This electronic supplement is divided into two main parts. The first part contains descriptions of the main attributes in the database with a brief description of what they are, why they have been collected, and issues encountered in importing them into a formal database structure. What is essential in all these entries is to preserve the information as it was contained in the original sources. This has been challenging at times, especially with earlier entries. The second part is an Appendix with the complete list of attributes contained in the database.
The New Zealand Tsunami Database (NZTD) is organized around New Zealand locations where tsunamis have impacted in the past. In the table view of the online database, observations are listed according to their location description (place name) and date. The latitude and longitude for each observation (for use in map view and listed on the overview tab of each entry) were estimated based on the locations discussed in the descriptive accounts, except for tide gauges for which the exact location is generally known. The accuracy of the coordinates depends on the amount of detail provided within the descriptive accounts as well as on the extent of the impacted area observed. Many accounts refer to generalized locations such as a settlement, beach, or bay. Coordinates were placed to approximately reflect the location at which the observations were most likely to have been made. In situations in which the location referred to a wide area, a default location was designated. For example, observations made in Auckland with no further specific locational information have a default location of Mechanics Bay. Default locations are placed at a central point within the area referred to, typically near the coast where observations are commonly made, such as ports. Because of the variability within the descriptive accounts and the necessity of defining a location as a single point, the coordinates provided are considered to be only indicative of the impacted location.
The date of wave arrivals should be a straightforward piece of information to collect. Indeed, it has been for most entries. However, earlier entries have large uncertainties. Because of this, some interpretations, such as using the context in which an observation was made were made to extract this information. The uncertainty in this attribute is part of preserving the original description of time (e.g., morning, evening, etc.). Some of these symbols or descriptions, as provided by eyewitnesses, are also subjective and might not be interpreted the same way by all. For example, how does one interpret “early morning”? Therefore, a variety of entries such as question marks, year ranges, or notes were used in the column that provided the date of the events. Currently, only the year appears in the table view of the online database, but the month and day can be accessed under the impact tabs.
Once users double-click on an entry of interest in the table view, they are taken to further details under six different tabs as outlined below.
Various summary location and date details are listed in the overview tab and, in addition, an indication of the likelihood that the entry refers to a tsunami is given. A validity ranking of 0–4 is assigned for each entry, with 4 being a definite tsunami and 0 being an erroneous tsunami report (Table 1 in the main article). Erroneous tsunami reports (validity 0) are included in the database to show that the records in question have been assessed and found not to be tsunamis, and future research may help in clarifying the origin of these events. Doubtful tsunamis (validity 1) are those that may have been described in early records as tsunamis but further analysis indicates that they are more likely to be storm surge or other processes. Possible tsunamis (validity 2) also have some uncertainty about their origin, which is usually discussed in the tsunami impact summary paragraphs in the database. Probable tsunamis (validity 3) have descriptive accounts and impacts, which are consistent with tsunamis, but there is no known source for them.
Another attribute that appears in the overview tab is “Tsunami Observed” with a Y for yes or an N for no next to it. If an N appears, then there may be an additional attribute called “Seiche Observed” with a Y for yes, next to it. There may also be an N next to Tsunami Observed if no eyewitness observation was made at that location. This attribute contained two columns in the original excel spreadsheet, with one column corresponding to tsunami observations and the other to meteotsunamis or seiches, hence the slightly nonintuitive presentation of this attribute in the online version (see Data and Resources). Occasionally, a question mark or other notes accompanied the entry, reflecting uncertainty in the corresponding record.
The impact measurements are those derived from sea-level or tide-gauge records, so the nineteenth-century entries do not have information under this tab. Attributes include time and date of first arrival, time elapsed from generation, maximum effective peak–trough wave height, first motion (up or down or withdrawal or uplift of water), duration of noticeable surges, availability of tide records whether digital or analog, etc. This tab is currently dominated by listing of different time attributes; because of the column format of the Excel database, year, month, and time currently appear on separate lines, but this presentation will be stream-lined in future. Time is a very important feature of the database, and avoiding errors in times of observations is crucial to avoid confusion in referring to wave arrival times and event occurrence and also to distinguish tsunami events from other events, especially those that occurred simultaneously or close to each other chronologically. Times are provided in New Zealand Mean Time (NZMT) or New Zealand Standard Time (NZST) as well as UTC. NZMT existed between 1868 and 1941 and was 11.5 hrs ahead of Greenwich mean time (GMT). Since 1941, New Zealand has operated under NZST that is 12 hrs ahead of GMT (and 12 hrs ahead of UTC). During summer (late September to early April), it is 13 hrs ahead and called New Zealand Daylight Time (NZDT). The Chatham Islands have been set 45 min ahead of New Zealand since 1868. The Chatham Islands’ time zone is Chatham Standard Time (CHAST) or Chatham Daylight Time (CHADT) during summer.
This tab contains some similar attributes to the Impact Measurements tab, but these are based on observations and descriptive accounts. A number of different measures of tsunami size are listed as potential attributes (Appendix) to cover the options of what information was available and to enable the best use of descriptive accounts, for example, maximum observed effective peak–trough, maximum water elevation at or close to shore, maximum height reached at the limit of inundation relative to expected sea level at the time, and so on, but for many events only a few of these size descriptors are populated.
For each location impacted by an event, a grade indicating tsunami intensity was assigned based on the New Integrated Tsunami Scale (ITIS-2012) proposed by Lekkas et al. (2013). ITIS-2012 is divided into 12 grades, ranging from I (not felt) to XII (completely devastating). The six categories of criteria on which the scale is based are physical quantities, such as tsunami wave height, flow depth, and inundation extent; impact on humans; impact on mobile objects, such as boats and cars; impacts on infrastructure; environmental effects; and impact on structures. The grades were assigned based on the descriptive accounts and measured quantities where applicable (see Fig. 2, Table 1 in the main article).
Using the tsunami intensity scale has its own challenges, due to the varying amount of observations and quality of information available for historical events, compared to the large amount of data available for modern events. In many cases, grades were assigned based on limited information that did not correlate perfectly with the category descriptions. The assignments were intended to keep consistency between similar intensities within the database. The tsunami height or flow depth was found to have a very poor association with the other listed parameters across the various categories, typically being below the observed wave height corresponding to the impacts for a given grade. Therefore, in our assignment of grades, we downweighted the importance of height and flow depth over impacts.
The Impact Observations tab also includes an attribute that categorizes the observed water levels according to the threat levels developed by Power and Gale (2011). Threat levels for New Zealand are defined as listed in Table S1.
The primary attribute under this tab is the cause code. These codes identify what caused the tsunami or what likely combination of events caused the tsunami. Following this are attributes related to the time of occurrence, location, and magnitude of the source event (i.e., the event that caused the tsunami) which, for distantly triggered tsunamis, could all relate to events quite far away from New Zealand. Therefore, accurate collection of this information often relies on the international literature or websites of overseas institutions. For the larger and/or more widely reported source events, a full summary of the event is provided. The source class or proximity to nearest New Zealand coast is under this tab. As an operational tool for emergency management, this classification is based on the travel time of a tsunami to its first arrival at the New Zealand coast and consists of (a) local-source tsunami (<1 hr travel time to New Zealand coast, i.e., no possibility of getting a warning out to the public), (b) regional-source tsunami (1–3 hrs travel time; maybe time to issue a public warning), and (c) distant-source tsunami (>3 hrs travel time; probably time to issue a public warning).
In this tab, some of the primary historical source material is provided. This is usually abridged and referenced, but full transcripts of the source material, including nonprimary accounts, have been compiled and are available as separate files from the database custodian.
This tab contains references used in the compilation of event information, notes about the completeness of the data, and listings of the previous databases in which each event appeared.
The map view of the online database currently enables each tsunami impact entry to be displayed as a point on a map of New Zealand according to four different attributes: source type, intensity, validity, and year of impact. Additionaly, an overlay shows the number of impacts at any one location and the impact locations for some of the most significant tsunamis in New Zealand’s history (Table 1 in the main article).
Event Identifier: Date of First Arrival/Observation New Zealand Time
Tsunami Validity
Location of Observation
Tsunami, Meteotsunami, or Seismic seiche?
TSUNAMI IMPACT DATA: Sea-Level/Tide Gauge Records
Time of first arrival (NZST)
NZST or NZMT (pre-1941); 24 hrs time
Universal Time (UTC); 24 hrs time
Time elapsed from generation (hrs)
Maximum effective peak–trough (m) [Maximum amplitude × 2]
Time of maximum effective peak–trough (NZST or NZMT [pre-1941])
NZST or NZMT (pre-1941); 24 hrs time
Universal Time (UTC); 24 hrs time
Time elapsed from generation (hrs)
Maximum effective zero-peak (m) [Maximum amplitude]
Time of maximum effective zero-peak (NZST)
NZST or NZMT (pre-1941); 24 hrs time
Universal Time (UTC); 24 hrs time
Time elapsed from generation (hrs)
Predominant periods (min)
First motion (u, up; d, down)
Duration of noticeable surges (hrs)
Tidal records available? Digital or graphic?
Reference or data source
Approximate expected travel time of first arrival (to nearest 0.3–0.5 hrs, calculated using International Tsunami Database 2004, unless specified)
TSUNAMI IMPACT DATA: Observations/Descriptive Accounts
Expected time of first arrival
NZST or NZMT (pre-1941); 24 hrs time
Time of first arrival/observation
NZST or NZMT (pre-1941); 24 hrs time
Universal Time (UTC); 24 hrs time
Time elapsed from generation
Comments
Maximum observed effective peak–trough (m) [Maximum amplitude × 2]
Time of maximum effective peak–trough or vice versa
NZST or NZMT (pre-1941); 24 hrs time
Universal Time (UTC); 24 hrs time
Maximum water elevation at or close to the shore (m), that is, zero to peak
Time of maximum water elevation at or close to the shore (m), that is, zero to peak
NZST or NZMT (pre-1941); 24 hrs time
Universal Time (UTC); 24 hrs time
Maximum height reached at the limit of inundation relative to expected sea level at time (m)
Time of maximum height reached at the limit of inundation relative to sea level at time (m)
NZST or NZMT (pre-1941); 24 hrs time
Universal Time (UTC); 24 hrs time
PROXY (maximum if range) for water height at maximum inundation relative to sea level, etc.
PROXY (minimum if range given) for water height at maximum inundation relative to sea level, etc. Type as in the previous column
Measurement type
Threat level as defined by Power et al. for New Zealand
Maximum onshore height reached relative to sea level at the time, that is, peak level between shore and limit of inundation (probably only known in modern postevent surveys) (m)
Maximum onshore flow depth (m)
Maximum horizontal extent of inundation measured from high water (m)
Maximum observed height of inundation above mean sea level, that is, maximum impact, not necessarily largest waves (sometimes the only parameter known in historical events, when tide level at the time not known)
Time of maximum observed height of inundation above mean sea level, that is, maximum impact (not necessarily largest waves)
NZST or NZMT (pre-1941); 24 hrs time
Universal Time (UTC); 24 hrs time
Distance traveled up stream/river (km)
Maximum height of bore in river/stream
Number of significant (>1 m) waves
Predominant period (min)
First motion (u, up; d, down)
Duration of noticeable surges (hrs)
RRF, rapid rise and fall; BW, breaking wave; BC, bore/wall of smooth water on the coast; BS, bore in coastal stream or river
Number of fatalities; Injuries
Damage: Built; marine; environmental
Photographs of tsunami effects, inundation maps available? Reference?
References
SOURCE PARAMETERS
Cause code
Cause code comment
Origin time of source event
Universal Time (UTC); 24 hrs time
NZST or NZMT (pre-1941) pre-1868, local longitude time was used, here converted to NZMT
Source location, including landslide or volcanic source if known
Source location reference
Earthquake source local magnitude ML
Earthquake source ML reference
Earthquake source surface-wave magnitude Ms
Earthquake source Ms reference
Earthquake source moment magnitude Mw
Earthquake source Mw reference
Earthquake source mantle magnitude Mm
Earthquake source Mm reference
Source tsunami magnitude Mt
Source Mt reference
Nonspecific earthquake source magnitude M
Earthquake source M reference
Source class/proximity to nearest New Zealand coast, including Chatham Islands
Source event summary of distant and regional source tsunamis, and larger local earthquakes/tsunamis
PRIMARY DESCRIPTIVE ACCOUNTS/SOURCE MATERIAL (ABRIDGED AND REFERENCED)
Discussion of observational data and other relevant data
Tsunami impact/sea level disturbance in New Zealand summary
References
Completeness of data
Previous database listing?
Table S1. Threat levels of tsunamis in the New Zealand Tsunami Database.
All data used and attributes defined in this electronic supplement came from published sources listed in the references and from the New Zealand Tsunami Database (NZTD; http://data.gns.cri.nz/tsunami/, last accessed December 2016).
Lekkas, E. L., E. Andreadakis, I. Kostaki, and E. Kapourani (2013). A proposal for a new integrated tsunami intensity scale (ITIS-2012), Bull. Seismol. Soc. Am. 103, no. 2B, 1493–1502.
Power, W., and N. Gale (2011). Tsunami forecasting and monitoring in New Zealand, Pure Appl. Geophys. 168, 1125–1136.
[ Back ]