As they flew along the newly formed Greendale Fault in September 2010, there were many things that Dr Simon Cox and his GNS Science colleagues could see were clearly awry during the first aerial reconnaissance. But among the offset roads, bent hedgerows and fences, torn and cracked fields, and damaged houses, there was one curiosity that specifically caught Simon’s attention. Many of the wells now had groundwater flowing out of them where normally it would have been at least 40 metres underground. Having worked on the formation of quartz veins and exploration of gold deposits, which are developed by earthquakes deep in the earth’s crust, here was a real-life example of processes he had studied for his PhD and worked on for years, but never actually observed as they happen deep in the earth’s crust. Not only that, but he was seeing it in full realisation there would be many recordings of what had happened in monitoring bores all over the Canterbury plains. While there was much to do initially for earthquake recovery, this was a serendipitous opportunity for new science that rarely comes along.
The weird things that seemed to happen to groundwater during the earthquakes also caught the attention of Dr Helen Rutter from Aqualinc, as well as many people from Environment Canterbury and regional councils throughout New Zealand. Emails quickly brought collaborators together and observations gathered. One of the first things that seemed apparent was the earthquakes generated a rapid increase in groundwater pressures immediately during the shaking, followed by a drop in the deeper aquifer levels and rise in shallow aquifers immediately after the earthquake. In many places there had been an upwards flow that coincided with emergence of new springs and a pulse of water flowing down the rivers. The effects were also far-reaching, affecting aquifers over 1000 km-away in Northland and throughout New Zealand. In a paper1 in the New Zealand Journal of Geology and Geophysics, which subsequently won the New Zealand Geophysics Prize, Cox and his colleagues postulated that release of artesian groundwater pressure and groundwater flow also played pivotal roles in Christchurch liquefaction.
Above: ‘Haywire hydrology’ as seen from an earthquake reconnaissance flight on 4 September 2010. Offset on the Greendale fault disturbed the flow of Hororata River. Photo: Richard Jongens/GNS Science.
Above: An irrigation channel dammed and partially offset by the Greendale Fault. As as seen from an earthquake reconnaissance flight on 4 September 2010. Photo: Richard Jongens/GNS Science.
With support from the Royal Society Marsden fund and the Natural Hazards Research Platform, a portfolio of ‘Earthquake Hydrology’ research began to grow. Hydrologic responses in groundwater are a well-documented phenomenon, observed for thousands of years, but of renewed interest due to implications for water supplies, engineering and hazards2. But although it appears permeability is best considered a dynamic variable that can change with stress and strain, exact causal mechanisms are yet to be perfectly understood3,4. The density of groundwater and seismic monitoring networks in New Zealand, combined with regular earthquake activity, was a unique opportunity for local research to contribute to internationally. Albeit dominated by observations in permeable young aquifers, events of the 2009 Fiordland, 2010-2011 Canterbury, 2013 Cook Strait and 2016 Kaikoura earthquakes have now been particularly well-documented1,5-7, shown to produce changes to groundwater throughout New Zealand lasting locally for many years8. Other impactful New Zealand studies, all enabled through collaborations and support of regional councils and the GeoNet facility, include large ‘multiple-earthquake at multiple-site’ datasets of earthquake-responses that enable causative variables to be controlled and examined9,10, and the first international attempt to derive probabilistic models of groundwater change given varied levels of shaking11.
Above: Liquefaction at Kirstens Place as a results of the Mw6.2 Christchurch earthquake, 22 February 2010. Photo: Alun Davies/simplicitywebdesign.co.nz.
With aquifer leakage and compromised aquitards postulated to have somehow been involved in liquefaction1,6,12, Sjoerd van Ballegooy and other engineers from Tonkin & Taylor became involved in the research. One of the first challenges for understanding liquefaction was to map out exactly what happened within, across and between the aquifers throughout the earthquake sequence. A collaboration contributed maps of the water table position to help inform the Christchurch rebuild13. Mapping pressure changes during the earthquakes, the team found groundwater in aquifers beneath Christchurch rose locally in pressure an equivalent of > 5 m in hydraulic head during the Mw6.3 earthquake. This added substantially to the +5-10 m ‘above ground’ pressures that were normally contained by the overlying impermeable aquitard layers. The spatial correlations also appeared very clear – much of the worst liquefaction and surface flooding coincided with the areas that experienced highest groundwater pressures. But ‘correlation is not proof of causation’ so initial reviews of the work were critical that spatial relationships could have been coincidental because the highest pressures were also where the greatest thickness of liquefaction-prone sandy-silt were found. The science was still premature, and a new approach was needed.
Graphic figure: (A) Map of the occurrence of liquefaction as ejected to the surface after the 22 Feb 2011 Mw6.2 earthquake, overlain with coloured contours of total coseismic hydraulic head in aquifers. (B) Histograms of the number of 50 x 50 m grid cells of ejected sediment occurrence (red, left scale) and not observed (blue, right scale) within 1 m intervals of aquifer hydraulic head. (C) Spatial probability distributions for ejected sediment ‘occurrence’ and ‘not observed’ derived from the relative proportion of classified grid cells within each 1 m groundwater level in the study area. Reproduction of Figure 4 from Cox et al. (2021).
Enlisting further expertise of statistician David Harte, the team began unpicking the relationships between the geotechnical tests carried out by engineers on liquefaction vulnerability, actual observations of liquefaction during the earthquakes, and groundwater pressure changes that had occurred in aquifers about 20-40 m below each site. Where the first approach had been only at a regional-scale, relationships between liquefaction and groundwater, deliberately including areas of the Canterbury Plains to the west of Christchurch to span a wide range of aquifer pressures, it had potentially caused non-causative covariations in the data. The reanalysis looked only within the Christchurch urban area. When tests were grouped on the basis of liquefaction vulnerability and soil strength, irrespective of location, places where ‘minor’ and ‘moderate-severe’ liquefaction occurred during the 22 February Mw6.2 earthquake, had distinctly higher aquifer pressure than sites where liquefaction was not observed. Now with immediate relevance to engineering, a series of manuscripts were written and rewritten for the Journal of Engineering Geology.
‘It was an incredibly frustrating and long road that was bad for morale at the time’ says Simon Cox ‘but ultimately a process that greatly strengthened the science. Although it took a decade from initial observations, developing the hypothesis then teasing all the details out, it now seems unequivocal leakage of artesian groundwater contributed to near-surface liquefaction-induced ground damage.’
The final paper14 was published on the 10th anniversary of the Canterbury earthquake. It argues leakage and upwards flow from artesian aquifers promoted the ejection of liquefied sediment in Christchurch. It showed that hazard assessments need to consider hydrogeological setting and conditions, that Christchurch may have been one of the worst-case examples ever experienced, and there is a need to re-evaluate other examples of liquefaction worldwide. The paper can be downloaded for free from https://doi.org/10.1016/j.enggeo.2020.105982, while the associated data archive can be found at http://doi.org/10.5281/zenodo.4391461.
- Cox, S.C., Rutter, H.K., Sims, A., Manga, M., Weir, J.J., Ezzy, T., White, P.A., Horton, T.W., Scott, D., 2012. Hydrological effects of the Mw 7.1 Darfield (Canterbury) earthquake, 4 September 2010, New Zealand. New Zealand Journal of Geology and Geophysics 55(3), 231-247.
- Wang, C-Y., Manga, M. 2010a. Earthquakes and water: Lecture Notes in Earth Sciences, v. 114, Springer, New York 235 pages.
- Ingebritson, S.E., Manga, M., 2019. Earthquake hydrogeology. Water Resources Research 55, 5121-5216 doi:10.1029/2019WR025341.
- Manga, M., Beresnev, I., Brodsky, E.E., Elkhoury, J.E., Elsworth, D., Ingebritsen, S.E., Mays, D.C., Wang, C-Y., 2012. Changes in permeability caused by transient stresses: Field observations, experiments, and mechanisms. Reviews in Geophysics 50, RG2004.
- Cox, S.C., Menzies, C.D., Sutherland, R. Denys, P.H., Chamberlain, C., Teagle, D.A.H. 2015. Changes in hot spring temperature and hydrogeology of the Alpine Fault hanging wall, New Zealand, induced by distal South Island earthquakes. Geofluids, 15(1/2): 216-239
- Gulley, A., Dudley Ward, N.F., Cox, S.C., Kaipio, J.P., 2013. Groundwater responses to the recent Canterbury Earthquakes: a comparison. Journal of Hydrology 504, 171-181.
- Weaver, K.C.; Cox, S.C.; Townend, J.; Rutter, H.; Hamling, I.J.; Holden, C. 2019. Seismological and hydrogeological controls on New Zealand-wide groundwater level changes induced by the 2016 Mw7.8 Kaikōura earthquake. Geofluids, A9809458.
- Rutter, H.K., Cox, S.C., Dudley Ward, N.F., Weir, J.J., 2016. Aquifer permeability change caused by a nearfield earthquake, Canterbury, New Zealand. Water Resources Research 52, 8861-8878
- O’Brien, G.A.; Cox, S.C.; Townend, J. 2016 Spatially and temporally systematic hydrologic changes within large geoengineered landslides, Cromwell Gorge, New Zealand, induced by multiple regional earthquakes. Journal of Geophysical Research. Solid Earth, 121(12): 8750-8773.
- Weaver, K.C., Doan, M.-L., Cox, S.C., Townend, J., Holden, C., 2019. Tidal behavior and water-level changes in gravel aquifer in response to multiple earthquakes: a case study from New Zealand. Water Resources Research 55, doi:10.1029/2018WR022784.
- Weaver, K.C.; Arnold, R.; Holden, C.; Townend, J.; Cox, S.C., 2020. A probabilistic model of aquifer susceptibility to earthquake-induced groundwater-level changes. Bulletin of the Seismological Society of America 110(3), 1046-1063.Dudley Ward, N. 2015. On the mechanism of earthquake induced groundwater flow. Journal of Hydrology 530, 561-567.
- Dudley Ward, N. 2015. On the mechanism of earthquake induced groundwater flow. Journal of Hydrology 530, 561-567.
- van Ballegooy, S.; Cox, S.C.; Thurlow, C.; Rutter, H.K.; Reynolds, T.; Harrington, G.; Fraser, J.; Smith, T. 2014. Median water table elevation in Christchurch and surrounding area after the 4 September 2010 Darfield Earthquake. Version 2. Lower Hutt, NZ: GNS Science. GNS Science report 2014/18. 79 p. + appendices.
- Cox, S.C., van Ballegooy, S., Rutter, H.K., Harte D.S., Holden, C., Gulley A.K., Lacrosse V., Manga M. (2021) Can artesian groundwater and earthquake-induced aquifer leakage exacerbate the manifestation of liquefaction? Engineering Geology, 281: article 105982; doi:10.1016/j.enggeo.2020.105982.