Abstract

Record-setting rainfall occurred over the state of South Carolina in early October 2015, with maximum accumulations exceeding 500 mm. During the heavy rainfall, Hurricane Joaquin was located offshore to the southeast of the flooding event. Prior research, storm summaries, satellite imagery, and media accounts suggest that Joaquin played a major role in the flooding, mostly through the provision of additional water vapor. Here, numerical simulations are utilized to elucidate Joaquin’s role in the flooding and to diagnose moisture transport mechanisms. The South Carolina precipitation event and the track of Hurricane Joaquin are reasonably represented by two control simulations, a 36-km simulation without nesting and another with 12- and 4-km nests added; the latter improves upon a negative intensity bias for Joaquin. A band of intense moisture transport into the flooding region is associated with a narrow, diabatically produced cyclonic lower-tropospheric potential vorticity (PV) maximum. Simulations in which Joaquin is removed exhibit a similar moisture transport mechanism and also produce a band of heavy precipitation, though the axis of heaviest precipitation shifts northward into North Carolina, and there is a modest reduction (~7%) in area-averaged rainfall. Removing Joaquin produces negligible changes in regional total water vapor content but diminished upper-tropospheric diabatic outflow. The diminished outflow allows greater eastward progression of an upper-level trough, consistent with the northward precipitation shift and with weaker forcing for ascent. Changes in the upper jet associated with Joaquin appear to exert a greater influence on the flooding event than Joaquin’s contribution to water vapor content.

1. Introduction

Record flooding, featuring peak 5-day precipitation totals of over 500 mm, took place over the state of South Carolina during the first five days of October 2015. The flooding was the result of a complex synoptic weather pattern that featured a slow-moving upper-level trough over the southeastern United States, unseasonably warm, moist tropical air, and Hurricane Joaquin, which was located over the Bahamas at the time of the most intense flooding. Storm summaries and media coverage of the event suggested a link between Joaquin and the flooding, and in some cases indicated that Joaquin had served as a moisture source for the flooding rainfall (e.g., NWS 2015; NHC 2016). Other factors, such as a coastal front and Appalachian cold-air damming may have contributed as well, but these aspects are beyond the scope of this study. The goals of this paper are to 1) document the moisture transport mechanisms active during this event, and 2) quantify and elucidate Joaquin’s role in the flooding over South Carolina.

Heavy precipitation events that take place in the downstream vicinity of tropical cyclones have been dubbed predecessor rain events or (PREs) by Cote (2007), Galarneau et al. (2010), and others. PREs are defined as “meso- and subsynoptic-scale regions of high-impact heavy rainfall that occur well in advance of recurving tropical cyclones (TCs) over the eastern third of the United States” (Galarneau et al. 2010, p. 3272). Specific association between the accompanying TC and the rain event is included in the PRE definition: “A distinguishing feature of PREs is that they are sustained by deep tropical moisture that is transported poleward directly from the TC” (Galarneau et al. 2010, p. 3272). The TC is hypothesized to contribute to the heavy precipitation in several distinct ways: by acting as a moisture source, by enhancing synoptic-scale lift through diabatic strengthening of an upper outflow jet, and via warm advection or isentropic lift as the storm circulation encounters sloping isentropes in a baroclinic zone (e.g., Colle 2003; Srock and Bosart 2009). In some cases, it is not clear if the PRE is directly attributable to the TC, while in other events a connection is clear (e.g., Schumacher et al. 2011).

Because TCs feature strongly convergent flow in the lower troposphere, and because precipitation greatly exceeds evaporation in the vicinity of a TC (e.g., Kurihara 1975; Marks and Houze 1987; Braun 2006; Yang et al. 2011), it is unclear through what mechanisms TCs “export” moisture to downstream regions during a PRE. A broad TC wind field could enhance upward turbulent moisture fluxes within the planetary boundary layer over a large area, and the portion of this additional vapor that is not consumed by the TC could potentially enhance vapor content in remote locations, given favorable advection patterns. Budget studies, however, indicate that the majority of water vapor evaporated from the sea surface is fed into the TC and removed as precipitation (e.g., Kurihara 1975; Braun 2006; Fritz and Wang 2014; Huang et al. 2014). However, during extratropical transition or at landfall, TC water budget characteristics may change (Yang et al. 2011). At upper levels, outflow jets, which result from strong condensational heating associated with the TC, can contribute to synoptic-scale ascent, particularly in the right-jet entrance [see Figs. 5.1 and 5.2 in Cote (2007) or Fig. 24 in Galarneau et al. (2010)].

Our working hypothesis is that even without Joaquin, heavy precipitation would have taken place in the southeastern United States because of the exceptionally moist air that was in place ahead of a slow-moving upper-tropospheric trough. We address the following questions: Did Joaquin export water vapor to the region of heavy precipitation during the flooding event? And, did upper-level outflow from Joaquin enhance synoptic-scale and mesoscale lift over the flooding region? To test our hypothesis and address these questions, simulations of this event are performed with and without Joaquin using the Weather Research and Forecasting (WRF) Model (Skamarock and Klemp 2008).

2. Materials and methods

a. Data sources

To document large-scale moisture transport across the southeastern United States, we obtained 0.25° analyses from the Global Forecast System (GFS; see NOAA/National Centers for Environmental Prediction 2000) model, along with 0.7° reanalysis data from the European Centre for Medium-Range Weather Forecasts (ECMWF) interim reanalysis dataset (ERA-Interim, hereafter ERA-I; Dee et al. 2011). We also utilized analyses from the Rapid Update Cycle (RUC; Benjamin et al. 2004) model, interpolated to a 40-km grid, to provide an independent evaluation of our control simulation. For precipitation accumulation, we utilized the Stage-IV gridded precipitation product (Lin and Mitchell 2005). National Hurricane Center “best track” data were used in evaluation of the track and intensity of Hurricane Joaquin in our model simulations (Torn and Snyder 2012; NHC 2016).

b. Numerical model and configuration

We used the Advanced Research version of WRF (WRF-ARW), version 3.7.1, to conduct our numerical experiments. Initial simulations were performed on a regional domain using 36-km horizontal grid spacing, with 50 vertical levels and a model top of 10 hPa; the domain corresponds to the area shown in Fig. 1. To reduce a negative bias in the intensity of Joaquin, we also performed simulations with added 12- and 4-km two-way nests (Fig. 1a). Each simulation was initialized at 0000 UTC 1 October 2015 and run for 120 h. Longwave and shortwave radiation were parameterized using the Rapid Radiative Transfer Model for global climate models (RRTMG). The Thompson six-class microphysics scheme, which includes graupel and is coupled directly with the RRTMG radiation scheme, was used to represent microphysical processes (Thompson and Eidhammer 2014; Fovell et al. 2016), while the boundary layer was parameterized using the Yonsei University (YSU) scheme. Following Bassill (2014, 2015), the modified Tiedtke scheme, which accounts for convective momentum adjustment, was used to handle subgrid-scale convection on the outer two domains; no convective parameterization was employed on the 4-km domain.

Fig. 1.

Sea level pressure (black contours; interval 4 hPa), 500-hPa geopotential height (dashed, dark blue contours; interval 6 dam), and 950–500-hPa layer-average potential vorticity (PVU; 1 PVU = 10−6 K kg−1 m2 s−1; shaded as in legend at bottom of panels) at model initialization time of 0000 UTC 1 Oct 2015 for the (a) control and (b) No-TC simulation on the 36-km domain. Dotted red lines in (a) denote locations of 12- and 4-km nested domains.

Fig. 1.

Sea level pressure (black contours; interval 4 hPa), 500-hPa geopotential height (dashed, dark blue contours; interval 6 dam), and 950–500-hPa layer-average potential vorticity (PVU; 1 PVU = 10−6 K kg−1 m2 s−1; shaded as in legend at bottom of panels) at model initialization time of 0000 UTC 1 Oct 2015 for the (a) control and (b) No-TC simulation on the 36-km domain. Dotted red lines in (a) denote locations of 12- and 4-km nested domains.

Initial and lateral boundary conditions were provided by the National Centers for Environmental Prediction (NCEP) Global Data Assimilation System (GDAS) 0.25° Final Analysis (NOAA/National Centers for Environmental Prediction 2000). Each simulation also utilized the Diabatic Digital Filter Initialization (DDFI) feature in WRF (Peckham et al. 2016), which was especially important for establishing dynamical balance in simulations in which the model initial conditions were altered to remove Joaquin. We utilized a control simulation with the model configuration described above and an otherwise identical “No-TC” simulation in which Hurricane Joaquin was removed from the model initial conditions as described below. Both the control and No-TC simulations were run with the 36-/12-/4-km nesting configuration.

c. TC removal

Our experimental objective is similar to that of Grams and Blumer (2015) and Grams and Archambault (2016), but here, the TC vortex is isolated and removed using a combination of geostrophic vorticity inversion and anomaly removal. The inversion is performed using a standard successive overrelaxation technique. Geostrophic vorticity is partitioned and vorticity anomalies associated with Joaquin, defined here as geostrophic vorticity greater (less) than s−1 ( s−1) within 600 km of the analyzed position of the TC, are omitted from the inversion to recover the geopotential height field without Joaquin (Fig. 1b). The difference between the original geopotential height field and that associated with all non-Joaquin vorticity represents Joaquin. We experimented with larger and smaller radii to isolate the geostrophic vorticity associated with the TC and test sensitivity. A radius of 600 km is the smallest radius that 1) removes the TC when visually inspecting the geopotential height field at all standard isobaric levels after the inversion, and 2) precluded development of a TC when using the modified initial conditions in the No-TC simulation. Temperature, wind, sea level pressure, and surface pressure perturbations are computed as the deviations from the zonal average in each field. Here, the zonal averages are computed along each longitudinal band within the outermost domain. The temperature and wind perturbations are computed at each isobaric level while the surface and sea level pressure perturbations are computed at single levels. These fields, unlike geopotential height, are then weighted with an inverse exponential dependence within 600 km of the center of Joaquin, with weights given by

 
formula

where is the weight at a radial distance (km), from the center of Joaquin. The weighted perturbations are then subtracted from the full field for each respective variable in the initial conditions to remove Joaquin. Relative humidity is held constant in this process; through the removal of the positive temperature perturbations associated with warm-core Joaquin, atmospheric moisture content decreases in the No-TC simulation in accord with the Clausius–Clapeyron relation (~7% K−1). The use of DDFI ensures that the mass and thermodynamic fields are in balance at model initialization despite these alterations to the initial fields in the No-TC simulation. Comparison of sea level pressure, upper-tropospheric geopotential heights, and lower-tropospheric potential vorticity (PV) in the control and No-TC simulations indicates a successful removal of Joaquin from the initial conditions (Fig. 1).

Removing Joaquin from the initial conditions as early as we did allows us to not only investigate how much, if any, moisture is transported downstream from the TC, but also to identify the impacts of the TC outflow on the upper-tropospheric dynamics in this event. However, one inherent limitation of this technique is that it obscures possible contributions from Joaquin that evolved over time and prior to our model initialization time.

3. Results

a. Antecedent conditions and synoptic-scale evolution

Inspection of vertically integrated precipitable water (PWAT) and sea surface temperature (SST) anomalies, computed using data from ERA-I, reveal an antecedent environment conducive to heavy precipitation. Atmospheric water vapor content across the southeastern United States and adjacent western North Atlantic was anomalously high during the last week of September 2015 (Fig. 2). Consistent with this, persistent easterly flow and SST values that were 1°–3°C warmer than climatology (Fig. 3) aided in producing positive precipitable water anomalies over the southeastern United States.

Fig. 2.

PWAT anomalies, defined as departure from a weighted 1979–2012 monthly climatology (mm; shaded as in legend) and 850-hPa winds (barbs) from ERA-I for (a) 0000 UTC 26 Sep, (b) 0000 UTC 28 Sep, (c) 0000 UTC 30 Sep, and (d) 0000 UTC 2 Oct 2015.

Fig. 2.

PWAT anomalies, defined as departure from a weighted 1979–2012 monthly climatology (mm; shaded as in legend) and 850-hPa winds (barbs) from ERA-I for (a) 0000 UTC 26 Sep, (b) 0000 UTC 28 Sep, (c) 0000 UTC 30 Sep, and (d) 0000 UTC 2 Oct 2015.

Fig. 3.

As in Fig. 2, but for SST anomalies (K; shaded as in legend).

Fig. 3.

As in Fig. 2, but for SST anomalies (K; shaded as in legend).

A complex synoptic and mesoscale evolution played out during this event as summarized in Fig. 4, taken from the 36-km domain of the control simulation. Comparison of this simulation to observed radar and RUC analyses (Fig. 5) allows evaluation of the control simulation.

Fig. 4.

Synoptic overview of nested control simulation on the 36-km domain. (left) 250-hPa height (red; contour interval 12 dam), 250-hPa isotachs (m s−1; shaded as in legend), sea level pressure (black; contour interval 4 hPa), 250-hPa divergence (blue contours; interval 5 × 10−5 s−1, zero contour omitted), and ageostrophic wind (vectors; larger than 15 ms−1). (right) Sea level pressure (black; contour interval 2 hPa), simulated composite reflectivity (shaded as in legend), 2-m potential temperature (dashed maroon; contour interval 5 K), and near-surface frontogenesis [red solid contours; interval 5 K (100 km)−1 (3h)−1, computed using 2-m temperature and 10-m wind]. (a),(b) 24-h simulation valid at 0000 UTC 2 Oct; (c),(d) 60-h simulation valid at 1200 UTC 3 Oct; (e),(f) 72-h simulation valid at 0000 UTC 4 Oct; and (g),(h) 90-h simulation valid at 1800 UTC 4 Oct.

Fig. 4.

Synoptic overview of nested control simulation on the 36-km domain. (left) 250-hPa height (red; contour interval 12 dam), 250-hPa isotachs (m s−1; shaded as in legend), sea level pressure (black; contour interval 4 hPa), 250-hPa divergence (blue contours; interval 5 × 10−5 s−1, zero contour omitted), and ageostrophic wind (vectors; larger than 15 ms−1). (right) Sea level pressure (black; contour interval 2 hPa), simulated composite reflectivity (shaded as in legend), 2-m potential temperature (dashed maroon; contour interval 5 K), and near-surface frontogenesis [red solid contours; interval 5 K (100 km)−1 (3h)−1, computed using 2-m temperature and 10-m wind]. (a),(b) 24-h simulation valid at 0000 UTC 2 Oct; (c),(d) 60-h simulation valid at 1200 UTC 3 Oct; (e),(f) 72-h simulation valid at 0000 UTC 4 Oct; and (g),(h) 90-h simulation valid at 1800 UTC 4 Oct.

Fig. 5.

Observed composite radar reflectivity (dBZ; shaded as in legend) and RUC analyses, including sea level pressure (black; contour interval 2 hPa), 2-m potential temperature (dashed maroon; contour interval 5 K), and near-surface frontogenesis [red solid contours; interval 5 K (100 km)−1 (3 h) −1, computed using 2-m temperature and 10-m wind]. Valid at (a) 0000 UTC 2 Oct, (b) 1200 UTC 3 Oct, (c) 0000 UTC 4 Oct, and (d) 1800 UTC 4 Oct.

Fig. 5.

Observed composite radar reflectivity (dBZ; shaded as in legend) and RUC analyses, including sea level pressure (black; contour interval 2 hPa), 2-m potential temperature (dashed maroon; contour interval 5 K), and near-surface frontogenesis [red solid contours; interval 5 K (100 km)−1 (3 h) −1, computed using 2-m temperature and 10-m wind]. Valid at (a) 0000 UTC 2 Oct, (b) 1200 UTC 3 Oct, (c) 0000 UTC 4 Oct, and (d) 1800 UTC 4 Oct.

At hour 24 of the simulation, valid at 0000 UTC 2 October, the Carolinas and Virginia are located beneath the right-entrance region of an anticyclonically curved upper jet at the 250-hPa level, which is associated with ageostrophic flow toward lower geopotential height and divergence aloft over the Carolinas (Fig. 4a). A slow-moving upper-level trough is evident in the 250-hPa height field, with the axis at this time extending from Alabama into the northern Gulf of Mexico. In the lower troposphere, a frontal trough extends southwestward from near the Canadian Maritime provinces to the North Carolina coast, with precipitation falling along the cold side of the frontal zone (Fig. 4b). There is some evidence for cold-air damming over the Carolinas and Virginia, with an inverted ridge in the sea level pressure field and relatively cool near-surface potential temperature. Hurricane Joaquin is located near the Bahamas.

By hour 60 of the simulation, valid at 1200 UTC 3 October, the 250-hPa jet has advanced westward, accompanied by a band of divergence aloft that aligns with the right entrance (Fig. 4c). The upper-level trough exhibits a closed contour at this time, with strongly subgeostrophic flow in the base of this trough across the northeastern Gulf of Mexico. Consistent with the upper-level divergence, an elongated band of enhanced composite radar reflectivity extends from western North Carolina, across South Carolina, and offshore (Fig. 4d). Along the immediate coast of South Carolina, strong near-surface frontogenesis is observed, at this time oriented roughly perpendicular to the band of heavy rainfall. The inverted trough that was located offshore at hour 24 has progressed inland, with a zone of strong easterly or southeasterly flow now extending into the Carolinas, as implied by the sea level isobar pattern. By this time, Hurricane Joaquin has moved east of the Bahamas.

Between simulation hours 60 and 72 (1200 UTC 3 October–0000 UTC 4 October), the trough at the 250-hPa level remains nearly stationary, and the jet located along the northeastern flank of this trough weakens (Figs. 4c,e). A band of strong divergence aloft still extends across South Carolina, consistent with very heavy precipitation falling across the state at that time (Figs. 4e,f and 5b,c). Strong easterly geostrophic flow is present across North Carolina, and an elongated low pressure center has formed within the southeast–northwest-oriented trough extending across South Carolina.

By hour 90, valid at 1800 UTC 4 October, the upper trough has cut off at the 250-hPa level and is centered near the east coast of Florida (Fig. 4g). Heavy precipitation is still evident in the simulation at this time (Fig. 4h) and in observations (Fig. 5d) over eastern South Carolina.

The RUC-analyzed sea level pressure, near-surface frontogenesis, and observed composite radar reflectivity are displayed in Fig. 5 in order to allow an evaluation of our control simulation. Because our simulation utilizes initial and boundary conditions provided by GFS analyses, comparison with RUC analyses allows a more independent evaluation. Specifically, Fig. 5 corresponds to Figs. 4b, 4d, 4f, and 4h, although with a reduced domain owing to the lack of observed radar coverage offshore. Observed precipitation coverage across southern Georgia was greater in the observations than in the control simulation at 0000 UTC 2 October (Figs. 5a, 4b), but by 1200 UTC 3 October stronger agreement was evident, with a band of heavy rainfall covering central South Carolina in both observations and the simulation (Figs. 5b, 4d). The inland migration of the inverted trough that was noted in the control simulation is also clearly evident in the RUC-analyzed sea level pressure field (Fig. 5b). The band of coast-parallel near-surface frontogenesis seen in hour 60 of the control simulation is also evident in the RUC analysis valid at this time (Fig. 5b); this band rotates to become more parallel to the axis of heavier precipitation at both 0000 and 1800 UTC 4 October (Figs. 5c,d). The counterclockwise rotation and northeastward migration of the heavy precipitation band indicates qualitative consistency between the control simulation and observations (Figs. 4f,h and 5c,d).

b. Representation of Joaquin and heavy rainfall in the Carolinas in the control simulation

Comparison of Joaquin’s track in a 36-km nonnested control simulation with the National Hurricane Center (NHC) best track dataset indicates a root-mean-square track deviation error of ~75 km (Fig. 6a) and a large negative intensity bias (Fig. 6b). Therefore, we ran additional simulations featuring 12- and 4-km grid lengths. Examination of Fig. 6a indicates that the nonnested control simulation exhibited a slight westward bias, whereas the nested control simulation featured an eastward deviation. Initializing at earlier times (0000 UTC 30 September and 0000 UTC 29 September) resulted in larger track errors and featured Joaquin making landfall in the United States (not shown). The intensity of Joaquin, defined here using interpolated minimum sea level pressure, is significantly improved with the addition of 12- and 4-km nests (Fig. 6b). The negative intensity bias is likely in part due to insufficient storm intensity in the GFS-analyzed initial conditions.

Fig. 6.

(a) Track of Hurricane Joaquin in the analyzed NHC best track data (solid black line), nonnested control simulation (solid red line), and 36-km domain of the nested control simulation (solid green line) from 0000 UTC 1 Oct to 0000 UTC 6 Oct 2015. (b) As in (a), except for a time series of minimum SLP (hPa).

Fig. 6.

(a) Track of Hurricane Joaquin in the analyzed NHC best track data (solid black line), nonnested control simulation (solid red line), and 36-km domain of the nested control simulation (solid green line) from 0000 UTC 1 Oct to 0000 UTC 6 Oct 2015. (b) As in (a), except for a time series of minimum SLP (hPa).

Figure 7 presents a comparison of simulated precipitation from the 4-km grid with the Stage-IV precipitation analysis [a gridded multisensor precipitation estimate produced from radar- and gauge-based observations; Lin and Mitchell (2005)], for the period from 0000 UTC 1 October to 0000 UTC 6 October. Despite a northeastward shift of the heaviest rainfall in the simulation, this comparison demonstrates strong qualitative agreement. The control simulation reproduces a large swath of precipitation exceeding 200 mm through much of coastal South Carolina, with an axis of greater than 500 mm focused in northeastern South Carolina; this simulated precipitation maximum corresponds to the observed area of heavy rainfall that ultimately resulted in catastrophic flooding, albeit shifted slightly north and east in the simulation. Given that the foci of the present study are to investigate moisture transport and the role that Joaquin played in this extreme precipitation, the control simulation is deemed satisfactory for our purposes.

Fig. 7.

Total precipitation (mm) from 0000 UTC 1 Oct to 0000 UTC 6 Oct 2015 in the (a) Stage-IV analysis and (b) control simulation on the 4-km domain.

Fig. 7.

Total precipitation (mm) from 0000 UTC 1 Oct to 0000 UTC 6 Oct 2015 in the (a) Stage-IV analysis and (b) control simulation on the 4-km domain.

c. Moisture transport analysis of control simulation

The results presented here are largely consistent using data from the GFS, RUC, ERA-I, or our WRF control simulation, but we elect to present the latter to facilitate consistency in comparing with the No-TC simulation subsequently. At 0000 UTC 1 October, the time of model initialization, a region of PWAT in excess of 50 mm extends from the vicinity of Hurricane Joaquin in the Bahamas, north and east along the U.S. East Coast and over the New England coastal waters (Fig. 8a). By hour 48 of the simulation, valid at 0000 UTC 3 October, a band of PWAT exceeding 60 mm extends northward from the center of Joaquin into the Carolinas and Virginia (Fig. 8c). As Joaquin subsequently moves northward, the axis of greatest PWAT rotates counterclockwise, extending northwest, then west, and finally southwest from Joaquin into the Carolinas (Figs. 8d–f).

Fig. 8.

PWAT (mm; shaded as in legend), 500-hPa height (red contours; 6-dam interval), and sea level pressure (black contours; 4-hPa interval) for the 36-km domain of the nested control simulation: (a) initial condition (0000 UTC 1 Oct), (b) 24-h simulation valid at 0000 UTC 2 Oct, (c) 48-h simulation valid at 0000 UTC 3 Oct, (d) 72-h simulation valid at 0000 UTC 4 Oct, (e) 96-h simulation valid at 0000 UTC 5 Oct, and (f) 120-h simulation valid at 0000 UTC 6 Oct.

Fig. 8.

PWAT (mm; shaded as in legend), 500-hPa height (red contours; 6-dam interval), and sea level pressure (black contours; 4-hPa interval) for the 36-km domain of the nested control simulation: (a) initial condition (0000 UTC 1 Oct), (b) 24-h simulation valid at 0000 UTC 2 Oct, (c) 48-h simulation valid at 0000 UTC 3 Oct, (d) 72-h simulation valid at 0000 UTC 4 Oct, (e) 96-h simulation valid at 0000 UTC 5 Oct, and (f) 120-h simulation valid at 0000 UTC 6 Oct.

Also evident in Fig. 8 is the aforementioned inverted trough in the sea level pressure field, which advances inland between hours 48 and 72 of the simulation (0000 UTC 3 October–0000 UTC 4 October). The 500-hPa height field reveals the presence of a slow-moving upper trough over the southeastern United States; to the east of this trough, a veering geostrophic wind profile between the surface and 500-hPa level is evident, consistent with warm advection and synoptic-scale forcing for ascent (e.g., Figs. 8b,c).

The sequence of moisture flux plots from hours 48 to 96 (0000 UTC 3 October–0000 UTC 5 October), a period encompassing the heaviest rainfall in the flooding region, demonstrates a continuous, strong moisture flux into South Carolina (Fig. 9). A possible connection between Joaquin and the heavy precipitation over South Carolina is suggested by the continuous band of enhanced vapor flux between these two locations. The lower-tropospheric PV exhibits a structure similar to that seen in other diabatically driven low-level jets (LLJ), sometimes accompanying atmospheric river events; that is, the strongest moisture flux and LLJ are located immediately adjacent to a lower-tropospheric cyclonic PV maximum (Lackmann and Gyakum 1999; Lackmann 2002). The same general structure remains in place between hours 72 and 96 (Figs. 9b–d). Cross sections taken through these elongated PV maxima demonstrate that they are the result of condensational heating and that they are isolated from the stratospheric PV reservoir (Fig. 10). The diabatic PV tendency, computed from a simplified parameterization of condensational heating as in Lackmann [2002, see his Eqs. (1)–(4)] and Emanuel et al. (1987), confirms positive nonadvective PV tendencies collocated with the lower cyclonic maximum, with negative tendencies above (Fig. 10b).

Fig. 9.

Moisture flux diagnosis for the 36-km domain of the nested control simulation. The 900-hPa moisture flux (×10−2 m s−1; shaded as in legend), moisture flux vectors, and lower-tropospheric PV (brown contours; shaded every 0.5 PVU beginning at 0.75 PVU). (a) The 48-h simulation valid at 0000 UTC 3 Oct, (b) 72-h simulation valid at 0000 UTC 4 Oct, (c) 84-h simulation valid at 1200 UTC 4 Oct, and (d) 96-h simulation valid at 0000 UTC 5 Oct.

Fig. 9.

Moisture flux diagnosis for the 36-km domain of the nested control simulation. The 900-hPa moisture flux (×10−2 m s−1; shaded as in legend), moisture flux vectors, and lower-tropospheric PV (brown contours; shaded every 0.5 PVU beginning at 0.75 PVU). (a) The 48-h simulation valid at 0000 UTC 3 Oct, (b) 72-h simulation valid at 0000 UTC 4 Oct, (c) 84-h simulation valid at 1200 UTC 4 Oct, and (d) 96-h simulation valid at 0000 UTC 5 Oct.

Fig. 10.

Cross-sectional analysis corresponding to line A–B in Fig. 9b valid at hour 72 of the 36-km domain of the nested control simulation: (a) PV (red contours; PVU), section-normal moisture transport (×10−2 m s−1; shaded as in legend; negative values consistent with flow into section); (b) PV tendency diagnostics, including PV (contoured and shaded as in legend), parameterized latent heating (blue dashed contours every 100 × 10−5 K s−1), nonadvective PV flux vectors, and nonadvective PV tendency (×10−11 PVU s−1; green solid contours for positive values, brown dashed contours for negative values).

Fig. 10.

Cross-sectional analysis corresponding to line A–B in Fig. 9b valid at hour 72 of the 36-km domain of the nested control simulation: (a) PV (red contours; PVU), section-normal moisture transport (×10−2 m s−1; shaded as in legend; negative values consistent with flow into section); (b) PV tendency diagnostics, including PV (contoured and shaded as in legend), parameterized latent heating (blue dashed contours every 100 × 10−5 K s−1), nonadvective PV flux vectors, and nonadvective PV tendency (×10−11 PVU s−1; green solid contours for positive values, brown dashed contours for negative values).

d. Analysis of the role of Joaquin

The removal of Joaquin produced changes in the simulated intensity and location of heavy precipitation (Fig. 11). While the axis of heaviest precipitation in the control simulation (Fig. 7b) was centered over South Carolina, this feature shifts northeastward into North Carolina in the No-TC simulation (Fig. 11a). The region of heaviest precipitation in the No-TC simulation is also narrower relative to that in the control simulation (Figs. 7b, 11b).

Fig. 11.

As in Fig. 7, but for (a) the 36-km domain of the nested No-TC simulation and (b) change in precipitation with Joaquin removed (No-TC minus control; mm; shaded as in legend).

Fig. 11.

As in Fig. 7, but for (a) the 36-km domain of the nested No-TC simulation and (b) change in precipitation with Joaquin removed (No-TC minus control; mm; shaded as in legend).

To isolate Joaquin’s effect on water vapor content over the southeastern United States, we computed histograms of total precipitation (Fig. 12a) and PWAT (Fig. 12c) for the region shown in Figs. 7 and 11 on the 4-km simulation domain. A reduction in total precipitation is found when Joaquin is removed, with time–area-integrated precipitation for this region decreasing ~7%. A more dramatic change is seen in the heaviest precipitation totals as shown in Fig. 12b. Precipitation totals exceeding or equal to 350 mm are much less frequent in the No-TC simulation and decrease ~40% when integrated over time and area. This is somewhat consistent with what is seen in the histogram of PWAT (Fig. 12c), which shows that large PWAT values in the tail of the distribution become less frequent without Joaquin, while values within the center of the distribution become more frequent. PWAT values greater than or equal to 60 mm particularly decrease, which is consistent with less extreme precipitation in the No-TC simulation. However, computing “total moisture availability,” here defined as the sum of time–area-integrated precipitation and time–area-integrated PWAT, reveals little change (less than a tenth of a percent decrease). This finding suggests that the presence of Joaquin was crucial for the extreme values of precipitation and PWAT seen, but not for total moisture availability. To better understand the reduction of precipitation and extreme precipitation, we analyzed vertical motion within this same region for the 4-km domain. Analysis of vertical velocity in the 4-km domain is particularly useful here given that all precipitation is explicitly computed on the model grid (e.g., no convective parameterization is implemented on this domain). Overall, the histograms of 700-hPa vertical velocity (Fig. 12d) indicate that the largest values of upward vertical velocity (negative omega) are reduced in the No-TC simulation. Although the magnitude of time–area-integrated vertical velocity only decreases a meager ~1% in the No-TC simulation, more dramatic change is seen in the strongest upward vertical motion. Upward vertical velocities less than −0.75 Pa s−1 are integrated in area and time for both simulations to quantify differences in the most vigorous upward vertical motion. Consistent with the reduction in total and heavy precipitation, the calculated time–area integral of strongest upward vertical velocity is found to decrease in magnitude by ~8.5%. This provides evidence that the removal of Joaquin had modified the downstream environment in such a way to weaken upward vertical motion.

Fig. 12.

Histograms of (a) total precipitation (mm), (b) extreme values of total precipitation (mm), (c) PWAT (mm), and (d) 700-hPa vertical velocity (10−2 Pa s−1) within the region shown in Figs. 7 and 11 for the 4-km domain. Data associated with each simulation are as shown in the legends. Note that the ordinate axis in (d) utilizes a log scale.

Fig. 12.

Histograms of (a) total precipitation (mm), (b) extreme values of total precipitation (mm), (c) PWAT (mm), and (d) 700-hPa vertical velocity (10−2 Pa s−1) within the region shown in Figs. 7 and 11 for the 4-km domain. Data associated with each simulation are as shown in the legends. Note that the ordinate axis in (d) utilizes a log scale.

Negatively tilted troughs are present at the 250-hPa level in both the control (Fig. 13a) and No-TC simulation (Fig. 13b) at 1800 UTC 3 October. Jet streaks are oriented from southeast to northwest along the eastern flanks of each trough, with the control simulation exhibiting a stronger jet. The trough in the control simulation is visibly narrower in extent than the trough in the No-TC simulation, presumably a manifestation of diminished diabatic outflow eroding the eastern flank of the trough in the No-TC simulation and consistent with diabatic ridging to the north and west of Joaquin in the control simulation. As evidence for this interpretation, Fig. 14 presents a comparison of the irrotational wind component (e.g., Loughe et al. 1995; Archambault et al. 2013, 2015) on the dynamic tropopause during the time when pronounced differences in the upper trough evolution over the southeastern United States began to develop. At hour 18 (valid at 1800 UTC 1 October), the divergent signature of Joaquin on the dynamic tropopause is clearly evident over the Bahamas, with pronounced easterly flow over southern Florida (Fig. 14a). In contrast, the No-TC simulation shows weak flow over southern Florida at this time (Fig. 14b); a difference plot demonstrates that the western portion of Joaquin’s outflow coincides with a difference of over 9 K in tropopause potential temperature across central Florida and the western Gulf of Mexico (Fig. 14e). At hour 36, valid at 1200 UTC 2 October, the differences in the strength of the upper-level jet have become strongly evident between the control and no-TC run (Figs. 14c,d,f). The divergent upper-level outflow at this time features a cross-jet component, consistent with acceleration and the stronger jet extending from West Virginia across South Carolina. This divergent upper-level flow supports our interpretation that the diabatic outflow from Joaquin slowed and thinned the upper trough over the southeastern United States and aided in strengthening the jet and jet entrance region across the heavy precipitation region (Figs. 13 and 14c,d).

Fig. 13.

The 250-hPa geopotential height (dam; solid black contours) and wind speed (kt, where 1 kt = 0.51 m s−1; shaded as in legend) valid at 1800 UTC 3 Oct 2015 for the (a) control and (b) No-TC simulation on the 36-km domain of the nested simulations. The 700-hPa upward vertical velocities (e.g., ω < 0) are also contoured every 1 Pa s−1 (solid green contours). (c) Difference (control minus No-TC) in 250-hPa wind speed (kt; shaded as in legend). (d) As in (c), but for upward vertical velocities (Pa s−1; shaded as in legend).

Fig. 13.

The 250-hPa geopotential height (dam; solid black contours) and wind speed (kt, where 1 kt = 0.51 m s−1; shaded as in legend) valid at 1800 UTC 3 Oct 2015 for the (a) control and (b) No-TC simulation on the 36-km domain of the nested simulations. The 700-hPa upward vertical velocities (e.g., ω < 0) are also contoured every 1 Pa s−1 (solid green contours). (c) Difference (control minus No-TC) in 250-hPa wind speed (kt; shaded as in legend). (d) As in (c), but for upward vertical velocities (Pa s−1; shaded as in legend).

Fig. 14.

Comparison of dynamic tropopause at hours 18 and 36, valid at 1800 UTC 1 Oct and 1200 UTC 2 Oct 2015. (top),(middle) Potential temperature (K; shaded as in legend), irrotational flow (vectors; scale as shown), and isotachs (kt; green dashed contours beginning 50 kt) on the dynamic tropopause (defined as the 1.5-PVU surface): (a),(c) from the control simulation and (b),(d) from the No-TC simulation. (bottom) Difference in tropopause potential temperature (K; shaded as in legend) and vector difference in irrotational wind (vectors) for control minus No-TC simulation for hours (e) 18 and (f) 36.

Fig. 14.

Comparison of dynamic tropopause at hours 18 and 36, valid at 1800 UTC 1 Oct and 1200 UTC 2 Oct 2015. (top),(middle) Potential temperature (K; shaded as in legend), irrotational flow (vectors; scale as shown), and isotachs (kt; green dashed contours beginning 50 kt) on the dynamic tropopause (defined as the 1.5-PVU surface): (a),(c) from the control simulation and (b),(d) from the No-TC simulation. (bottom) Difference in tropopause potential temperature (K; shaded as in legend) and vector difference in irrotational wind (vectors) for control minus No-TC simulation for hours (e) 18 and (f) 36.

Difference fields of 250-hPa wind speed (Fig. 13c) demonstrate the impacts of Joaquin’s diabatic outflow on the positioning of the upper-level jet. There is a northeastward shift of the upper-level jet in the No-TC simulation, which is consistent with the shift seen in the axis of heaviest precipitation (Figs. 7b, 11a). Likewise, 700-hPa upward vertical velocities in each simulation are collocated with the right jet entrance, consistent with synoptic-scale forcing for ascent (Figs. 13a,b). The difference field of 700-hPa upward vertical velocity (Fig. 13d) exhibits a northeastward shift consistent with that seen in the 250-hPa wind speed, meaning that the previously noted shift in precipitation can be attributed to differences in the upper-level synoptic-scale jet related to Joaquin’s diabatic outflow, consistent with Fig. 14. We recognize that diabatic outflow is not limited to that associated with Joaquin, as strong condensational heating took place over South Carolina as well. Another test of this hypothesis was performed using an additional simulation that omitted the effects of latent heating (not shown); in this simulation, the upper trough is also more progressive, consistent with the above interpretation.

The band of heavy precipitation evident in the No-TC simulation can be interpreted as what would have happened in the absence of Joaquin if the TC had been abruptly removed from the atmosphere at 0000 UTC 1 October. The PWAT and moisture flux evolution shown in Figs. 15 and 16 is generally similar to that presented for the control simulation (Figs. 8 and 9) except for the absence of the large maxima associated with Joaquin. After hour 72, time series analysis indicates that the area-integrated cross-coast moisture flux is stronger in the control simulation (not shown), which could contribute to heavier overall rainfall accumulation in the control relative to the No-TC simulation. A lower-tropospheric PV maximum, albeit slightly weaker than in the control simulation, is again located along the immediate southwestern flank of the band of strongest moisture transport (e.g., Fig. 16c). A cross section taken in the same location as in the control simulation again demonstrates the diabatic nature of this lower cyclonic PV maximum (Fig. 17). Thus, the moisture transport mechanism is similar to that in the control simulation, although in the presence of a weaker upper-level jet entrance (Figs. 13a,b). As a result, the forcing for ascent is somewhat less intense in the No-TC simulation. Consistent with less precipitation and this weaker synoptic-scale forcing for ascent, the lower-tropospheric cyclonic PV maximum is weaker in the No-TC simulation, thus lessening the apparent diabatic feedback (e.g., Lackmann and Gyakum 1999; Lackmann 2002) between low-level moisture flux, precipitation, and condensational heating after 72 h.

Fig. 15.

As in Fig. 8, but for 36-km domain of the No-TC simulation.

Fig. 15.

As in Fig. 8, but for 36-km domain of the No-TC simulation.

Fig. 16.

As in Fig. 9, but for 36-km domain of the No-TC simulation.

Fig. 16.

As in Fig. 9, but for 36-km domain of the No-TC simulation.

Fig. 17.

As in Fig. 10, but for 36-km domain of the No-TC simulation.

Fig. 17.

As in Fig. 10, but for 36-km domain of the No-TC simulation.

To complement the Eulerian analysis presented to this point, we also computed Lagrangian trajectories for both the outflow from Joaquin and for inflow into the region of heavy precipitation (Fig. 18). Backward trajectories from 0000 UTC 4 October from the 950-hPa level in the heavy rainfall region demonstrate the dominance of westward trajectories from the western North Atlantic (Fig. 18a). Without Joaquin, the trajectories from the No-TC simulation are still mostly westward, although there are additional trajectories with a more southward origin (Fig. 18b). This may be largely attributable to a shift in the precipitation region; we did not relocate the trajectory endpoints in order to allow for a more direct comparison.

Fig. 18.

Trajectories generated using the Unidata IDV (http://www.unidata.ucar.edu/software/idv/), colored by specific humidity (kg kg−1; shaded as in legend), for the 36-km domain of the nested control simulation: (a) 84-h backward trajectories from the heavy precipitation region from 1200 UTC 4 Oct at the 950-hPa level; (b) as in (a), but for the No-TC simulation; and (c) 72-h forward trajectories from the vicinity of Joaquin originating at the 950-hPa level from 0000 UTC 1 Oct, colored by geopotential height (shaded in legend), with 250-hPa geopotential height (black solid contours) valid at 0000 UTC 4 Oct.

Fig. 18.

Trajectories generated using the Unidata IDV (http://www.unidata.ucar.edu/software/idv/), colored by specific humidity (kg kg−1; shaded as in legend), for the 36-km domain of the nested control simulation: (a) 84-h backward trajectories from the heavy precipitation region from 1200 UTC 4 Oct at the 950-hPa level; (b) as in (a), but for the No-TC simulation; and (c) 72-h forward trajectories from the vicinity of Joaquin originating at the 950-hPa level from 0000 UTC 1 Oct, colored by geopotential height (shaded in legend), with 250-hPa geopotential height (black solid contours) valid at 0000 UTC 4 Oct.

Trajectory analysis is also useful to establish the connection between Joaquin’s outflow, the upper-level jet over the eastern United States, and the progression of the upper-tropospheric trough. Forward trajectories originating from the 950-hPa level in the vicinity of Joaquin demonstrate that the outflow from Joaquin did reach the jet and upper-level ridge over the eastern United States (Fig. 18c). These trajectories, colored by geopotential height, indicate that trajectories reaching the downstream ridge region were far aloft. Another subset of trajectories passed over the heavy rainfall region, but at midlevel altitudes. The trajectory analysis is largely consistent with earlier results from an Eulerian perspective.

4. Conclusions

We present an investigation of the role of Hurricane Joaquin in the October 2015 flooding event in South Carolina. The WRF-ARW Model is used to simulate this event with and without the presence of Joaquin in order to determine if Joaquin acted as a moisture source for this event and to investigate whether the outflow associated with Joaquin enhanced the upper-level dynamical setting for prolonged, heavy precipitation. Removal of Joaquin resulted in a northward shift in the region of heavy precipitation and a modest (~7%) decrease in area-averaged total precipitation over the southeastern United States. Negligible changes were evident in regional calculations of total moisture, indicating that Joaquin did not act as a significant moisture source (or sink) for this event. Plots of 900-hPa moisture flux (Figs. 9 and 16) reveal a slight weakening and northward shift in the axis of maximum westward moisture flux (into the flooding region) without Joaquin. The reduction in precipitation without Joaquin is attributable to weaker upward vertical velocities over the southeastern United States for much of the duration of the No-TC simulation. This is due to the absence of diabatic outflow from Hurricane Joaquin in the No-TC simulation and a correspondingly weakened jet entrance circulation (Figs. 13 and 14). In the control simulation, this outflow enhanced the upper-level jet and strengthened synoptic-scale forcing for ascent. Similarly, the diabatic outflow thinned the upper-level trough and slowed its eastward progression such that when Joaquin is removed, the axis of precipitation is shifted northeastward, consistent with the hypothesis regarding Joaquin’s diabatic outflow (Fig. 14).

The control simulation featured a diabatically produced cyclonic PV maximum in the lower troposphere associated with a strong easterly and southeasterly LLJ into the Carolinas (Fig. 10). This basic picture is unchanged when Joaquin is removed, although the PV anomaly and axis of strongest vapor transport shifted northeastward into North Carolina in that simulation (Fig. 17). Backward trajectories from the region of heavy rainfall reveal predominantly westward trajectories in both the control and No-TC simulations (Fig. 18).

In our simulations of this event, Joaquin did not play a dominant role as a moisture source to the region of heavy precipitation. Rather, our results are consistent with prior studies in showing the importance of upper-level diabatic outflow from the TC, which influenced the location and intensity of downstream precipitation (e.g., Galarneau et al. 2010). In this case, Joaquin augmented downstream precipitation through the diabatic enhancement of a jet streak. While this work stops short of performing a full water budget, it demonstrates that the jet-level influence of diabatic outflow can affect the location and intensity of downstream precipitation. Our analysis has not examined other important aspects of TC–jet interactions, such as those identified by Grams and Blumer (2015) and Grams and Archambault (2016). Although it is difficult to generalize from a single case study, these results suggest that the definition of PRE events could be adjusted to loosen the requirement for water vapor export and perhaps to increase emphasis on the role of upper-level jet alterations.

Acknowledgments

This research was supported by NSF Grants AGS1007606 and AGS-1546743. The Weather Research and Forecasting Model is made available by the National Center for Atmospheric Research (NCAR), which is sponsored by the National Science Foundation. The gridded analyses used as boundary conditions for our simulations, along with the Stage-IV precipitation analyses, are made available by the National Centers for Environmental Prediction. The ERA-I data used in this study were obtained from NCAR Computational Information Systems Laboratory (CISL). We acknowledge the Unidata program for providing the Integrated Data Viewer (IDV) software package that was used to compute 3D trajectories. Constructive comments from two anonymous reviewers improved the manuscript substantially.

REFERENCES

REFERENCES
Archambault
,
H. M.
,
L. F.
Bosart
,
D.
Keyser
, and
J. M.
Cordeira
,
2013
:
A climatological analysis of the extratropical flow response to recurving western North Pacific tropical cyclones
.
Mon. Wea. Rev.
,
141
,
2325
2346
, doi:.
Archambault
,
H. M.
,
D.
Keyser
,
L. F.
Bosart
,
C. A.
Davis
, and
J. M.
Cordeira
,
2015
:
A Composite perspective of the extratropical flow response to recurving western North Pacific tropical cyclones
.
Mon. Wea. Rev.
,
143
,
1122
1141
, doi:.
Bassill
,
N. P.
,
2014
:
Accuracy of early GFS and ECMWF Sandy (2012) track forecasts: Evidence for a dependence on cumulus parameterization
.
Geophys. Res. Lett.
,
41
,
3274
3281
, doi:.
Bassill
,
N. P.
,
2015
:
An analysis of the operational GFS simplified Arakawa Schubert parameterization within a WRF framework: A Hurricane Sandy (2012) long‐term track forecast perspective
.
J. Geophys. Res. Atmos.
,
120
,
379
398
, doi:.
Benjamin
,
S. G.
,
G. A.
Grell
,
J. M.
Brown
,
T. G.
Smirnova
, and
R.
Bleck
,
2004
:
Mesoscale weather prediction with the RUC hybrid isentropic/terrain-following coordinate model
.
Mon. Wea. Rev.
,
132
,
473
494
, doi:.
Braun
,
S. A.
,
2006
:
High-resolution simulation of Hurricane Bonnie (1998). Part II: Water budget
.
J. Atmos. Sci.
,
63
,
43
64
, doi:.
Colle
,
B. A.
,
2003
:
Numerical simulations of the extratropical transition of Floyd (1999): Structural evolution and responsible mechanisms for the heavy rainfall over the northeast United States
.
Mon. Wea. Rev.
,
131
,
2905
2926
, doi:.
Cote
,
M. R.
,
2007
: Predecessor rain events in advance of tropical cyclones. M.S. thesis, Department of Atmospheric and Environmental Sciences, University at Albany, State University of New York, 200 pp., http://cstar.cestm.albany.edu/CAP_Projects/Project10/index.htm.
Dee
,
D.
, and Coauthors
,
2011
:
The ERA‐Interim reanalysis: Configuration and performance of the data assimilation system
.
Quart. J. Roy. Meteor. Soc.
,
137
,
553
597
, doi:.
Emanuel
,
K. A.
,
M.
Fantini
, and
A. J.
Thorpe
,
1987
:
Baroclinic instability in an environment of small stability to slantwise moist convection. Part I: Two-dimensional models
.
J. Atmos. Sci.
,
44
,
1559
1573
, doi:.
Fovell
,
R. G.
,
Y. P.
Bu
,
K. L.
Corbosiero
,
W.
Tung
,
Y.
Cao
,
H.
Kuo
,
L.
Hsu
, and
H.
Su
,
2016
: Influence of cloud microphysics and radiation on tropical cyclone structure and motion. Multiscale Convection-Coupled Systems in the Tropics: A Tribute to Dr. Michio Yanai, Meteor. Monogr., No. 56, Amer. Meteor. Soc., 11.1–11.27, doi:.
Fritz
,
C.
, and
Z.
Wang
,
2014
:
Water vapor budget in a developing tropical cyclone and its implication for tropical cyclone formation
.
J. Atmos. Sci.
,
71
,
4321
4332
, doi:.
Galarneau
,
T. J.
,
L. F.
Bosart
, and
R. S.
Schumacher
,
2010
:
Predecessor rain events ahead of tropical cyclones
.
Mon. Wea. Rev.
,
138
,
3272
3297
, doi:.
Grams
,
C. M.
, and
S. R.
Blumer
,
2015
:
European high‐impact weather caused by the downstream response to the extratropical transition of North Atlantic Hurricane Katia (2011)
.
Geophys. Res. Lett.
,
42
,
8738
8748
, doi:.
Grams
,
C. M.
, and
H. M.
Archambault
,
2016
:
The key role of diabatic outflow in amplifying the midlatitude flow: A representative case study of weather systems surrounding western North Pacific extratropical transition
.
Mon. Wea. Rev.
,
144
,
3847
3869
, doi:.
Huang
,
H.
,
M.
Yang
, and
C.
Sui
,
2014
:
Water budget and precipitation efficiency of Typhoon Morakot (2009)
.
J. Atmos. Sci.
,
71
,
112
129
, doi:.
Kurihara
,
Y.
,
1975
:
Budget analysis of a tropical cyclone simulated in an axisymmetric numerical model
.
J. Atmos. Sci.
,
32
,
25
59
, doi:.
Lackmann
,
G. M.
,
2002
:
Cold-frontal potential vorticity maxima, the low-level jet, and moisture transport in extratropical cyclones
.
Mon. Wea. Rev.
,
130
,
59
74
, doi:.
Lackmann
,
G. M.
, and
J. R.
Gyakum
,
1999
:
Heavy cold-season precipitation in the Northwestern United States: Synoptic climatology and an analysis of the flood of 17–18 January 1986
.
Wea. Forecasting
,
14
,
687
700
, doi:.
Lin
,
Y.
, and
K. E.
Mitchell
,
2005
: The NCEP Stage II/IV hourly precipitation analyses: Development and applications. 19th Conf. on Hydrology, San Diego, CA, Amer. Meteor. Soc., 1.2, https://ams.confex.com/ams/Annual2005/techprogram/paper_83847.htm .
Loughe
,
A. F.
,
C.-C.
Lai
, and
D.
Keyser
,
1995
:
A technique for diagnosing three-dimensional ageostrophic circulations in baroclinic disturbances on limited-area domains
.
Mon. Wea. Rev.
,
123
,
1476
1504
, doi:.
Marks
,
F. D.
, and
R. A.
Houze
,
1987
:
Inner core structure of Hurricane Alicia from airborne Doppler radar observations
.
J. Atmos. Sci.
,
44
,
1296
1317
, doi:.
NHC
,
2016
: National Hurricane Center Tropical Cyclone Report: Hurricane Joaquin (AL112015). National Hurricane Center Rep. AL112015, 36 pp., http://www.nhc.noaa.gov/data/tcr/AL112015_Joaquin.pdf.
NOAA/National Centers for Environmental Prediction
,
2000
: NCEP FNL Operational Model Global Tropospheric Analyses, continuing from July 1999 (updated daily). National Center for Atmospheric Research Computational and Information Systems Laboratory Research Data Archive, accessed 10 December 2015, doi:.
NWS
,
2015
: Historical rainfall and flooding, October 2015. National Weather Service, Columbia, SC, accessed 1 March 2016, http://www.weather.gov/cae/HistoricFloodingOct2015.html.
Peckham
,
S. E.
,
T. G.
Smirnova
,
S. G.
Benjamin
,
J. M.
Brown
, and
J. S.
Kenyon
,
2016
:
Implementation of a digital filter initialization in the WRF Model and its application in the Rapid Refresh
.
Mon. Wea. Rev.
,
144
,
99
106
, doi:.
Schumacher
,
R. S.
,
T. J.
Galarneau
Jr.
, and
L. F.
Bosart
,
2011
:
Distant effects of a recurving tropical cyclone on rainfall in a midlatitude convective system: A high-impact predecessor rain event
.
Mon. Wea. Rev.
,
139
,
650
667
, doi:.
Skamarock
,
W. C.
, and
J. B.
Klemp
,
2008
:
A time-split nonhydrostatic atmospheric model for weather research and forecasting applications
.
J. Comput. Phys.
,
227
,
3465
3485
, doi:.
Srock
,
A. F.
, and
L. F.
Bosart
,
2009
:
Heavy precipitation associated with southern Appalachian cold-air damming and Carolina coastal frontogenesis in advance of weak landfalling Tropical Storm Marco (1990)
.
Mon. Wea. Rev.
,
137
,
2448
2470
, doi:.
Thompson
,
G.
, and
T.
Eidhammer
,
2014
:
A study of aerosol impacts on clouds and precipitation development in a large winter cyclone
.
J. Atmos. Sci.
,
71
,
3636
3658
, doi:.
Torn
,
R. D.
, and
C.
Snyder
,
2012
:
Uncertainty of tropical cyclone best-track information
.
Wea. Forecasting
,
27
,
715
729
, doi:.
Yang
,
M.
,
S. A.
Braun
, and
D.
Chen
,
2011
:
Water budget of Typhoon Nari (2001)
.
Mon. Wea. Rev.
,
139
,
3809
3828
, doi:.

Footnotes

© 2017 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).