For granulometric, geochemical and mineralogical analyses, approximately 25 ml of each surface sample was frozen for 24 hours and subsequently lyophilized using a Christ BETA 1-8 LDplus (Martin Christ Gefriertrocknungsanlagen GmbH, Osterode am Harz, Germany). The freeze-dried samples were homogenized and split into two aliquots. The other aliquot of the freeze-dried surface samples was ground to <63 µm with a Planetary Mill Pulverisette 5 (FRITSCH GmbH, Idar-Oberstein, Germany) and used for mineralogical and geochemical analyses. The bulk mineralogy was determined on powder samples using an X-ray diffractometer (D8 Discover; Bruker, Billerica, MA, USA) with a Cu X-ray tube (λ = 1.5418 Å, 40 kV, 30 mA) and a LYNXE XE detector (opening angle = 2.9464°). The spectrum from 3° to 100° 2-theta was measured in 0.02° steps at 1 second exposure time. Mineral identification was carried out using the software packages SEARCH (Stoe and Cie GmbH, Darmstadt, Germany) and Match! (Crystal Impact 2014, Bonn, Germany), supported by the data base pdf2 (ICDD 2003, Newton Square, PA, USA). The concentration of the minerals was evaluated using the program TOPAS Rietveld (Coelho Software, Brisbane, Australia), which yields a standard deviation of analyzed minerals varying from +/- 2 % (for quartz) to +/- 5 to 10 % (for feldspars and clay minerals; Środoń et al., 2001; Vogt et al., 2002). For the clay mineral group illite the error range can be even higher (Scott 1983). Given these uncertainties, a detection limit of 5 % is considered in the discussion of the mineralogical composition of the surface sediments.