Keywords Liquefaction, risk assessment, planning, regional New Zealand
1. Introduction
Since the Canterbury Earthquake Sequences (CES), liquefaction and its effects have been a focus for engineers and councils. There has been a significant amount learnt from these events in New Zealand due to the disproportionate damage and loss of building stock. Consequently, several guidelines have been produced to assist designers and owners to construct suitably resilient structures. In addition, changes from the November 2019 Building Code Update have revised B1/AS1 to ensure new buildings are built safe and strong enough to withstand liquefaction effects.
As part of the update, the Ministry of Business, Innovation and Employment (MBIE) has required zoning by councils to provide guidance in their territory boundaries to define liquefaction risk to better assist design and consenting based on the above changes. Typically for councils, this zoning has been undertaken by various organisations or consultants across the country.
The zoning approach is easier for main centres as there is usually sufficient readily available data to help with these types of assessments and/or there are local based cone penetration test (CPT) / borehole contractors to allow for easy testing. However, in smaller centres such as Invercargill City Council (ICC) there is less readily available geotechnical information, and it is also costly and difficult to get cone penetration tests or machine borehole rigs to undertake site specific testing especially for small scale projects compounding the area wide lack of data. A review of the New Zealand Geotechnical Database (NZGD) shows that readily available deep testing information around Invercargill as an example is concentrated along large government/Council funded linear infrastructure or isolated larger commercial developments.
Although the Invercargill region has a lower seismic risk compared to other parts of the country, the experience of the author indicates that some parts of the region have a high liquefaction risk and need to be appropriately assessed. Invercargill City Council has previously engaged Tonkin & Taylor (2022) to undertake a regional assessment of risk. The T&T assessment defined two zones for Invercargill; “Liquefaction Damage is Possible” and “Liquefaction Category is Undetermined”. Consequently, the Invercargill City Council (ICC) has then made rules requiring deep geotechnical investigations to define the risk for the “Liquefaction Damage is Possible” zones and to confirm the risk in the “Liquefaction Damage is Undetermined” areas for all developments regardless of size or scope.
The author considers that the generalised approach may not be warranted for low risk small scale projects and proposes a simplified assessment methodology based on his experience within the region, for typical single or double storey lightweight residential houses, without having to undertake expensive CPTs and/or boreholes. The paper presents a review of relevant data and proposes an alternative approach to qualifying liquefaction risk around Invercargill as opposed to the binary approach currently adopted. Key important sources of relevant data are presented that should be reviewed prior to engaging expensive out of town CPT/borehole contractors.
2. Geological Setting and Geotechnical Information
Invercargill City Council region covers an area of the south of the South Island. It contains the main city area in the centre with smaller areas such as Lorneville to the north, Otatara to the west, Awarua and Bluff to the south with a series of other smaller areas between.
The city of Invercargill is built upon Early to Late Quaternary alluvial and swamp sediments that overlie Oligocene to Pliocene sedimentary rock. The hills around Bluff are formed of mafic to ultramafic igneous rock of Permian age (GNS, 2013).
Readily available geotechnical information has been historically sparse but more data is slowly being uploaded on the NZGD including deep testing like boreholes and CPTs. What is less obvious but also freely available is the significant number of borehole records held by Environment Southland. However, the Environment Southland data is currently not on a readily available and open system like a lot of other regional councils. The above free data is in addition to the information internally held by experienced local consultants such as GeoSolve who have been working in the region for a significant amount of time.
In terms of seismic risk, Invercargill is importantly in a lower seismic region in New Zealand. The 25, 100 and 500 year design accelerations according to the MBIE and New Zealand Geotechnical Society Inc (NZGS) earthquake geotechnical engineering practice (2021) are 0.05g, 0.11g and 0.21g respectively with a magnitude 6.1 event. This is significantly less than other regions in New Zealand. However, the new national seismic model accelerations are slightly greater typically in the 0.15g to 0.45g range with a median value around 0.31g (assuming a Vs30 of 300 m/s to be consistent with the MBIE/NZGS values). Therefore, it is possible that there is a slightly higher risk than currently stated as presented later in the paper.
3. Typical Ground Conditions Around Invercargill
The author has reviewed over 300 CPTs and boreholes, as well as significant number of shallow testing and simple deep probe results in addition to the relevant geological maps. This includes data held by GeoSolve, the NZGD and Environment Southland. Based on the review and even though there are difference geological zoning across most of the Invercargill area, the basic ground conditions are very similar with the exception of the Bluff Area. The ground conditions can generally be grouped into three typical zones as discussed below. The main difference between these typical ground conditions is depth of ground water level and the start of the dense to very dense alluvial gravel layer. However, in the Bluff area, there is a mix of volcanics, alluvial soils and reclaimed soil conditions so the area is more complex.
3.1 Typical zone 1: Central Invercargill
Boreholes drilled show approximately 2-3m of historical fill overlying sands and silts. Gravels grading of various portions of sand and silt are encountered at depth. SPT values are generally 50+ within the gravels below the sands/silts. The gravels extend to depth. Groundwater is within the gravel unit. Based on this profile, liquefaction risk is calculated to be limited and equivalent to MBIE TC1, according to Table 3.1 of MBIE guidelines (2018).
3.2 Typical zone 2: Eastern Invercargill
CPTs in this area infer silts and sands to about 5 m and they are typically penetrated 8 m. Environment Southland borehole information around the area shows similar ground conditions with the gravel unit extending to depth. Groundwater is within the upper sands and silts. Liquefaction triggering is about a 100-year earthquake and shows minor to moderate risk in the MBIE/NZGS Module 1 ULS earthquake event. Based on this profile, liquefaction risk is calculated to be equivalent to MBIE TC2, although no liquefaction is calculated in the SLS earthquake event.
3.3 Typical Zone 3: Western Invercargill
CPTs in this area infer sands to at least 10 m depth with layers of silt. Water is relatively shallow. Triggering is again about a 100-year earthquake and shows moderate to major risk in the MBIE/NZGS Module 1 ULS earthquake given shallow groundwater and loose material at near ground surface which suggests high risk of liquefaction damage to the near surface. Based on this profile, liquefaction risk is calculated to be equivalent to MBIE TC2. Although given shallow ground water and potentially liquefiable soils, there is a high risk of ground damage and direct interaction of a shallow foundation and liquefiable layers. So given the risk of shallow layers liquefying, MBIE TC2/TC3 categorisation is considered more appropriate.
3.4 Concluding remarks
Based on the three zones presented above, it is obvious that the liquefaction risk in the area is a result of the ground conditions and groundwater. This is true of general liquefaction but the distinct ground conditions and groundwater regime in the three zones lends to the potential to easily classify the wider area into three distinct zones in terms of liquefaction risk. To validate the hypothesis, a detailed liquefaction analysis was undertaken across the region based on readily available data as presented in Section 4 below.
4. Detailed Liquefaction/Sensitivity Analysis
4.1 Liquefaction damage potential from existing data
To provide further insight, all readily available CPTs (NZGD and internally held by GeoSolve) have been used to run a liquefaction assessment for a range of earthquakes events using the method of Boulanger and Idriss (2014). Due to the size of the data set and known similar upper ground conditions in the region, the author has not looked at zoning sites based on geological mapping. Instead, the entire dataset has been reviewed as a whole. In the data set used, there is approximately 200 tests. The analysis did not include CPT data with early refusals or old data not supported by recent nearby data. For example, there are 40 relatively shallow tests around the wastewater treatment ponds which due to their age encountered effective refusal at relatively shallow depth compared to recent surrounding testing. And as discussed before, the bluff area has not been considered as there is a complicated transition between reclaimed/alluvial soils into volcanics and there is currently not good enough data set to look at properly. Other areas of reclaimed land especially around the south-western side of the city have been considered. Boreholes have not been considered in the analysis for ease of the assessment. Lateral spreading has also not been considered in this assessment and would require a site specific assessment.
A probability of liquefaction (PL) equal to 15% has been used. Groundwater has been adopted based on recorded levels adjusted where appropriate. As the author is not aware of any large enough laboratory data to justify any adjustment, a fines content correction factor (CFC) of 0 and soil classification index (IC) cut off at 2.6 was adopted. Soil profiles have been cut off at 10 m in line with MBIE Guidelines for site classification and no additional layers are calculated below the refusal of any CPTs if 10 m has not been reached.
To show changing liquefaction risk, plots of settlement based on Zhang et al (2002) and Liquefaction Severity Number (LSN) based on Tonkin & Taylor (2013) are shown in Figure 1 below. Grey, yellow and blue zones have been shown on the settlement plot showing MBIE Technical Categories TC1, TC2 and TC3 respectively. Blue, green and red zones have been shown on the LSN plot showing none to minor, minor to moderate and moderate to severe inferred liquefaction-induced ground damage risk.
Figure 1: Indexed settlement and LSN against peak ground acceleration. Red, orange and yellow solid lines indicate median for the entire data set, western data set and central/eastern data set respectively. Dashed lines indicate upper and lower quartiles.
Based on the above, the step change where liquefaction occurs is between a peak ground acceleration (PGA) 0.1g and 0.15 g which is approximately a 1/100 year earthquake. Full liquefaction is about a 0.2g to 0.35 g which is either equal to or above the MBIE/NZGS ULS earthquake event. Therefore, although the relatively low seismic risk inferred in Invercargill, a relatively higher proportion of liquefaction at the MBIE/NZGS ULS earthquake event is possible based on the analysis. It can also be seen that the updated New Zealand Seismic Model values are about 60%-70% higher than the MBIE/NZGS earthquake.
As there is quite a lot of scatter in the data, the data has been split into two zones; the western side of Invercargill and the central/eastern side covering the rest of Invercargill, and calculated median and upper/lower quartile values of the data shown on the same plots. The main difference between these two zones is the gravel unit depth. Based on the results of assessment, it can be seen that the liquefaction risk in the central/eastern side of the Invercargill is considerably lower with relatively low risk occurring in the lower 75% of the data.
From this, a distinct difference between the two zones can be observed, with the western zone having higher risk than the central and eastern zone. For reference, liquefaction induced vertical effects show the western zone predominately having an equivalent MBIE TC2 category whereas the central/eastern zones have an equivalent MBIE TC1/TC2 category. This also makes sense based on the regional geology and gravel unit depth discussed before.
4.2 Effect of groundwater table
The other main aspect is the effect of groundwater depth especially as groundwater or point of saturation is relatively easy to gauge on site. Therefore, a second sensitivity study was undertaken to assess the groundwater depth at which liquefaction risk would be relatively low, as shown in Figure 2 below.
Figure 2: Indexed settlement and LSN against groundwater level at MBIE ULS earthquake. Red, orange and yellow solid lines indicate median entire data set, western data set and central/eastern data set respectively. Dashed lines indicative of upper and lower quartiles.
When groundwater depth decreases, liquefaction risk for a shallow founded structures will decrease which is seen on the site. However, a clearer drop in risk is seen on the central/eastern side of the area due to groundwater depth getting closer to the gravel unit. For the central/eastern side, there is relatively low risk for at least 75% of the data when the groundwater table is deeper than 2 m depth. Whereas for the western side, there isa relatively low risk for at least 75% of the data when the groundwater table is deeper than 4-5 m depth. Therefore, it is possible that groundwater depth could be used to infer liquefaction risk at the site.
For completeness, the groundwater sensitivity analysis was carried out for 0.31 g using information from the updated New Zealand seismic hazard model, as shown in Figure 3. These show a similar trend as above but groundwater depths resulting in relatively low risk are at greater depths.
Figure 3: Indexed settlement and LSN against groundwater level at updated New Zealand Seismic Model ULS earthquake. Red, orange and yellow solid lines median entire data set, western data set and central/eastern data set respectively. Dashed lines upper and lower quartiles.
5. Discussion
Even though Invercargill is considered to be a region of lower seismic risk, based on the analysis as presented in this paper, there is a liquefaction risk in moderate to major earthquakes in areas of high groundwater levels. However, when groundwater is deeper, especially in the central and eastern areas of Invercargill the liquefaction risk can be relatively low.
Looking back at the entire dataset and including boreholes, shallow investigations and simple probe tests, the difference between the western and central/eastern parts of Invercargill is the depth of the main gravel unit compared to the ground water level. Therefore, this assessment effectively shows that if groundwater can be confirmed to be within gravel units that liquefaction risk would be limited. It then follows that if testing was to be undertaken such as deep test pits to confirm the gravel and groundwater depth, then the liquefaction potential could be qualified without needing expensive deep CPT or borehole testing. However, caution is required to ensure that the consistency of the gravel unit is confirmed as these gravel units can be relatively thin at times and be underlain by sand and gravel layers.
Based on the above, the author proposes that it is possible that current mapping could be adjusted to account for the presence of relatively shallow gravels and deep groundwater levels in parts of Invercargill. The assessment indicates that the seismic risk and associated liquefaction risk could increase once the updated New Zealand seismic model is incorporated into codes. Therefore, some allowance should be made for this in assessments in the region in the meantime given the potentially large change in the ULS earthquake event and showed liquefaction results.
There is also no reason why a similar type of assessment could not be undertaken in other areas with lower quantities of readily available high quality data focusing on groundwater and units of high/low liquefaction risk.
6. Implications for Piled Foundations
As previously mentioned, the upper soils can contain poor historical fill which can include high organic content or other unsuitable soils. In other areas upper soils are very soft to soft. For this reason, it is not uncommon in Invercargill to pile structures into the gravel unit including smaller standard structures like houses. It should be noted that the assessment presented in this paper is more focused on shallow foundation systems than piled systems as the focus was on indicators for ground damage. Where deep piled foundations are to be used and founded within the non-liquefiable gravel unit, deep testing is required to confirm embedment depths and consistency of founding layer below pile toe.
7. Conclusion
A review of readily available geotechnical and geological information has been undertaken for the Invercargill area to re-assess the liquefaction risk. A sensitivity assessment has been undertaken using CPT information to assess vertical liquefaction effects. The assessment shows that the liquefaction risk is not significant in areas where shallow gravels and/or deep groundwater is present. Therefore, based on this assessment the author proposes that current mapping could be adjusted to account for the presence of relatively shallow gravels and deep groundwater levels in parts of Invercargill.
In the short term, the proposed generalised approach of confirming groundwater and gravel depths may also be used for low risk small scale projects without having to undertake expensive CPTs and/or boreholes.
References
Boulanger, R. W. and Idriss, I. M. (2014). CPT and SPT based liquefaction triggering procedures. Centre for Geotechnical Modelling, Department of Civil and Environment Engineering, University of California, Davis, California.
GNS Science (2023). Urban Geological Maps, Invercargill. https://www.gns.cri.nz/our-science/land-and-marine-geoscience/te-riu-a-maui-our-continent/geology-of-new-zealand/urban-geological-maps/invercargill/
Ministry for the Environment (MfE) and Ministry of Business, Innovation and Employment (MBIE), (2017). Planning and engineering guidance for potentially liquefaction-prone land, Resource Management Act and Building Act aspects.
Ministry of Business, Innovation and Employment (MBIE), (2018). Repairing and rebuilding houses affected by the Canterbury earthquakes.
Ministry of Business, Innovation and Employment (MBIE) and New Zealand Geotechnical Society Inc (NZGS), (2021). Earthquake geotechnical engineering practice, module 1, overview of the guidelines.
Tonkin & Taylor Ltd. (2013). Liquefaction Vulnerability Study.
Tonkin & Taylor Ltd, (2022). Liquefaction vulnerability assessment.
Zhang, G., Robertson, P. K., and Brachman, R. W. I. (2002). Estimating liquefaction-induced ground settlements from CPT for level ground.
