1. Introduction and data analysis
The low-lying Turks and Caicos Islands (TCI) are the most southeastern of the Bahamas island chain situated on shallow banks about 200 km north of mountainous Hispaniola. For most of the year, TCI experiences Atlantic trade winds (Klingel 1961; Halkitis et al. 1980) moderated by the warm Antilles Current. Yet the climate is remarkably dry with a mean rainfall of ~1 m yrā1 (Reed 1926) and evaporation of ~2 m yrā1 (Little et al. 1977) that is unfavorable to farming and the recharge of groundwater. Sea surface temperatures (SSTs) reach 29.1Ā°C in late summer, and localized convection arises from cloud bands oriented west of islands (Little et al. 1977; Sullivan 1981). Flood-producing weather disturbances include tropical cyclones and frontal troughs infused with moisture from the Caribbean (Portig 1965; Pagney 1966; Hastenrath 1967; Dimego et al. 1976; Garcia et al. 1978). Bosart and Schwartz (Bosart and Schwartz 1979) found substantial gradients in mean rainfall across the southern Bahamas, and modern datasets reflect a wide range at TCI, 1.2ā2.9 mm dayā1 (gauge), 1.8ā3.0 mm dayā1 (satellite), and 2.1ā3.1 mm dayā1 (model), that motivate this work.
Early reports on the northern Antilles climate stress uniformity of temperature and consistency of easterly winds and isolated showers due to the North Atlantic anticyclonic ridge (Ward 1931) and its subtropical humid air mass. Later research sought to understand how airflow over small flat islands is modified by friction, diurnal heating, and moisture convergence. Even small Antilles islands shed cloud bands due to eddy fluxes of momentum, heat, and moisture (Malkus and McCasland 1949; Malkus 1963; Garstang 1967; Garstang 1972). Conceptual and numerical models have been used to understand the interaction of convection and the island-perturbed flow (Malkus and Stern 1953; Smith 1955; Smith 1957; Estoque and Bhumralkar 1969; Takeda 1971). Aircraft observations of surface temperatures over a Bahamian island show a 7Ā°C increase at noon during undisturbed summer conditions (Bhumralkar 1973) that contribute to sea-breeze convergence and afternoon convection. An analysis of the climate of Hispaniola is provided by Jury and Chiao (Jury and Chiao 2011), wherein the wind flow around the island and its influence on rainfall is outlined.
This paper uses existing mesoscale datasets to describe the mean climate of the TCI area (20Ā°ā24Ā°N, 74Ā°ā70Ā°W). These include North American Regional Reanalysis (NARR) winds [Black 1994; Mesinger et al. 2006; National Oceanic and Atmospheric Administration (NOAA)/Earth System Research Laboratory (ESRL) website: http://www.esrl.noaa.gov/]; National Centers for Environmental PredictionāU.S. Department of Energy Global Reanalysis 2 (NCEP-2) and Coupled Forecast System (CFS) reanalysis surface temperature and winds (Kanamitsu et al. 2002; Saha et al. 2010); Interim European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis rain minus evaporation, air temperature, and winds (Dee et al. 2011; Climate Explorer website: http://climexp.knmi.nl); Quick Scatterometer (QuikSCAT) satellite winds (Risien and Chelton 2008); Moderate Resolution Imaging Spectroradiometer (MODIS) satellite sea and land surface temperature and color [Esaias et al. 1998; Donlon et al. 2002; National Aeronautics and Space Administration (NASA) Giovanni website: http://disc.sci.gsfc.nasa.gov/giovanni]; aircraft profiles via NOAA aircraft meteorological data reports (AMDAR; Moninger et al. 2003); local weather station records via Wunderground; Global Precipitation Climatology Project (GPCP) and Tropical Rainfall Measuring Mission (TRMM)āGlobal Precipitation Climatology Centre (GPCC) rainfall (Adler et al. 2003; Joyce et al. 2004; Liu et al. 2012); Simple Ocean Data Assimilation (SODA) ocean reanalysis currents, temperature, and salinity in the surface layer [Carton and Giese 2008; International Research Institute for Climate and Society (IRI) Climate Library website: http://iridl.ldeo.columbia.edu/]; Weather Research and Forecasting Nonhydrostatic Mesoscale Model (WRF-NMM) weather modelāanalyzed boundary layer height and winds [JanjiÄ 2003; National Operational Model Archive and Distribution System (NOMADS) National Climatic Data Center (NCDC) website: http://nomads.ncdc.noaa.gov/]; Geostationary Operational Environmental Satellite (GOES) infrared cloud temperatures from NASA [Goddard Space Flight Center (GSFC) Giovanni website: http://giovanni.gsfc.nasa.gov/giovanni/]; Food and Agriculture Organization (FAO) of the United Nations marine catch (http://www.fishbase.org); and records of hurricane intensity and tracks since 1890 [Landsea 1993; Coastal Services Center (CSC) of NOAA website: http://csc.noaa.gov/hurricanes/]. While most of the mesoscale fields are analyzed for mean conditions in the period 1990ā2010, some start ~2000 (QuikSCAT, MODIS, and TRMM). A case study TCI cloud band is analyzed on 2 July 2012, with supporting data averaged for JuneāJuly 2012. Climatic controls are addressed by analysis of spatial gradients in mean fields and by calculation of temporal cross correlations between anomaly series smoothed with a 3-month running mean. Wavelet spectral analysis is applied to smoothed SST anomalies for the TCI area: 21.4Ā°ā22.3Ā°N, 72.5Ā°ā71.5Ā°W. Apart from a generalized description of the mean climate, the following scientific questions are posed: What are the main controls on TCI climate? What is the relationship between rainfall and thermodynamic/kinematic variables? What is the climatic influence of Hispaniola? How could the ocean climate affect early colonization? What are the environmental influences on marine resources?
2. Results
2.1. Atmospheric climate and annual cycle
The spatial map of mean surface winds from NARR (Figure 1a) exhibits a distinct area of strong easterly flow (>7 m sā1) located at 21.5Ā°N, 71Ā°W. This wind maximum is associated with a Venturi effect of mountainous island of Hispaniola and an inversion that caps the flow near 2 km (Jury and Chiao 2011). The annual cycle of air temperatures (Figure 1b) rise from 24Ā°C in late winter to 28Ā°C in late summer, a narrow range typical of a subtropical marine air mass. It is constrained by an even narrower range for SST (26Ā°ā28Ā°C) that keeps 1Ā°C above air temperatures throughout the year. Precipitation minus evaporation (P ā E) remains negative over the year, especially in NovemberāFebruary. There is a small peak in May and a broad peak in SeptemberāOctober (P ā E ~ 0) due to seasonal storms, separated by a weak deficit in July.
TCI climate represented by (a) NARR wind vectors and speed (shaded) and (b) annual cycle of ECMWF precipitation minus evaporation, 2-m air temperature, and SST. Annual cycle of (c) CFS zonal and meridional wind and SW radiation and (d) satellite vegetation (NDVI) and aerosol optical depth fraction and surface salinity from SODA. (e) TRMM + GPCC rainfall climatology. (f) Tracks of hurricanes > category 3 within 100 km of TCI since 1890 (red lines). The panels (a)ā(d) are based on 1990ā2010 averages; (e) is averaged for 2000ā10; and (f) is hurricane cases.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
TCI climate represented by (a) NARR wind vectors and speed (shaded) and (b) annual cycle of ECMWF precipitation minus evaporation, 2-m air temperature, and SST. Annual cycle of (c) CFS zonal and meridional wind and SW radiation and (d) satellite vegetation (NDVI) and aerosol optical depth fraction and surface salinity from SODA. (e) TRMM + GPCC rainfall climatology. (f) Tracks of hurricanes > category 3 within 100 km of TCI since 1890 (red lines). The panels (a)ā(d) are based on 1990ā2010 averages; (e) is averaged for 2000ā10; and (f) is hurricane cases.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
TCI climate represented by (a) NARR wind vectors and speed (shaded) and (b) annual cycle of ECMWF precipitation minus evaporation, 2-m air temperature, and SST. Annual cycle of (c) CFS zonal and meridional wind and SW radiation and (d) satellite vegetation (NDVI) and aerosol optical depth fraction and surface salinity from SODA. (e) TRMM + GPCC rainfall climatology. (f) Tracks of hurricanes > category 3 within 100 km of TCI since 1890 (red lines). The panels (a)ā(d) are based on 1990ā2010 averages; (e) is averaged for 2000ā10; and (f) is hurricane cases.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
Key aspects of TCI's seasonal climate are analyzed in Figures 1c,d. Surface zonal winds are consistently ā6 m sā1 and strengthen to almost ā8 m sā1 in July as the North Atlantic anticyclone extends toward the Caribbean in summer. Meridional winds are generally weak (Ā±1 m sā1) and negative from November to April, indicating a northerly component and cool air advection. In summer (JuneāSeptember), meridional winds are positive, representing a southerly component and warm air advection. Incoming shortwave radiation follows the sun angle but has a plateau from April to July near 300 W mā2 and values below 200 W mā2 in DecemberāJanuary. The satellite vegetation [normalized difference vegetation index (NDVI)] fraction remains ~0.1 most of the year, with a slight increase at the end of the rainy season in October. Such a low value is expected from the P ā E deficit and sandy, low-lying soils. The atmospheric aerosol optical depth varies from 0.15 in winter months to 0.35 in JuneāJuly, when Saharan air layers sweep across the Atlantic (Jury and Winter 2009). Surface layer salinity rises in spring [36.5 parts per thousand (ppt)] and declines by the end of the rainy season (36.4 ppt), an annual cycle that is opposite to the vegetation index. Salinity exhibits a narrow range due to the persistent Antilles Current and absence of rivers.
The mean rainfall map (Figure 1e) reveals low values between Haiti and Inagua (21Ā°N, 73Ā°W) that increase to the northeast. The heavier rains derive from passing seasonal weather systems. Tracks of the eight most intense hurricanes affecting TCI in the past century are shown in Figure 1f, based on National Weather Service (NWS) Hurricane Center records. Months with intense hurricanes tend to have SST anomalies > +0.5Ā°C east of Bermuda, which amplifies the upper high and westward track of these storms. The average return period of hurricanes above category 3 is ~20 years: 18 August 1893, 25 July 1926, 16 September 1926, 15 September 1928, 14 September 1945, 7 September 1960, 2 September 2004, and 7 September 2008. Most follow a west-northwest path following development east of the Antilles and passage north of Puerto Rico. Many of these have reshaped the low-lying islands via sustained winds up to 250 km hā1 and storm surges up to 10 m. Reports of hurricanes in the early 1800s āendingā colonial farming efforts and āfreeingā the population played a role in shaping the culture of today. Many intense hurricanes passing TCI continue toward Miami, Florida (Figure 1f).
2.2. Ocean climate
The mean ocean climate is analyzed in Figure 2. Currents sweep northwestward to the east of TCI (72Ā°W), while to the southwest there is a counterclockwise gyre between Haiti and Inagua (Figure 2a). On the eastern flank of the gyre, surface layer currents are northward partly because of Ekman transport beneath the wind maximum. Southward currents on the western flank of the gyre reflect a pulling action by throughflow in the Windward Passage (Jury and Chiao 2011). Temperature and salinity have opposing spatial patterns (Figures 2b,c). Surface layer temperatures are maximum (28Ā°C) north of Haiti and decrease (1Ā°C) northward, as expected from a lower sun angle. Temperature gradients are stronger near TCI. Salinity is highest (36.5) in the Bahamas and decrease southward toward Haiti (36.2). Considering the local atmospheric water budget, it might be expected that salinity would decrease northward. However runoff from Hispaniola and more distant rivers (OrinocoāAmazon) could be a factor.
TCI ocean climate represented by SODA 5ā15-m (a) current vectors (maximum = 0.15 m sā1; the arrow is a gyre) with island vegetation fraction (marshes; reefs are blue), (b) temperature, and (c) salinity. (d) QuikSCAT surface wind climatology.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
TCI ocean climate represented by SODA 5ā15-m (a) current vectors (maximum = 0.15 m sā1; the arrow is a gyre) with island vegetation fraction (marshes; reefs are blue), (b) temperature, and (c) salinity. (d) QuikSCAT surface wind climatology.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
TCI ocean climate represented by SODA 5ā15-m (a) current vectors (maximum = 0.15 m sā1; the arrow is a gyre) with island vegetation fraction (marshes; reefs are blue), (b) temperature, and (c) salinity. (d) QuikSCAT surface wind climatology.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
2.3. QuikSCAT winds and heat budget
A different view on mean winds is afforded by the QuikSCAT satellite (Figure 2d) and confirms strong trade winds (>5 m sā1) between TCI and Hispaniola. Yet a substantial wind shadow zone (<2 m sā1) extends west of Caicos Island past Inagua (21.5Ā°N), indicating frictional drag on the prevailing trades. Using the bulk aerodynamic formula for latent heat flux Qe = Ļ Ce L (U) (qs ā qa)(10ā3), where Ļ is 1.15 kg mā3, the exchange coefficient Ce is ~1.4 Ć 10ā3, the heat of vaporization L is 2.5 Ć 106 J kgā1 , and qs and qa are specific humidity at sea level (22.0) and in the air (15.5 g kgā1); the change in evaporation is from ~150 W mā2 in the (6 m sā1) trade winds to ~50 W mā2 in the (2 m sā1) wind shadow (cf. Figure 8). The predicted increase of SST = (dQe dt)/(M Cp) is ~1Ā°C using residence time dt ~ 4 Ć 105 s, water mass M ~ 104 kg, and heat capacity Cp ~ 4 Ć 103 J kgā1 Kā1. Although small, it occurs near a threshold temperature (27.5Ā°C; Figure 1b) that can transform fair-weather cumulus into thunderstorms, as shown below.
2.4. Ocean vertical structure
The vertical structure of ocean climate is analyzed in TCI longitudes as a section using SODA data averaged over 1990ā2008. The temperature section (Figure 3a) shows a rather level thermocline around 180-m depth. Isotherms tend to fan out northward, suggesting more stable conditions near Hispaniola. A similar pattern is found in salinity (Figure 3b), but there is a fresh layer in the upper 40 m near Hispaniola and a salty layer from 140 to 210 m near TCI. The zonal current structure shows an eastward (westward) flow of 0ā100 m (200ā400 m) within 50 km of Hispaniola. In the vicinity of the Caicos Bank, zonal currents are weak (ā0.01 m sā1), as expected by friction. North of TCI there is broad westward flow that reaches ā0.06 m sā1 in the surface layer (Figure 3c) and from 80- to 360-m depth. The meridional circulation (Figure 3d) is generally northward at depth, with a downwelling (upwelling) component south (north) of TCI below 100 m. In the surface layer the meridional currents are northward (0.03 m sā1), and there is minor uplift in the 25ā75-m layer near Hispaniola.
Ocean climate from SODA cross section averaged over TCI longitudes of (a) temperature, (b) salinity, (c) zonal currents, and (d) meridional currents and vertical motion (exaggerated). The Caicos Bank is represented.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
Ocean climate from SODA cross section averaged over TCI longitudes of (a) temperature, (b) salinity, (c) zonal currents, and (d) meridional currents and vertical motion (exaggerated). The Caicos Bank is represented.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
Ocean climate from SODA cross section averaged over TCI longitudes of (a) temperature, (b) salinity, (c) zonal currents, and (d) meridional currents and vertical motion (exaggerated). The Caicos Bank is represented.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
2.5. Diurnal features
The mean daytime and nighttime SST and land temperature is mapped using MODIS data. A southwestānortheast warm (28Ā°C) to cool (26Ā°C) SST gradient is found (Figures 4a,b), and dayānight differences of SST are ~0.3Ā°C, except in the lee of the Caicos Islands (~1.5Ā°C), where solar radiation (cf. Figure 1c; average of 240 W mā2) reflects off a 60 km Ć 30 km sand bank with depths of 7ā15 m (cf. Figure 8). There is a line of higher SST in deep water on 70.2Ā°W that may be related to downwelling associated with bathymetric undulations and a ābow waveā in trade winds next to Hispaniola (Jury and Chiao 2011). Land temperatures (Figures 4c,d) exhibit greater diurnal range. Mean MODIS dayānight differences are ~5Ā°C, while Providenciales International Airport (PLS) station maximum and minimum temperatures average 31.7Ā° and 21.1Ā°C, respectively. Windwardāleeward gradients appear weak in MODIS land temperatures, possibly due to winter season westerly wind events in the 10-yr record.
Mean daytime and nighttime (a),(b) SST and (c),(d) land temperature from MODIS climatology. TCI station values given in (c),(d) at location of circle.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
Mean daytime and nighttime (a),(b) SST and (c),(d) land temperature from MODIS climatology. TCI station values given in (c),(d) at location of circle.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
Mean daytime and nighttime (a),(b) SST and (c),(d) land temperature from MODIS climatology. TCI station values given in (c),(d) at location of circle.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
2.6. Atmospheric vertical structure
The vertical structure of the lower atmosphere plays a role in surface climate, so northāsouth cross sections of CFS reanalysis data are calculated over 74Ā°ā70Ā°W (TCI). Sinking motions are prevalent (Figure 5a) in the subtropics and especially notable next to Hispaniola 20Ā°ā21.5Ā°N at 850 mb (1.5 km). This is related to the wind maximum and overturning circulations next to Hispaniola (Jury and Chiao 2011). A zonal wind ājetā is exhibited (Figure 5b) below 900 mb (1 km) at 21.5Ā°N (<ā7 m sā1). The CFS climatology locates the wind jet farther north than NARR (cf. Figure 1a), despite similar horizontal resolution. Easterly winds decrease with height, so by 600 mb (4 km) the mean zonal flow is about ā1 m sā1. This wind shear is attributable to the trade wind inversion and upper westerly flow during winter. The sinking motions and trade wind jet near Hispaniola induce a shallow (900 mb, 1 km) moist layer (Figure 5c) at 21Ā°N. The axis of reduced moisture and boundary layer height inhibits rainfall (cf. Figure 1c) and thus groundwater resources in support of island populations.
Vertical cross section of vertical motion (omega; Pa sā1), zonal wind (m sā1), and relative humidity (%) from CFS reanalysis averaged JuneāJuly 2012 in the longitudes 74Ā°ā70Ā°W. The topographic feature on the lower axis is TCI, reaching 12 m.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
Vertical cross section of vertical motion (omega; Pa sā1), zonal wind (m sā1), and relative humidity (%) from CFS reanalysis averaged JuneāJuly 2012 in the longitudes 74Ā°ā70Ā°W. The topographic feature on the lower axis is TCI, reaching 12 m.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
Vertical cross section of vertical motion (omega; Pa sā1), zonal wind (m sā1), and relative humidity (%) from CFS reanalysis averaged JuneāJuly 2012 in the longitudes 74Ā°ā70Ā°W. The topographic feature on the lower axis is TCI, reaching 12 m.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
2.7. Case study: 2 July 2012
An aspect of particular interest is the frictional drag of Caicos Island on the wind and the subsequent effect on thermodynamic forcing of convection. Although Caicos Island is flat and surface roughness is low, its northwestāsoutheast orientation protects a shallow bankācreating a āfootprintā in wind and SST fields (cf. Figures 2d, 4a). The case study of 2 July 2012 is characterized by undisturbed summer trade wind conditions. AMDAR aircraft profiles at PLS (Figures 6aāc) reveal wind directions from ~100Ā° at speeds increasing from 5 to 12 m sā1 up to 1500 ft. Temperatures decrease at the moist adiabatic rate (~7Ā°C kmā1) with few discontinuities. Back trajectories arriving at TCI (Figure 6d) reflect steady airflow from the east-southeast with a sinking component contributed by the background Hadley circulation. Weather station records during the morning of 2 July 2012 (Figure 6e) exhibit temperatures and dewpoints of 29Ā° and 24Ā°C, respectively. Winds were steady from the east at speeds of 20 km hā1 with gusts to 35 km hā1. Intermittent dips in the wind, pressure, and temperature traces reflect the passage of convective showers between 1200 and 1600 local standard time (LST) (there is no detailed rain record).
Aircraft profiles at TCI in the period 1ā2 Jul 2012 for (a)ā(c) wind direction, speed, and temperature. (d) Airflow from back trajectories arriving at TCI at 1400 LST 2 Jul 2012. (e) Weather station records at TCI on 2 Jul 2012 (location in Figure 7b).
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
Aircraft profiles at TCI in the period 1ā2 Jul 2012 for (a)ā(c) wind direction, speed, and temperature. (d) Airflow from back trajectories arriving at TCI at 1400 LST 2 Jul 2012. (e) Weather station records at TCI on 2 Jul 2012 (location in Figure 7b).
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
Aircraft profiles at TCI in the period 1ā2 Jul 2012 for (a)ā(c) wind direction, speed, and temperature. (d) Airflow from back trajectories arriving at TCI at 1400 LST 2 Jul 2012. (e) Weather station records at TCI on 2 Jul 2012 (location in Figure 7b).
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
The (WRF-NMM) mesoscale weather model analysis at 1400 LST 2 July 2012 is given in Figure 7a. The island's friction and heat fluxes deepen the boundary layer to 1300-m coincident with a wind wake < 6 m sā1. The cloud band in GOES imagery (Figure 7b) has an east-southeast orientation with top temperatures reaching <ā50Ā°C at 1300 LST 2 July 2012. Windward of Caicos Island the SST is 28Ā°C compared with 29Ā°C leeward. The 60 km Ć 30 km pool of warmer water (Figure 7c) coincides with a shallow sand bank (ā10 m) and lower evaporation. This extends the island footprint where confluence and convection interact. However, the 12-km grid spacing of the weather model is barely able to resolve these features. The MODIS vegetation index (Figure 7d) exhibits contrasts in land color at 5-km resolution that suggest windwardāleeward effects and low values of 0.02ā0.4.
(a) Mesoscale weather model analysis of smoothed boundary layer height (shaded; m) and winds (arrows; contours; m sā1) at 1400 LST 2 Jul 2012. (b) GOES infrared cloud temperatures at 1300 LST 2 Jul 2012 (Ā°C) and weather station (circle). Mean (c) MODIS SST and (d) vegetation fraction averaged JuneāJuly 2012. Maps here cover 1Ā° Ć 1Ā°.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
(a) Mesoscale weather model analysis of smoothed boundary layer height (shaded; m) and winds (arrows; contours; m sā1) at 1400 LST 2 Jul 2012. (b) GOES infrared cloud temperatures at 1300 LST 2 Jul 2012 (Ā°C) and weather station (circle). Mean (c) MODIS SST and (d) vegetation fraction averaged JuneāJuly 2012. Maps here cover 1Ā° Ć 1Ā°.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
(a) Mesoscale weather model analysis of smoothed boundary layer height (shaded; m) and winds (arrows; contours; m sā1) at 1400 LST 2 Jul 2012. (b) GOES infrared cloud temperatures at 1300 LST 2 Jul 2012 (Ā°C) and weather station (circle). Mean (c) MODIS SST and (d) vegetation fraction averaged JuneāJuly 2012. Maps here cover 1Ā° Ć 1Ā°.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
2.8. Climate, colonization, and marine resources
Environmental controls on colonization include sea level, soils and aridity, and navigation factors (Sears and Sullivan 1978; Keegan and Mitchell 1986; Keegan 1993). Easterly trade winds prevail most of the year and sustain the Antilles Current, which flows through the islands. It is theorized that present conditions were similar to those encountered by initial colonists, with a few exceptions. Archaeological evidence suggests that Stone Age peoples first visited the southern Bahamas from northern Cuba more than 2000 years ago when sea levels were lower (Hodell et at. 1991; Granberry 1993; Fitzpatrick 2009; Cooper and Peros 2010). At that time, the Caicos would have been a single island larger than Inagua. Early colonization was driven by exploration with adequate boat technology and navigation. Marine transportation networks were formed according to currents, winds, and waves; island size, configuration, and safe harbors (Keegan and Diamond 1987; Keegan 1992; Callaghan 2001; Hofman et al. 2007); and seasonal availability of marine resources (Sullivan 1980; Sullivan 1981; Winter et al. 1985; Berman and Gnivecki 1995).
It is suggested here that early colonization routes between Cuba, Hispaniola, and TCI made use of the counterclockwise current gyre (Figure 2a) with northward drift on its east flank due to Ekman transport and gradients in the wind field (Figure 1a). Weak southward currents on the west flank of the gyre near Inagua enabled return voyages for resupply of materials and people and exchange of information (Anthony 1990). There is little evidence to prove this theory, only the assumption that the configuration of large Antilles Islands and regional climate was similar to conditions of 2000 years ago, with exception of lower sea level and contiguous TCI.
Ancient mariners would have noticed the turquoise color of the shallow Caicos Bank against the adjacent deep blue ocean (Figure 8). Other navigational aids would have included island-distorted winds (Figure 2d) and cloud bands (Figure 7b) and the celestial positions (Lewis 1972). Sailing northāsouth is crosswind over much of the area (Figure 1a) and thus weak currents in the area enabled primitive boats to pass from Cuba and Hispaniola to TCI, in both ancient and modern times. Changing trends in TCI population (Figure 9a) suggest renewed migration.
Visible satellite image of the Caicos Islands illustrating the shallow bank of turquoise water.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
Visible satellite image of the Caicos Islands illustrating the shallow bank of turquoise water.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
Visible satellite image of the Caicos Islands illustrating the shallow bank of turquoise water.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
(a) Annual marine catch in TCI waters and population. Regression of conch catch onto (b) ECMWF zonal winds and (c) CFS surface temperature in the period 1979ā2007.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
(a) Annual marine catch in TCI waters and population. Regression of conch catch onto (b) ECMWF zonal winds and (c) CFS surface temperature in the period 1979ā2007.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
(a) Annual marine catch in TCI waters and population. Regression of conch catch onto (b) ECMWF zonal winds and (c) CFS surface temperature in the period 1979ā2007.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
Most Turks and Caicos islanders live from the sea through either fisheries or tourism (Sadler 1997). Conch is the main species of commercial catch, accounting for >90% of the total. It oscillates from 3000 to 7000 tons yrā1 since 1975 (Figure 9a), with significant cycling at 2.6ā3.0 yr in the period 1980ā95. Making a regression of its time series on annual environmental data for the surrounding area, it is found that stronger easterly winds and cooler SST correspond with years of higher conch catch (Figures 9b,c). It is conjectured that such conditions 1) coincide with less frequent storms and greater catch effort and 2) provide sustained ventilation beneficial to the conch physiology. Further in-depth analysis is needed to clarify these points.
2.9. Climatic controls and interrelationships
Environmental influences on TCI climate are analyzed using mean spatial fields in a 5 Ć 5 matrix (~12 degrees of freedom) and via smoothed anomaly time series (1990ā2010; ~22 degrees of freedom). The spatial analysis investigates controls on the rainfall gradient by regridding the datasets to a common 1Ā° resolution over the study area 20Ā°ā24Ā°N, 74Ā°ā70Ā°W, then calculating the cross correlations between two rainfall fields and a number of thermodynamic and kinematic fields (defined in Table 1). It is found that lower rainfall in the south is related to stronger trade winds (āU) accompanied by a more northerly (āV) sinking (āW) component. Conversely, higher rainfall in the north is related to weaker trade winds with a more southerly rising component. The rainfall gradient is opposed to the sea temperature, salinity, and precipitable water fields; hence, there is little thermodynamic control. Instead, the kinematic effect of wind acceleration next to Hispaniola is a key factor producing the gradients in rainfall, as described in scatterplots (Figures 10aād). Some environmental relationships are ambiguous, as seen by contrasting correlations between the two rainfall fields (cf. Table 1, divergence and vorticity).
Correlation of spatial fields. Correlation coefficients are generated from a 5 Ć 5 xāy matrix of 1Ā° gridded mean values over the period of record: cm Rain denotes the multisatellite gauge merged rainfall; gp Rain is the GPCP satellite rainfall; ec tauX and tauY are the ECMWF zonal and meridional wind stress; soda S and T are the SODA ocean reanalysis 5ā15-m salinity and temperature; cfs W 850 denotes the CFS reanalysis vertical motion at 850 mb; cfs U and V 850 are the CFS reanalysis zonal and meridional winds at 850 mb; nc div and vort 850 denote the NCEP reanalysis divergence and vorticity at 850 mb; and nc Pwat is the NCEP reanalysis precipitable water. Bold values are significant above 90% confidence for ~12 degrees of freedom: r > 0.46. Italic values have a sign āopposite to expected.ā Scatterplots for key cross correlations are given in Figure 10.
Scatterplots between rainfall (y axis) and kinematic variables from intercomparisons of (a)ā(d) spatial fields corresponding with Table 1 and (e),(f) temporal anomaly series corresponding with Table 2. (a),(b) GPCP rainfall (mm dayā1) and CFS vertical motion (omega; Pa sā1) and zonal wind at 850 mb (m sā1). (c),(d) Satellite/gauge merged rainfall and ECMWF meridional wind stress (N mā2) and CFS meridional wind at 850 mb. (e) ECMWF precipitation minus evaporation (mm dayā1) and NCEP wind stress vorticity (Ć10ā7 sā1) and (f) GPCP rainfall and Hadley sea level pressure (mb). Linear fit and equations are given. Correlations for these and other variables are listed in Tables 1 and 2.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
Scatterplots between rainfall (y axis) and kinematic variables from intercomparisons of (a)ā(d) spatial fields corresponding with Table 1 and (e),(f) temporal anomaly series corresponding with Table 2. (a),(b) GPCP rainfall (mm dayā1) and CFS vertical motion (omega; Pa sā1) and zonal wind at 850 mb (m sā1). (c),(d) Satellite/gauge merged rainfall and ECMWF meridional wind stress (N mā2) and CFS meridional wind at 850 mb. (e) ECMWF precipitation minus evaporation (mm dayā1) and NCEP wind stress vorticity (Ć10ā7 sā1) and (f) GPCP rainfall and Hadley sea level pressure (mb). Linear fit and equations are given. Correlations for these and other variables are listed in Tables 1 and 2.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
Scatterplots between rainfall (y axis) and kinematic variables from intercomparisons of (a)ā(d) spatial fields corresponding with Table 1 and (e),(f) temporal anomaly series corresponding with Table 2. (a),(b) GPCP rainfall (mm dayā1) and CFS vertical motion (omega; Pa sā1) and zonal wind at 850 mb (m sā1). (c),(d) Satellite/gauge merged rainfall and ECMWF meridional wind stress (N mā2) and CFS meridional wind at 850 mb. (e) ECMWF precipitation minus evaporation (mm dayā1) and NCEP wind stress vorticity (Ć10ā7 sā1) and (f) GPCP rainfall and Hadley sea level pressure (mb). Linear fit and equations are given. Correlations for these and other variables are listed in Tables 1 and 2.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
Temporal variability is analyzed by extracting 3-month smoothed anomalies in the period 1990ā2010 (N = 263) for a 1Ā° grid box at TCI (21.4Ā°ā22.3Ā°N, 72.5Ā°ā71.5Ā°W) for variables defined in Table 2. Cross correlations are calculated with respect to GPCP rainfall and ECMWF P ā E, and scatterplots are analyzed (Figures 10e,f). Only a few local variables exhibit significant negative relationships (<ā0.34): sea level pressure and wind stress curl. While there is a significant 4.3-yr cycle in the smoothed SST anomaly time series after 1995 (Figures 11a,b), the relationship with P ā E is weakly positive. The SLP relationship indicates that lower pressure corresponds with higher rainfall via storm activity as expected. The curl relationship indicates a more cyclonic (ā) wind flow pattern corresponds with higher rainfall. However the kinematic variables at the surface (Us, Vs) and upper level (200 mb) reveal weak correlation in the study area. Hence, year-to-year fluctuations of TCI climate are likely controlled at larger scales via North Atlantic and eastern Pacific oceanāatmosphere coupling (Jury and Malmgren 2011).
Correlation of temporal series. Variables consist of TCI-area 3-month-smoothed anomaly time series in the period 1990ā2010: ec P ā E is the ECMWF precipitation minus evaporation; oi SST denotes the optimal interpolated NOAA sea surface temperature (cf. Figure 11a); ha SLP is the Hadley Centre sea level pressure; na Hc denotes the NOAA 0ā500-m ocean heat content; ec Us and Vs are the ECMWF surface zonal and meridional wind; ec Lflux is the ECMWF latent heat flux; cf U 200 and V 200 are the CFS reanalysis zonal and meridional winds at 200 mb; and nc Curl is the NCEP reanalysis wind stress vorticity. Bold values are significant above 90% confidence for ~22 degrees of freedom: r > 0.34. Scatterplots for key cross correlations are given in Figure 10.
TCI area (a) 3-month running mean SST anomaly and (b) wavelet spectral analysis, with power shaded above 90% confidence of validity.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
TCI area (a) 3-month running mean SST anomaly and (b) wavelet spectral analysis, with power shaded above 90% confidence of validity.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
TCI area (a) 3-month running mean SST anomaly and (b) wavelet spectral analysis, with power shaded above 90% confidence of validity.
Citation: Earth Interactions 17, 18; 10.1175/2012EI000490.1
3. Conclusions
This brief study of Turks and Caicos Islands (TCI) climate has uncovered the following key points:
The zone of accelerated trade winds north of Hispaniola enhances anticyclonic subsidence and forms a dry-south/wet-north pattern of rainfall.
The mean rainfall gradient is opposed to atmospheric and oceanic thermodynamic patterns.
Ocean currents east of TCI are northwestward and consistent with an Antilles Current; however, surface currents southwest of TCI exhibit a counterclockwise gyre that may have supported colonization from Haiti.
Conch catch, the dominant marine resource, fluctuates by a factor of 2 from year to year, and high catch coincides with stronger trade winds and cooler SST.
The dry climate of TCI limits groundwater resources and, together with poor soils, inhibits food production (cf. vegetation fraction of ~0.1 and population density of 65 inhabitants per square kilometer).
Caicos Island has a shallow sand bank below the wind wake that boosts SST and local convection. Even operational weather models and QuikSCAT observations reflect this feature.
References
Adler, R. F., and Coauthors, 2003: The Version-2 Global Precipitation Climatology Project (GPCP) monthly precipitation analysis (1979āpresent). J. Hydrometeor., 4, 1147ā1167.
Anthony, D. W., 1990: Migration in archeology: The baby and the bathwater. Amer. Anthropol., 92, 895ā914.
Berman, M. J., and P. L. Gnivecki, 1995: The colonization of the Bahama archipelago: A reappraisal. World Archaeol., 26, 421ā441.
Bhumralkar, C. M., 1973: An observational and theoretical study of atmospheric flow over a heated island: Part II. Mon. Wea. Rev., 101, 719ā730.
Black, T. L., 1994: The new NMC mesoscale Eta model: Description and forecast examples. Wea. Forecasting, 9, 265ā284.
Bosart, L. F., and B. E. Schwartz, 1979: Autumnal rainfall climatology of the Bahamas. Mon. Wea. Rev., 107, 1663ā1672.
Callaghan, R. T., 2001: Ceramic age seafaring and interaction potential in the Antilles: A computer simulation. Curr. Anthropol., 42, 308ā313.
Carton, J. A., and B. S. Giese, 2008: A reanalysis of ocean climate Using Simple Ocean Data Assimilation (SODA). Mon. Wea. Rev., 136, 2999ā3017.
Cooper, J., and M. Peros, 2010: The archaeology of climate change in the Caribbean. J. Archaeol. Sci., 37, 1226ā1232.
Dee, D. P., and Coauthors, 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553ā597.
Dimego, G. J., L. F. Bosart, and G. W. Endersen, 1976: An examination of the frequency and mean conditions surrounding frontal incursions into the Gulf of Mexico and Caribbean Sea. Mon. Wea. Rev., 104, 709ā718.
Donlon, C. J., P. J. Minnett, C. Gentemann, T. J. Nightingale, I. J. Barton, B. Ward, and M. J. Murray, 2002: Toward improved validation of satellite sea surface skin temperature measurements for climate research. J. Climate, 15, 353ā369.
Esaias, W. E., and Coauthors, 1998: An overview of MODIS capabilities for ocean science observations. IEEE Trans. Geosci. Remote Sens., 36, 1250ā1265.
Estoque, M. A., and C. M. Bhumralkar, 1969: Flow over a localized heat source. Mon. Wea. Rev., 97, 850ā859.
Fitzpatrick, S. M., 2009: Saladoid seafarers: On the origins and migrations of early ceramic age Amerindian populations. Proc. Symp. Caribbean Basin before Columbus, Washington., D.C., Pre-Columbian Society, 1ā18.
Garcia, O., L. F. Bosart, and G. J. Dimego, 1978: On the nature of winter season rainfall in the Dominican Republic. Mon. Wea. Rev., 106, 961ā982.
Garstang, M., 1967: The role of momentum exchange in flow over a heated island. Proc. Conf. Tropical Meteor., Miami, FL, U.S. Army, 178ā195.
Garstang, M., 1972: Review of hurricane and tropical meteorology. Bull. Amer. Meteor. Soc., 53, 612ā630.
Granberry, J., 1993: The people who discovered Columbus: The prehistory of the Bahamas, a review and commentary. Fla. Anthropol.,6, 56ā60.
Halkitis, M., S. Smith, and K. Rigg, 1980: The Climate of the Bahamas. Bahamas Geographical Association, 122 pp.
Hastenrath, S. L., 1967: Rainfall distribution and regime in Central America. Arch. Meteor. Geophys. Bioklimatol., B15, 201ā241.
Hodell, D. A., J. H. Curtis, G. A. Jones, A. Higuera-Gundy, M. Brenner, M. W. Binford, and K. T. Dorsey, 1991: Reconstruction of Caribbean climate change over the past 10,500 years. Nature, 352, 790ā793.
Hofman, C. L., A. J. Bright, A. Boomert, and S. Knippenberg, 2007: The web of social relationships and interaction networks in the pre-Columbian Lesser Antilles. Lat. Amer. Antiq., 18, 243ā268.
JanjiÄ, Z. I., 2003: A non-hydrostatic model based on a new approach. Meteor. Atmos. Phys., 82, 271ā285.
Joyce, R. J., J. E. Janowiak, P. A. Arkin, and 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, 487ā503.
Jury, M. R., and A. Winter, 2009: Warming of an elevated layer over the Caribbean. Climatic Change, 99, 247ā259, doi:10.1007/s10584-009-9658-3.
Jury, M. R., and S. Chiao, 2011: Mesocirculation associated with summer convection over the central Antilles. Earth Interact., 15, doi:10.1175/2011EI391.1.
Jury, M. R., and B. A. Malmgren, 2011: Joint modes of climate variability across the inter-Americas. Int. J. Climatol., 32, 1033ā1046, doi:10.1002/joc.2324.
Kanamitsu, M., W. Ebisuzaki, J. Woollen, S.-K. Yang, J. J. Hnilo, M. Fiorino, and G. L. Potter, 2002: NCEPāDOE AMIP-II Reanalysis (R-2). Bull. Amer. Meteor. Soc., 83, 1631ā1643.
Keegan, W. F., 1992: Lucayan settlement patterns and recent coastal changes in the Bahamas. Paleoshorelines and Prehistory: An Investigation of Method, L. L. Johnson, Ed., CRC Press, 5ā14.
Keegan, W. F., 1993: Inagua archaeology. Florida Museum of Natural History Department of Anthropology Rep. 51, 29 pp.
Keegan, W. F., and S. W. Mitchell, 1986: Lucayan Arawak artifact distributions, Bahama Islands. J. Field Archaeol., 13, 255ā258.
Keegan, W. F., and J. M. Diamond, 1987: Colonization of islands by humans: A biogeographical perspective. Advances in Archaeological Methods and Theory, Vol. 10, M. B. Schiffer, Ed., Academic Press, 49ā91.
Klingel, G., 1961: The Ocean Island (Inagua). Natural History Press, 415 pp.
Landsea, C. W., 1993: A climatology of intense (or major) Atlantic hurricanes. Mon. Wea. Rev., 121, 1703ā1713.
Lewis, D., 1972: We, the Navigators: The Ancient Art of Landfinding in the Pacific. University Press Hawaii, 346 pp.
Little, B. G., D. K. Buckley, R. Cant, P. W. T. Henry, A. Jefferiss, J. D. Mather, J. Stark, and R. N. Young, Eds., 1977: Land resources of the Bahamas: A summary. Ministry of Overseas Development Land Resource Study 27, 143 pp.
Liu, Z., D. Ostrenga, W. Teng, and S. Kempler, 2012: Tropical Rainfall Measuring Mission (TRMM) precipitation data and services for research and applications. Bull. Amer. Meteor. Soc., 93, 1317ā1325.
Malkus, J. S., 1963: Tropical rain induced by a small natural heat source. J. Appl. Meteor., 2, 547ā556.
Malkus, J. S., and K. McCasland, 1949: Observational studies of convection. Woods Hole Oceanographic Institution Tech. Rep. 3, 44 pp.
Malkus, J. S., and M. E. Stern, 1953: The flow of a stable atmosphere over a heated islandāPart 1. J. Meteor., 10, 30ā41.
Mesinger, F., and Coauthors, 2006: North American Regional Reanalysis. Bull. Amer. Meteor. Soc., 87, 343ā360.
Moninger, W. R., R. D. Mamrosh, and P. M. Pauley, 2003: Automated meteorological reports from commercial aircraft. Bull. Amer. Meteor. Soc., 84, 203ā216.
Pagney, P., 1966: Le Climat des Antilles. Institut des Hautes Etudes de l'Amerique Latine, 486 pp.
Portig, W. H., 1965: Central American rainfall. Geogr. Rev., 15, 68ā90.
Reed, W. W., 1926: Climatological data for the West Indian Islands. Mon. Wea. Rev., 54, 133ā160.
Risien, C. M., and D. B. Chelton, 2008: A global climatology of surface wind and wind stress fields from eight years of QuikSCAT scatterometer data. J. Phys. Oceanogr., 38, 2379ā2413.
Sadler, H. E., 1997: Turks Islands Landfall: A History of the Turks and Caicos Islands. United Cooperative Printers, 299 pp.
Saha, S., and Coauthors, 2010: The NCEP Climate Forecast System Reanalysis. Bull. Amer. Meteor. Soc., 91, 1015ā1057.
Sears, W. H., and S. D. Sullivan, 1978: Bahamas prehistory. Amer. Antiq., 43, 3ā25.
Smith, R. C., 1955: Theory of air flow over a heated land mass. Quart. J. Roy. Meteor. Soc., 81, 382ā395.
Smith, R. C., 1957: Air motion over a heated land mass-part 2. Quart. J. Roy. Meteor. Soc., 83, 248ā256.
Sullivan, S. D., 1980: An overview of the 1976 to 1978 archeological investigations in the Caicos Islands. Fla. Anthropol., 33, 120ā151.
Sullivan, S. D., 1981: Prehistoric patterns of exploitation and colonization in the Turks and Caicos Islands. PhD thesis, University of Illinois at UrbanaāChampaign, 460 pp.
Takeda, T., 1971: Numerical simulation of a precipitating convective cloud: The formation of long-lasting cloud. J. Atmos. Sci., 28, 350ā376.
Ward, R., 1931: A summer cruise in the West Indies. Mon. Wea. Rev., 59, 331ā339.
Winter, J., J. Granberry, and A. Leibold, 1985: Archaeological investigations within the Bahamas. Proc. 10th Int. Congress for the Study of Pre-Columbian Cultures of the Lesser Antilles, Montreal, Canada, Bahamas Historical Society, 83ā92.
