Abstract

This study considers eastern Antilles (11°–18°N, 64°–57°W) weather and climate interactions in the context of the 2013 Christmas storm. This unseasonal event caused flash flooding in Grenada, St. Vincent, St. Lucia, Martinique, and Dominica from 24 to 25 December 2013, despite having winds <15 m s−1. The meteorological scenario and short-term forecasts are analyzed. At the low level, a convective wave propagated westward while near-equatorial upper westerly winds surged with eastward passage of a trough. The combination of tropical moisture, cyclonic vorticity, and uplift resulted in rain rates greater than 30 mm h−1 and many stations reporting 200 mm. Although forecast rainfall was low and a few hours late, weather services posted flood warnings in advance. At the climate scale, the fresh Orinoco River plume brought into the region by the North Brazil Current together with solar radiation greater than 200 W m−2, enabled sea temperatures to reach 28°C, and supplied convective available potential energy greater than 1800 J kg−1. Climate change model simulations are compared with reference fields and trends are analyzed in the eastern Antilles. While temperatures are set to increase, the frequency of flood events appears to decline in the future.

1. Introduction

Convective weather systems in the Caribbean often originate from African easterly waves that are modulated by the large-scale circulation and thermal energy from the sea (Wang et al. 2006; Jury et al. 2007; Kossin et al. 2010). Tropical cyclogenesis is frequent in the eastern (Lesser, Windward) Antilles Islands (11°–18°N, 64°–57°W), but trade wind upwelling north of Venezuela inhibits development there (Shieh and Colucci 2010). Intraseasonal oscillations affect these systems via near-equatorial winds from the Pacific (Wang and Rui 1990; Maloney and Hartmann 2000; Maloney and Sobel 2004; Frank and Roundy 2006; Molinari et al. 2007; Kiladis et al. 2009; Camargo et al. 2009). Troughs and cyclones often come under westerly shear in the autumn while being sustained by warm waters; hence, the systems tend to move slower or recurve poleward. One example was Hurricane Tomas (Pasch and Kimberlain 2011), which caused extensive damage as it passed westward near St. Lucia (14°N, 61°W) on 31 October 2010, generating winds >33 m s−1 and rainfall ~500 mm. Its low-latitude path was over a zone of unusually warm (+1°C) sea surface temperature (SST) anomalies related to diminished trade wind evaporation (Bell et al. 2011; Blunden et al. 2011). Following the hurricane, international agencies provided assistance to upgrade St. Lucia’s hydrological monitoring network and flood warning system.

A unique feature of the regional oceanography is a freshwater plume from the Orinoco and Amazon Rivers that spreads across the eastern Caribbean via the North Brazil Current. The low salinity [<34.6 parts per thousand (ppt)] plume is buoyant and forms a shallow stable layer. Discharges from the Orinoco and Amazon Rivers average ~2.5 × 105 m3 s−1 (Perry et al. 1996; Pailler et al. 1999) and tend to peak in summer. As the plume spreads into the east Antilles it encounters incoming solar radiation of 230 W m−2 and a trade wind shadow next to the islands. Surface evaporation declines from 145 (east) to 132 W m−2 (west of the islands) and mean SSTs increase to 28°C. Along the coast of Venezuela, there is a zone of upwelling where SSTs <25°C extend 100 km offshore. The North Brazil Current divides these two areas: north of its 0.3 m s−1 axis that passes near Grenada lie the warm fresh waters that invigorate marine convection (Ffield 2007; Vizy and Cook 2010; Balaguru et al. 2012; Grodsky et al. 2012).

The meteorological event of interest here is the 2013 Christmas storm that brought flash floods to many eastern Antilles islands. The primary objectives of this work are to (i) analyze the synoptic-scale weather scenario and regional impacts of the 24–25 December 2013 event in the eastern Antilles (11°–18°N, 64°–57°W), (ii) study the synoptic-scale meteorological forcing of this system, and (iii) place the event in context of marine climate variability and change. To achieve these objectives, a variety of regional observations and global model analyses are considered, distinguished as transient weather and stationary ocean forcing. In section 2, the data and methods are outlined. Section 3 provides results on the storm, its forcing, and climate context, whereas section 4 summarizes the results.

2. Data and analysis methods

The regional weather scenario and impacts of the 2013 Christmas storm are diagnosed using eastern Antilles observations and meteorological maps, while the climate context is studied via atmospheric data assimilation and forecasts, ocean reanalysis fields, and coupled climate model projections. The analysis covers thermodynamic and kinematic advection patterns contributed by the atmosphere and the background marine climate governed by the ocean. The intensity of convection in the eastern Antilles region is quantified in the period 24–25 December 2013 using 4-km Geostationary Operational Environmental Satellite (GOES) infrared cloud temperatures at 30-min interval, 25-km multisatellite microwave rain rate at 3-h intervals [Climate Prediction Center (CPC) morphing technique (CMORPH); Joyce et al. 2004], and CPC 50-km interpolated gauge rainfall. Surface winds at 20-km resolution are composited from two overpasses of the Advanced Scatterometer (ASCAT) satellite in the area 11°–18°N, 64°–57°W on the afternoon of 24 December. Local weather station records are consulted via official reports and from Weather Underground. Hourly observations are available during the daytime at airports in St. Vincent, St. Lucia, Martinique, and Dominica (cf. Figure 1d), but most storm impacts were in the evening. One automatic station in Martinique had rain rate, temperature, pressure, and winds. Rain estimates and forecasts are available from the hourly 20-km Coupled Forecast System (CFS; Saha et al. 2010), 60-km National Aeronautics and Space Administration (NASA) Modern-Era Retrospective Analysis for Research and Applications (MERRA) reanalysis (Rienecker et al. 2011), and the 3-hourly 50-km operational Global Forecast System (GFS) assimilation. Time series in a 1° × 1° grid box over St. Lucia (14°N, 61°W) are considered at 0-, 1-, and 2-day leads to evaluate forecasts. It is understood that area-averaged rainfall from satellites and models may have lower values than point measurements from gauges.

Figure 1.

(a)–(c) GOES infrared cloud temperature sequence. (d) CMORPH satellite-derived total rainfall (mm) for 24–25 Dec 2013 with island labels and maximum rate (mm h−1) and with time per island indicating southward progression. (e) Composite ASCAT-derived surface winds on the afternoon of 24 Dec 2013.

Figure 1.

(a)–(c) GOES infrared cloud temperature sequence. (d) CMORPH satellite-derived total rainfall (mm) for 24–25 Dec 2013 with island labels and maximum rate (mm h−1) and with time per island indicating southward progression. (e) Composite ASCAT-derived surface winds on the afternoon of 24 Dec 2013.

The meteorological scenario of the 2013 Christmas storm is analyzed via GFS assimilation fields of winds and vertical motion, convective available potential energy (CAPE), precipitable water, and vorticity at 0000 UTC 25 December 2013. Vertical atmospheric profiles are obtained from radiosonde at Curacao (12°N, 68°W) and Barbados (13°N, 59°W) and from two Aircraft Meteorological Data Relay (AMDAR)-capable aircraft (Moninger et al. 2003) departing from St. Lucia and Martinique at 1800 and 2000 UTC 24 December 2013. The radiosonde profiles include temperature, dewpoint, wind speed, and wind direction, while the AMDAR aircraft profiles include wind and vertical gust. Zonal propagating waves are studied by Hovmöller time–longitude plots on 14°N of 6-hourly 180-km-resolution National Centers for Environmental Prediction–U.S. Department of Energy (NCEP–DOE) Atmospheric Model Intercomparison Project version 2 (AMIP-II) reanalysis (NCEP-2; Kanamitsu et al. 2002) precipitable water and 200-mb absolute vorticity, 3-hourly CMORPH rain rate, and hourly 20-km-resolution CFS surface zonal wind.

The ocean’s influence on convection is studied using December 2013 mean SST composited from 4-km NASA satellite data. National Oceanic and Atmospheric Administration (NOAA) Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) ensemble back trajectories arriving at 2000 m over St. Lucia are analyzed during 24 December 2013 to provide a Lagrangian perspective. The annual cycle is studied via daily GPCP rainfall (Huffman et al. 2009) and satellite SST (Reynolds et al. 2007) averaged over the eastern Antilles (11°–18°N, 64°–57°W) in the period 1997–2012. Late summer climate variability is evaluated by correlating east Antilles November–December rainfall with 60–70-km-resolution MERRA and European Centre for Medium-Range Weather Forecasts (ECMWF; Dee et al. 2011) reanalysis fields of wind, surface temperature, and vertical motion in the period 1980–2010. The mean climate is studied via NCEP-2 data averaged over the period 1980–2010 with a focus on the Hadley circulation and relative humidity in a north–south vertical slice over the southeastern Caribbean (5°–25°N, 64°–57°W). Background on the Orinoco River plume is provided by 1–100-m depth-averaged ocean reanalysis salinity and current data for December 2013 and over a wider region averaged for 1980–2010 from the NOAA 100 km × 30 km resolution Global Ocean Data Assimilation System (GODAS) and the 50-km-resolution Simple Ocean Data Assimilation (SODA) reanalysis (Behringer 2007; Carton and Giese 2008).

The 2013 Christmas storm prompted suggestions that climate change may enhance flood-producing weather systems. To this end, a number of phase 5 of the Coupled Model Intercomparison Project (CMIP5) models were analyzed for their ability to simulate past climatology of air pressure, ocean currents, sea temperatures, and salinity. The Model for Interdisciplinary Research on Climate (MIROC) developed in Japan (Hasumi and Emori 2004) is found to closely match the observed patterns for sea level pressure (SLP), Ts (surface temperature), and rain. It has a sophisticated 100-km-resolution atmospheric model coupled to an eddy-resolving (30 km) ocean model and boundary layer/surface flux schemes similar to the Geophysical Fluid Dynamics Laboratory (GFDL) model (Meehl et al. 2007; Taylor et al. 2012). The ability of this general circulation model (GCM) to represent the low salinity Orinoco River plume and its advection into the eastern Antilles by the North Brazil Current is of particular interest. MIROC model simulations are intercompared with observed mean fields for 1980–2010 and projected trends of the daily frequency of extreme temperature and rainfall are analyzed up to 2050 using representative concentration pathway (RCP) 4.5 and 8.5 scenarios (van Vuuren et al. 2011). The websites used for data analysis are listed in the acknowledgments.

3. Results

3.1. Observations

Station and satellite observations during the 24–25 December 2013 storm are considered in this section. The sequence of 4-km GOES infrared cloud images (Figures 1a–c) reveals a large cluster of thunderstorms extending from 12° to 16°N and from 62° to 60.5°W, with the coldest cloud tops of −70°C. Convection progressed from north to south, with maximum CMORPH satellite rain rates of 12.5 mm h−1 at Dominica at 1030 UTC 24 December, 11.0 mm h−1 at Martinique at 1900 UTC 24 December, 29.3 mm h−1 at St. Lucia at 2230 UTC 24 December, 18.1 mm h−1 at St. Vincent at 0130 UTC 25 December, and 27.3 mm h−1 at Grenada at 0730 UTC 25 December 2013. The CMORPH map of rainfall over the 2-day period (Figure 1d) has south-southwest–north-northeast alignment over the chain of islands. Surface easterly winds estimated by ASCAT (Figure 1e) were >10 m s−1 to the east of the islands and <5 m s−1 to the west, on the afternoon of 24 December 2013. Hence, deceleration of the flow contributed to convergence and upward motion over the islands that was further enhanced by orography. Official reports indicated some mountain stations received over 200 mm of rainfall, while coastal stations had ~100 mm of rainfall and winds up to 14 m s−1.

Figure 2a illustrates the time series of rainfall over St. Lucia from satellites and models. The GFS forecast at a 2-day lead gave 30 mm late on 25 December; at a 1-day lead, the rain forecast was 50 mm early on 25 December. Both forecasts were “too little, too late.” The CFS analysis shows a sustained peak of 60 mm on the afternoon of 24 December, consistent with the observed coastal rainfall, although a few hours too early. MERRA was closer to CMORPH with a sudden increase of rainfall and a gradual decline. The automatic station on Martinique (Figure 2b) recorded a sharp increase in rain rate from 1800 to 2200 UTC 24 December 2013, followed by an increase of wind from 2 to 6 m s−1. Temperatures declined to ~20°C from evaporation-cooled thunderstorm downdrafts. Winds over Martinique became northerly during the evening of 24 December, following the band of heavy rain toward St. Lucia. The MERRA model rainfall map over the 2-day period (Figure 2c) had the correct intensity in comparison with satellite and gauge but was too far northwest. The Roseau River in Dominica experienced flash flooding, but streamflow gauges in St. Lucia were unserviceable during the event because of maintenance problems. Météo-France Antilles radar data over St. Lucia (Figure 2d) illustrate a band of >200-mm rainfall lying southwest–northeast across the island that triggered flash floods.

Figure 2.

Time series of (a) 1° × 1° area rainfall over St. Lucia by CFS and MERRA assimilation, CMORPH satellite, and GFS forecasts at 1- and 2-day leads. (b) Martinique station wind speed, temperature, and rainfall with gap filled by airport observations. (c) MERRA model (shaded) and CPC interpolated gauge (numbers) rainfall for 24–25 Dec 2013. (d) Météo-France Antilles radar estimated rainfall over St. Lucia in the period 1800–2400 UTC 24 Dec 2013 with millimeter scale.

Figure 2.

Time series of (a) 1° × 1° area rainfall over St. Lucia by CFS and MERRA assimilation, CMORPH satellite, and GFS forecasts at 1- and 2-day leads. (b) Martinique station wind speed, temperature, and rainfall with gap filled by airport observations. (c) MERRA model (shaded) and CPC interpolated gauge (numbers) rainfall for 24–25 Dec 2013. (d) Météo-France Antilles radar estimated rainfall over St. Lucia in the period 1800–2400 UTC 24 Dec 2013 with millimeter scale.

Vertical profiles were available from radiosonde at surrounding stations and AMDAR-capable aircraft departing St. Lucia and Martinique (Figures 3a–d). The 1200 UTC 24 December 2013 profile at Curacao, located 600 km to the southwest, reveals a moist layer up to 700 mb and northwesterly winds of 40 m s−1 at 225 mb. The profile at Barbados, situated 100 km to the southeast, has a moist layer up to 450 mb and southwesterly winds above 700 mb, reaching 25 m s−1 at 350 mb. CAPE at Barbados was 1867 J kg−1 and within the storm inflow. The aircraft AMDAR wind profiles at 1800 and 2000 UTC 24 December 2013 both illustrate 10–20 m s−1 speeds from 2 to 7 km that rotated from easterly through southerly. Vertical gusts up to 8 m s−1 were recorded near 3 km.

Figure 3.

Radiosonde profiles at 1200 UTC 24 Dec at (a) Curacao (12°N, 69°W) and (b) Barbados (13°N, 59°W) with CAPE = 1867. (c),(d) Wind speed and direction profiles for aircraft departing St. Lucia and Martinique at 1800–2000 UTC 24 Dec, where height is in meters and red numbers are vertical gusts (m s−1).

Figure 3.

Radiosonde profiles at 1200 UTC 24 Dec at (a) Curacao (12°N, 69°W) and (b) Barbados (13°N, 59°W) with CAPE = 1867. (c),(d) Wind speed and direction profiles for aircraft departing St. Lucia and Martinique at 1800–2000 UTC 24 Dec, where height is in meters and red numbers are vertical gusts (m s−1).

3.2. Regional model analyses

Operational weather data assimilations of the 2013 Christmas storm are considered here. The GFS 0000 UTC 25 December 2013 700- and 200-mb winds and CAPE are striking (Figures 4a–d). There was a deep cyclonic circulation near Puerto Rico and a 40 m s−1 westerly jet streak over Venezuela. The upper wave was asymmetrical: 200-mb streamlines were confluent to the west and diffluent to the east of the trough axis on 64°W. Low-level 700-mb southerly flow over the eastern Antilles was fed by tropical Atlantic easterlies. CAPE exceeded 1600 J kg−1 east of the Antilles, where stronger surface winds were noted (cf. Figure 1e). The surface weather chart for 0000 UTC 25 December 2013 (Figure 4c) shows the subtropical anticyclone northeast of the region; surface isobars curve southeasterly and dewpoint temperatures were ~22°C. The equatorward pressure gradient fed surface easterlies into the storm.

Figure 4.

GFS 0000 UTC 25 Dec 2013 analysis of (a) 700-mb winds, (b) 200-mb wind streamlines (thin) and isotachs (thick, m s−1), (c) NHC surface weather map, and (d) CAPE (J kg−1).

Figure 4.

GFS 0000 UTC 25 Dec 2013 analysis of (a) 700-mb winds, (b) 200-mb wind streamlines (thin) and isotachs (thick, m s−1), (c) NHC surface weather map, and (d) CAPE (J kg−1).

The vertical structure of the meridional circulation during the storm is analyzed (Figure 5a). Southerly wind anomalies >20 m s−1 developed in the layer of 600–200 mb at 14°–18°N. Rising motion was greatest from 300 to 500 mb from 10° to 15°N. The poleward flow on the eastern flank of the trough generated an overturning circulation, opposite to the Hadley cell that is usually established by December. Cyclonic vorticity of upper-level flow was a key feature (Figure 5b) over the eastern Caribbean at 0000–1200 UTC 25 December 2013 and satellite outgoing longwave radiation (OLR) anomalies of <180 W m−2 were located on the eastern flank of the trough along 60°W. These weather features would indicate a major storm event to a local forecaster.

Figure 5.

NCEP plots for 0000–1200 UTC 25 Dec 2013: (a) vertical section averaged 64°–57°W of meridional and vertical wind (vector maximum = 20 m s−1; W is exaggerated) and relative vorticity (shaded; 10−5 s−1) and (b) map of 200-mb relative vorticity (shaded) and NOAA satellite OLR (dashed; <220 W m−2 plotted).

Figure 5.

NCEP plots for 0000–1200 UTC 25 Dec 2013: (a) vertical section averaged 64°–57°W of meridional and vertical wind (vector maximum = 20 m s−1; W is exaggerated) and relative vorticity (shaded; 10−5 s−1) and (b) map of 200-mb relative vorticity (shaded) and NOAA satellite OLR (dashed; <220 W m−2 plotted).

The propagation of zonal atmospheric waves is studied by Hovmöller analysis on 14°N in Figures 6a–d. The 200-mb absolute vorticity revealed eastward movement of a cyclonic signal starting over Central America (85°W) on 22 December, intensifying on 25 December at 65°W, and continuing into the tropical Atlantic to 45°W by 30 December 2013. The vorticity signal reflects a surge of upper westerly winds near the equator, propagating eastward at 5° day−1 or ~6 m s−1. Alternatively, variables such as precipitable water and surface zonal wind show westward propagation of a convective wave from 40°W on 20 December, intensifying on 25 December at 60°W and continuing toward Central America by 29 December 2013. The low-level westward system moved at about the same speed as the upper-level eastward system, and the two crossed paths over the eastern Antilles as the storm was generated. An interesting feature of the CFS zonal wind Hovmöller is the topographically anchored St. Lucia wind wake and weakened trade winds associated with the wave trough. The CMORPH rain rate increased on 25 December 2013 over the eastern Antilles (Figure 6c), but convection dissipated as the two atmospheric waves moved apart.

Figure 6.

Hovmöller plots on 14°N of (a) NCEP-2 6-hourly 200-mb absolute vorticity (x10−4 s−1) and (b) precipitable water (mm), (c) CMORPH 3-hourly rain rate (mm h−1), and (d) CFS hourly surface zonal wind (m s−1), illustrating wave propagation in the period 20–30 Dec 2013 (time upward).

Figure 6.

Hovmöller plots on 14°N of (a) NCEP-2 6-hourly 200-mb absolute vorticity (x10−4 s−1) and (b) precipitable water (mm), (c) CMORPH 3-hourly rain rate (mm h−1), and (d) CFS hourly surface zonal wind (m s−1), illustrating wave propagation in the period 20–30 Dec 2013 (time upward).

Part of the upper westerly wind surge appears related to a high phase (+2.3 anomaly in December 2013) Arctic Oscillation, causing the subtropical jet stream to shift southward. Figure 7a is a correlation map between ECMWF 200-mb zonal winds in December and the NCEP Arctic Oscillation for 1980–2012. There is a positive region denoting faster westerlies over the Antilles (12°N) and a negative region at ~25°N. Usually the subtropical jet stream lies poleward of the Antilles in December, thereby generating anticyclonic vorticity and sinking motion. However, the December 2013 anomalies show a negative north–positive south pattern in upper zonal winds (Figure 7b), consistent with an expanded polar vortex and southward shift of the subtropical jet. When the jet lies south of the Antilles, cyclonic vorticity and upward motion are generated.

Figure 7.

(a) Correlation of December Arctic Oscillation with ECMWF 200-mb zonal wind (1980–2012). (b) NCEP-2 vertical section of December 2013 zonal wind anomaly averaged at 64°–57°W.

Figure 7.

(a) Correlation of December Arctic Oscillation with ECMWF 200-mb zonal wind (1980–2012). (b) NCEP-2 vertical section of December 2013 zonal wind anomaly averaged at 64°–57°W.

3.3. Climatic background

In this section, December 2013 average maps are studied to determine the climatic background of this storm event. The high-resolution SST field is given in Figure 8a together with SST anomalies and low-level back trajectories arriving at 2000 m over St. Lucia. SSTs exceeded 27.6°C over a broad northwest–southeast swath passing through the eastern Antilles. Storm inflow toward St. Lucia passed over warm waters lying south of Barbados. Reduced evaporation in the wind shadows west of the Antilles Islands resulted in small patches of SST > 28°C that could have enhanced storm convection. SSTs were +0.5°C above normal during this period, especially near Grenada.

Figure 8.

December 2013 mean ocean climate: (a) SST (shading) and anomalies (contours) overlaid with ensemble back trajectories arriving at 2000 m over St. Lucia from 1200 UTC 24 Dec to 0000 UTC 25 Dec; (b) 1–100-m depth-averaged salinity (ppt) and ocean currents (max 0.2 m s−1) from NOAA GODAS; and north–south depth section of (c) temperature with solar radiation (inset white values W m−2) and (d) salinity and meridional circulation (vectors) averaged over longitude.

Figure 8.

December 2013 mean ocean climate: (a) SST (shading) and anomalies (contours) overlaid with ensemble back trajectories arriving at 2000 m over St. Lucia from 1200 UTC 24 Dec to 0000 UTC 25 Dec; (b) 1–100-m depth-averaged salinity (ppt) and ocean currents (max 0.2 m s−1) from NOAA GODAS; and north–south depth section of (c) temperature with solar radiation (inset white values W m−2) and (d) salinity and meridional circulation (vectors) averaged over longitude.

To study the factors maintaining high SST, December 2013 mean incoming solar radiation, ocean temperature, salinity and currents are analyzed in Figures 8b–d as a depth section. Solar radiation was > 225 W m−2 across the warm swath, coincident with a fresh upper layer signifying the Orinoco River plume. The North Brazil Current advected this water northwestward at ~0.2 m s−1. GODAS fields indicate the salinity minimum was located near Grenada at 13°N. The depth section illustrates a meridional overturning circulation comprised of upwelling near the coast of South America (11°N) and poleward transport in the upper 30 m. The fresh layer was about 70 m deep in December 2013 with a temperature > 27°C. North of the fresh axis with the North Brazil Current, there was downwelling, especially in the latitudes of St. Lucia and Martinique (14°–15°N).

3.4. Mean climate, drivers, and trends

The climate of the eastern Antilles region and its regional drivers and long-term trends are studied in this section. Figures 9a,b show the mean annual cycle of SST and rainfall, which decline to a minimum in February–March followed by a persistent rise in summer. The SST makes an upward step after midsummer and crests in the hurricane season, August–September. On the other hand, rainfall continues to rise in October–November as SST starts to fall. The gap between maximum and mean rainfall is larger in the second half of the year, yet the maximum − mean for SST is larger in the first half.

Figure 9.

Annual cycle of daily (a) SST and (b) rainfall averaged over the eastern Antilles area [box in (f)], illustrating the mean and top 2.5%. Correlation maps (1980–2012) of November–December eastern Antilles GPCP rainfall and (c) surface temperature, (d) surface zonal wind from MERRA reanalysis, (e) 700-mb meridional wind, and (f) 500-mb vertical motion (omega) from ECMWF reanalysis. The season is denoted by the bracket in (b), and arrows in (d),(e) highlight wind direction.

Figure 9.

Annual cycle of daily (a) SST and (b) rainfall averaged over the eastern Antilles area [box in (f)], illustrating the mean and top 2.5%. Correlation maps (1980–2012) of November–December eastern Antilles GPCP rainfall and (c) surface temperature, (d) surface zonal wind from MERRA reanalysis, (e) 700-mb meridional wind, and (f) 500-mb vertical motion (omega) from ECMWF reanalysis. The season is denoted by the bracket in (b), and arrows in (d),(e) highlight wind direction.

November–December GPCP rainfall in the eastern Antilles is correlated over 1980–2012 with reanalysis fields to study regional influences (Figures 9c–f). Autumn rains are sensitive to surface warming along the coast of South America, corresponding with westerly surface wind anomalies. Such a feature points to a slowing of the upwelling circulation near Venezuela. Stronger trade winds and cooler SSTs in the northern Antilles (blue shades in Figures 9c,d) correspond with high rainfall in the study area. Correlations with respect to 700-mb meridional winds form a cyclonic trough with equatorward flow in 75°–65°W and poleward flow in 60°–50°W. The correlation map for 500-mb vertical motion yields the expected signal of uplift over the study area. Its zonal axis extends across the coast of South America and the tropical Atlantic, suggesting expansion of the equatorial trough.

Much of the research has focused on the December 2013 weather scenario. Let us consider the 30-yr mean climate patterns and variance. Using NCEP reanalysis and GCM simulations from the MIROC model, the atmospheric meridional overturning (Hadley) circulation and relative humidity is studied (Figures 10a,b). The mean climate suppresses convection via large-scale sinking motion joined by equatorward flow aloft. There is a rotary cell centered at 800 mb and 10°N in both analyses that divides sinking motion over the eastern Antilles from rising motion over the Orinoco River valley. The relative humidity slice reveals a shallow moist layer: 2.5 km in NCEP and 1.5 km in MIROC. Both show a deepening over South America, but the dominant feature is a dry layer < 30% RH that extends from 700 to 300 mb (6 km deep) that suppresses marine convection most of the time.

Figure 10.

Vertical section of 1980–2010 mean Hadley circulation averaged at 64°–57°W from (a) NCEP-2 and (b) the MIROC GCM, illustrating the (left) meridional wind/vertical motion and (right) relative humidity. Mean maps of (left) ocean circulation and (right) salinity in the upper 100 m from (c) SODA-2.6 reanalysis and (d) the MIROC GCM.

Figure 10.

Vertical section of 1980–2010 mean Hadley circulation averaged at 64°–57°W from (a) NCEP-2 and (b) the MIROC GCM, illustrating the (left) meridional wind/vertical motion and (right) relative humidity. Mean maps of (left) ocean circulation and (right) salinity in the upper 100 m from (c) SODA-2.6 reanalysis and (d) the MIROC GCM.

The ocean climate is analyzed using 30-yr mean fields from SODA reanalysis and MIROC GCM currents and salinity in the upper 100-m layer (Figures 10c,d). Both exhibit northwestward currents along the coast of South America. The MIROC model simulates a weaker North Brazil Current, especially near Trinidad, and a fresher North Atlantic. Yet its freshwater plume is well developed and the MIROC salinity pattern is relatively close to SODA, despite lower values near the Orinoco and Amazon River mouths.

The MIROC GCM performs better than most CMIP5 models in respect of air pressure distributions (and trade winds), SST, and rainfall in the eastern Caribbean (Figures 11a–c). Its bias in terms of high pressure and cool SST is +1 mb and −1°C, respectively. Yet marine rainfall is well simulated; only the coast of South America has a dry bias. Computing the trends in the same parameters using the RCP 4.5 scenario from 1980 to 2050 (Figures 11d–f), it is found that the MIROC projection is for a reduction in the pressure gradient (low north–high south). There is widespread warming that is faster over the Orinoco River valley (+0.03°C yr−1), but rainfall trends are relatively weak. These features are consistent with the CMIP5 average of all 40 models.

Figure 11.

MIROC model minus observation maps (1980–2010) of (a) air pressure (Hadley-2 reanalysis; mb), (b) surface air temperature (ECMWF reanalysis; °C), and (c) rainfall (GPCP; mm day−1). (d)–(f) The MIROC trend analyses over 1980–2050 (rate per year).

Figure 11.

MIROC model minus observation maps (1980–2010) of (a) air pressure (Hadley-2 reanalysis; mb), (b) surface air temperature (ECMWF reanalysis; °C), and (c) rainfall (GPCP; mm day−1). (d)–(f) The MIROC trend analyses over 1980–2050 (rate per year).

Extracting the time series of MIROC projections for the eastern Antilles region (Figures 12a,b), it is projected that the number of hot days each year may increase from ~5 to 30 days, regardless of scenario. The number of flood days each year may decrease slightly, from 3 days in the past to 2 days in the future. This result may be related to the fact that the atmosphere is warming faster than the ocean (Jury and Winter 2009), which induces a down trend in heat transfer. Other factors in the drying trend include an accelerating Hadley circulation (Perez and Jury 2013). The projection suggests that global warming may not induce more frequent flood events in the eastern Antilles, in agreement with past trends (Chenoweth and Divine 2008).

Figure 12.

MIROC model simulation/projection of the daily number of days per year of (a) high temperatures and (b) heavy rainfall, using 4 and 8 W m−2 global warming scenarios. Linear trend regressions are given.

Figure 12.

MIROC model simulation/projection of the daily number of days per year of (a) high temperatures and (b) heavy rainfall, using 4 and 8 W m−2 global warming scenarios. Linear trend regressions are given.

4. Summary

This study has considered weather and climate coupling in the eastern Antilles during the 2013 Christmas storm that caused flash flooding in Grenada, St. Vincent, St. Lucia, Martinique, and Dominica. The storm was unusually intense for the holiday season, causing impacts similar to Hurricane Tomas in 2010 (Pasch and Kimberlain 2011). The analysis highlighted the transient atmospheric thermodynamic and kinematic wave structure using hourly to daily data, against the background stationary pattern of marine climate using monthly data. At the lower level a moist convective wave propagated westward, while at the upper level a jet stream trough surged eastward, as shown in Hovmöller analyses (cf. Figures 6a,b). The transient atmospheric forcing crossed paths, triggering the storm. The combination of tropical moisture, cyclonic vorticity, and uplift resulted in rain rates >30 mm h−1 and many stations reporting 200 mm. The convection progressed southward (cf. Figure 1d) because of the orientation of the atmospheric waves. Although operational model rain forecasts were on the low side and a few hours late, weather services had posted flood warnings in advance. Public uptake of forecasts was hampered by electricity and media disruptions between agencies and an expectation of diminished impacts in the holiday season. It is recommended that mesoscale observations and models be extended across the region to augment existing hydrology, meteorology, and oceanography data assimilation systems and afford improved short-term forecasts.

At the climate scale, the fresh Orinoco River plume advected into the region by the North Brazil Current coincided with sea temperatures >27.6°C that supplied convective available potential energy >1800 J kg−1 in December 2013 (cf. Figures 8b,a and 4d). Expansion of the polar vortex triggered a low-latitude acceleration of the jet stream (cf. Figure 7a) and consequent cyclonic vorticity (cf. Figures 5a,b). Correlation maps indicate that a weakening of Venezuela coastal upwelling and trade winds anticipates a wet November–December season (cf. Figures 9c,d). Climate model simulations from MIROC were compared with reference fields from NCEP and SODA, and trends were analyzed in the eastern Antilles (cf. Figure 12). While temperatures are set to increase ~+0.02°C yr−1, the frequency of flood events may decline in future.

Acknowledgments

References

References
Balaguru
,
K.
,
P.
Chang
,
R.
Saravanan
, and
C. J.
Jang
,
2012
:
The barrier layer of the Atlantic warm pool: Formation mechanism and influence on the mean climate
.
Tellus
,
64A
, 18162, doi:.
Behringer
,
D. W.
,
2007
. The Global Ocean Data Assimilation System (GODAS) at NCEP. Preprints, 11th Symp. Integrated Observing and Assimilation Systems for the Atmosphere, Oceans, and Land Surface, San Antonio, TX, Amer. Meteor. Soc., 3.3. [Available online at https://ams.confex.com/ams/pdfpapers/131381.pdf.]
Bell
,
G. D.
,
E. S.
Blake
,
T. B.
Kimberlain
,
C. W.
Landsea
,
J.
Schemm
,
R. J.
Pasch
, and
S. B.
Goldenberg
,
2011
:
Tropical cyclones in the Atlantic [4.2 in “State of the Climate in 2010”]
.
Bull. Amer. Meteor. Soc.
,
92
,
S115
S121
.
Blunden
,
J.
,
D. S.
Arndt
, and
M. O.
Baringer
, Eds.,
2011
:
State of the Climate in 2010
.
Bull. Amer. Meteor. Soc.
,
92
,
S1
S266
, doi:.
Camargo
,
S. J.
,
M. C.
Wheeler
, and
A. H.
Sobel
,
2009
:
Diagnosis of the MJO modulation of tropical cyclogenesis using an empirical index
.
J. Atmos. Sci.
,
66
,
3061
3074
, doi:.
Carton
,
J. A.
, and
B. S.
Giese
,
2008
:
A reanalysis of ocean climate using Simple Ocean Data Assimilation (SODA)
.
Mon. Wea. Rev.
,
136
,
2999
3017
, doi:.
Chenoweth
,
M.
, and
D.
Divine
,
2008
:
A document-based 318-year record of tropical cyclones in the Lesser Antilles, 1690–2007
.
Geochem. Geophys. Geosyst.
,
9
,
Q08013
, doi:.
Dee
,
D. P.
, and Coauthors
,
2011
:
The ERA-Interim reanalysis: Configuration and performance of the data assimilation system
.
Quart. J. Royal Meteor. Soc.
,
137
,
553
597
, doi:.
Ffield
,
A.
,
2007
:
Amazon and Orinoco River plumes and NBC rings: Bystanders or participants in hurricane events?
J. Climate
,
20
,
316
333
, doi:.
Frank
,
W. M.
, and
P. E.
Roundy
,
2006
:
The role of tropical waves in tropical cyclogenesis
.
Mon. Wea. Rev.
,
113
,
647
658
, doi:.
Grodsky
,
S. A.
, and Coauthors
,
2012
:
Haline hurricane wake in the Amazon/Orinoco plume: AQUARIUS/SACD and SMOS observations
.
Geophys. Res. Lett.
,
39
, L20603, doi:.
Hasumi
,
H.
, and
S.
Emory
,
2004
: K-1 coupled model (MIROC) description. University of Tokyo,Center for Climate System Research Tech. Rep. 1, 38 pp. [Available online at http://ccsr.aori.u-tokyo.ac.jp/~hasumi/miroc_description.pdf.]
Huffman
,
G. J.
,
R. F.
Adler
,
D. T.
Bolvin
, and
G.
Gu
,
2009
:
Improving the global precipitation record: GPCP version 2.1
.
Geophys. Res. Lett.
,
36
,
L17808
, doi:.
Joyce
,
R. J.
,
J. E.
Janowiak
,
P. A.
Arkin
, and
P. P.
Xie
,
2004
:
CMORPH: A method that produces global precipitation estimates from passive microwave and infrared data at high spatial and temporal resolution
.
J. Hydrometeor.
,
5
,
487
503
, doi:.
Jury
,
M. R.
, and
A.
Winter
,
2009
:
Warming of an elevated layer over the Caribbean
.
Climatic Change
, 99 (1–2), 247–259, doi:.
Jury
,
M. R.
,
A.
Winter
, and
B.
Malmgren
,
2007
:
Sub-regional precipitation climate of the Caribbean and relationships with ENSO and NAO
.
J. Geophys. Res.
,
112
,
D16107
, doi:.
Kanamitsu
,
M.
,
W.
Ebisuzaki
,
J.
Woollen
,
S. K.
Yang
,
J. J.
Hnilo
,
M.
Fiorino
, and
G. L.
Potter
,
2002
:
NCEP–DOE AMIP-2 Reanalysis (R-2)
.
Bull. Amer. Meteor. Soc.
,
83
,
1631
1643
, doi:.
Kiladis
,
G. N.
,
M. C.
Wheeler
,
P. T.
Haertel
,
K. H.
Straub
, and
P. E.
Roundy
,
2009
:
Convectively coupled equatorial waves
.
Rev. Geophys.
,
47
, doi:.
Kossin
,
J. P.
,
S. J.
Camargo
, and
M.
Sitkowski
,
2010
:
Climate modulation of North Atlantic hurricane tracks
.
J. Climate
,
23
,
3057
3076
, doi:.
Maloney
,
E. D.
, and
D. L.
Hartmann
,
2000
:
Modulation of hurricane activity in the Gulf of Mexico by the Madden-Julian oscillation
.
Science
,
287
,
2002
2004
, doi:.
Maloney
,
E. D.
, and
A. H.
Sobel
,
2004
:
Surface fluxes and ocean coupling in the tropical intraseasonal oscillation
.
J. Climate
,
17
,
4368
4386
, doi:.
Meehl
G. A.
,
C.
Covey
,
T.
Delworth
,
M.
Latif
,
B.
McAvaney
,
J. F. B.
Mitchell
,
R. J.
Stouffer
, and
K. E.
Taylor
,
2007
:
The WCRP CMIP3 multimodel dataset: A new era in climate change research
.
Bull. Amer. Meteor. Soc.
,
88
,
1383
1394
, doi:.
Molinari
,
J.
,
K.
Lombardo
, and
D.
Vollaro
,
2007
:
Tropical cyclogenesis within an equatorial Rossby wave packet
.
J. Atmos. Sci.
,
64
,
1301
1317
, doi:.
Moninger
,
W. R.
,
R. D.
Mamrosh
, and
P. M.
Pauley
,
2003
:
Automated meteorological reports from commercial aircraft
.
Bull. Amer. Meteor. Soc.
,
84
,
203
216
, doi:.
Pailler
,
K.
,
B.
Bourlès
, and
Y.
Gouriou
,
1999
:
The barrier layer in the western tropical Atlantic Ocean
.
Geophys. Res. Lett.
,
26
,
2069
2072
, doi:.
Pasch
,
R. J.
, and
T. B.
Kimberlain
,
2011
: Tropical Cyclone report: Hurricane Tomas. National Hurricane Center Rep., 20 pp.
Perez
,
C. R.
, and
M. R.
Jury
,
2013
:
Spatial and temporal analysis of climate change in Hispañola
.
Theor. Appl. Climatol.
,
113
,
213
224
, doi:.
Perry
,
G. D.
,
P. B.
Duffy
, and
N. L.
Miller
,
1996
:
An extended data set of river discharges for validation of general circulation models
.
J. Geophys. Res.
,
101
,
21
339
21
349
, doi:.
Reynolds
,
R. W.
,
T. M.
Smith
,
C.
Liu
,
D. B.
Chelton
,
K. S.
Casey
, and
M. G.
Schlax
,
2007
:
Daily high-resolution blended analyses for sea surface temperature
.
J. Climate
,
20
,
5473
5496
, doi:.
Rienecker
,
M. M.
, and Coauthors
,
2011
:
MERRA: NASA’s Modern-Era Retrospective Analysis for Research and Applications
.
J. Climate
,
24
,
3624
3648
, doi:.
Saha
,
S.
, and Coauthors
,
2010
:
The NCEP Climate Forecast System Reanalysis
.
Bull. Amer. Meteor. Soc.
,
91
,
1015
1057
, doi:.
Shieh
,
O. H.
, and
S. J.
Colucci
,
2010
:
Local minimum of tropical cyclogenesis in the eastern Caribbean
.
Bull. Amer. Meteor. Soc.
,
91
,
185
196
, doi:.
Taylor
,
K. E.
,
R. J.
Stouffer
, and
G. A.
Meehl
,
2012
:
An overview of CMIP5 and the experiment design
.
Bull. Amer. Meteor. Soc.
,
93
,
485
498
, doi:.
van Vuuren
,
D. P.
, and Coauthors
,
2011
:
The representative concentration pathways: An overview
.
Climatic Change
,
109
,
5
31
, doi:.
Vizy
,
E. K.
, and
K. H.
Cook
,
2010
:
Influence of the Amazon/Orinoco Plume on the summertime Atlantic climate
.
J. Geophys. Res.
,
115
,
D21112
, doi:.
Wang
,
B.
, and
H.
Rui
,
1990
:
Dynamics of the coupled moist Kelvin–Rossby wave on an equatorial β-plane
.
J. Atmos. Sci.
,
47
,
397
413
, doi:.
Wang
,
C.
,
D. B.
Enfield
,
S.-K.
Lee
, and
C. W.
Landsea
,
2006
:
Influences of the Atlantic warm pool on Western Hemisphere summer rainfall and Atlantic hurricanes
.
J. Climate
,
19
,
3011
3028
, doi:.