Elsevier

Atmospheric Research

Volume 216, 1 February 2019, Pages 160-166
Atmospheric Research

The Amazon basin as a moisture source for an Atlantic Walker-type Circulation

https://doi.org/10.1016/j.atmosres.2018.10.009Get rights and content

Highlights

  • Over the large extension of the Pacific ocean, the Walker circulation is formed.

  • The Atlantic ocean, with a smaller surface, does not give place to this structure.

  • humidity present in the Amazon basin, adds water surface to the region, allowing to distinguish a Walker-type circulation.

Abstract

The Amazon basin constitutes the most developed rainforest in the world, accounting for 15-20% of the global freshwater input into the oceans. The low level flow over this region is climatologically dominated by the Atlantic anticycslone and the trade winds. This yields an incoming oceanic moist air to the continent from the East, which is forced to lift up over the Andes range at the West. The confluence of the entrance of humidity, heat, evaporation and strong rainfall results in an accumulation of water vapor in this region. There is a statistically significant surplus of humidity over land as compared to over ocean (the largest difference is found during austral summer). This turns the Amazon basin into one of the most important heat sources for the tropical atmosphere, feeding a global pattern like the Atlantic Walker-type circulation, where the ascent stage is not over ocean but over land. The Global Positioning System radio occultation data show to be an excellent tool to observe the accumulated water vapor above the Amazon basin.

Introduction

The atmospheric transport of heat and momentum from the Equator to the poles is mainly explained by conceptual models of meridional tropospheric circulations known as Polar, Ferrel and Hadley cells. Longitudinally, the transport is given by zonal circulations forced by thermal gradients over the surface of the seas. The dominant circulation, called the Walker circulation (after G. T. Walker) by Bjerknes (1969), is mainly located over the equatorial Pacific Ocean. According to Barry and Chorley (2009), during the phases corresponding to no ENSO (El Niño Southern Oscillation) conditions, high sea surface temperatures over the western Pacific favor strong convection and its associated upward motion. Over the Eastern Pacific, the air carried by the westerly upper level flow descends over the cold water. Finally, the Walker cell is closed with the easterly trade winds over the surface (Fig. 1a, red full line). The narrower width of the Atlantic Ocean (West-East direction), in turn, hinders the formation of a Walker-type circulation (i.e. Wang, 2005). However, if we consider the South American continent as part of the circulation, we find a similar pattern, composed of an upward path over the Amazon basin and a downward trajectory over the Atlantic ocean (Fig. 1a, blue full line). The cause of this effect lies in the Amazon region, which is one of the biggest tropical heating centers of the Earth, with the most developed rainforest in the world. The basin covers an area close to 40% of South America (about 650 million hectares) roughly delimited in longitude and latitude by 75-45W and 0-20S. This region accounts for 15-20% of the global freshwater input into the oceans and its large annual regional precipitation is one of the important heat sources for the tropical atmosphere (Nobre et al., 2004). This feature of the Amazon basin may be explained by its geographical characteristics (latitude range, topography and location between two oceans), as well as by its regional low level atmospheric circulation (Fig. 1b). The latter factor plays a major role in transporting water vapor, since nearly half of it resides below an altitude of about 1.5 km and less than 5% is above 5 km (Seidel, 2002). In this case, the Atlantic anticyclone, the trade winds and monsoon flux from the North provide an incoming low level moist flow which, when located over the continent, is mostly channeled to the south at the eastern side of the Andes range (e.g. do Nascimento et al., 2016 and references therein). This acts as a major barrier for the surface trade winds (Fig. 1b, white curve over the continent), giving place to upward motion and strong precipitation over its slopes. This rainfall with the available heat then yields a strong evaporation, which, together with the arriving humidity contribute to the different stages of the water cycle in this region.

South America is characterized by its large oceanic area and the presence of the Andes range from north to south, which make difficult to set up a good radiosonde network over the region. The development of continuous satellite measurements with a regional coverage, allowed to observe thermal and dynamic processes over South America that were possible only with model data in the past. This is the case of the Global Positioning System (GPS) radio occultation (RO). It is a unique satellite method regarding the retrieval of global and continuous atmospheric data and it is not affected by meteorological conditions. In April 2006, the Constellation Observing System for Meteorology, Ionosphere and Climate (COSMIC) launched six GPS RO satellites, which significantly increased the number of daily soundings at that time. For several years it provided about 2000 daily profiles of water vapor, temperature, refractivity and pressure from the lower troposphere up to the middle stratosphere. The present study analyzed the post-processed data version 2013.3520 provided by the COSMIC Data Analysis and Archive Center. A good concordance between radiosonde data and the COSMIC mission water vapor distribution (50°S-50°N) was found by Kishore et al. (2011). The water vapor pressure measured by GPS RO can be easily converted into specific humidity q (e.g. Hierro et al., 2012). This is a representative tropospheric water vapor variable which relates the mass of water vapor with the total mass of an air parcel.

Space- and ground-based GPS technqiues have in general shown to be a useful tool for studying several meteorological features of the lower atmosphere (e.g. Kuleshov et al., 2016; Bonafoni and Biondi, 2016). In the case of humidity measurements, the integrated water vapor from both sources has been validated against other measurement methods and reanalysis data (e.g. Huang et al 2013, Calori et al., 2016). In particular, the space-based GPS RO technique has demonstrated to be a powerful tool for studying the global distribution of water vapor over land and sea with a moderate/high spatial and temporal resolution. The tropospheric mean patterns of humidity at different levels have been reproduced by Hierro et al. (2012). They showed that this technique is able to detect the behavior of this variable at a regional scale in South America. Oscillation modes of the integrated specific humidity derived from GPS RO over the Amazon basin were analyzed by Hierro et al. (2013). Llamedo et al. (2016), in turn, have shown the usefulness of this technique to study the water vapor over land and sea with a moderate/high spatial and temporal resolution, as well as its anomalies.

GPS RO retrievals usually provide temperature, refractivity and water vapor against height or pressure. The uncertainty of the former two variables has been assessed e.g. by Alexander et al. (2014) but will not be further considered as they are both not used in this study. The reliability of GPS RO water vapor or specific humidity has been evaluated in several works. Chou et al. (2009) compared COSMIC and AIRS (Atmospheric Infrared Sounder) tropospheric q globally and also specifically in the tropical Pacific and found discrepancies within 15%, whereas global validations against NCEP (National Centers for Environmental Prediction) reanalysis showed differences around 30% in oceanic areas above 300 hPa, but it should be kept in mind that those heights typically make a small contribution to the total vertical column amount of water vapor. According to Ho et al. (2010) the COSMIC water vapor profiles exhibit discrepancies within 0.2 g/kg in the lowest 2 km when compared with ECMWF (European Centre for Medium-Range Weather Forecasts) global analyses and 0.05 g/kg above. In addition, Teng et al. (2013) have shown a very good agreement of q from COSMIC and NCEP reanalysis in the lowest 5 km over the tropical Pacific. Vergados et al. (2018) have shown that RO COSMIC specific humidity profiles have a 10-20% precision within the layer 400-700 hPa. Teng et al. (2013) estimated that on average the biases of precipitable water (PW), which is essentially the vertical integral of specific humidity in the troposphere, remain within 2 mm. Also, Burgos Fonseca et al. (2018) validated GPS RO PW against values derived from ground-based GPS retrievals (see e.g. Calori et al. 2016) and found a global mean difference around 1 mm, which represented on average a discrepancy around 5%. Finally, notice that besides uncertainty issues, GPS RO data have the significant advantage over reanalysis that they are experimental values. On the contrary, PW and q belong to the reanalysis output variables category B (Kalnay et al., 1996), which means that although observational information affects their value, the numerical model has a strong influence on them. PW from reanalyses neither are usually given an associated uncertainty nor should they be considered standard benchmark data. In addition, Burgos Fonseca et al. (2018) have shown that PW over land from models may be significantly biased if their representation of the terrain height is not accurate enough. Among the diversity of observational instruments providing PW and q, satellite platforms are the only ones capable of providing permanent and global soundings over land and over ocean.

Section snippets

Water vapor and circulation

Fig. 2 shows the latitudinal distribution of q derived from GPS RO (COSMIC) data, averaged during the period 2006-2014 between 70-60W, which covers the central part of South America. A maximum of humidity is observed close to the surface around the equator, slightly towards the Southern Hemisphere. Fig. 3 shows the longitudinal distribution of q, seasonally averaged between latitudes 0-20S. One interesting feature observed is the already mentioned stopping effect of the Andes Range on the

Conclusion

The Walker cell is an ocean-based system of air circulation which is the result of a difference in surface pressure and temperature over the western and eastern equatorial Pacific Ocean. Nowadays the term also usually refers to other similar but smaller and weaker cells in the tropics over the Atlantic and the Indian Ocean. We here show that the quality of GPS RO water vapor data is good enough to demonstrate that the moisture for the Atlantic Walker-type cell stems from the land humidity in

Acknowledgements

The study has been supported by grants CONICET PIP 11220120100034 and ANPCYT PICT 2013-1097. We acknowledge NCEP and ERA Interim atmospheric model data provided by www.mmm.ucar.edu. GPS RO data downloaded from cdaac-www.cosmic.ucar.edu/cdaac/products.html and ground-based GPS data from rda.ucar.edu/datasets/ds721.1.

References (23)

  • R. Hierro et al.

    Oscillation modes of humidity over the Amazon basin derived from GPS RO profiles

    J. Geophys. Res.

    (2013)
  • Cited by (0)

    View full text