Abstract. Acid layers of volcanic origin detected in polar snow and ice layers are commonly used to document past volcanic activity on a global scale or, conversely, to date polar ice cores. Although most cataclysmic eruptions of the last two centuries (Pinatubo, El Chichon, Agung, Krakatoa, Cosiguina, Tambora, etc.) occurred in the tropics, cold tropical glaciers have not been used for the reconstruction of past volcanism. The glaciochemical study of a 137 m ice core drilled in 1999 close to the summit of Nevado Illimani (Eastern Bolivian Andes, 16 • 37' S, 67 • 46' W, 6350 m asl) demonstrates, for the first time, that such eruptions are recorded by both their tropospheric and stratospheric deposits. An 80-year ice sequence and the Tambora years have been analyzed in detail. In several cases, ash, chloride and fluoride were also detected. The ice records of the Pinatubo (1991), Agung (1963) and Tambora (1815) eruptions are discussed in detail. The potential impact of less important regional eruptions is discussed.

Sublimation and melt disturb the environmental information obtained from ice core records in the Andes. In two case studies we demonstrate to what extent these post-depositional processes may remove major parts of the accumulated snow cover. Dark ash layers from the Tungurahua eruption changed the albedo of surface snow on Chimborazo glacier (6268 m, 1 • 30 S, 78 • 36 W, Ecuador) between two ice core drilling campaigns and forced substantial melt. Redistribution and washout of the chemical constituents shifted the concentration profiles obtained in December1999 as compared to an equivalent core drilled in December 2000. The stable isotope records showed that approximately the water equivalent (weq) of an annual layer had melted, and that the percolating melt water penetrated within the firn layer to a depth of at least 16.5 m without refreezing. In the second example, from a site on the dry axis between the tropical and extra-tropical precipitation belts, significant loss of accumulated snow layers occurred by sublimation. A surface experiment at Cerro Tapado glacier (5536 m, 30 • 08 S, 69 • 55 W, Chile) revealed that losses of ≈2 mm weq (≈5 mm snow) per day occurred during the dry period following the 1997/98 El Niño. This loss generally included the entire surface layer enriched in stable isotopes, and thus caused minimal disturbance of the isotopic signature (and hence climatic information) of the net accumulation, yet chemical constituents again experienced considerable changes in concentration. From annual layer counting and direct dating it is obvious that the major part of the accumulated ice on both glaciers is younger than 100 years; however, isotopic and chemical variations at least in the basal ice from Cerro Tapado clearly point to climate conditions different from the recent centuries. This evidence is supported by mass balance considerations derived from a glacier-climate model. The possibility of a third type of disturbance aside from sublimation and melting -in this case a significant hiatus in the environmental chronology -also deserves consideration for other ice core records from this region. Potential disruptions or discontinuities need to be carefully evaluated given the profound changes in climatic and glaciological conditions since the Last Glacial Maximum throughout Holocene times.

Sublimation and melt disturb the environmental information obtained from ice core records in the Andes. In two case studies we demonstrate to what extent these post-depositional processes may remove major parts of the accumulated snow cover. Dark ash layers from the Tungurahua eruption changed the albedo of surface snow on Chimborazo glacier (6268 m, 1 • 30 S, 78 • 36 W, Ecuador) between two ice core drilling campaigns and forced substantial melt. Redistribution and washout of the chemical constituents shifted the concentration profiles obtained in December1999 as compared to an equivalent core drilled in December 2000. The stable isotope records showed that approximately the water equivalent (weq) of an annual layer had melted, and that the percolating melt water penetrated within the firn layer to a depth of at least 16.5 m without refreezing. In the second example, from a site on the dry axis between the tropical and extra-tropical precipitation belts, significant loss of accumulated snow layers occurred by sublimation. A surface experiment at Cerro Tapado glacier (5536 m, 30 • 08 S, 69 • 55 W, Chile) revealed that losses of ≈2 mm weq (≈5 mm snow) per day occurred during the dry period following the 1997/98 El Niño. This loss generally included the entire surface layer enriched in stable isotopes, and thus caused minimal disturbance of the isotopic signature (and hence climatic information) of the net accumulation, yet chemical constituents again experienced considerable changes in concentration. From annual layer counting and direct dating it is obvious that the major part of the accumulated ice on both glaciers is younger than 100 years; however, isotopic and chemical variations at least in the basal ice from Cerro Tapado clearly point to climate conditions different from the recent centuries. This evidence is supported by mass balance considerations derived from a glacier-climate model. The possibility of a third type of disturbance aside from sublimation and melting-in this case a significant hiatus in the environmental chronology-also deserves consideration for other ice core records from this region. Potential disruptions or discontinuities need to be carefully evaluated given the profound changes in climatic and glaciological conditions since the Last Glacial Maximum throughout Holocene times.