KAIKOURA EARTHQUAKE AND RESILIENCE ON THE MAIN NORTH LINE Daniel Headifen, BE Civil (Hons), CPEng NZ, CEng UK, MIPENZ, MICE

KiwiRail Wellington Railway Station Wellington New Zealand daniel.headifen@kiwirail.co.nz 0064 21926873

NUMBER OF WORDS: 5032

ABSTRACT The November 2016 Kaikoura Earthquake was the largest disaster to ever strike New Zealand’s railway in terms of amount of damage. Large parts of the Main North Line (MNL) Railway were engulfed in huge slips or thrown into the sea. Bridges were destroyed and tunnels broken. The damage to the rail and State Highways disconnected New Zealand’s transport system. The subsequent response and rebuild has challenged not just KiwiRail but New Zealand’s construction industry overall. It has resulted in the biggest rail project in the South Island of New Zealand since the Second World War.

As part of that project, the need for considering resilience has had to be faced. How this would be defined, the analysis undertaken to quantify it and the work done to achieve it are part of this paper. In doing so, this paper will also outline some of the internationally award winning works that have been undertaken on the railway to achieve reopening the line in only 10 months after the earthquake, despite been hit by several tropical cyclones and then what has been done to continue to improve service levels after that and keep trains moving across New Zealand.

INTRODUCTION This paper covers the events that occurred in the Kaikoura Earthquake, the makeup of the Main North Line (MNL) and the damage that the earthquake caused on it. It details some of the works undertaken to repair that damage. It then reviews how resilience for these repair works was considered, planned for and then undertaken in the rebuild. It breaks down resilience into both engineering works for robustness and also operational planning for responsiveness.

WHAT HAPPENENED IN THE KAIKOURA EARTHQUAKE? th At 00:02am on 14 November 2016, a Mw 7.8 (Richter Scale) earthquake occurred in the northeast of the South Island of New Zealand. The approximately 2 minutes of shaking caused significant damage across this part of the South Island and also in New Zealand’s capital, Wellington, at the bottom of the North Island. Two fatalities resulted and numerous people were injured. A number of buildings were damaged to the extent that they either collapsed or were no longer suitable for use.

In the course of over twenty faults rupturing, the ground shaking progressed in a northern direction reaching Peak Ground Accelerations of >3g at one location. Thousands of landslides were caused, several over 1.3 million cubic feet (1 million cubic meters) in volume. The large number of fault ruptures and the manner in which those interrelated has led to this earthquake been classed as the most complex ever to have been studied.

The South Islands transport links were severely affected, in particular by hundreds of the landslides. New Zealand’s State Highway 1 (SH1) and the sole rail link in the upper South Island, the MNL, were severed at numerous locations. It was due to the cutting of these transportation links to the popular coastal tourist town of Kaikoura, that this earthquake has become known as the Kaikoura Earthquake. All road, rail and coastal links to Kaikoura were closed due to earthquake damage. Hundreds of tourists from around the world were evacuated by an air lift operation as well as New Zealand navy craft with assistance, by chance, from several visiting international navies, including the United States of America’s. However, once tourists were evacuated, Kaikoura’s 2,000 residents were still effectively cut off from the rest of the world with only limited links been restored by the end of 2016.

Figure 1: Northern South Island with its major geological faults identified. The Kaikoura earthquake initiated near Waiau in North Canterbury, at a point marked by the red star. GNS Science (1)

MAKE UP OF THE MNL The MNL is a single track narrow gauge (42 inches (1067mm)) railway running 217miles (350km) between the ferry port of Picton by the Cook Strati to the South Island’s largest city, Christchurch. It commenced construction in 1870 however encountered numerous delays and was not completed till 1945. The majority of the section affected by the earthquake was built between 1930 and 1945. It covers a variety of geography but the sections affected by the earthquake were generally where the route tracks along a narrow coastal bench at the bottom of significant mountain ranges as well as through several river valleys. It is for this reason that it passes through 21 tunnels, over 183 bridges and past several significant seawall structures.

The MNL is a predominately a freight route carrying over 1,000,000 tones of mixed freight a year but it also carries a daily tourist focused passenger service. The majority of freight runs north to south, on the Auckland (the country’s largest city in the north of the North Island) to Christchurch route where the rail service bridges the Cook Strait between the islands utilizing rail ferries that KiwiRail employs. The MNL is therefore predominately a transiting route with little freight generation on its length.

Figure 2: Key Damage Location to the MNL

DAMAGE TO THE MNL Over 900 different sites were affected, with the scale varying from track buckles to the complete destruction of reinforced concrete bridges, the dislocation of tunnel structures by ground movement and huge slips of several hundred thousand cubic feet (cubic meters) picking up the rail track and throwing it bodily into the Pacific Ocean.

Figure 3: Rail pushed into the sea by a slip at Half Moon Bay’, north of Kaikoura

While the slips were the most noticeable damage features, with over two dozen major ones affecting the MNL, the emergency inspections by railway staff and consultants in the 3 weeks following the earthquake found the less readily visible damage to 60 of the MNL’s bridges. Bridge 129 MNL over the Tirohanga Stream was unique in that its damage was certainly very visible – the Kekerengu Fault moved 9.85yards (9m) transversely across the bridge while also dropping vertically nearly 1.1yards (1m) on the northern side of the fault. This meant that the formerly straight and evenly graded bridge was massively distorted and a new lake formed around it as the streams outlet was now blocked.

Figure 4: The destroyed Tirohanga Bridge. Pre earthquake it was straight and no lake existed

Less visible again was the damage to the MNL’s tunnels. Twenty of these suffered damage. While several shallow type tunnels suffered severe distortion due to ground movement of their shallow cover, the most noticeable tunnel damage was suffered in Tunnel 18 at the south end of Half Moon Bay. Here the Hope Fault sheared the tunnel barrel with a reverse upward movement that lifted the northern 87.5 yards (80m) of the tunnel by 13.8 inches (350mm) but only ruptured a 1.1 yards (1m) length of the unreinforced 1930 era concrete lining.

Figure 5. Tunnel 18 fault damage from scan survey – Elevation

Figure 6. Tunnel 18 fault damage from scan survey - Plan

a. Post-quake damage at fault c. After track lowering

b. Timber set displacement in sidewall d. After last pour of concrete Figure 7. Tunnel 18 fault damages and repair works

In all, a length of 112miles (180km) of the MNL suffered damage.

REBUILDING THE MNL While not covered in this paper, the damage to roading infrastructure was even greater than that of rail. While SH1 differs significantly in many areas from the route that the MNL takes, in the areas of greatest earthquake damage between the Clarence River in the North and Oaro in the South, it hugs a very similar coastal route. The damage to either road or rail was unprecedented in New Zealand’s history. Together they created a need to respond in a unique fashion. This response was the creation of the first joint alliance between New Zealand’s road and rail transport providers. This was the North Canterbury Transport Infrastructure Recovery (NCTIR). It represents the New Zealand Transport Agency (NZTA) who provide state highway road infrastructure and KiwiRail, who provide rail infrastructure, and the majority of rail operations. Both are government owned entities. This alliance to repair the road and rail networks also included New Zealand’s four largest infrastructure companies – HEB, Fletcher Construction, Fulton Hogan and Downers.

NCTIR SCOPE NCTIR came into being in late December 2016 and commenced major works in February 2017. A summary of its overall scope is: • Establish access to Kaikoura via SH1 south and Inland Road (Route 70 – a mountainous local road) to reconnect the Kaikoura community. • Strengthen and manage the roading infrastructure on the SH1 alternate route (via the less robust state highways through the center north of the South Island) to cope with the extra traffic that could no longer travel on SH1 or the MNL (all rail traffic between Wellington and Christchurch being diverted either to road or coastal shipping). • Re-connect the road and rail links from Kaikoura to Picton. • Restoration of the Transport System along the coastal route between Cheviot and Clarence (located north and south of Kaikoura respectively). • Design and construct safety, resilience and tourist amenity improvements on SH1 between Oaro and Clarence. • Reinstate a safe, functioning Kaikoura harbor – the small local fishing and tourist facilities having been badly affected by uplift out of the sea. • Provide environmental, planning, consenting, and stakeholder management for the above works.

For the MNL repairs, NCTIR was tasked with undertaking all the major civil and structural works with KiwiRail undertaking the rail specific track and signaling works. NCTIR has inducted over 8,000 people at the time of writing. The peak number of workers it employed at any one time was 1,700, near the end of 2017, with closer to 1,000 through the majority of 2018 and into 2019.

Several ambitious milestones were targeted for rail, road and harbor works. While work still continues and will do so into 2020, the following milestones have been set, and successfully achieved, to date:

9 June 2017 – first work train reaches Kaikoura from the south. 8 August 2017 – last weld installed in damage areas reforming continuous rail track 15 September 2017 – night time freight restart on the MNL so that works can continue during the day. 14 November 2017 – Kaikoura Harbor is reopened. 15 December 2017 – SH1 reopens north of Kaikoura for daytime use. 9 October 2018 – daytime freight trains recommence on the MNL. 1 December 2018 – passenger train services recommence on the MNL.

The success of the works to date has been recognized with numerous New Zealand awards as well as two prestigious international awards – the Rail Technical Society of Australasia’s (RTSA) Biennial 2018 Project Award and the United Kingdom’s Institution of Civil Engineers (ICE) 2018 People Choice Award.

RESILIENCE – WHAT IS IT? For the rebuild of the MNL and other transport infrastructure, as well as the requirements of the scope listed in “NCTIR Scope” section, KiwiRail requested that NCTIR consider the resilience of both the assets created and the route as a whole. A key desire was to understand the level of service that the transport corridor would be returned to once works were complete. To do this, a Resilience Study was launched in June 2017, taking place in parallel with the recovery works that were already underway (2). The Study defined resilience as the ability of the transport corridor to recover quickly and return to original form following an adverse event, demonstrated with a simple equation:

Resilience = Robustness + Redundancy + Response

In this definition: • ‘Robustness’ reflects the ability of assets to withstand a level of stress without suffering degradation or loss of function • ‘Redundancy’ considers the extent to which assets are capable of satisfying functional requirements in the event of disruption, degradation, or loss of functionality. An example of this would be the use of an alternate route (either locally or regionally) in the case of damage to the primary route. • ‘Response’ includes the ability to sense and anticipate hazards, develop a forewarning of disruption threats, and have the ability to resource quickly in the event that a particular hazard occurs.

Mason & Brabhaharan (3) define ‘Resilience’ as the ability of infrastructure to recover quickly and return to original form following an adverse event. The resilience of a transport corridor is dependent on the loss of quality or serviceability, and the time taken to bring the corridor back into its original usage state.

Figure 8. Network Resilience (3)

HOW WAS RESILIENCE PLANNED FOR ON THE MNL? It was now important to transfer the resilience concept as a meaningful methodology of assessment for NCTIR’s work area. The scope was chosen to cover the corridor between Oaro to Clarence. While this didn’t encompass NCTIR’s entire physical scope area, for the railway, and to a somewhat lesser extent the road, it covered the majority of the badly earthquake damaged infrastructure.

Asset Types Considered The Study looked at all the primary assets that would have significant consequences to the railway and highway if they failed. The details for rail only are given here.

Bridges Twenty two rail bridges exist in the Study area. As noted previously, these bridges were of 1930s era design and construction. All were reinforced concrete sub and super structure and of ballast deck design. The majority of these bridges had the same structural features with additional spans been added to match the required bridge length. Most bridges exist to span waterways of varying widths but there are two rail over road bridges and one road over rail bridge within the Study area.

Nine bridges needed substantial repairs due to the earthquake, varying from concrete crack repairs to underpinning with new piers. Complete replacement is still being considered for two of these bridges.

Embankments A number of relatively large embankments have been constructed for rail in the Study area, Compaction, while likely better than 19th century rail fills, would still not have been to modern engineering requirements.

In general terms, embankment fills showed evidence of fill settlement with some slumping during the Kaikoura earthquake including for a large number of rail bridge abutments.

Tunnels There are 17 rail tunnels in the Study area. These tunnels typically have a cast in-situ unreinforced concrete lining with thickness varying from >20inchs (500mm) to <4inchs (100mm). Some tunnel sections would be considered more as rock shelters as they were constructed, in part, as cut and cover rather than fully mined tunnels.

In general terms, all of the rail tunnels have performed well with the notable exceptions to several tunnels in their “cut and cover” sections and Tunnel 18 (see “Damage to the MNL” section).

Unsupported Slopes Defining of natural slopes as “assets” is something that KiwiRail has done proactively across its network in almost its entirety over the last decade. The physical dimensions of a slope asset are defined by the natural terrain where an area of potential instability can be delineated from a separate feature and are not governed by human works.

Over 80 landslides occurred within the Study area as a result of the Kaikoura earthquake. In many areas, hillside and ridge cracking occurred during the earthquake without evacuative failure. At least two significant failures developed on slopes post the earthquake with one of these being of a major nature.

A significant number of debris flow sites have developed over the time interval since the Kaikoura earthquake. These failures typically appear to mobilize in short duration, high intensity rainfall events.

Timeframes Considered As the Resilience Study was happening in parallel with works happening and the transport corridor was planned to become operational in a staged manner, it was realized that the study needed to account for several “slices of time”. This was so that the owners and operators could have visibility of the changing level of service in the future. The Study therefore at the following dates: • 1 August 2017 - resumption of work trains for KiwiRail in the Resilience Study area • 15 December 2017 - reopening of State Highway 1 to public traffic • End of NCTIR - completion of all works as currently proposed

Natural Hazard Event Types Considered For each of the above dates, the resilience of the transport corridor was assessed against a number of natural hazard events. More than one level of intensity of each event type was considered.

Rain/Storm Events Return periods and characteristics as follows: • 1 in 2 year high intensity short duration rainfall (less than 1 hour) • 1 in 5 year return period long duration (24 hour) rainfall • 1 in 25 year, long duration rainfall • 1 in 100 year, long duration rainfall

Earthquake Events Two event types were considered: • A large aftershock of the Kaikoura Earthquake Sequence (assumed as Modified Mercalli, MMV), the return period for which was based on New Zealand’s Geological and Nuclear Science institute (GNS) Aftershock Forecasts • A large regional earthquake approximately equivalent to an Ultimate Limit State event. This event will have an assumed return period of around 1:250 to 1:500 and felt intensity of MMVII or higher, approximately equivalent to the return period for rupture of the Hope Fault.

Service Levels Required Service level requirements are outlined in Table 1 below. For rail, these were created by the KiwiRail Professional Heads based on KiwiRail’s documented risk appetite.

Road Rail

Duration of outage Target Return Period Duration of outage (for Target Return freight services) Period

2-4 hours 1 1-3 hrs 3

1 day 0.2 (5 years) 3-12 hrs 1

3 - 5 days 0.1 (10 years) 12-48 hrs 0.2 (5 years)

6 - 14 days 0.04 (25 years) 2-5 days 0.1 (10 years)

15 - 49 days 0.02 (50 years) > 5 days 0.04 (25 years)

50 - 120 days 0.02 (50 years)

more than 120 days 0.01 (100 years) Table 1. Service Level Requirements

These target Service Level Requirements could be directly transferable to the Studies different natural hazard return periods.

METHOD OF RESILIENCE ANALYSIS How the Study undertook its analysis was done in several pieces of work. Whilst it, as shown, considered a number asset classes, the majority of the focus of the Study became on the expected performance of unsupported slopes. This was because demonstrably the greatest disruption to the transport corridor as a result of the Kaikoura earthquake has been from large scale land sliding simply due to the number and size of slope failures that occurred.

Review Discussions with NCTIR technical leads for the various assets in regard to their predictions of asset performance at the specified dates were undertaken. A literature review was carried out reviewing historical records of slope failure within the subject section of the transport corridor and of existing papers in regard to observed asset performance following earthquakes. For slope instability for example, Nomura et al (2014) indicate that Japanese practice is to scale down the criterion for rainfall induced land sliding by 20 to 50% to allow for seismic disturbance of slopes.

Assessment For unsupported slopes, assessment of the likelihood of future slope failure and the impacts on the transport corridor. This has involved:

• Statistical assessment to define the probability of a landslide of a given area and volume sourcing from different combinations of slope angle and material type, at given levels of earthquake shaking and rainfall.

• Geomorphological assessment of landforms susceptible to slope instability. This involved a desk-top interpretation of landslide and other slope instability features apparent in NCTIR’s GIS (Geographic Information System) database of ground hazards and event information and included utilising of interpretation of landforms apparent in LiDAR (Light, Image, Detecting and Ranging) survey.

• Review of NCTIR’s database of slope hazards and events to understand the geotechnical conditions at each slip site and proposed remedial works.

• Discussions with NCTIR earthworks/operations staff about the influencing factors and variability in the recovery rates for clearing landslides – ground-truthing the above findings where possible.

Production of ‘heat maps’ for the subject section of the transport corridor then took place. In addition to an assessment of risk level, the heat maps reflected NCTIR’s expectations of the level of resilience for each asset, and how this is expected to change over time.

Figure 9: An excerpt of the MNL Heat Maps – 3 different timeframes are compared against specific assets in a linear fashion. “Hotter” colors represent longer outage periods.

Figure 10: Slope Specific Resilience Heat Map, plus LiDAR of the slopes with vegetation removed

Interpretation Based on the data assessment, evaluation of the probabilities of annualized outage of the transport corridor under the various hazard scenarios over the assessed timeframes was undertaken.

OUTCOMES OF RESILIENCE STUDY The area of focus was shown by the Study to have increased levels of susceptibility for the hill slopes along the coastal transport corridor to subsequent failure from ground shaking or rainfall. In this area a ground shaking intensity of MMVII after the earthquake can be assumed as the triggering threshold for large-scale coseismic land sliding of the earthquake-damaged slopes.

The number of events caused by significant slope instability is now expected to increase for an extended period of time. This is in the order of a 50% to 75% increase per year over the next 5 to 10 years, mainly due to a reduction of rainfall thresholds (4).

The Study sought to identify geological “known unknowns” that may be subject to such instability. As noted in the “Asset Types Consider; Embankments” section, two of these had already happened in the months following the earthquake. A total of nearly 4 dozen sites within the Study area were identified as having landform features consistent with such large-scale slope instability.

The study also noted that tunnels and bridges are likely to be relatively ‘binary’ in nature. In other words, these assets are expected to have a reasonable level of robustness under all but very large, low frequency events. In contrast, slopes are expected to be non-binary in behavior, with failure of some slopes and initiation of debris flows likely to occur under relatively small, high frequency rainfall events. The storm events likely to significantly damage seawalls and other structures are also equally or more likely to detrimentally effect slope performance (5). Total resilience of the transport corridor is therefore likely to be more sensitive to the hazard posed by unsupported slopes than to any other asset. Unsupported slopes are expected to fail more frequently compared to other assets and there are more of them.

CREATING RESLIENCE As noted the earthquake resulted in a significant decrease in the resilience of the MNL. Hence if repairs of the damage found was the only works done, then that resilience would still remain less than pre- earthquake.

The next steps to be taken were understanding which “levers” could be pulled to help improve the MNL’s resilience. Of the three factors noted in the “Resilience – What is it?” section, robustness, and response were targeted. This was because, given that the MNL is a single track point to point railway, it has very little redundancy. Indeed the only realistic redundancy as a transport corridor is the use of other transport means such as coastal shipping or road haulage. The immediate aftermath of the Kaikoura Earthquake demonstrated this.

Robustness In the initial recovery works following the earthquake, immediate decisions were made to create slope stability related assets below or on failure sites and in some instances to realign the transport corridor away from several large slope failures. The Resilience Study identified these same slope failures as parts of the near four dozen potential large-scale slope instability sites. Additionally, the form of assets being built in these locations was reviewed given the findings of the Study.

Figure 11: Realigning rail (and road) away from a large slip by Blue Duck Stream, north of Kaikoura

Some of the slope stability mitigation assets used to date by NCTIR on the MNL are: • Sluicing, scaling to remove unstable material. • Earthworks to re-profile slopes and create a create level of slope stability. • Catch ditches, bunds and engineered catch fences at downslope locations to trap material before it reaches the rail. • Debris barriers and catch fences on upslope locations to catch material in an upslope location. • Horizontal drainage drilled into slope faces to lower the ground saturation levels. • Earthworks to create benches on slope faces for improving slope face angle and also creating areas of catch for future instabilities above. • Rock bolting and rock fall mesh of slope faces to secure specific rock faces or areas of looser rock material. • Rock fall drapes and attenuator fences (open mouthed drapes) to slope downward momentum of material that may come loose off a slope face.

For other potential large-scale slope instabilities, the Study helped further the thinking on future works needed. The work that resulted from this thinking included: • Construction of numerous additional slope stability assets, as per the above list. • Several additional realignments of rail and road alignments away from slips thus creating more run out room for future large scale slope failures. • Construction of several rock fall shelters at the ends of rail tunnels. Two such rock fall shelters have been built to date while the design of two more is, at the time of writing, underway.

Figure 12: Construction of rock fall shelter at the south end of Tunnel 14, south of Kaikoura

• Construction of debris flow basins and bridges to channel large amounts of debris under the MNL and SH1. This debris flows were expected to result from several huge slip faces that, while having not carried material onto the transport corridor in the main earthquake, had shown the potential to discharge tens of thousands of cubic feet (cubic meters) of material in future large rainstorm events.

Figure 13: Debris flow bridges, circled, to carry material from the large slip site above

The lack of resilience issue that the above work is trying to treat is illustrated by a site known as “Jacobs Ladder”. The Resilience Study identified this as one of the potential issue sites, recommended additional works and scoping of a solution was commenced in late 2017. In February 2018, post both road and rail corridors been reopened, the Kaikoura Coast was hit by Tropical Cyclone Gita. Such was the intensity of the event that it produced the second highest recorded rainfall quantity in areas of the Kaikoura Coast’s. Figure 13 compares the site post earthquake in April 2017 with the debris flow that then resulted in Cyclone Gita. The amount of debris that came down from the hills totaled several hundred thousand cubic feet (cubic meters). The MNL was shut for 13 days while this debris was cleared and the rail track and embankment rebuilt. Work is currently underway building a new 4.3ydx4.3yd (4mx4m) box culvert under both rail and road to carry debris from future events and a large debris basin with bunds surrounding it to contain the likely excess of material.

While a large amount of the robustness works have now been completed on the MNL, the true test of them will be in the coming decades. However initial signs are that the resilience works completed to date have had a measurable and positive effect. In the 21 months that the MNL has been running trains again since September 2017, the number of days that slip events or debris flows have affected the MNL has progressively reduced. While KiwiRail continues to use operational controls based on weather forecasts and actual measured rain events, as detailed further in the following section “Response”, the tolerances of those interventions have been able to be increased several times with the successful completing of robustness works.

Figure 14: Jacobs Ladder Debris flow – April 2017 and February 2018

Response While the initial focus of NCTIR has been on creating robust infrastructure, KiwiRail realized that the operational aspect of resilience would be very important – that of response. KiwiRail has therefore invested significantly in such response focused aspects.

Operation and Maintenance Because a large amount of the slope assets built by NCTIR have been funded by both NZTA and KiwiRail, those two agencies are in the process of developing a joint rail and road operation and maintenance agreement. This is a first for New Zealand road and rail infrastructure. By utilizing the resources that both transport providers are able to bring, it will greatly increase their responsiveness.

Slope Monitoring KiwiRail, with NCTIR, developed a scale increase in the use of remote monitoring on the Kaikoura Coast. Trip wire fences and tilt meters to detect ground movement and rock falls were the most common items used. These provide early warning of either issues developing or failures having occurred.

Weather forecasting & Operational Controls KiwiRail and NCTIR have developed complex and comprehensive weather forecasting methods linked to Trigger Action Response Plans (TARPs). From these, and in increasing order of weather severity forecast, KiwiRail puts operational staff and geo-specialists on call, instigates additional trackside and helicopter inspections, reduces train transit speeds through high risk areas and, for the worst of forecasts, stops train running till the risk period has passed. An important element of this process is that the TARP covers the period of antecedent rainfall – the days after a storm when slopes are at greater risk of instability.

Operational Controls vs Robustness Solutions As noted in the Robustness section, as physical works have been progressively completed on site and once monitoring to determine the success of their performance, some of these operational response controls have been able to be amended. Generally this has involved making intervention levels less conservative around weather events and removing remote monitoring devices where physical works have removed the risk being monitored. However the same is allowed for in reverse – if slope behaviors appear to presenting new risks, the application of such operational controls is now available for KiwiRail to utilize.

CONCLUSION The level of rail infrastructure damage that the Kaikoura Earthquake caused was unprecedented. The speed of recovery in rebuilding the line has been likewise so. The Resilience Study that NCTIR undertook has ensured that ongoing and long term asset and corridor resilience has been taken into account for defined levels of service. The Study helped shape large aspects of what has and is being built on the MNL. The success of the works completed in delivering the desired level of service pay a debt to that thinking around resilience. It has also helped KiwiRail consider how it plans for response to future events too. Finally, it has reminded us that the Kaikoura Coast will remain a challenging area for both rail infrastructure and operations in the foreseeable future.

REFERENCES

1. Davies AJ, Sadashiva V, Aghababaei M, Barnhill D, Costello SB, Fanslow B, Headifen D, Hughes MW, Kotze R, Mackie J, Ranjitkar P, Thompson J, Troitino DR, Wilson TM, Woods S and Wotherspoon LM. (2017) Transport infrastructure performance and management in the South Island of New Zealand, during the first 100 days following the 2016 Mw 7.8 “Kaikōura” Earthquake 2. Mason D, Justice R, McMorran T (2018) Assessment of the Resilience of the North Canterbury Transport Network following the 2016 Kaikoura Earthquake. IPWEA Conference 2018 3. 3. Mason, D & Brabhaharan, P (2016) Resilience of State Highways: Recommended Regional Assessment Methodology for Low Frequency Hazard Exposure. NZTA September 2016 4. 4. Nomura, A, Okamoto, A, Kuramoto, K and Ikeda, H. (2014) Landslide-triggering Rainfall Thresholds after Major Earthquakes for Early Warning. International Journal of Erosion Control Engineering Vol. 7, No. 2, 2014 5. Justice R, Saul G, Mason D (2018) Kaikoura Earthquake Slope Hazards – Risk Mitigation and Network Resilience. NZ Geomechanics News, December 2018 Issue 96

FIGURES Figure 1: Northern South Island with its major geological faults identified. Figure 2: Key Damage Location to the MNL Figure 3: Rail pushed into the sea by a slip at Half Moon Bay’, north of Kaikoura Figure 4: The destroyed Tirohanga Bridge. Figure 5. Tunnel 18 fault damage from scan survey – Elevation Figure 6. Tunnel 18 fault damage from scan survey - Plan Figure 7. Tunnel 18 fault damages and repair works Figure 8. Network Resilience (3) Figure 9. An excerpt of the MNL Heat Maps Figure 10: Slope Specific Resilience Heat Map, plus LiDAR of the slopes with vegetation removed Figure 11: Realigning rail (and road) away from a large slip by Blue Duck Stream, north of Kaikoura Figure 12: Construction of rock fall shelter at the south end of Tunnel 14, south of Kaikoura Figure 13: Debris flow bridges Figure 14: Jacobs Ladder Debris flow – April 2017 and February 2018

TABLES Table 1. Service Level Requirements

Main North Line Resilence – Kaikoura Earthquake

23 September 2019 2 What Caused The Earthquake?

3 Closer Up…

4 Peak Ground Accelerations

5 Damage – Where It Happened

6 Damage To Slopes

7 Damage To Bridges

8 Damage To Tunnels

9 What Is Resilience?

• Using our Resilience Study

Resilience = Robustness + Redundancy + Response 10 State of Asset Knowledge pre EQ?

11 Key Asset Knowledge Pre EQ

70% of the highest rated slopes between the two worst damaged areas by major slips SLOPE RATING

INDIVIDUAL SLOPES ON MNL

12 Safety Considerations

• KiwiRail set client safety requirements for train running. • Assets such as bridges more binary – they either meet “code” or not. • Assets such as slopes require a greater degree of engineering judgement and utilise comparative scoring systems. • KiwiRail made use of So Far as is Reasonably Practical for its Assurance Process for defining “safety” opening of the MNL. • A key feature of this is operational rail controls used in addition to engineering ones – rainfall forecasting, remote monitoring, speed restrictions.

13 Reliability Considerations

• A weakened state around assets, particularly slopes, will exist for some years. • Compared to pre earthquake, outages from storm events are going to be higher. • Likewise heightened seismic activity will likely cause additional outages.

14 But…Is It Resilient?

NCTIR undertook a study in mid July 2017 to answer this. There were three time horizons considered: • August target for re-opening for commissioning Trains • December 2017 target for SH1 re-opening • End of NCTIR and Reinstatement Works

In broad terms: • Resilience improves as time moves on and permanent repairs are completed • Risk management controls increasingly moved to engineering controls rather than the high levels of operational controls that were used initially • There will be outages when bad weather is forecast and some clean up work may be required after an event - but the length of outage is important!

15 Heat Maps

16 Time Based Lidar Analysis

17 Building Robust Infrastructure

18 Building Robust Infrastructure

19 Response Planning Is Critical

20 Conclusions

• Knowing your assets • Knowing your network • Limits to how “robust” you can be • Planning for response

21 Questions?

22