| Title | Investigating the variance in activity of the Lady Knox Geyser, Waiotapu Thermal Wonderland, New Zealand |
|---|---|
| Authors | B.R. Rachman, S.J. Zarrouk, K. Luketina,J. McLeod, C. Kotze, A.W. Kempa-Liehr |
| Year | 2025 |
| Conference | New Zealand Geothermal Workshop |
| Keywords | geothermal springs, geysers, geothermal monitoring, geothermal resource management, Waiōtapu, machine learning, feature engineering, time-series analytics |
| Abstract | The Lady Knox Geyser is within the tourist complex of Waiotapu Thermal Wonderland (WTW), and is soap-induced to erupt for the entertainment of tourists. Currently, the height and duration of the eruption vary. The Thermal Park Management deems the eruption duration to have become shorter. The WTW management and the Waikato Regional Council would like to have the variability investigated. Understanding Lady Knox Geyser's behaviour is part of the landowners, the Iwi of Ngati Tahu – Ngati Whaoa, commitment to Kaitiakitanga (Guardianship, Conservation, and Sustainability). This study examines hot-spring temperature time series data from the Lady Knox Geyser. The geyser shows a distinct peak around the same time every day, as it was induced by the Park Rangers using soap to trigger the eruption. During its quiescent period, the spring temperature mean at several metres depth is 103°C. Detailed scrutiny of data shows that a different pattern of temperature data is distinguished between a short eruption duration and a long one. However, some eruptions do not abide by this distinct pattern. The correlation between peak spring temperature and weather data was analysed to determine whether there are any correlations between eruption duration and intrinsic factors and/or external perturbations. The peak spring temperature and air pressure are correlated with Lady Knox Geyser's eruption duration during the 3 months the geyser was investigated. In addition to correlations, time-series spring temperature datasets were used to build a machine-learning model that would predict the duration of eruption events in the future. A reasonable model would allow the WTW Management to know the duration of the eruption before it is induced and to help decide any actions to preserve the Lady Knox Geyser better. |