Everything shaded on the map is council or GNS Science data, unchanged. The walking routes, the aiming points and the barriers are mine, and they are the parts most likely to be wrong.
Kāpiti's three tsunami evacuation zones were defined by GNS Science using COMCOT, a hydrodynamic model that simulates a wave from its source across the ocean and onto the land. The council publishes them as spatial data, and this page draws that data directly rather than tracing it.
One thing worth knowing, because it is not obvious from the map: the three zones are adjacent bands, not nested rings. Red sits against the shore, orange behind it, yellow furthest inland, and they abut rather than containing one another. Being outside the yellow polygon is not the same as being out of danger unless you have also cleared the other two — which is why this tool routes to the far edge of all three.
| Zone | Modelled from | Recorded height |
|---|---|---|
| Red — Shore Exclusion | 22 distant-source scenarios at the 1 m threat level, checked against the national rule of land under 2 m above high tide | up to 1.2 m wave; runup ≈ 2.5 m above MHWS |
| Orange — CDEM Evacuation | 29 scenarios at the 3 m threat level and 20 at 5 m, all distant sources | up to 5.0 m wave height |
| Yellow — Self Evacuation | 50 non-uniform slip scenarios on the Hikurangi subduction interface, 8 on the Fisherman Fault, 1 on the Manaota Fault | up to 10.0 m wave height |
National guidance asks the yellow zone to cover at least the inundation expected once in 2,500 years, at 84% confidence. GNS reported that where the modelled scenarios overlap, the return period could be well beyond 7,300 years. For comparison they estimated the Wellington Harbour yellow zone at around 6,000 years.
So it is a rare event. It is also the one with no warning: the shaking is the warning, and the council puts a local-source arrival at as little as three minutes.
GNS computed maximum flow depths for every modelled cell. Only the inundation extents were released. That means neither residents nor the council can tell from the published data how deep the water would be at any particular place inside a zone — which is why this tool routes to the edge of the zone rather than to a chosen elevation, and why no height on this site is presented as a safe height.
The pink and purple shading on the map is how high the land is above sea level. It is not how high a wave would be. The two are easy to mix up here, because the same numbers turn up on both sides of the question: the yellow zone is modelled on waves of up to 10 m, and the ground shading also starts at 10 m. One is the danger and the other is the way out of it.
The shading comes from the LINZ 1 m LiDAR elevation model, not from any tsunami modelling, and it is drawn light for lower ground and deep for higher ground. It is there to show where the land rises, not to mark anywhere as safe — as the section above explains, the published data does not say how deep the water would be at any given place, so no height on this site is presented as a safe height.
The faults near Kāpiti button under the map shows known active faults on land and close offshore, and all of New Zealand shows them across the country. The lines come from GNS Science's model of active faulting in New Zealand. They are drawn at a regional scale: good for seeing where the faults run, not for telling whether one runs under a particular house. For that, see the council's fault rupture hazard maps linked below.
Most of these faults move very slowly, a few millimetres a year, and have big quakes hundreds or thousands of years apart. What is known about the main ones nearby:
| Fault | Last big quake | How often, on average |
|---|---|---|
| Ohariu | about 1,000 years ago | every 1,300–3,800 years |
| Northern Ohariu | about 300–1,000 years ago | every 1,000–3,000 years |
| Ōtaki Forks | not known | every 3,500–5,000 years |
| Wellington (Wellington–Hutt Valley) | between about 1640 and 1780 | about a 1 in 10 chance in the next 100 years |
| Wairarapa | 1855, magnitude 8.2 | about every 1,200 years |
Most of the faults on land would shake Kāpiti hard, but they would not push up the sea. A tsunami needs the sea floor to move. For Kāpiti's yellow zone, GNS modelled two kinds of source:
On the map, the two faults GNS used, Fisherman and Mana–Otaheke, are drawn in solid blue. GNS also tested six other faults, drawn in dashed blue: Ohariu, Okupe, Onepoto, Mascarin, Rangitīkei Offshore and Wairau in Cook Strait. The waves they could make were smaller than the Hikurangi and Fisherman ones, so they did not change the zones. GNS also used the Manaota Fault. In the fault data this map draws from it is called the Mana–Otaheke Fault, and it runs offshore between the Fisherman Fault and the coast. It is very rare (about once in 21,000 years) but, like Fisherman, close. (Source: GNS Science Consultancy Report 2017/200, Table 3.1 and section 4.1.)
That is why the rule is the same whatever the fault: if an earthquake is long or strong, do not wait for a warning. Go.
Fault lines: Litchfield, N.J. et al. (2014). A model of active faulting in
New Zealand. New Zealand Journal of Geology and Geophysics 57(1): 32–56, via the
GEM Global Active Faults Database, CC BY-SA 4.0.
Ohariu, Northern Ohariu and Ōtaki Forks figures:
Kāpiti Coast District Council, earthquake fault rupture hazards
(from the GNS Earthquake Fault Trace Survey, 2003, updated 2007).
Wellington and Wairarapa figures: GNS Science, It's Our Fault programme;
Rhoades et al. (2011), Bulletin of the NZ Society for Earthquake Engineering.
Hikurangi: East Coast LAB.
Many people have heard that Kāpiti Island would block a tsunami. It would not, and the reason is something you may have seen in a school science class: the ripple tank on an overhead projector. When ripples meet a block in the water, they bend around its ends and close in again behind it. A tsunami does the same, only far more so.
So why do people believe it? Because the island really does calm everyday waves. That is why the coast bulges out at Paraparaumu and Raumati: sand builds up in the island's shelter. Everyday waves are short, so the island stops them. A tsunami is not an everyday wave.
Videos from GNS Science (now Earth Sciences New Zealand):
The council's own data records the method behind each zone. Kāpiti Coast District Council has one entry: COMCOT hydrodynamic modelling, 2017. Every other territorial authority in the Wellington region has a mixture, and most still include zones drawn with the 2008 attenuation rule — a geometric rule of thumb rather than a simulation. On this measure Kāpiti has the best-modelled evacuation zones in the region.
That provenance is only discoverable by querying the map service directly. It is not shown on the council's public map.
Zone boundaries: Mueller, C.; Power, W.L.; Wang, X. (2017).
Tsunami modelling and evacuation zones for Porirua and Kāpiti Coast. Lower Hutt
(NZ): GNS Science. GNS Science Consultancy Report 2017/200.
Elevation: LINZ Wellington–Kāpiti Coast 1 m LiDAR DEM (2024), CC BY 4.0. Addresses: LINZ
NZ Addresses. Walking routes: OpenStreetMap via the Valhalla routing engine. Basemap:
© OpenStreetMap contributors.
It works out a walking route from an address to the far edge of the zone, and then on to the nearest assembly area. It prefers hand-chosen aiming points where local knowledge beats geometry, and it refuses routes that head towards the water or cross a barrier.
It knows nothing about locked gates, private land, fences that went up last year, paths that exist on a map but not on the ground, watercourses in flood, or damage after an earthquake. Only someone who has walked it can tell the difference — which is the whole reason for the feedback form, and for walking your route before you need it.