synoptic environment during the January 2010 event was much less conducive for quasigeostrophic forcing for ascent compared to the December 2014 case; the 250-hPa jet is substantially less amplified and weaker in magnitude, and there is no prominent trough at 500-hPa (Figs. 4b,d,f,i). Thus, the MFW in this case did not form in close proximity to any significant upper-level features (Fig. 4f), suggesting its development was likely driven by diabatic processes in the lower atmosphere. A PV transect through the MFW at the time of development (Fig. 4e) indicates that while a trough signature is evident, the tropopause does not penetrate below ~350 hPa and therefore, has much less influence on the circulations at the surface compared to the December 2014 event (not shown). In the lower atmosphere, a PV tower with a maximum exceeding 2 PVU extends from the surface to just below 700 hPa, indicating a significant diabatic contribution to the development of the MFW (Fig. 4g).
For the Russian River watershed, the January 2010 event ranked as an AR1 on the AR scale (Ralph et al. 2019); the event maximum IVT was ~360 kg m−1 s−1 and AR conditions lasted for 36 h (Table 1). Of the 81 mm (~3.2 in.) of precipitation that fell over the watershed from 24 to 26 January 2010 (Table 1), over half (~55 mm, 2.2 in.) is attributable to the second wave of AR conditions from 1200 UTC 25 January to 0600 UTC 26 January (Fig. 3b). Additionally, during this secondary period of AR conditions, the Russian River at Guerneville, California, rose ~3.9 m (~12.8 ft) from 3.8 m (12.5 ft) to ~7.7 m (25.3 ft) (Fig. 3b) before reaching its peak of ~8.2 m (~27 ft) (Table 1) at 1400 LT 26 January, just below monitor stage (8.8 m, 29 ft).
Similar forecast challenges were evident during the January 2010 event in that the timing and magnitude of maximum IVT and precipitation over the Russian River watershed were misforecast, even at day 0 (initialized at 1200 UTC 24 January; Fig. S1b,d). Spatially, precipitation forecast errors were larger in some areas compared to others, namely to the northeast and on the western edge of the Russian River watershed, but precipitation was generally underforecast throughout the region (Fig. S3). While the Russian River at Guerneville, California did not exceed monitor or flood stage during this event, nor were the precipitation totals as high as in the December 2014 case (Fig. 3b; Table 1), the AR–MFW interaction during the 24–26 January 2010 event highlights how the presence of an MFW of any strength can considerably alter storm-total precipitation on a watershed scale and consequently, introduce significant forecast challenges.
4. Comparison of CNTL and noLH simulations: AR, MFW, and landfalling precipitation
a. 10–13 December 2014
At initial landfall—0000 UTC 10 December 2014, 24 h into the simulation—the AR in the CNTL and noLH simulations are fairly similar with AR conditions occurring along the Pacific Northwest coast and offshore maximum IVT ~1200–1400 kg m−1 s−1 in both instances. The SLP trough along the northern edge of the AR and the surface high pressure system to the southeast, while present in the noLH simulations, are slightly weaker (Figs. 5a–c). AR conditions reach the Russian River watershed around 1800 UTC 10 December with a maximum IVT over the watershed ~400 kg m−1 s−1 in the CNTL and noLH runs (Figs. 6a–b). By 2100 UTC 10 December, the MFW has developed in the CNTL simulation, noted by the SLP trough and secondary IVT maximum ~36°N, 140°W. Neither a depression in the SLP field nor the signature IVT “cusp” are apparent in the noLH run (Figs. 5d–e). Consistent with MFW development and associated secondary IVT branch offshore, the AR is more expansive in CNTL at this time, but IVT values are stronger along the Pacific Northwest coastline in the noLH experiment (Figs. 5f). Differences in timing of the AR propagation over the Pacific Northwest account for the elevated IVT values in noLH (not shown).
SLP (hPa; contoured every 4 hPa) and IVT (kg m−1 s−1; shaded) for the Dec 2014 (left) CNTL and (center) noLH simulations and (right) IVT magnitude difference (noLH minus CNTL; kg m−1 s−1) at (a)–(c) 0000 UTC 10 Dec, (d)–(f) 2100 UTC 10 Dec, (g)–(i) 1200 UTC 11 Dec, and (j)–(l) 0000 UTC 12 Dec 2014. The Russian River watershed (~39°N, 123°W) is indicated with a white star in all panels.
Citation: Monthly Weather Review 149, 8; 10.1175/MWR-D-20-0364.1
(top) 6-hourly accumulated mean areal precipitation (mm) ending at the time indicated on the abscissa for the Russian River watershed (blue bars) and 6-hourly mean areal IVT (kg m−1 s−1) for the Russian River watershed at the end of the precipitation accumulation period (black line) for the Dec 2014 (a) CNTL and (b) noLH simulations. Red asterisks represent the maximum IVT (kg m−1 s−1) over the Russian River watershed at each time. (bottom) 72-h accumulated precipitation (mm; 1200 UTC 10 Dec–1200 UTC 13 Dec) for the Dec 2014 (c) CNTL, (d) noLH, and (e) noLH minus CNTL simulations. The 72-h storm total for the Russian River watershed is indicated in the bottom left of (c) and (d). The noLH minus CNTL absolute and percentage differences are indicated in the bottom left of (e). The Russian River watershed (~39°N, 123°W) is outlined in black in (c)–(e).
Citation: Monthly Weather Review 149, 8; 10.1175/MWR-D-20-0364.1
In the ~12 h leading up to MFW formation in the December 2014 event, differences in upper-level forcing between CNTL and noLH in the vicinity of the frontal wave are qualitatively minimal; the eventual MFW that forms in CNTL is in a similar position relative to the 250-hPa jet and shallow 500-hPa trough in both simulations (Figs. 7a–b,e–f,i–j). There are clear differences, however, between CNTL and noLH in the lower atmosphere. In CNTL, at 0600 UTC 10 December, low-level cyclonic vorticity maxima associated with enhanced lower-tropospheric moisture are evident ~35°N, 156°W, indicating small-scale instabilities present on the poleward side of the AR entrance region (Fig. 7c). Through enhanced moisture convergence and associated feedbacks (e.g., Demirdjian et al. 2020a), deformation along the secondary frontal zone increased, and over the next 12 h, this circulation strengthened (Figs. 7c,g,k), becoming strong enough to penetrate down to the surface and create a MFW ~2100 UTC 10 December (Fig. 5d). This initial region of low-level moisture is present in both the CNTL and noLH simulations through ~1200 UTC 9 December (not shown). Between 1200 UTC 9 December and 0000 UTC 10 December, weaker flow and moisture divergence in noLH, however, caused the moisture to dissipate by 0600 UTC 10 December (Fig. 7d). Without the enhanced lower-tropospheric moisture, moisture convergence, and associated diabatic feedbacks on the atmospheric circulation, the MFW was unable to develop in the noLH simulation (Fig. 5e).
500-hPa height (m; contoured every 60 m) and 250-hPa wind speed (m s−1; shaded) for the MPAS-A (a),(e),(i) CNTL and (b),(f),(j) noLH simulations; and 850-hPa potential temperature (θ; K; gray contours every 1 K), winds (m s−1; barbs), specific humidity (g kg−1; shaded), and relative vorticity (10−5 s−1; purple contours every 10 × 10−5 s−1 from 10 × 10−5 to 30 × 10−5 s−1 for the MPAS-A (c),(g),(k) CNTL and (d),(h),(l) noLH simulations at (a)–(d) 0600 UTC Dec, (e)–(h) 1200 UTC 10 Dec, and (i)–(l) 1800 UTC 10 Dec 2014. The approximate location of MFW development at 2100 UTC 10 Dec 2014 in CNTL is labeled “W” in all panels for reference. The predevelopment area of interest is outlined in black in (c), (g), and (k).
Citation: Monthly Weather Review 149, 8; 10.1175/MWR-D-20-0364.1
AR conditions continue to propagate eastward through 0600 UTC 11 December in both simulations, penetrating farther inland over Northern California while beginning to dissipate over Washington and Oregon (not shown). Interestingly, AR conditions over the Russian River watershed peak at this time in the noLH simulation with a maximum IVT ~853 kg m−1 s−1; the maximum IVT over the Russian River watershed in the CNTL run is ~636 kg m−1 s−1 (Figs. 6a,b). This difference in IVT magnitude is partially due to the strengthening frontal wave in the CNTL simulation which intensifies the AR offshore and slows its propagation toward the coast. The timing of this simulated lull in AR conditions over the Russian River watershed between 0000 and 0600 UTC 11 December (Fig. 6a) matches the “hiatus” identified by Martin et al. (2019, their Fig. 3) at the Atmospheric River Observatory in Bodega Bay, CA, which the authors attribute to the AR stalling offshore. Therefore, IVT over the Russian River watershed in CNTL decreases slightly through 0600 UTC 11 December and does not peak until 1200 UTC 11 December (Fig. 6a). The overall maximum IVT through the event is substantially higher in CNTL (~1024 vs ~853 kg m−1 s−1 in noLH; Table 1). By 1200 UTC 11 December in the noLH run, AR conditions are beginning to dissipate over the watershed (Fig. 6b). On the other hand, the MFW in CNTL has intensified into a strong secondary cyclone, helping to sustain AR conditions over Northern California and the Russian River watershed (Fig. 5g). In the noLH run, however, no prominent SLP feature exists off the coast of Northern California. Instead, IVT values are comparatively weaker over Northern California and offshore throughout the primary AR corridor (Fig. 5h), although a slight southward shift results in strengthened IVT on the equatorward edge of the AR over Central California (Fig. 5i). AR conditions dissipate over the region by 0600 UTC 12 December (Figs. 5k, 6b) in the noLH experiment, but persist over the Russian River watershed until ~1200 UTC 12 December in the CNTL simulation (Fig. 6a). This early termination of the event shortens the duration of AR conditions over the watershed from 36 h in the CNTL run to 30 h in the noLH simulation (Table 1). The shorter duration combined with a lower maximum IVT over the Russian River watershed weakens the event from an AR4 in CNTL to an AR3 in the noLH experiment.
Due to the weaker and shorter duration AR in the noLH simulation, precipitation totals over the Russian River watershed are reduced by ~64% from ~148 mm (~5.8 in.) in CNTL to ~53 mm (~2.1 in.) (Table 1, Figs. 6c–e). Differences in landfalling precipitation are not limited to the Russian River watershed. Consistent with the equatorward shift and strengthened IVT to the south in the noLH experiment (Figs. 5i,l), enhanced precipitation occurs over the central Sierra Nevada (Fig. 6e). Over much of the Central Valley and northern Sierra Nevada, however, precipitation is drastically reduced in the noLH experiment (Figs. 6c–e). Therefore, while an AR still occurs in the noLH simulation, and precipitation falls over much of California, including the Russian River watershed, the timing, intensity, location, and duration of the AR event are all impacted by the absence of diabatic processes and consequently, a lack of frontal wave and secondary cyclone formation.
b. 24–26 January 2010
At the time of initial landfall in the January 2010 case, 1200 UTC 24 January, there are slight differences in the CNTL and noLH AR; IVT values are stronger directly offshore of the Pacific Northwest coast in CNTL while the noLH AR propagates slower, evident by the positive–negative IVT difference dipole over the central Pacific. The cyclone to the northwest and maximum IVT within the AR core are of similar strength in both instances, albeit the AR as a whole is stronger in CNTL (Figs. 8a–c).
SLP (hPa; contoured every 4 hPa) and IVT (kg m−1 s−1; shaded) for the Jan 2010 (left) CNTL and (center) noLH simulations and (right) IVT magnitude difference (noLH minus CNTL; kg m−1 s−1) at (a)–(c) 1200 UTC 24 Jan, (d)–(f) 0000 UTC 25 Jan, (g)–(i) 1200 UTC 25 Jan, and (j)–(l) 0000 UTC 26 Jan 2010. The Russian River watershed (~39°N, 123°W) is indicated with a white star in all panels.
Citation: Monthly Weather Review 149, 8; 10.1175/MWR-D-20-0364.1
AR conditions reach the Russian River watershed ~0000 UTC 25 January in both the CNTL and noLH simulations with comparable strengths (Figs. 9a–b). Offshore, however, significant differences are evident. First, the extratropical cyclone and associated IVT are weaker in the noLH simulation. Second, the maximum IVT within the core of the AR is less intense in noLH with a maximum IVT between 500 and 600 kg m−1 s−1 compared to 600–700 kg m−1 s−1 in CNTL. Last, the low pressure system in the central Pacific ~155°W is stronger in the CNTL simulation, creating a more organized IVT maximum in the area (Figs. 8d–f).
(top) 6-hourly accumulated mean areal precipitation (mm) ending at the time indicated on the abscissa for the Russian River watershed (blue bars) and 6-hourly mean areal IVT (kg m−1 s−1) for the Russian River watershed at the end of the precipitation accumulation period (black line) for the Jan 2010 (a) CNTL and (b) noLH simulations. Red asterisks represent the maximum IVT (kg m−1 s−1) over the Russian River watershed at each time. (bottom) 54-h accumulated precipitation (mm; 1200 UTC 24 Jan–1800 UTC 26 Jan) for the Jan 2010 (c) CNTL, (d) noLH, and (e) noLH minus CNTL simulations. The 54-h storm total for the Russian River watershed is indicated in the bottom left of (c) and (d). The noLH minus CNTL absolute and percent differences are indicated in the bottom left of (e). The Russian River watershed (~39°N, 123°W) is outlined in black in (c)–(e).
Citation: Monthly Weather Review 149, 8; 10.1175/MWR-D-20-0364.1
By 1200 UTC 25 January, the initial AR has dissipated over Northern California and the Russian River watershed, leaving a small IVT plume offshore in both experiments. In CNTL, an MFW has formed along the remnant AR ~35°N, 128.5°W, which in conjunction with the extratropical cyclone to the west and its support of convergence along the existing cold-frontal zone, has created two distinct IVT corridors (Fig. 8j). The AR closest to California moves inland in CNTL, creating a second, stronger wave of AR conditions over the Russian River watershed with a maximum IVT ~393 kg m−1 s−1 (Fig. 9a). Neither the MFW nor the extratropical cyclone are evident in the noLH experiment. As a result, one broad, continuous AR extends over much of the central Pacific rather than two separate ARs (Fig. 8k). This single AR propagates eastward toward Northern California in noLH, but without the frontal wave aiding in the enhancement of IVT, AR conditions do not penetrate inland; the maximum IVT over the Russian River watershed associated with this second pulse is only ~227 kg m−1 s−1 (Fig. 9b).
In the 12 h leading up to MFW formation, a stronger 250-hPa jet in the noLH simulation (Figs. 10a,b) suggests an environment potentially more conducive for quasigeostrophic forcing for ascent; however, the jet offshore of Southern California has largely propagated out of the MFW development region by 0900 UTC 25 January (Figs. 9i,j). As with the December 2014 case, the most substantial differences between the CNTL and noLH simulations prior to MFW formation are in the lower troposphere with enhanced moisture, vorticity, and deformation along the frontal zone just upstream of the eventual MFW development location in CNTL (~31°N, 131°W) and lack thereof in noLH (Figs. 10k,l). While small maxima of lower-tropospheric cyclonic vorticity are evident in noLH around the CNTL MFW development area (Fig. 10l), the lack of sufficient moisture, moisture convergence, and associated diabatic feedbacks on the atmospheric circulation prevent the instabilities from strengthening enough to develop a MFW circulation at the surface (Fig. 8h).
500-hPa height (m; contoured every 60 m) and 250-hPa wind speed (m s−1; shaded) for the MPAS-A (a),(e),(i) CNTL and (b),(f),(j) noLH simulations; and 850-hPa potential temperature (θ; K; gray contours every 1 K), winds (m s−1; barbs), specific humidity (g kg−1; shaded), and relative vorticity (10−5 s−1; purple contours every 10 × 10−5 s−1 from 10 × 10−5 to 30 × 10−5 s−1 for the MPAS-A (c),(g),(k) CNTL and (d),(h),(l) noLH simulations at (a)–(d) 2100 UTC 24 Jan, (e)–(h) 0300 UTC 25 Jan, and (i)–(l) 0900 UTC 25 Jan 2010. The approximate location of MFW development at 1200 UTC 25 Jan 2010 in CNTL is labeled “W” in all panels for reference. The predevelopment area of interest is outlined in black in (k) and (l).
Citation: Monthly Weather Review 149, 8; 10.1175/MWR-D-20-0364.1
In both the CNTL and noLH simulations, there are two peaks in IVT over the Russian River watershed; the second peak, however, is weaker than the first in noLH, and likely due to the eastward propagation of the large-scale system (Figs. 9a,b). Therefore, while the MFW does not appear to be the sole reason for the second pulse of elevated IVT over the watershed, it is clear that the MFW and diabatic processes enhanced and extended AR conditions during this second wave. Over the course of the event, AR conditions exist over the Russian River watershed for only 6 h in the noLH experiment compared to a total of 30 h in CNTL (Table 1; Figs. 9a,b). With the exception of a slight increase in precipitation over the central Sierra Nevada and northern coast of the Russian River watershed, precipitation is substantially less across the region (Figs. 9c–e). Consequently, watershed precipitation totals are reduced by almost half (~49%) from ~70 mm (~2.6 in.) in CNTL to ~36 mm (~1.4 in.) in the noLH experiment (Table 1; Figs. 9c–e).
5. AR–MFW interactions in IVT and precipitation forcing
Since water vapor transport is most concentrated in the lower troposphere (Ralph et al. 2005), we can use the simulated 850-hPa wind speed and specific humidity to separately evaluate the differences in wind-driven (i.e., dynamic) and moisture-driven (i.e., thermodynamic) IVT components, respectively (e.g., Lavers et al. 2015). Doing so allows us to quantify which response dominates, and if the response is consistent throughout the events. Here, we consider the time of MFW development (2100 UTC 10 December 2014, 1200 UTC 15 January 2010) and time of maximum IVT over the Russian River watershed (1200 UTC 11 December 2014, 1800 UTC 25 January 2010) in each case.
In the vicinity of the MFW at the time of development for the December 2014 event, the 850-hPa wind speed and moisture content are reduced by ~19% (~4 m s−1) and ~28% (~1 g kg−1), respectively, indicating a larger thermodynamic component to the IVT differences at this time (Figs. 11i,m). IVT differences surrounding the MFW in the January 2010 case, on the other hand, are primarily dynamic in nature, as shown by the ~15% (~2 m s−1) reduction in 850-hPa wind speed compared to the ~4% (~0.2 g kg−1) reduction in moisture (Figs. 11k,o). Later in the event evolution, at the time of peak IVT over the Russian River watershed, we examine wind speed and moisture differences in a 5° × 5° latitude–longitude area surrounding the watershed. The difference in 850-hPa wind speed surpasses the difference in specific humidity, suggesting a larger wind-driven component to differences in IVT for both cases at this time; wind speeds are decreased by ~56% (~14 m s−1) and ~42% (~5 m s−1) in the December 2014 and January 2010 cases, respectively, compared to ~46% (~3 g kg−1) and ~23% (~1 g kg−1) reductions in specific humidity (Figs. 11j,l,m,p). Thus, removing diabatic effects alters IVT through dynamic and thermodynamic responses, although it is case and time dependent on which component dominates the change.
850-hPa specific humidity (shaded; g kg−1) and 850-hPa winds (barbs; m s−1) at the time of MFW development for the Dec 2014 (a) CNTL and (e) noLH simulations and for the Jan 2010 (c) CNTL and (g) noLH simulations. 850-hPa specific humidity (shaded; g kg−1) and 850-hPa winds (barbs; m s−1) at the time of peak IVT over the Russian River watershed for the Dec 2014 (b) CNTL and (f) noLH simulations and for the Jan 2010 (d) CNTL and (h) noLH simulations. (i)–(l) Difference (noLH minus CNTL) in 850-hPa wind speed (m s−1) at each time and for each event. (m)–(p) Difference (noLH minus CNTL) in 850-hPa specific humidity (g kg−1) at each time and for each event. Average percentage differences within an 8° × 8° latitude–longitude box surrounding the MFW are reported in (i), (k), (m), and (o). Average percentage differences within a 5° × 5° latitude–longitude box surrounding the Russian River watershed are reported in (j), (l), (n), and (p).
Citation: Monthly Weather Review 149, 8; 10.1175/MWR-D-20-0364.1
Moisture transport is a suitable indicator of orographic precipitation (e.g., Neiman et al. 2002, 2009; Ralph et al. 2013) and is primarily concentrated ~850 hPa (Ralph et al. 2005). The 850-hPa level (~1.5 km above mean sea level) is well above the topographic influence on wind convergence for the coastal mountain ranges of the Russian River watershed. Consequently, lower-tropospheric moisture convergence can be considered as a marker of dynamical precipitation forcing due to its generation of low-level PV and subsequent vertical motion, as shown by Demirdjian et al. (2020a). Therefore, to inform why precipitation differences occur between the CNTL and noLH experiments, we follow Demirdjian et al. (2020a) to break down precipitation forcing into orographic and dynamic components using the 850-hPa moisture transport and moisture convergence, respectively.
For the December 2014 case, the average 850-hPa moisture transport within the 5° × 5° latitude–longitude box surrounding the Russian River watershed (e.g., Fig. 11j) is generally higher in CNTL; over the course of the event, the 850-hPa moisture transport is ~26% greater (Figs. 12a,b). Similarly, the 850-hPa moisture convergence is consistently higher in CNTL with an average difference of ~43% over the 72-h precipitation accumulation period (Figs. 12c,d). Therefore, while removing the effects of latent heating weakens both orographic and dynamic precipitation forcings, the dynamic forcing is affected to a greater degree. Qualitatively, this result agrees with the absence of the secondary cyclone (and its support of rising motion) in the noLH run.
(top) Time series of average 850-hPa moisture transport (10−5 g kg−1 s−1) within the 5° × 5° latitude–longitude box surrounding the Russian River watershed (e.g., Fig. 11j) for CNTL (black line), noLH (blue line), and percentage differences (noLH minus CNTL; dashed red line) for (a),(b) Dec 2014 and (e),(f) Jan 2010 cases. (bottom) Time series of average 850-hPa moisture convergence (10−5 g kg−1 s−1) within the 5° × 5° latitude–longitude box surrounding the Russian River watershed (e.g., Fig. 11j) for CNTL (black line), noLH (blue line), and percentage differences (noLH minus CNTL; dashed red line) for (c),(d) Dec 2014 and (g),(h) Jan 2010 cases. Time-averaged percentage differences are reported in the top right of (b), (d), (f), and (h).
Citation: Monthly Weather Review 149, 8; 10.1175/MWR-D-20-0364.1
The January 2010 event also generally shows reduced 850-hPa moisture transport and moisture convergence over the 54-h precipitation accumulation period in noLH. The average percentage differences, however, are more comparable between the two forcings; moisture transport is reduced by ~26% in noLH compared to a ~28% reduction in moisture convergence (Figs. 12e,h). Consistent with Demirdjian et al. (2020a), it is clear latent heating impacts orographic precipitation forcing by altering moisture transport within the 850-hPa controlling layer as well as dynamic precipitation forcing through modifications in 850-hPa moisture convergence. As with changes in IVT, however, the dominating factor is not consistent across the evolution of a single event nor is the response the same between the two events presented here.
6. Summary and conclusions
It is well known that atmospheric rivers (ARs) represent a double-edged sword for many areas worldwide; these systems provide essential water resources, but threaten severe impacts if too extreme (e.g., Ralph et al. 2006; Dettinger et al. 2011; Lavers et al. 2011; Nayak and Villarini 2017; Viale et al. 2018). Additionally, it is well established that diabatic processes are crucial to the formation, evolution, and strengthening of many phenomena (e.g., Lackmann 2002; Davis 1992; Stoelinga 1996; Posselt and Martin 2004), including ARs (e.g., Cannon et al. 2020; Demirdjian et al. 2020a). Here, we focus specifically on how latent heating affects the interactions between mesoscale frontal waves (MFWs) and ARs, and ultimately the impacts on landfalling precipitation over the flood-vulnerable Russian River watershed in Northern California. We compare two AR events with different characteristics: 1) an extreme AR4 (Ralph et al. 2019) from 10 to 13 December 2014 accompanied by an MFW with upper-level support as well as low-level diabatic forcing that resulted in 178 mm (~7 in.) of precipitation over the Russian River watershed, causing the Russian River at Guerneville, California, to exceed flood stage at 10 m (33 ft), and 2) a weak AR1 (Ralph et al. 2019) from the 24–26 January 2010 that brought 81 mm (~3.2 in.) of precipitation to the Russian River watershed and was supported by a weak, primarily diabatically driven MFW. Despite their vastly different evolutions and impacts, both cases suffered from forecast errors related to the timing and magnitude of integrated vapor transport (IVT) and precipitation over the watershed region.
Conducting MPAS-A simulations with the effects of latent heating turned off in the microphysics, convective, and planetary boundary layer parameterization schemes removed or significantly weakened the MFW in each case, and consequently, diminished the AR–MFW interactions. As a result, both events were weaker in IVT magnitude, shorter in duration, and overall less-impactful ARs for the Russian River watershed. For the December 2014 case, AR conditions persisted for 6 h less and the peak IVT occurred 6 h earlier and was reduced by ~171 kg m−1 s−1, weakening the event to an AR3 (Ralph et al. 2019). Low-level circulations surrounding the MFW and secondary cyclone were considerably reduced, as was the lower-tropospheric moisture content, which together contributed to a decrease in moisture transport, moisture convergence, and IVT. In this case, differences in IVT were not consistently dynamic (i.e., wind-driven) or thermodynamic (i.e., moisture-driven); the moisture-driven component was stronger during MFW development while wind-driven differences dominated during landfall. As a result, orographic and dynamic precipitation forcing were weakened, and precipitation totals drastically reduced; precipitation over the Russian River watershed was reduced by 95 mm (~3.7 in.). Interestingly, due to the equatorward shift in the AR without latent heating and MFW development, stronger IVT occurred over areas in central and Southern California in the December 2014 noLH simulation.
Similarly, the January 2010 event was 24 h shorter in duration and precipitation totals were reduced by 33 mm (~1.2 in.) in noLH. While the first peak in IVT was only slightly weaker, the second peak was reduced by ~166 kg m−1 s−1 to subAR conditions (i.e., IVT < 250 kg m−1 s−1) and the timing was delayed by 18 h. No noticeable shift in the AR was evident during landfall, but a larger, more continuous IVT corridor over the central Pacific resulted in strengthened AR conditions offshore. As with the December 2014 case, removing latent heating weakened the lower-tropospheric circulations and moisture content in the vicinity of the MFW and Russian River watershed at landfall, thus reducing moisture transport, moisture convergence, IVT, and consequently, landfalling precipitation. In this instance, however, wind speed differences exceeded changes in moisture during MFW development and peak IVT intensity over the Russian River watershed; the percent changes in orographic and dynamic precipitation forcings were approximately equal.
As demonstrated, removing diabatic processes resulted in both dynamic and thermodynamic changes which influenced the development of mesoscale features, such as MFWs, and their interactions with surrounding systems, including ARs. The result was two events with vastly different characteristics and significantly underestimated impacts compared to reality, underlining the importance of diabatic processes in AR–MFW interactions. Consistent with theoretical frameworks, it is clear that the presence of instabilities along frontal zones that are strengthened via feedbacks between moisture convergence and latent heat release (e.g., Demirdjian et al. 2020b) are critical to frontal wave development. Even the December 2014 MFW with sufficient upper-air support failed to develop in the absence of latent heating, further reinforcing the importance of diabatic processes in the formation of these features. Therefore, sufficient representation of variables and processes contributing to latent heat release (e.g., water vapor content, vertical motion, microphysical processes, and their subsequent effects and feedbacks), and by extension, accurate initial conditions of the aforementioned variables, is imperative for the accurate representation of MFWs, AR evolution, and precipitation estimates, especially on a watershed scale. Moreover, the large diabatic component to MFW development and the AR–MFW interactions, at least for the two events presented here, has implications for possible increased frequency of AR–MFW events in a warmer, moister climate, which is an active area of future work.
To expand on and further quantify the results presented herein, future work will examine a larger sample of AR–MFW cases to better generalize the effects of latent heating on MFW development and on the AR–MFW interaction, as well as more closely examine the multiple growth mechanisms leading to MFW formation along ARs. Furthermore, additional observations of MFWs through field campaigns such as Atmospheric River Reconnaissance (AR Recon; Ralph et al. 2020) would be beneficial in furthering our understanding of AR–MFW interactions and thus informing future improvements to global forecast models.
Acknowledgments
This research was supported by the U.S. Army Corps of Engineers Forecast-Informed Reservoir Operations and California Department of Water Resources AR Program. M.A.F. was supported by the Rhodium Group as part of the Climate Impact Lab consortium. The MPAS-A and NCAR Command Language (NCL) are made available by the National Center for Atmospheric Research (NCAR), sponsored by the National Science Foundation (NSF). High-performance computing support from Cheyenne (https://doi.org/10.5065/D6RX99HX) was provided by NCAR’s Computational and Information System Laboratory, also sponsored by the NSF. Special thanks to Dr. Anna Wilson and Brian Kawzenuk for their help in the initial identification of AR–MFW events, and to Drs. Anna Wilson and Forest Cannon for fruitful discussions and constructive comments on earlier versions of this manuscript. We are grateful to Laura Fowler at NCAR for her guidance on removing latent heating from the MPAS-A simulations. We also thank three anonymous reviewers whose constructive feedback improved the quality of this manuscript.
Data availability statement
Model output from the simulations presented in this manuscript is located on the local CW3E computing cluster. Please contact the corresponding author for details on accessing these data.
REFERENCES
Adusumilli, S., M. A. Fish, H. A. Fricker, and B. Medley, 2021: Atmospheric river precipitation contributed to rapid increases in surface height of the West Antarctic Ice Sheet in 2019. Geophys. Res. Lett., 48, e2020GL091 076, https://doi.org/10.1029/2020GL091076.
Appenzeller, C., and H. C. Davies, 1996: PV morphology of a frontal-wave development. Meteor. Atmos. Phys., 58, 21–40, https://doi.org/10.1007/BF01027554.
Baehr, C., B. Pouponneau, F. Ayrault, and A. Joly, 1999: Dynamical characterization of the FASTEX cyclogenesis cases. Quart. J. Roy. Meteor. Soc., 125, 3469–3494, https://doi.org/10.1002/qj.49712556117.
Bjerknes, J., and H. Solberg, 1922: Life cycle of cyclones and the polar front theory of atmospheric circulation. Geophys. Publ., 3, 1–18.
Blier, W., and R. M. Wakimoto, 1995: Observations of the early evolution of an explosive oceanic cyclone during ERICA IOP 5. Part I: Synoptic overview and mesoscale frontal structure. Mon. Wea. Rev., 123, 1288–1310, https://doi.org/10.1175/1520-0493(1995)123<1288:OOTEEO>2.0.CO;2.
Bouniol, D., A. Protat, and Y. Lemaitre, 1999: Mesoscale dynamics of a deepening secondary cyclone in FASTEX IOP16: Three-dimensional structure retrieved from dropsonde data. Quart. J. Roy. Meteor. Soc., 125, 3535–3562, https://doi.org/10.1002/qj.49712556120.
Browning, K. A., and C. W. Pardoe, 1973: Structure of low-level jet streams ahead of mid-latitude cold fronts. Quart. J. Roy. Meteor. Soc., 99, 619–638, https://doi.org/10.1002/qj.49709942204.
Buzzi, A., N. Tartaglione, and P. Malguzzi, 1998: Numerical simulations of the 1994 Piedmont flood: Role of orography and moist processes. Mon. Wea. Rev., 126, 2369–2383, https://doi.org/10.1175/1520-0493(1998)126<2369:NSOTPF>2.0.CO;2.
Cannon, F., J. M. Cordeira, C. W. Hecht, J. R. Norris, A. Michaelis, R. Demirdjian, and F. M. Ralph, 2020: GPM satellite radar observations of precipitation mechanisms in atmospheric rivers. Mon. Wea. Rev., 148, 1449–1463, https://doi.org/10.1175/MWR-D-19-0278.1.
Chaboureau, J., and A. J. Thorpe, 1999: Frontogenesis and the development of secondary wave cyclones in FASTEX. Quart. J. Roy. Meteor. Soc., 125, 925–940, https://doi.org/10.1002/qj.49712555509.
Chang, S. W., T. R. Holt, and K. D. Sashegyi, 1996: A numerical study of the ERICA IOP 4 marine cyclone. Mon. Wea. Rev., 124, 27–46, https://doi.org/10.1175/1520-0493(1996)124<0027:ANSOTE>2.0.CO;2.
Cordeira, J. M., F. M. Ralph, and B. J. Moore, 2013: The development and evolution of two atmospheric rivers in proximity to western North Pacific tropical cyclones in October 2010. Mon. Wea. Rev., 141, 4234–4255, https://doi.org/10.1175/MWR-D-13-00019.1.
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.
Dacre, H. F., and S. L. Gray, 2006: Life-cycle simulations of shallow frontal waves and the impact of deformation strain. Quart. J. Roy. Meteor. Soc., 132, 2171–2190, https://doi.org/10.1256/qj.05.238.
Davis, C. A., 1992: A potential-vorticity diagnosis of the importance of initial structure and condensational heating in observed extratropical cyclogenesis. Mon. Wea. Rev., 120, 2409–2428, https://doi.org/10.1175/1520-0493(1992)120<2409:APVDOT>2.0.CO;2.
Davis, C. A., and K. A. Emanuel, 1991: Potential vorticity diagnostics of cyclogenesis. Mon. Wea. Rev., 119, 1929–1953, https://doi.org/10.1175/1520-0493(1991)119<1929:PVDOC>2.0.CO;2.
Demaria, E. M. C., F. Dominguez, H. Hu, G. von Glinski, M. Robles, J. Skindlov, and J. Walter, 2017: Observed hydrologic impacts of landfalling atmospheric rivers in the Salt and Verde River basins of Arizona, United States. Water Resour. Res., 53, 10 025–10 042, https://doi.org/10.1002/2017WR020778.
Demirdjian, R., J. D. Doyle, C. A. Reynolds, J. R. Norris, A. C. Michaelis, and F. M. Ralph, 2020a: A case study of the physical processes associated with the atmospheric river initial-condition sensitivity from an adjoint model. J. Atmos. Sci., 77, 691–709, https://doi.org/10.1175/JAS-D-19-0155.1.
Demirdjian, R., J. R. Norris, A. Martin, and F. M. Ralph, 2020b: Dropsonde observations of the ageostrophy within the pre-cold-frontal low-level jet associated with atmospheric rivers. Mon. Wea. Rev., 148, 1389–1406, https://doi.org/10.1175/MWR-D-19-0248.1.
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.
Du, Q., V. Faber, and M. Gunzburger, 1999: Centroidal Voronoi tessellations: Applications and algorithms. SIAM Rev., 41, 637–676, https://doi.org/10.1137/S0036144599352836.
Eiras-Barca, J., S. Brands, and G. Miguez-Macho, 2016: Seasonal variations in North Atlantic atmospheric river activity and associations with anomalous precipitation over the Iberian Atlantic margin. J. Geophys. Res., 121, 931–948, https://doi.org/10.1002/2015JD023379.
Fehlmann, R., and H. C. Davies, 1999: Role of salient potential-vorticity elements in an event of frontal-wave cyclogenesis. Quart. J. Roy. Meteor. Soc., 125, 1801–1824, https://doi.org/10.1002/qj.49712555716.
Fish, M. A., A. M. Wilson, and F. M. Ralph, 2019: Atmospheric river families: Definition and associated synoptic conditions. J. Hydrometeor., 20, 2091–2108, https://doi.org/10.1175/JHM-D-18-0217.1.
Gershunov, A., T. Shulgina, F. M. Ralph, D. A. Lavers, and J. J. Rutz, 2017: Assessing the climate scale variability of atmospheric rivers affecting western North America. Geophys. Res. Lett., 44, 7900–7908, https://doi.org/10.1002/2017GL074175.
Gorodetskaya, I. V., M. Tsukernik, K. Claes, F. M. Ralph, W. D. Neff, and N. P. M. Van Lipzig, 2014: The role of atmospheric rivers in anomalous snow accumulation in East Antarctica. Geophys. Res. Lett., 41, 6199–6206, https://doi.org/10.1002/2014GL060881.
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.
Hecht, C. W., and J. M. Cordeira, 2017: Characterizing the influence of atmospheric river orientation and intensity on precipitation distributions over North Coastal California. Geophys. Res. Lett., 44, 9048–9058, https://doi.org/10.1002/2017GL074179.
Hersbach, H., and Coauthors, 2020: The ERA5 global reanalysis. Quart. J. Roy. Meteor. Soc., 146, 1999–2049, https://doi.org/10.1002/qj.3803.
Hewson, T. D., 2009: Diminutive frontal waves—A link between fronts and cyclones. J. Atmos. Sci., 66, 116–132, https://doi.org/10.1175/2008JAS2719.1.
Hirota, N., Y. N. Takayabu, M. Kato, and S. Arakane, 2016: Roles of an atmospheric river and a cutoff low in the extreme precipitation event in Hiroshima on 19 August 2014. Mon. Wea. Rev., 144, 1145–1160, https://doi.org/10.1175/MWR-D-15-0299.1.
Hjelmfelt, M. R., and R. R. Braham Jr., 1983: Numerical simulation of the airflow over Lake Michigan for a major lake-effect snow event. Mon. Wea. Rev., 111, 205–219, https://doi.org/10.1175/1520-0493(1983)111<0205:NSOTAO>2.0.CO;2.
Hoskins, B. J., M. E. McIntyre, and A. W. Robertson, 1985: On the use and significance of isentropic potential vorticity maps. Quart. J. Roy. Meteor. Soc., 111, 877–946, https://doi.org/10.1002/qj.49711147002.
Joly, A., and A. J. Thorpe, 1990: Frontal instability generated by tropospheric potential vorticity anomalies. Quart. J. Roy. Meteor. Soc., 116, 525–560, https://doi.org/10.1002/qj.49711649302.
Joly, A., and Coauthors, 1997: The Fronts and Atlantic Storm-Track Experiment (FASTEX): Scientific objectives and experimental design. Bull. Amer. Meteor. Soc., 78, 1917–1940, https://doi.org/10.1175/1520-0477(1997)078<1917:TFAAST>2.0.CO;2.
Kamae, Y., W. Mei, and S. Xie, 2017: Climatological relationship between warm season atmospheric rivers and heavy rainfall over East Asia<