Every year huge quantities of soil dust are carried by winds from Africa, across the Atlantic, and to the Caribbean. No other ocean region is so extensively and persistently impacted by such high concentrations of dust, a region that extends over 7000 km from the coast of Africa to the Caribbean and to the bounding continental shores of the Americas. In effect, the Caribbean Basin can be thought of as the “receptor” site of the Saharan dust “source,” a source that accounts for over half of global dust emissions. It is generally recognized that on a global scale mineral dust can affect many aspects of climate, marine biogeochemical processes, soil fertility, air quality, and human health. However, it is difficult to assess the impact on the Caribbean Basin because of the dearth of studies in this region. A better understanding is needed of the factors that affect the transport of dust, the physical and chemical properties of the transported materials, and how these might change during transport. Of particular interest is how climate change might affect dust transport in the future. Presented here is a brief overview of research relevant to the region and issues in dust-related research that need to be addressed.
The huge quantities of African dust carried into the Caribbean Basin warrant the formation of an observational program that tracks the impacts of this transport and future changes linked to climate variability.
AFRICAN DUST TRANSPORT, METEOROLOGY, AND CLIMATE.
The arid regions of North Africa are estimated to emit about 800 Tg yr−1 of soil dust each year, 70% of the global total and six times more than the next largest source, Asia (Huneeus et al. 2011). A large fraction of these emissions are carried to the west over the Atlantic Ocean. Satellite images of individual dust outbreaks (Fig. 1) often show dust extending in a continuous plume from the coast of Africa into the Caribbean Basin. Satellite measurements of aerosol properties (e.g., aerosol optical depth; Hsu et al. 2012) clearly show the temporal and spatial distribution of dust over this huge region. While satellites serve as a valuable tool for studying dust transport, our detailed understanding of the long-term record of this transport, its link to climate, and the specific impact on the Caribbean Basin rests largely on the aerosol records accumulated at Barbados (Prospero and Lamb 2003), Puerto Rico (Gioda et al. 2013; Reid et al. 2003a), and Miami (Prospero 1999) . The longest records are from Barbados, where studies began in 1965, and from Miami, starting in 1974. These measurements continue to this day. A portion of these records, from 2004 to 2009, is shown in Fig. 2. Concentrations are greatest in Barbados and show a strong seasonal cycle with a summer maximum and winter minimum. Miami dust concentrations track those in Barbados in a general way. The Miami record differs in that concentrations are generally less than those in Barbados and the dust transport season is shorter. There are large year-to-year differences between the two records, which are attributed to changes in large-scale winds and removal by precipitation. The similarities and differences between these two sites provide a sense of what we might expect to see across the entire basin.
The Barbados measurements show that large changes have occurred over the 48-yr record. Initially these changes seemed to be linked to African climate (Prospero and Lamb 2003). Concentrations were low in the late 1960s, at the end of a long wet phase in West Africa. Transport dramatically increased in the early 1970s in apparent response to drought and again in the early 1980s when drought was most intense. Over this period trade-wind dust concentrations were highly anticorrelated to Sahel precipitation, used as a proxy for general changes in source-region climate (Prospero and Lamb 2003). There were also suggestions of relationships to major climate indices, such as El Niño–Southern Oscillation (ENSO) (Prospero and Lamb 2003), the Atlantic multidecadal oscillation (AMO) (Evan et al. 2011; Wang et al. 2012), and the North Atlantic Oscillation (NAO) (Ginoux et al. 2004; Evan et al. 2006). Since the late 1980s, rainfall has improved in the Sahel, and in some years has been plentiful. Indeed, there is evidence of “greening” across the region (de Jong et al. 2011). Nonetheless, Barbados dust concentrations have remained higher than pre-drought levels. Moreover, in contrast to the first 25 years of the record, there is no clear relationship to Sahel rainfall or to the cited (and other) climate indices. The absence of such linkages makes it difficult to predict how dust emissions and transport might change over the coming decades as climate changes. The problem is exacerbated by the inability of models (Seneviratne et al. 2012) to agree on future rainfall trends over large areas of North Africa (including the Sahel) that are known to be major dust sources today and in the recent past.
However, there are factors other than rainfall and dust source activity in Africa that could impact transport to the Caribbean (Engelstaedter et al. 2009). If we wish to understand present-day variability and to predict future trends, we need a better understanding of the entire dust cycle, starting with the processes that affect dust mobilization in Africa, the meteorological environment that controls transport including the transit time, and finally the removal en route, especially by precipitation.
There is continuing debate about the nature of the most active dust sources and the relative importance of small-scale sources (dust “hot spots”) (Ginoux et al. 2012; Okin et al. 2011) such as dry lakes, playas, and wadis as contrasted to larger-scale terrain characteristics and climate (Bullard et al. 2011; Muhs 2013; Prospero et al. 2002). Of particular importance, and still highly uncertain, is the role of anthropogenic land disturbance, which could account for a large fraction of current emissions (Ginoux et al. 2012; Mahowald et al. 2010). There have been a number of major field campaigns in North Africa during the past decade that have led to a better grasp of dust mobilization processes, most notably the African Monsoon Multidisciplinary Analyses program (AMMA; Mari et al. 2011) and the Saharan Mineral Dust Experiment (SAMUM; Heintzenberg 2009). While these and other studies have greatly broadened our knowledge of dust processes in North Africa, we nonetheless continue to lack a good understanding of the meteorological process that are the most important drivers of dust emissions (e.g., frontal systems, haboobs, mesoscale cyclonic systems, and small-scale turbulent processes; Knippertz and Todd 2012). The Cloud–Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO) dust product shows a strong link to the seasonal migration of the intertropical convergence zone over West Africa and monsoon dynamics (Adams et al. 2012). All these processes could play a fundamental role in dust emissions; the relative importance of these processes could change with season and with changing climate (Engelstaedter and Washington 2007; Williams 2008).
Moreover, dust can itself impact atmospheric and ocean processes to a degree that could affect weather and climate over the Atlantic and the Caribbean Basin. Of particular interest is the evidence that high dust concentrations and the meteorological environment associated with Saharan air outbreaks could modulate the growth of tropical cyclones (Dunion and Velden 2004). Evan et al. (2011) suggest that variations in dust transport over time are negatively correlated to sea surface temperature changes over the tropical Atlantic. These changes, in turn, could be linked to the suggested negative correlation between Atlantic dustiness and hurricane activity. Cloud microphysics could also be affected by African dust, which can serve as both condensation (Twohy et al. 2009) and freezing nuclei (Cziczo et al. 2013; Heymsfield et al. 2009). Because of the complexity of the ocean–atmosphere processes in this region and the impact on cloud processes (Rauber et al. 2007) it is difficult to quantify the role of dust in this context at this time. Nonetheless, there is a clear need to focus on these relationships.
Models are essential to the development of an understanding of the entire dust cycle. However, dust models are in an early stage of development. A recent intercomparison of 15 global models in the Aerosol Comparisons between Observations and Models (AEROCOM) project (Huneeus et al. 2011) shows large disparities. Estimates of global emissions span a wide range, from 514 to 4313 Tg yr−1, and those from North Africa range from 204 to 2888 Tg yr−1. On regional scales, models had difficulties in reproducing concentrations and deposition, most notably the annual dust cycle at Barbados and Miami and the southward shift of the plume in boreal winter that results in transport to South America (Prospero et al. 1981) and the Amazon basin (Swap et al. 1992). Model development is hampered by the dearth of dust measurements over the oceans. Given the wide impact of dust over the Atlantic and the Caribbean Basin, measurements in this region would contribute greatly to model development both to better constrain emission estimates in North Africa and to better characterize transport and deposition over the receptor region.
Improved model performance will require a more accurate description of all processes controlling dust mobilization and distribution including the evolution and persistence of the strong stratification of dust layers. Early research (Carlson and Prospero 1972; Prospero and Carlson 1972) postulated the existence of the Saharan air layer (SAL)—a hot, dry, elevated layer within which the highest dust concentrations are usually found (Reid et al. 2003a). The SAL is a persistent feature in sonde profiles over the region in summer (Dunion 2010). Long-term lidar studies on Barbados carried out by the University of Miami (http://mplnet.gsfc.nasa.gov/index.html; select “Ragged Point”) and the Max Planck Institute for Meteorology in Hamburg (http://barbados.zmaw.de/) confirm the often-layered structure of dust events although there is considerable variability in dust vertical distributions. An example of a dust event (Fig. 3a) shows a complex but well-defined series of layers above the surface mixed layer. Sea-salt aerosol dominates the lidar backscatter below about 500-m altitude although dust is also present based on measurements at Ragged Point. The lidar data coupled with concurrent aerosol optical depth measurements (http://aeronet.gsfc.nasa.gov/; select “Ragged Point”) enables the extraction of the vertical profiles of extinction; in the example (Fig. 3b) the low-level sea-salt layer and the elevated dust layers are clearly distinguished. Concurrent sonde data from the local airport show the presence of a classical SAL profile of temperature and water vapor.
On a larger scale, the complex and dynamic relationship between dust source activity in North Africa, vertical distributions, and meteorological fields is clearly evident in the CALIPSO lidar aerosol product (Adams et al. 2012). Figure 4 shows the distribution of dust in boreal summer over Africa, the Atlantic, and Caribbean. Dust appears primarily as an elevated plume that is consistent with the canonical concept of a SAL distribution. The layer is clearly seen to persist, although progressively weakened, across the Atlantic and into the western Caribbean. The persistence of this layer over such a great distance presents challenges to our understanding of how dust and other aerosols are transported over great distances.
Dust models are being used to produce routine forecasts of dust emissions and transport. The WMO Sand and Dust Storm Warning Advisory and Assessment System (SDSWAS) website (http://sds-was.aemet.es/forecast-products/dust-forecasts) has links to 14 daily forecast products. Of these, 10 include the Caribbean Basin and the Americas in their coverage. These products, coupled with ground-based aerosol measurements in the region, could serve as the basis of an air quality alerting system as discussed below.
DUST PARTICLE CHARACTERIZATION.
The impact of dust particles on climate, ocean productivity, terrestrial ecosystems, and human health will ultimately depend on the concentrations of dust and the chemical and physical properties of individual particles. These properties will change during transit, most notably because of the loss of large particles and the shift of the particle size spectrum to smaller sizes, but also because of chemical processing (“aging”) during transit (Prather et al. 2008; Usher et al. 2003). The impact of these processes would be most noticeable in dust-laden air masses that mix with pollutants (Shi et al. 2012; Meskhidze et al. 2005)—for example, European pollutants mixing with dust over the Sahara. Of particular interest is the effect on the solubility of Fe in dust particles and, after deposition, the impact on ocean primary productivity, the carbon cycle, and climate (Jickells et al. 2005). At present, this assessment is limited by our lack of understanding of the properties that render dust-Fe “bio-available.” In the absence of a specific understanding of those factors, Fe solubility is often used as a proxy. However, there is no general agreement as to a specific protocol for determining solubility; in practice, aqueous solutions of various compositions and pH levels are used in an effort to simulate natural environments ranging from that of seawater to relatively acid cloud droplets to wetted aerosol particles (Buck et al. 2010). African dust is also a significant, often major, contributor to soil formation in the Caribbean and the Bahamas (Muhs et al. 2007, 2012). Trace species, especially phosphorus, could contribute to soil fertility. In particular, African dust is believed to supply critical nutrients to the Amazon basin (Swap et al. 1992); dust could be playing a similar role throughout the Caribbean. Therefore, there is a great need for a more coordinated approach that combines laboratory, field observation, and modeling experiments to test hypotheses regarding the mechanisms of iron dissolution and to assess their relative importance in different environments. The Caribbean, as the receptor of African dust, is an ideal location for the study of properties of trace metals and other nutrients in dust and how they might change with time in the atmosphere; the Atlantic transit time is relatively long—over a week—and during transit the dust-laden air in the SAL is relatively isolated from mixing with inputs from other sources (Trapp et al. 2010), which could complicate the interpretation of results.
IMPACT OF DUST ON HEALTH.
The concentration of African dust and other aerosols in the Caribbean Basin often exceeds the air quality standards linked to health effects (Prospero and Lamb 2003). The inhalation of mineral dust is known to produce a wide range of physiological responses and impact human health (Plumlee et al. 2006; Morman and Plumlee 2013). On a worldwide basis, mineral dust is estimated to be a major health threat (Liu et al. 2009). However, in the basin there has been little research on health effects. Of particular interest is the possible role of dust on asthma. Asthma rates are high throughout the Caribbean, comparable to those in urban/industrial environments. Despite this evidence, there has been little research on the causative factors. Furthermore, there are few data on aerosol properties that might help to identify linkages to health. We lack even the simplest of metrics for “respirable” particles as defined by the U.S. Environmental Protection Agency (EPA) [i.e., particles less than 2.5-μm diameter (PM 2.5) and those less than 10-μm diameter (PM 10) as determined by EPA criteria; www.epa.gov/air/criteria.html]. Measurements made in field programs (Li-Jones and Prospero 1998; Prospero et al. 2001; Reid et al. 2003b) show that about half of the African dust mass conforms to PM 2.5 and over 90% to PM 10. Based on the daily measurements on Barbados we would expect that dust concentrations will frequently exceed the EPA and World Health Organization (2006) 24-h guidelines for PM 2.5 and PM 10 aerosols.
There is a clear need for controlled studies of human health effects at the population level and at the individual level. These should focus on vulnerable populations (i.e., the young and the aged) for both acute and chronic exposures. Also of interest are the mechanisms by which dust and other aerosols act to impact health and the identification of bio-active components: allergens, biological materials, fungal spores, metals (iron, aluminum, arsenic, lead, cadmium), endotoxins, and organics. Despite the widespread occurrence of dust around the globe, a literature survey by de Longueville et al. (2013) shows remarkably few studies of the health impacts in dust-rich natural environments, globally a total of 50 papers. Of these, 11 dealt with African dust, and almost all focused on the impact on Europe. Health studies in North Africa coupled with parallel studies in the Caribbean Basin would be useful in assessing the scope of dust impacts. To this end, there should be an effort to devise protocols that could be used in both regions so that the results can be compared and contrasted in a quantitative manner.
Finally, we need to develop means of communicating risk to impacted populations and to design interventions based on the human health effects. Approaches include the issuance of air quality alerts based on real-time monitoring networks, the use of remote sensing data, modeling forecasts such as those cited above, and the development of response plans.
SUMMARY AND CONCLUSIONS.
In light of the changes observed in the Barbados dust record, there is a clear need for long-term measurements in the Caribbean Basin that focus on characterizing trends in aerosol concentrations and in critical aerosol properties. Sites should be established at various locations to define the temporal and spatial variability of African dust and also of aerosols from other sources (e.g., pollutants, biomass burning). Sampling should conform to PM 2.5 and PM 10 protocols so that the potential for health impacts can be better assessed. Precipitation collectors should be collocated with the aerosol samplers so as to relate deposition rates and chemistry to aerosol concentrations and properties.
It is important to link measurements of aerosol properties over the Caribbean Basin to those made in the source region, North Africa. These measurements should be related to those made in other continental regions where pollutant species are usually dominant, and also to other dusty regions so as to better elucidate dust-specific health effects.
To this end, aerosol “super sites” should be established with instrumentation comparable to the super sites designed for air quality studies such as those in the United States (Hersey et al. 2011). One site should be located on, or near, the coast of Africa to better characterize the “source” aerosol and the meteorological processes related to source activity. There are several sites where research facilities and infrastructure support are available: at Dakar, Senegal (Drame et al. 2011); at the Surface Ocean Lower Atmosphere (SOLAS) station on Sao Vicente, Cape Verde Islands (Müller et al. 2010); and on Tenerife, Canary Islands, where there is a well-developed World Meteorological Organization (WMO) site that follows an extensive aerosol protocol (Rodríguez et al. 2011). In addition to aerosol measurements, lidars and aerosol optical depth measurements would provide detailed information on vertical distributions and the relationship to the meteorological setting associated with dust outbreaks (Drame et al. 2011).
A second site should be located in the Caribbean to characterize aerosols in the receptor region. The preferred site would be Barbados because of its location in the axis of the dust plume in summer, the season when the entire Caribbean basin is impacted by dust (Adams et al. 2012). Barbados is also favored because of its long aerosol record and the presence of infrastructure and active research facilities. In addition to the University of Miami presence at Ragged Point, Barbados, at an adjacent site the Max Planck Institute for Meteorology (Hamburg) carries out cloud and aerosol studies (http://barbados.zmaw.de/) that are expected to continue for many years. It is also important for the aerosol community to establish working relationships with those focusing on environmental health. Barbados is of particular interest; because of the relatively homogeneous population, it is increasingly becoming the focus of research on the genetic aspects of health, including asthma (Mathias et al. 2010).
There is a need to facilitate education and training in aerosol measurements in the region. The University of Puerto Rico–Río Piedras could serve this function based on their long experience and ongoing research activities. These take place at two sites: the aerosol monitoring program at Cape San Juan site located on the easternmost end of Puerto Rico (in operation since 2004; www.esrl.noaa.gov/gmd/aero/net/cpr/index.html) and a cloud-forest site, Pico del Este, in El Yunque National Forest (1051 m ASL) (Gioda et al. 2013), which focuses on aerosol–cloud interactions. The Caribbean Institute of Meteorology and Hydrology (CIMH), located on Barbados, could play a role in research and education through its close links with the meteorological community. CIMH, an organ of the Caribbean Meteorological Organization (CMO), coordinates the joint scientific and technical activities of the meteorological and hydro-meteorological services of 16 Caribbean countries; it provides training to this community and also, more broadly, through its relationship with the University of the West Indies system, located throughout the Caribbean.
Given the scope and complexity of the African dust phenomenon, it is clear that the full assessment of its impact on the Caribbean Basin will require a coordinated research effort. We see the need for an organizational mechanism that would facilitate such a broad-scale integrated research program across the basin that, ideally, should be closely linked to the extensive research programs now being carried out in North Africa. Such a program will be essential to characterize and track the changes in dust aerosol emissions in Africa that we might expect with climate change.
This paper is based in part on the proceedings of the First International Workshop on the Long-Range Transport and Impacts of African Dust on the Americas, October 6-9, 2011, San Juan, Puerto Rico. We thank the International Global Atmospheric Chemistry Project (IGAC), the primary sponsor of the workshop, and the University of Puerto Rico for hosting the workshop (UPR's Resource Center for Science and Engineering, Center for Hemispherical Cooperation in Research and Education in Engineering and Applied Science, Research Initiative for Scientific Enhancement, and Long-Term Ecological Research program). We also acknowledge the workshop support of the NOAA Center for Atmospheric Science and Merck. O. L. Mayol-Bracero and J. M. Prospero acknowledge support from U.S. National Science Foundation Grants AGS 0936879 and AGS-0962256, respectively.