the case of Mt Ruapehu, New Zealand

Setting up a probabilistic eruption forecasting scheme at an open conduit volcano with a crater lake: the case of Mt Ruapehu, New Zealand

Laura Sandri , Gill Jolly , Jan Lindsay , Brad Scott , Steve Sherburn , Art Jolly , Nico Fournier , Harry Keys , and Warner Marzocchi 1 Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Bologna, Via Donato Creti 12, 40128 Bologna, Italy; 2 GNS Science, Wairakei Research Centre, Taupo, New Zealand; 3 School of Environment, The University of Auckland, Auckland, New Zealand;

4 Department of Conservation, Turangi, New Zealand; 5Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Roma1, Via di Vigna Murata 605, 00143 Roma, Italy

Mt Ruapehu and its activityMt Ruapehu is located at the southern end of the Taupo Volcanic Zone in central North Island, New Zealand. It is an andesitic stratovolcano rising to 2797 m elevation above sea level, the highest point in North Island. There is a well-developed hydrothermal system at the summit with a permanent crater lake sitting on top of the magmatic conduit. Mo-nitoring of Mt Ruapehu is undertaken by GNS Science under the GeoNet project, a New Zealand Earthquake Commission funded project to monitor and collect research data on geological hazards across the country. Currently, a network of broadband and short period seismome-ters, microbarometers, cGPS and tilt instruments combined with monthly sampling of the crater lake and airborne gas measurement constitutes the monitoring program for the volcano. Mt Ruapehu has had numerous small eruptions in historic time, with an average of one eruption capable of producing lahars approximately every ten years. Smaller eruptions tend to be phreatic or phreatomagmatic with minor juvenile material, and usually occur with few or no precursors (Kilgour et al. 2010; Jolly et al. 2010). Fully magmatic eruptions lasting over se-veral months occurred in 1945 and 1995-6. The latter was preceded by significant variations in crater lake chemistry and temperature, but little obvious precursory seismicity.In recent years, two small phreatic or phreatomagmatic eruptions occur-red at Mt Ruapehu in 2006 and 2007. These were the first significant eruptions since the last magmatic eruption in 1996. On 5 October2006, a small phreatic eruption through the crater lake resulted in seve-ral metres wash up along the sides of the lake. Airborne gas measure-ments made soon after the event showed that CO2 and SO2 output had increased significantly, but no obvious precursors in any of the data streams were noted.On 25 September 2007, a much larger event resulted in a wet surge across the summit of the volcano and the generation of two lahars: one to the east and one to north-west (Kilgour et al. 2010). The latter travel-led through the top of the skifield. Fortunately the event occurred at 8.26 pm, so there were no skiers on the field. A groomer driver was in the valley at the time and managed to take evasive action. Two clim-bers were camped on the summit of the volcano, one of whom was se-verely injured by the surge. His injuries were almost fatal and he subse-quently had one leg amputated.In the present study, we first look at time series of monitoring data to determine whether there are any significant long term precursors prior to small phreatomagmatic eruptions, and then use this information to build Bayesian Event Trees for determining long and short term proba-bilities of eruption for forecasting and assessment of consequent hazar-dous phenomena.

Time serie analysisDetailed analysis of the monitoring data immediately prior to the 2007 event showed some very small volcanotectonic earthquakes a few minutes before the main eruption (Jolly et al. 2010), but this type of signal is not anomalous on Mt Ruapehu. Otherwise, there were no obvious indications of an impending eruption.

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Change points in chemestry/seismic parameters at Crater Lake

MgClMg/ClSO2COSPECH2SSO4Volc Eqs 30daysVT 30daysVolc Eqs 7daysVT 7daysCumrsam 7days

all data T>Tmean T<Tmean T>Tmean+ T>Tmean- T>90prctile T<10prctile Trend UP Trend DOWNsize 1 79% 57% 23% 26% 5% 19% 3% 51% 25%size 2 11% 8% 3% 3% 1% 3% 1% 4% 7%size 3 6% 4% 2% 3% 1% 3% 1% 4% 2%size 4 2% 1% 1% 1% 0% 1% 0% 1% 1%size 5 2% 2% 1% 2% 0% 2% 0% 2% 1%

TOT 182 128 54 64 13 51 8 110 65

We have looked at the time series of the last 40 years of monitoring data on Mt Ruapehu to see if there are any long term indicators of increase in activity, possibly related to influx of magma to the conduit. In particular, there has been some debate about whether the temperature of the crater lake can be used to indicate whether a phreatic or phrea-tomagmatic eruption is more likely.

In particular, there has been some debate about whether the temperature of the crater lake can be used to indicate whether a phreatic or phreatomagmatic eruption is more likely. We have plotted the occurrence of eruptions in conjunction with temperature of the crater lake and investigated whether eruptions are more or less likely to occur when the crater lake is hot, cold or whether the trend is to hotter or colder temperatures. This work suggests that the tem-perature of the crater lake does not show good correlation with eruption occurrence.

1 2 3 4 50

0.5

1

1.5

2

2.5Frequency Size distribution of eruptions

Size

Log1

0(N

umbe

r of e

vent

s) (n

ot c

umul

ated

)

allall except 1995 96above meanbelow meanhotcoldupdown

2000 2001 2002 2003 2004 2005 2006 2007 20080

5

10

15

20

25

30

Cumulative number of VT ALL DATA

over 1 dayover 1 weekover 1 month

2000 2001 2002 2003 2004 2005 2006 2007 20080

20

40

60

80

100

120

140

160

180

200

Cumulative number of VO ALL DATA

over 1 dayover 1 weekover 1 month

We have analyzed all the systematically monitored parameters since 1999 and looked for change points in their time series. Although these show no clear indicators, there is an interesting cluster of change points in both seismic and geochemical data from late 2004 to late 2005. The reason for these change points merits further investigation, although it is possible that both the hydrothermal system and crater lake may act to mask or change monitoring signals (if present) that magma produces deeper in the edifice.

1965 1970 1975 1980 1985 1990 1995 2000 2005 10

20

30

40

50

60

Eruption SizeSize 1Size 2Size 3Size 4Size 5

Temperature in Crater Lake and eruption occurrence

Tem

pera

ture

in C

rate

r Lak

e

time

BET applicationIn the present study, we attempt to apply BET_EF to Mt Ruapehu, a very active and well-monitored volcano exhibiting the typical features of open conduit volcano-es. We think that an attempt to apply BET_EF at Mount Ruapehu is worthwhile, for several reasons. First, moni-toring data at Mt Ruapehu can be helpful in forecasting major events, especially if a large amount of magma is intruded into the edifice and becomes available for ph-reatomagmatic or magmatic eruptions, as for example in 1995-96. Secondly, in setting up BET_EF for Mt Rua-pehu we are forced to define quantitatively what the background activity is. This will result in a quantitative evaluation of what changes in long time monitored pa-rameters may influence the probability of future erup-tions. Further, we would like to extend BET_EF probabi-lities into a more complete BET_VH application, in order to assess the hazard from lahars invading the skifields.

“Volcano Observatory Best Practices Workshop: Near-Term Eruption Forecasting”, Erice, 11-15 September 2011

ERUPTION LOCATION SIZE/TYPE

NODE3

NODE4

NODE5

ORIGIN

NODE2

UNREST

NODE1

BET_EFEruption Size 1

No Eruption Size i

Location 1

Location N

Location j

Location 2

...

...

Size M

...

...

Magmatic

Hydrothermal

Unrest

No unrest

ERUPTION LOCATION SIZE/TYPE PHENOMENA REACHINGAREA

OVERCOMINGTHRESHOLD

BET_VHNODE1-2-3

NODE4

NODE5

NODE6

NODE7

NODE8

Eruption Size 1

No Eruption Size i

Location 1

Location N

Location j

...

...

......

...

Lahar

Size M Lava Flow

Pyrocl. Flow

Area 1

Area L

Area 2

No

Yes...

...

1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979

20

40

60

Temperature in Crater Lake

Tem

pera

ture

(C)

1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994

20

40

60

Tem

pera

ture

(C)

1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

20

40

60

Tem

pera

ture

(C)

1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979

1000

2000

3000

Mg in Crater Lake

Mg

(ppm

)1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994

1000

2000

3000

Mg

(ppm

)

1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

1000

2000

3000

Mg

(ppm

)

1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 19790.5

11.5

22.5

x 104 Cl in Crater Lake

Cl (

ppm

)

1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 19940.5

11.5

22.5

x 104

Cl (

ppm

)

1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 20090.5

11.5

22.5

x 104

Cl (

ppm

)

1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 19790.05

0.1

0.15

0.2Mg/Cl ratio in Crater Lake

Mg/

Cl

1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 19940.05

0.1

0.15

0.2

Mg/

Cl

1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 20090.05

0.1

0.15

0.2

Mg/

Cl

definition of background and of anomalies

Conclusions- We have investigated past and current activity at Mt Ruapehu to deter-mine likely precursor signals prior to eruptions, even for small scale, and yet hazardous, events. Further work is needed to understand whether change point indicators are significant.

- We are trying to define quantitatively the background and unrest states at Mt Ruapehu, in order to apply BET_EF on a well monitored, frequen-tly active, open conduit volcano.

- We plan to evaluate the hazard posed by lahars on Whakapapa ski fields.

References- A.D. Jolly , S . Sherburn, P. Jousset, G. Kilgour (2010) Eruption source processes derived from seismic and acoustic observations of the 25 Sep-tember 2007 Ruapehu eruption - North Island, New Zealand. Journal of Volcanology and Geothermal Research, 191:33–45- G. Kilgour , V. Manville, F. Della Pasqua, A. Graettinger , K.A. Hodgson, G.E. Jolly (2010) The 25 September 2007 eruption of Mount Ruapehu, New Zealand: Directed ballistics, surtseyan jets, and ice-slurry lahars. Journal of Volcanology and Geothermal Research, 191:1–14- W. Marzocchi, L. Sandri, J. Selva (2008) BET_EF: a probabilistic tool for long- and short-term eruption forecasting. Bulletin of Volcanology 70:623–632 - - W. Marzocchi, L. Sandri, J. Selva (2010) BET_VH: a probabilistic tool for long-time volcanic hazard assessment. Bulletin of Volcanology 72:705–716- S. Sherburn, C.J. Bryan, A.W . Hurst, J.H. Latter , B.J. Scott (1999) Sei-smicity of Ruapehu volcano, New Zealand, 1971–1996: a review. Journal ofVolcanology and Geothermal Research, 88:255–278

1 2 22223 4 5