Elsevier

Remote Sensing of Environment

Volume 123, August 2012, Pages 90-97
Remote Sensing of Environment

Comparison of satellite based observations of Saharan dust source areas

https://doi.org/10.1016/j.rse.2012.03.019Get rights and content

Abstract

Satellite remote sensing products such as Meteosat Second Generation (MSG) Infra Red (IR) dust index and Ozone Monitoring Instrument (OMI) Aerosol Index (AI) are commonly used to infer dust source areas. Here, two methods for dust source identification are compared, (1) a “back-tracking” method applied to 15-minute MSG IR dust index, and (2) a “frequency” method applied to daily OMI AI and daily MODIS DeepBlue Aerosol Optical Thickness (AOT) data.

Using the “back-tracking” method, dust source areas are inferred by tracking individual dust plumes back to their place of origin, allowed by the high temporal resolution of the MSG images. OMI AI and MODIS Deep Blue AOT products are available on daily resolution only, which does not allow for back-tracking of individual dust plumes. Thus, dust source areas are identified by relating the frequencies of occurrence of high dust loadings to source areas.

The spatial distribution of inferred dust source areas not only from the two methods, but also from the two satellite products, shows significant differences. The MSG back-tracking method highlights frequent dust emission from sources within complex terrain, while frequencies of high OMI AI values emphasise topographic basins as important dust source areas. Dust source areas retrieved from DeepBlue AOTs are generally further south towards the Sahel region. This study shows that the temporal resolution of satellite dust products is a key issue in identifying dust source areas. Both, the spatial distribution of dust sources and their annual cycle strongly depend on the acquisition time related to the start of dust emission.

Highlights

► Comparison of maps of dust sources retrieved from different satellite products. ► Temporal resolution of satellite products is key issue in retrieving dust sources. ► Spatio-temporal distribution depends on observation and start of dust emission.

Introduction

Mineral dust is a significant contributor to atmospheric aerosol (e.g., Shao et al., 2011) and affects life on Earth in various ways. It impacts directly and indirectly on radiation fluxes and nutrient cycles (IPCC report, 2007). Because of its optical properties and its mineral composition, dust aerosol interacts with radiation by back-scattering, absorbing, and re-emitting at different parts of the wavelength spectrum ultimately affecting directly the atmosphere's radiation budget (e.g., Miller and Tegen, 1999, Sokolik and Toon, 1996, Sokolik et al., 2001, Tegen and Lacis, 1996). Dust particles also affect the atmospheric radiation budget indirectly. Acting as a cloud condensation nuclei dust particles impact on cloud formation processes (e.g., Hoose et al., 2008, Lohmann, 2002) and cloud optical properties as well as precipitation rates (e.g., DeMott et al., 2003, Rosenfeld et al., 2001, Wurzler et al., 2000). Besides its role for the Earth's radiation budget, clouds are an essential part in the global water cycle. In addition to altering cloud properties and therefore changing the radiation budget indirectly, layers of atmospheric dust cool the skin surface by scattering and absorbing incoming solar radiation. While less solar radiation reaches the ground underneath a dust layer, the layer itself warms because of thermal re-emission. This re-distribution of radiative energy results in a stabilised atmosphere (Miller and Tegen, 1999, Perlwitz et al., 2001) ultimately affecting vertical transport processes relevant for dust emission (Heinold et al., 2008), but also convective processes (Miller & Tegen, 1999).

Also our day-to-day life is affected by airborne dust. Low horizontal visibilities due to dust plumes cause roads and airport closures (Criado & Dorta, 2003), and an increase in human diseases like asthma and meningitis (e.g., Griffin and Kellogg, 2004, Sultan et al., 2005). Due to its dimming effect, solar economy is affected by dust plumes due to reduced energy supply (Breitkreuz et al., 2007).

Mineral dust particles are lithogenic soil particles and characterised by the same mineralogical properties as their source. Because of an iron content of around 4% for North Africa (Wedepohl, 1995), dust transported from sources over North Africa acts as a transport medium for iron, a micro-nutrient for terrestrial and marine ecosystems (e.g., Fung et al., 2000, Jickells et al., 2005, Mahowald et al., 2005, Sarthou et al., 2003).

Therefore, the knowledge of dust sources, including their location, emission activity, and geomorphological characteristics, is a prerequisite for understanding the regional dust transport pathways and estimating regional dust impacts on human life, economy and climate. Due to the vast extent and the limited accessibility of deserts, human exploration of every individual dust source area is not feasible. In general different remotely operating approaches have been used for identifying dust source areas, i. e.: (1) mineralogy of dust samples, (2) satellite remote sensing techniques, (3) horizontal visibility, and (4) Lagrangian back-trajectory. The first links mineral composition of dust samples to potential source areas considering transport pathway and source geomorphology. A recent overview on analytic methods is given by Formenti et al. (2011). Remote sensing techniques use the radiative signature of dust at different wavelength bands to identify airborne dust above source areas (e.g., Christopher et al., 2011). Earlier studies also analyse measurements of horizontal visibility as provided by synoptic stations (Mahowald et al., 2007, N'Tchayi Mbourou et al., 1997). Although these stations provide a long-term record of local visibility, the observation sites are located close to human settlements rather than in dust source areas. Furthermore, the observation network over the Sahara is sparse, local and transported dust affect the measurements. Thus, the interpretation of visibility data with regard to identifying dust sources is difficult (Mahowald et al., 2007). Lagrangian back-trajectories are commonly used to link dust source areas to remote dust samples. Thereby, the travel path of a dust loaded air parcel is followed backwards to the region of dust entrainment (e.g., Alonso-Pérez et al., 2012, Escudero et al., 2011).

Due to the high spatio-temporal coverage of satellite remote sensing data sets, this approach has the potential to provide the most accurate localization of dust source areas. Several dust products exist from different satellite instruments flying in different orbits. Here, the difference of two commonly applied techniques inferring dust source areas from satellite products will be discussed. We compare dust source areas inferred from 15-minute MSG SEVIRI IR dust product (Schepanski et al., 2007), and dust areas inferred from daily noon-time Aqua MODIS DeepBlue AOT and Aura OMI AI using frequencies of high AOT and AI values as applied by e.g. Middleton and Goudie (2001) and Prospero et al. (2002) to TOMS (Total Ozone Mapping Spectrometer) AI values, the predecessor of the OMI AI product.

Section snippets

Satellite remote-sensing products

To retrieve information on the atmospheric dust content over bright surfaces like the Sahara desert, a contrast in wavelength dependent radiative signature of the airborne dust and the surface is required. Due to a similar signature of dust and barren soil at visible wavelengths airborne dust cannot be distinguished from the surface with a satisfying quality. At ultra-violet (UV), near-UV (deep blue), and infra-red (IR), the radiative signature of dust can be separated from the signature of

Mapping Saharan dust sources

In general, active dust sources are identified by linking the presence of airborne dust to a specific source location. Due to different temporal availabilities of dust products, the actual accuracy in linking the atmospheric dust burden to a source location varies. Here, two different approaches are discussed: (1) MSG IR dust index images are used for visual back-tracking of individual dust plumes to their source, and (2) MODIS DeepBlue AOT and OMI AI are used to identify dust source regions by

Discussion

The differences between the “back-tracking” and “frequency” approaches are due to both the temporal and spatial resolution. The temporal resolution is the critical factor when comparing the different source area retrievals. Due to the lag of up to 6 h between start of dust emission and satellite overpass for the low-orbiting satellites, detected dust plumes have propagated downwind the source area. Dust emitted during the morning hours is mixed up over the entire boundary layer and transported

Summary and conclusion

The present study summarizes the commonly used methods for inferring dust source areas from satellite products. In particular due to differences in temporal resolution (daily noon time versus 15-minute), different dust source areas are retrieved by the different methods. Daily satellite products only show an instantaneous distribution of atmospheric dust content and therefore a statistical interpretation in terms of dust source relates areas with high dust loads to recently emitted dust plumes

Acknowledgements

The authors thank NASA GES DISC for development and maintenance of Giovanni online system providing MODIS Deep Blue AOT and OMI AI data. Many thanks also to the ECMWF for providing ERA-Interim re-analysis data. KS also acknowledge funding from European Research Council Grant no. 257543 “Desert Storms”. Many thanks to the reviewers whose comments and suggestions helped to improve the manuscript.

References (54)

  • C. Cavazos et al.

    Numerical model simulation of the Saharan dust event of 6–11 March 2006 using the Regional Climate Model version 3 (RegCM3)

    Journal of Geophysical Research

    (2009)
  • S.A. Christopher et al.

    Multi-sensor satellite remote sensing of dust aerosol over North Africa during GERBILS

    Quarterly Journal of the Royal Meteorological Society

    (2011)
  • P.J. DeMott et al.

    African dust aerosols as atmospheric cloud ice nuclei

    Geophysical Research Letters

    (2003)
  • S. Engelstaedter et al.

    Atmospheric controls on the annual cycle of North African dust

    Journal of Geophysical Research

    (2007)
  • P. Formenti et al.

    Recent progress in understanding physical and chemical properties of African and Asian mineral dust

    Atmospheric Chemistry and Physics

    (2011)
  • I. Fung et al.

    Iron supply and demand in the upper ocean

    Global Biogeochemical Cycles

    (2000)
  • J. Giles

    Climate science: The dustiest place on Earth

    Nature

    (2005)
  • P. Ginoux et al.

    Sources and distributions of dust aerosols simulated with the GOCART model

    Journal of Geophysical Research

    (2001)
  • P. Ginoux et al.

    Empirical TOMS index for dust aerosol: Application to model validation and source characterization

    Journal of Geophysical Research

    (2003)
  • P. Ginoux et al.

    Identification of anthropogenic and natural dust sources using Moderate Resolution Imaging Spectroradiometer (MODIS) Deep Blue level 2 data

    Journal of Geophysical Research

    (2010)
  • D.W. Griffin et al.

    Dust storms and their impact on ocean and human health: Dust in Earth's atmosphere

    EcoHealth

    (2004)
  • B. Heinold et al.

    Dust radiative feedback on Saharan boundary layer dynamics and dust mobilization

    Geophysical Research Letters

    (2008)
  • B. Heinold et al.

    Regional modelling of Saharan dust and biomass-burning smoke Part I: Model description and evaluation

    Tellus

    (2011)
  • C. Hoose et al.

    The global influence of dust mineralogical composition on heterogeneous ice nucleation in mixed-phase clouds

    Environmental Research Letters

    (2008)
  • N.C. Hsu et al.

    Aerosol properties over bright-reflecting source regions

    IEEE Transactions on Geoscience and Remote Sensing

    (2004)
  • N.C. Hsu et al.

    Deep Blue retrievals of Asian aerosol properties during ACE-Asia

    IEEE Transactions on Geoscience and Remote Sensing

    (2006)
  • IPCC

    Climate Change 2007: The physical science basis

  • Cited by (0)

    View full text