1. Petropavlovsk-Kamchatsky, August 2004 RECURRENT CALDERA-FORMING ERUPTIONS: KSUDACH CASE STUDY Preliminary results of the ongoing petrological and experimental study Pavel Izbekov 1, James Gardner 2, Ivan Melekestsev 3, and John Eichelberger 1 1 Alaska Volcano Observatory, Geophysical Institute, University of Alaska Fairbanks, Alaska 2 University of Texas at Austin, Texas 3 Institute of Volcanology and Seismology, Petropavlovsk-Kamchatsky

2. Petropavlovsk-Kamchatsky, August 2004 Introduction The development of volcanic centers is often cyclic: cone – caldera – cone – caldera... The development of volcanic centers is often cyclic: cone – caldera – cone – caldera... Ksudach caldera complex in Kamchatka appears to represent an exemplary case where at least three such cycles have been completed since middle Pleistocene. Ksudach caldera complex in Kamchatka appears to represent an exemplary case where at least three such cycles have been completed since middle Pleistocene. Landsat 7 ETM & SRTM DEM

3. Petropavlovsk-Kamchatsky, August 2004 We use petrological and experimental approaches to determine the relationships of magmas erupted at Ksudach and to infer their pre-eruptive conditions, which potentially may illuminate the mechanism of such cyclicity, as well as the mechanism of the recurrent caldera- forming eruptions. Introduction cont.

4. Petropavlovsk-Kamchatsky, August 2004 Location Southern part of the Eastern Volcanic Front of Kamchatka Ca. 150 km to the SW from PK Japan Russia Alaska

5. Petropavlovsk-Kamchatsky, August 2004 Background 5 caldera-forming eruptions at one single volcanic center: 5 caldera-forming eruptions at one single volcanic center: 2 larger Pleistocene calderas and 2 larger Pleistocene calderas and 3 Holocene calderas (8800 BP, 6300- 6000 BP, and 1800 BP) 3 Holocene calderas (8800 BP, 6300- 6000 BP, and 1800 BP) Cycles of a cone-building effusive activity and caldera-forming events: Cycles of a cone-building effusive activity and caldera-forming events: (1) The first caldera has destroyed an older Pleistocene shield volcano. (2) Then a new cone has been constructed inside the caldera and was subsequently destroyed by a collapse of the second Pleistocene caldera. (3) The intra-caldera cone has been constructed again and was destroyed by the first Holocene caldera-forming eruption KS-4 (8800 BP). Then three nested calderas were formed sequentially with only a subtle intra- caldera effusive activity between the caldera- forming eruptions. (4) After KS-1 eruption (1800 BP) the cone- building activity resumed again and culminated in the explosive eruption of 1907 AD. from Volynets et al. (1999)

6. Petropavlovsk-Kamchatsky, August 2004 Questions We focused first on the composition of the most evolved magmas erupted during a sequence of three caldera-forming eruptions occurred at Ksudach during the Holocene: the dacite of the initial fall deposit of KS-4 (8800 yr. BP, 67.4 wt % SiO2), the KS-3 rhyolite (~6000 yr. BP, 70.3 wt % SiO2), and the KS-1 rhyolites (1800 yr. BP, 71.5-72.1 wt % SiO2). 1. Do the products of the Holocene caldera- forming eruptions at Ksudach represent snapshots of a single magma reservoir? 2. What are the pre-eruptive conditions for the last Holocene caldera-forming eruption (1800 BP)?

7. Petropavlovsk-Kamchatsky, August 2004Samples

8. Methods Mineral and glass compositions were determined by electron microprobe (3-10 micron, 10 nA, 15 kV beam). Mineral and glass compositions were determined by electron microprobe (3-10 micron, 10 nA, 15 kV beam). Pre-eruptive conditions for KS-1 were reproduced experimentally in Rene and TZM pressure vessels (NNO, water-saturated, 100 MPa). Pre-eruptive conditions for KS-1 were reproduced experimentally in Rene and TZM pressure vessels (NNO, water-saturated, 100 MPa). Whole rock compositions were determined by XRF and ICP-MS Whole rock compositions were determined by XRF and ICP-MS

9. Petropavlovsk-Kamchatsky, August 2004 WR chemistry

10. Petropavlovsk-Kamchatsky, August 2004 WR chemistry cont.

11. Petropavlovsk-Kamchatsky, August 2004 Mineral compositions Magnetite-ilmenite thermometry indicates that the dacite of KS-4 was last equilibrated at 914-924  C and fO 2 of NNO+0.4, KS-3 rhyolite at 894-927  C and fO 2 of NNO+0.1, KS-1 rhyolite at 870-907  C and NNO+0.6. It appears that there is a weak overall cooling trend from older to younger magmas of similar silicic composition.

12. Petropavlovsk-Kamchatsky, August 2004 Mineral compositions Plagioclase is a dominant phenocryst phase in the studied KS-1, KS-3, and KS-4 products. Most show oscillatory-zoning. Their compositions range from An 52.3  3 in KS-4 to An 43  6 in KS-1, with KS-3 plagioclase compositions being intermediate, which is consistent with the view that these magmas have been derived from a single, slowly cooling source. However, the presence of “dusty-zoned” plagioclases, in which resorbed An 46-49 cores are mantled by An 72-75 inclusion-rich rims, as well as rare anorthite xenocrysts (An 93-96 ) indicate that the composition of silicic magmas was likely modified by mixing with basalt.

13. Petropavlovsk-Kamchatsky, August 2004 Experimental results: KS-1 P-T phase diagram for KS-1 rhyolite based on the phase diagram for Karymsky dacite, which has a similar bulk composition. Red symbols indicate the conditions of experiments, which used the natural KS-1 rhyolite pumice (sample 02IPE45). The equilibrium mineral phases observed in KS-1 experiments match those in Karymsky experiments with the exception of one run at 900ºC, where plagioclase is stable.

14. Petropavlovsk-Kamchatsky, August 2004 Experimental results cont. Shaded areas correspond to the average composition (±1 sigma) of natural plagioclase rims and the average temperature estimated using Fe-Ti oxides (±1 sigma). 100 MPa, NNO buffer

15. Petropavlovsk-Kamchatsky, August 2004 Experimental results cont. Shaded areas correspond to the average composition (±1 sigma) of matrix glass and the average temperature estimated using Fe-Ti oxides (±1 sigma) for the KS-1 natural sample. 100 MPa, NNO buffer Variations of experimental melt composition as a function of temperature

16. Petropavlovsk-Kamchatsky, August 2004 Summary of experiments Compositions of natural plagioclase, orthopyroxene, clinopyroxene, and matrix glass of KS-1 has been reproduced experimentally at 892±9°C, 100 MPa, and oxygen fugacity near NNO buffer.Compositions of natural plagioclase, orthopyroxene, clinopyroxene, and matrix glass of KS-1 has been reproduced experimentally at 892±9°C, 100 MPa, and oxygen fugacity near NNO buffer. Water in glass inclusions averages 3.8±1 wt.%, which matches saturation at ~100 MPa (Andrews et al., AGU- 2003, V31G-01). The pre-eruptive pressure corresponds to approximately 3-4 km depth.Water in glass inclusions averages 3.8±1 wt.%, which matches saturation at ~100 MPa (Andrews et al., AGU- 2003, V31G-01). The pre-eruptive pressure corresponds to approximately 3-4 km depth.

17. Petropavlovsk-Kamchatsky, August 2004 Possible scenarios: Model 1 A single, long-lived, fractionating magma chamber continually generates silicic magmas, which periodically ascend and erupt. Eruptions may be triggered by basaltic replenishments, which do not however, prevent the magma system from a gradual cooling. A single, long-lived, fractionating magma chamber continually generates silicic magmas, which periodically ascend and erupt. Eruptions may be triggered by basaltic replenishments, which do not however, prevent the magma system from a gradual cooling. Pro’s: The volume of silicic material appears to increase in time, as well as the concentration of SiO 2 in magmas and apparent period of quiescence between eruptions. Pro’s: The volume of silicic material appears to increase in time, as well as the concentration of SiO 2 in magmas and apparent period of quiescence between eruptions.

18. Petropavlovsk-Kamchatsky, August 2004 Possible scenarios: Model 2 Silicic inputs originate from a partially molten basement. They segregate and ascend periodically, flushing an evolved, shallow-level andesitic reservoir. Basaltic magmas from even deeper source transport heat and matter, and may trigger volcanic eruptions Silicic inputs originate from a partially molten basement. They segregate and ascend periodically, flushing an evolved, shallow-level andesitic reservoir. Basaltic magmas from even deeper source transport heat and matter, and may trigger volcanic eruptions Pro’s: Magmas of contrasting composition are involved in each eruption. The younger KS-1 rhyolite has lower REE concentrations as compared to the slightly less evoloved KS-3 and KS-4 rhyodacites. Also, the oxidation state of the magmas appears to vary with time. Pro’s: Magmas of contrasting composition are involved in each eruption. The younger KS-1 rhyolite has lower REE concentrations as compared to the slightly less evoloved KS-3 and KS-4 rhyodacites. Also, the oxidation state of the magmas appears to vary with time.

19. Petropavlovsk-Kamchatsky, August 2004 Conclusions The rhyolite of KS-1, the most recent caldera forming eruption at Ksudach, was last equilibrated at 892±9°C, 100 MPa, at water saturated conditions. The pre-eruptive pressure corresponds to approximately 3-4 km depth, which coincides with the regional stratigraphic boundary between Cretaceous metamorphic basement and an overlying volcano-sedimentary layer. The rhyolite of KS-1, the most recent caldera forming eruption at Ksudach, was last equilibrated at 892±9°C, 100 MPa, at water saturated conditions. The pre-eruptive pressure corresponds to approximately 3-4 km depth, which coincides with the regional stratigraphic boundary between Cretaceous metamorphic basement and an overlying volcano-sedimentary layer. The silicic magmas of the Holocene caldera-forming eruptions at Ksudach most likely has originated from the same reservoir and then their temperatures may reflect its cooling from ca. 920°C at 8800 BP to ca. 890°C at 1800 BP. This reservoir, however, may not be at the depth, where the erupted magmas has last equilibrated. The silicic magmas of the Holocene caldera-forming eruptions at Ksudach most likely has originated from the same reservoir and then their temperatures may reflect its cooling from ca. 920°C at 8800 BP to ca. 890°C at 1800 BP. This reservoir, however, may not be at the depth, where the erupted magmas has last equilibrated.