Introduction: the desert soil ecosystem

Deserts constitute the most abundant and persistent terrestrial biome on Earth (Peel and Finlayson 2007; Thomas 2011; UNEP 1992). The term desert is used here interchangeably with drylands, and used to delineate regions in moisture deficit as indicated by a precipitation to potential evapotranspiration ratio of <0.65 (precipitation supplies less than 65 % of the moisture needed to sustain optimum plant growth) (UNEP 1992). The environmental stresses imposed by moisture limitation, plus thermal, UV, and oligotrophic pressures, have resulted in landscapes that support relatively sparse vascular plant and animal life. Instead, drylands are dominated biologically by soil-rock-surface communities (SRSCs) comprising microorganisms, mosses and lichens (Pointing and Belnap 2012).

Desert soils are dominated by biological soil crusts (biocrusts). These are cohesive soil surface structures; morphologically variable with climate (Belnap et al. 2003). They are dominated by cyanobacteria (particularly Microcoleus sp. (Garcia-Pichel et al. 2013) plus heterotrophic bacteria, algae, fungi, mosses and lichens. They are ubiquitous in dryland soils of all but the most extreme deserts worldwide (Belnap et al. 2003). Desert surfaces in extreme hyper-arid locations are often dominated by desert pavement, where pebbles are embedded in the surface soil matrix. These surfaces are characterized by hypolithic communities (Chan et al. 2012; Pointing et al. 2007; Warren-Rhodes et al. 2006, 2007; Wong et al. 2010). Hypoliths support extensive cyanobacteria-dominated (notably Chroococcidiopsis sp.) biofilms that colonize the ventral surface of translucent stones (mainly quartz), and may facilitate extensive microbial and moss colonization in surrounding soils (Chan et al. 2012). Open soils in drylands support a largely heterotrophic community dominated by Actinobacteria (Pointing and Belnap 2012). These poikilohydric communities extend over soils and rock surfaces, covering 70 % or more in dryland landscapes (Belnap et al. 2003). They comprise ancient phylogenetic lineages (Bahl et al. 2011) and may be very long lived. Estimates of longevity ranging into hundreds of years for lichens (Lange 1990) and thousands of years for hypoliths in extreme arid landscapes (Warren-Rhodes et al. 2006).

Globally SRSCs account for ~7 % of terrestrial productivity and almost 50 % of terrestrial nitrogen fixation (Elbert et al. 2012). They perform many critical ecosystem roles and can define the ‘critical zone’ of biological interaction in the ecosystem (Pointing and Belnap 2012). There is mounting evidence that biological soil crusts are important in regulating multi-functionality in dryland ecosystems (Bowker et al. 2013, 2011; Maestre et al. 2012). Local hydrological cycles are strongly influenced by biocrusts as they mediate the transport of water, gases, nutrients, heat and light down into underlying soil as well as movement outwards from the soil (Delgado-Baquerizo et al. 2012; Delgado-Baquerizo et al. 2013; Maestre et al. 2013) and to vascular plants (Harper and Belnap 2001; Harper and Pendleton 1993). Biocrusts and hypoliths trap incoming dust, increasing soil fertility (Chan et al. 2012; Field et al. 2009). Most importantly, biocrusts and hypolithic communities promote soil stability; indeed, where they are well-developed, soils can be almost totally resistant to wind and water erosion (Belnap and Gillette 1997; Chan et al. 2012; Zhang et al. 2006).

Here we highlight the regional scale of impacts that arise from disturbance to desert soils. We identify the causes of disturbance, and emphasize the mobilization of dust to the atmosphere as a major driver of impacts that manifest remotely. Examples of negative environmental consequences are given for terrestrial and marine ecosystems, including potential threats to biotic and human health.

Disturbance to desert soil surface ecosystems

Desert SRSCs are at risk from human activities (e.g., agriculture [cropping and livestock grazing], human habitation, energy and mineral extraction, fire, and recreational use), environmental stochasticity (particularly drought), and predicted long-term climate change (Belnap 2003; Pointing and Belnap 2012). The increase in disturbance by human activity is of special concern, as it is potentially catastrophic. Subsistence and large-scale agricultural use are ironically both threatened by and a causal agent of desertification. Approximately 1 billion people rely on drylands for their livelihood, and it is estimated that of the 5.2 billion hectares of dryland used for agriculture, 69 % is degraded or undergoing desertification (Gilbert 2011). Clear differences in the level of subsistence usage, large-scale exploitation, and the level of conservation effort are evident between under-developed and developed regions (Anon 2005).

Disturbance to SRSCs at a local scale results in substrate degradation and loss of soil stability and fertility, plus changes to hydrological and heat regimens in soils that can be irreversible (Pointing and Belnap 2012). This occurs naturally in stochastic ‘pulses’ due to desert storms (Kellogg and Griffin 2006), and increasingly from human disturbance. When this severely impacts local SRSC and plant covers, biological feedbacks may be activated towards increased desertification (Rietkerk et al. 2004; Schlesinger et al. 1990). A major factor that compounds the issue of SRSC disturbance is that recovery of communities is extremely slow (Belnap and Eldridge 2003). In addition, large soil losses can lead to fundamental changes to soil properties that impede re-colonization and succession of these organisms (Bowker et al. 2013; Pointing and Belnap 2012; Requena et al. 2001).

A major outcome is mobilization of soil previously bound in biocrusts and associated with hypolith communities of desert pavements (Chan et al. 2012; Field et al. 2009; Middleton 1989). This may manifest as massive dust storms or chronic smaller events (Fig. 1a) (Bullard et al. 2011; McTainsh and Strong 2007). The trajectory of transport for desert dust is well established (Fig. 1b) (Kellogg and Griffin 2006; Lawrence and Neff 2009; Smith et al. 2013) and includes dispersal across trans-continental trajectories (Collyer et al. 1984; De Deckker et al. 2008; Derimian et al. 2006; Lim 2011; Reynolds 2001) and trans-oceanic distances (Collyer et al. 1984; Marx et al. 2005; Prospero et al. 2005; Prospero and Lamb 2003). There is growing concern that the frequency and magnitude of such dust events is increasing (McTainsh et al. 1998; Middleton 1989; Neff et al. 2008; Safriel 2006).

Fig. 1
figure 1

a Image illustrating the magnitude of dust storms originating from disturbed dryland soils, depicted location: Xinjiang, China. b Map illustrating extent of drylands globally, and major trajectories for airborne desert dust (blue arrows)

Full size image

Impacts of aerosolized desert dust

Biodiversity

Wind-blown dust may physically damage organisms, for example by scratching the cuticles of plants and insects, leading to increased water loss and thus mortality (Belnap 2003). Dust deposited on plant leaves can clog stomata, reducing photosynthetic rates, and in sufficient amounts, bury young seedlings or low stature plants (Collyer et al. 1984; Field et al. 2009). Dust is also a vector for airborne microbial and plant propagule dispersal (Womack et al. 2010), and so threats to local endemic biodiversity may arise due to invasive species and pathogen introduction. Microorganisms have been recovered from desert dust and estimated to increase microbial loading in aerosols by an order of magnitude (Favet et al. 2013; Griffin et al. 2006; Griffin 2007; Jeon et al. 2011; Kellogg et al. 2004; Lim 2011; Prospero et al. 2005; Schlesinger et al. 2006). Diversity assessments have indicated a likely desert origin for many of the microorganisms deposited in remote dust sinks (Abed et al. 2012; De Deckker et al. 2008; Favet et al. 2013; Griffin et al. 2006; Jeon et al. 2011; Lim 2011; Prospero et al. 2005; Woo et al. 2013).

Terrestrial ecosystems

Dust transport can lead to regional impacts on hydrology, with far-reaching consequences for hydrological regimes of distant catchments (Fig. 2). Wind-blown soils from desert sources have been traced to deposits on snowpack in Colorado in the USA (Painter et al. 2010). Similarly Australian desert dust has frequently crossed the Tasman Sea and deposited on New Zealand’s South Island (Marx et al. 2005; Marx and McGowan 2005; McGowan et al. 2005). The deposition of dust on snow causes radiative forcing due to enhanced absorption of solar radiation, and results in disturbance to the magnitude and timing of runoff from snowpack (Painter et al. 2010). The increased snowmelt enhances evapo-transpiration from exposed soils and plants and this in turn reduces freshwater input to lower catchments. For example, reduced input to the Colorado River has been estimated at up to 5 % of the annual average (Painter et al. 2010). This can have direct economic consequences at dust sinks (e.g., ski resorts) and where the water is used for agriculture and domestic supplies. Dust can also introduce toxic compounds and into water catchments. For example mercury contamination of a Florida lake has been explained as a consequence of desert dust originating in West Africa (Holmes and Miller 2004). Nutrient input from dust to soils (Okin et al. 2004) and lakes (Mahowald 2003) has been linked to altered aquatic biogeochemical function. Invasive microorganisms may also be introduced with dust, as postulated for African dust in European alpine lakes (Hervas et al. 2009).

Fig. 2
figure 2

Schematic illustration of local and regional impacts from desert dust mobilization: 1 at local scales disturbance removes soil crust and hypolith cover and impedes re-colonization of soils, 2 local disturbance mobilizes soils that may become aerosolized or enter waterways, 3 dust deposition on snowpack impacts regional hydrology and reduces freshwater input to lowland catchments, 4 dust input to marine systems impacts nitrogen fixation by phytoplankton and is detrimental to coral reefs, 5 desert dust storms have major impacts on human infrastructure, and dust is linked to respiratory allergies and pathogen dispersal. Yellow arrows denote airborne transport, whilst red arrows denote soil/aquatic transport.

Full size image

Marine ecosystems

Mobilization of desert dust results in significant deposition to ocean basins (Mackie et al. 2008) (Fig. 2). Here the massive increases in iron input from dust has directly contributed to significant impacts on global patterns of marine nitrogen fixation by phytoplankton, due to localized removal of iron limitation to diazotrophy (Jickells et al. 2005; Sohm et al. 2011). Additional impacts on marine ecosystems arise from deposition of dust to coral reefs, where physical coverage of coral by dust reduces zooxanthellae symbioses (Garrison 2003). In addition to this physical disturbance, dust-borne microorganisms have also been implicated as causal agents in coral diseases and death (Rypien 2008; Weir-Brush et al. 2004).

Human health

Dust deposition on human settlements has clear impacts on infrastructure and safety (Fig. 2) (Bener et al. 1996; Griffin 2007; Pöschl 2005). The fine particulate nature of desert dust easily clogs machinery and impaired visibility can lead to fatal highway accidents during large dust storms. Globally, chronic exposure to desert dust is estimated to cause 1.7 % of deaths due to lung cancer and cardio-pulmonary disease, although in latitudes across Africa, the middle East and Asia that support extensive deserts this figure is estimated to be between 15 and 50 % (Giannadaki et al. 2013). Even short-term exposure may be harmful, a survey of Afghan and gulf war veterans deployed during 2003–2004 revealed almost 70 % reported respiratory illness that was attributed to dust (Sanders et al. 2005). African desert dust falling on Caribbean islands after trans-Atlantic transport has also been implicated in respiratory illness among local populations (Gyan et al. 2005).

Human pathogenic bacteria and fungi are transmitted via aerosols and concern is rising that they may travel vast distances with a desert dust vector (Griffin 2007). Under normal conditions the outdoor aerosols of urban settlements support very low levels of pathogenic microorganisms (Brodie et al. 2007; Woo et al. 2013). Putative bacterial and fungal pathogens have been identified in desert dust worldwide (reviewed in Griffin 2007), commonly encountered taxa were gram-negative bacteria and the fungal genus Aspergillus. Desert dust is unlikely to support harmful levels of human pathogens under normal circumstances, although dust sources are often in regions where living standards are low and outbreaks of highly contagious disease are not uncommon. The potential role of dust as a long distance microbial vector warrants more research attention.

Conclusions

In this short review we have identified the connection between disturbance of desert soils and negative environmental consequences at remote dust deposition sites. The worldwide airborne dispersal of desert dust makes this a truly global problem that affects both terrestrial and aquatic biomes. A major concern is that the magnitude and frequency of dust storms appear to be increasing due to climate change (Prospero and Lamb 2003) and anthropogenic causes (Neff et al. 2008). Indeed the very nature of dust mobilization into the atmosphere may also create a positive feedback where transfer of CO2 from soils to atmosphere may exacerbate global warming and its impacts on desertification, in addition to regional carbon stock reductions (Reichstein et al. 2013) and increased inputs to downwind carbon sinks (Chappell et al. 2013). A greater focus on understanding and monitoring the causes, trajectory and impacts of dust events is therefore critical to understanding global ecology. We highlight the need for greater understanding of microbial dispersal in airborne dust as an area in need of urgent research attention, since this likely impacts ecosystem health and biogeography on a global scale.