LIQUEFACATION IN THE BULLER REGION IN THE 1929 AND 1968 EARTHQUAKES

This article describes the results of a search for sites of liquefaction in the 1929 M=7.6 Murchison and the 1968, M=7.l Inangahua earthquakes in the Buller region of New Zealand. Evidence of liquefaction was found for nine sites in each earthquake; two sites were common to both events. Widespread ejection of sand occurred in the epicentral regions of both earthquakes. Liquefaction was more sporadic at larger epicentral distances, except in the North Beach area of Westport, the most distant 1968 site, where sand boils occurred over several hectares, together with the toppling of utility poles and some lateral mass movement. This area corresponded to the very young beach sands deposited since extension of the Westport harbour breakwaters towards the end of last century. Liquefaction was much less common in the river-lain sands in the same vicinity.


INTRODUCTION
A general study of historical cases of seismic liquefaction in New Zealand (Fairless, 1984;Fairless and Berrill, 1984) suggests that liquefaction has been common in large New Zealand earthquakes, and that it occurred quite widely in the 1968, M 7.1 Inangahua earthquake.Prompted by the number of cases found in the literature by Fairless, a more intensive search was mounted for occurrences of liquefaction in the Buller region, the scene of the 1968 earthquake.Several new cases were found, together with additional information about previously known ones.
As well, a number of cases from the 1929, M = 7.6 Murchison earthquake were discovered, some sites having liquefied in both earthquakes.
Further, two sites near Westport possibly liquefied during the 1962 Westport earthquake sequence (Adams and Le Fort, 1963) but we wish to obtain corroboration of this before publishing details.
The aim of this article is to report the evidence of past liquefaction found in the Buller region and to pin-point sites for further investigation, and for inspection following future earthquakes.Sources of information include reports in the literature and press interviews with eyewitnesses and aerial photographs taken a few days after the Inangahua earthquake.Even in 1968, sand liquefaction did not attract the interest that it does today and most references to liquefaction in the literature are cursory, with few details or specific locations given.
For 1929 no mention of liquefaction effects such as sand ejection or movement of near-flat slopes could be found in the technical literature.
Figure l shows isoseismal patterns for the earthquakes.
Both were shallow focus events with surface faulting, predominantly in a thrust mode.
Brief seismological descriptions of the earthquakes are given in the appendix and further geological and seismological references may be found from the bibliography.
It is of interest to note that most of the Inangahua aftershocks lie within the MM IX isoseismal, indicating this was the region of rupture, with rupture propagating southwestwards from a northern hypocentre.

General Remarks
Figure 2 shows the sites of liquefaction that have been positively identified.These sites are listed in

Newspaper report
The main evidence for liquefaction in all cases was the ejection of sand.At one site, Kilkenny Park, Westport, this was accompanied by large-scale mass movement and by the toppling of electric power poles.
By far the most extensive occurrences of liquefaction were found in the flood plains of the Buller River at Three Channel Flat, about 3km northeast of Inangahua Junction, and at Walker's Flat, about 5km west of Inangahua Junction.
Several hundreds of sand boils show up on aerial photographs taken on May 29 and 30 and in terrestrial photographs taken by various observers.In both places, the ejection of sand was quite pervasive.Figure 3 gives the flight plan of the post-earthquake aerial photography flown by New Zealand Aerial Mapping Ltd.

The pilot was
Mr Cyril Whittaker who confirms that runs A and B shown in the plan were not flown, due to poor weather.
Note that at least two sites, Three Channel Flat and Keoghan's Farm liquefied in more than one earthquake.Furthermore, one site near Meddows' Farm, Walkers Flat, liquefied in an aftershock of the 1968 earthquake as well as in the main shock.
Three unconfirmed sites are also marked in Figure 2.
For these there is some indefinite evidence for liquefaction in 1968 in the form of vaque word-of-mouth reports or apparently ejected sand in aerial photographs that could not be confirmed on the ground.
In the following paragraphs each site will be discussed in more detail.

Three Channel Flat
Extensive ejection of sand was observed at Three Channel Flat following the 1968 earthquake.The flat comprises a system of low terraces of Recent alluvium.The photograph shown in Figure 4 was taken on the day after the earthquake.The present owner of the land, Mr Warren Inwood, also remembers seeing the ejected sand, although he did not own the land himself at the time.The extent of sand ejection is clear from Figure 5, a vertical aerial photograph taken 5 days after the earthquake.
The flat comprises a system of low terraces of Recent alluvium.
Extensive hand auger boring at the south end of the flat shows a layer of silt 3 to 4m thick overlying fine sands of variable thickness over gravels (Ooi, 1987;Adlam, 1988).Thickness of the sand layer increases towards the outside of the abandoned south channel evident in Figure 5, and decreases towards the inside of the channel, marked by the absence of sand boils.It appears that where there is sand, it was ejected in 1968.
The paddocks to the north of the central vehicle track have been extensively regraded since 1968 and their contours changed markedly.But to the south of the track the terrain is essentially the same, although some fences have been moved.The two paddocks at the south end of the flat (in the foreground of Figure 4) have been ploughed two or three times since 1968 yet 177 in them it is still possible to pick up lenses of fine grey sand, presumably the ejecta, starting at depths of up to 200mm below the ground surface.After ploughing, while the soil is still moist, Mr Inwood reports that the patches of grey sand stand out markedly from the darker silts.The sand possibly owes its being buried to the ploughing, but possibly also to a build up of silt since 1968, as the paddocks are flooded about once a year.
Of the various sites of liquefaction in 1968, Three Channel Flat was the most extensively studied at the time.Sutherland (1970) notes that patches of ejected sand occurred either in a linear arrangement •with a fissure linking the flows or in a random pattern with no linking fissures.Presumably the formation of the linear fissure with sand ejection indicated some lateral mass movement, while the unconnected random pattern of boils, which he notes were found at Three Channel Flat and Walkers Flat, mark areas of complete liquefaction on virtually level ground.Dodd and Dunlop (Dodd, 1970;Dunlop, 1968) investigated sand boils at the south west of the flats with auger.borings and sieve analysis.
The log of their boring, starting from the bottom of a recently excavated trench, is shown in Figure 6; grain size distributions of the ejected and subsurface sand strata are shown in Figure 7.
Note the similarity of the sand gradings.Presumably the ejected sand came from the lower layer, below the water table.
Liquefaction also occurred at Three Channel Flat in the 1929 Murchison earthquake according to an article in the June 19, 1929 issue of the Wellington "Evening Post", which reports "Paddocks at Three Channel Flat are covered with white sand".

Sand Ejection near Inangahua Township
The confluence of the Buller and Inangahua rivers near the Inangahua township is marked by large areas of young alluvial river terraces.There are several contemporary reports (Lensen and Suggate, 1968;Adams et al., 1968, Sutherland, 1970;Dodd, 1970) of sand boils on these terraces in 1968; and patches of sand show up well on the aerial photographs.Figure 8, drawn up from the photographs, shows the areas in which ejected sand is apparent.Sutherland (1970) reports a witness's account of water gushing to a height of 3m in the fields immediately to the north of Inangahua and that the "gusher" was active for at least an hour and a half after the earthquake.

Walkers Flat
Aerial photographs from runs c and D of the 1968 post earthquake photography show ejected sand over a large part of Walkers Flat. Figure 8 shows the extent of the 1968 sand boils.1.65mThe eyewitness accounts and photographs taken near Mr G. Meddows' farmhouse at the east end of the flat are of particular interest.The photographs in Figure 9 show sand boils in the paddock immediately to the north of the farmhouse.Mr Meddows remembers sand and water being ejected from vents in the vicinity of the farmhouse and cowshed during some aftershocks as well as the main shock.But he did not notice any sand ejection in the area following the 1962 Westport earthquake sequence.

Keoghan Farm, Sergeants Hill
Mr Noel Keoghan's farm is situated on the banks of the Orowaiti Ri~er, south east of Sergeants Hill.
His house is marked on Figure 10.Mr Keoghan observed sand boils in the paddocks to the north of his house following the 1929 earthquake and again in 1968.
From 1929, when he was a boy of 11, Mr Keoghan remembers geysers of sand and water 5 to 6 feet high, leaving sand cones 2 to 3 feet high.
In 1968, smaller cones were formed, about 6 inches high, in the same general area.
Small sand boils were also observed by Mr Keoghan in the bed of the orowaiti River immediately to the south of his house.
Mr Keoghan's daughter, Mrs Wayne Meddows of Whitecliffs, also observed the•l968 boils in the Orowaiti River, known locally as Giles creek.

Mr
Keoghan also observed premonitory behaviour amongst his livestock immediately before the 1968 earthquake.
At the time, he was bringing in his cows for their morning milking.
It was dark and he was passing a shed housing pigs.
First the pigs started to squeal; about 3 seconds later the cows stampeded; then Mr Keoghan was thrown to the ground by the earthquake.Following the earthquake there was a strong smell of sulphur and the sound of many landslides in the surrounding hills.Also, cracks opened in the ground passing through the Gaynors' house at 14 Shelswell Street and others on the south side of Shelswell Street.
The cracks were roughly parallel to Shelswell St, (and to the nearby beach) and were of the order of a few hundred millimetres wide.
Mr Gaynor, employed by the Westport Borough Council at the time, remembers having to add 1.2m of pipe to reconnect the broken water main running approximately north-south along Derby St, near Kilkenny Park.Taken together with the cracking, the broken water main suggests that a large mass of surficial soil had moved towards the coast on a layer of liquefied sand.To the west of Patterson Park at lOa Derby St, sand erupted under a house with sufficient force to lift .. it, rotate it and leave it skewed on its foundations.
(The house was subsequently demolished. The position of the 1968 boil and house was in front of the present house at number lOa).
Another sand boil was seen by Mr James Fischer to erupt in the grass verge in front of the house at 2 Salisbury Street, in 1968.A less definite report also had water coming out of the ground somewhere near 23 Romily Street, in the same general area.
Figure 12 shows tilted electric power poles on the west side of Derby st adjacent to Kilkenny Park, also pointing to liquefaction at a shallow depth.
No reports of liquefaction effects were found for sites to the south of this area in the town of Westport in 1968.This may be attributed to a greater susceptibility to liquefaction of the dune sands found to the north of Orowiti Road compared with the alluvial deposits found to the south.(Suggate and Wood, 1979) report that the boundary between dune sands and river-born sediments lies roughly along Orowaiti Road).It should also be noted that the dune sands from about Kilkenny Park north have been deposited very recently, following extensions to the harbour breakwaters in the 1890's.
~orby Estate, Sed~q_nvp.1_~1-1-~2-~Mrs Corby remembers seeing sand boils on the Mohikinui River terraces between the Corby homestead and the Seddonville road.She could not locate the boils pre-cisely, after 57 years, but they were in the area shown in Figure 13.

Little Wanganui area, 1929
Mr and Mrs R. Duncan of Little Wanganui can both remember seeing sand boils following the 1929 earthquake, at the approximate locations shown in Figure 14.

Karamea area, 1929
Sand boils were observed at several places around the Karamea township and further up the Karamea River valley.Also, in numerous places, water and sand was exuded from cracks in the ground, especially in road and railway embankments, indicating liquefaction-driven spreading phenomena.

Karamea School
The most frequently commented-on site was the Karamea School and the adjacent Domain sports field, shown in Figure 15.The 1929 earthquake occurred during school hours and many of the elder present-day residents of the district were then at school.Three eyewitnesses independently referred us to a site immediately in front of the old school building, where sand and some gravel was ejected.
An auger boring made at the school in May 1987 (Bienvenu, 1988) found a gravelly coarse sand, answering the general description of the ejecta, from a depth of about 2 to 3m and overlain by silts and clays.The water table was at a depth of 2.2m in May 1987.
Sand boils and the ejection of sand from cracks were also reported elsewhere in the school grounds and in the paddock across the road to the north of the school.

Anderson's Farm, Arapito, 1929
Mr Cyril Lineham of Arapito, then a school boy at the Arapito school, remembers seeing sand and water together with some gravel ejected in the land to the south of the present road at Arapito at the location labelled "Anderson's Farm" (after the present owner) in Figure 15.Auger borings at the site indicated by Mr Lineham found coarse sand, overlain by sandy silts from a depth of about 2m.
The coarse sand contained some small pebbles, from 15 to 30mm in size, consistent with Mr Lineham's description of the ejected material.

Kongahu Mudflats, Karamea, 1929
Although he was away from Karamea at the time, Mr Karl Jones clearly remembers his grandfather describing sand boils erupting around him in the estuary between Kongahu and Maori Point, where he was fishing at the time.The rough location is shown in Figure 15.

Near Karamea Golf Course, 1929
As well as remembering sand ejection in the school grounds, Mr Johnson also recalls a line of fence posts toppled just to the west of the pond between the aerodrome and the coast, immediately south of the present golf course.
The approximate location is shown in Figure 15.

Grigg•s Gravel Cone
A booklet by Grigg (1947) shows a photograph of a heap of gravel some 2 m high, which is claimed to have been produced by the 1929 earthquake.However a marginal note in a document shown to us by Mr and Mrs Monahan states that the gravel heap existed before the earthquake, and was presumably a stockpile of roading material.

Greymouth, 1929
The "Evening Post" of June 19, 1929 reports "At the south of the Grey Lagoon mud and sand were exuded during the big shock, and formed little hillocks on grass-covered areas.The fluid is straw-coloured, while the sand is of a different kind to that of the vicinity.One vent forced out a fence post."This clearly indicates the ejection of sand from well below the ground surface.

Unconfirmed Reports
Indefinite, generally third-hand reports were obtained of sand boils in 1929 between carter's Beach and Cape Foulwind, near Westport, and in 1968 on Van Vught's Farm at Iron Bridge near Lyell.Also, the 1968 aerial photographs G 8 and 9 show white patches that may be ejected sand beside the Inangahua-Reefton road about 10 km south of Inangahua Junction.This land was owned by a Mr McLelland at the time of the earthquake, but he could not be traced and, as for the two other cases, no eyewitness confirmation of liquefaction could be found.These sites are marked in Figure 2.

MAGNITUDE-DISTANCE RELATIONSHIP
The 18 cases of liquefaction found in the Buller region are plotted on a magnitudedistance diagram in Figure 17, together with the Kuribayashi and Tatsuoka (1977) upper bound relationship for distance R to site of furthest liquefaction.max For the 1968 data, it is seen that liquefaction might have been expected at greater distances, especially towards the south, since the epicentre was at the northern end of the rupture surface.Enquiries were made of residents in the Reefton-Ikamatua area, but no further cases were found.
Amongst the 1929 cases, the most distant site, at Greymouth, plots on the bounding distance line, suggesting that this would be an especially interesting case to investigate further.And, in fact, work is in progress at the first two sites, to establish the soil properties there and to check the results of various predictive liquefaction models against the known occurrence of liquefaction.

1 23
Dept of Civil Engineering, University of Canterbury, Christchurch, New Zealand Societe d'Aplication du Texsol, Orsay, France; formerly, Dept of Civil Engineering, University of Canterbury, Christchurch, New Zealand Works Consultancy Services, Westport LOG OF TRENCH AND AUGER BORING BY DODD AND DUNLOPIN AUGUST  1968.

FIGURE
FIGURE 9A: SAND BOILS IN THE "BULL PADDOCK" MEDDOWS' FARM.AT NORTHEAST END OF WALKERS FLAT, MAY 1968.

FIGURE 11 :
FIGURE 11: AREAS OF LIQUEFACTION IN 1968 IN THE KILKENNY PARK-NORTH BEACH AREA OF WESTPORT,

FIGURE
FIGURE 12: DERBY ST, WESTPORT, LOOKING SOUTH FROM KILKENNY PARK,

FIGURE
FIGURE 13: APPROXIMATE BOILS SEEN EARTHQUAKE.LOCATION AFTER OF THE SAND 1929

TABLE 1
and searching newspaper accounts, 9 cases were confirmed for the 1968 earthquake and 9 also for the 1929 earthquake.Two sites, Three Channel Flat, Inangahua and Keoghan's Farm, Westport, liquefied in both events.In the epicentral areas of both earthquakes extensive liquefaction occurred with sand being ejected over large areas of Walkers and Three Channel Flats in 1968 and over much of the Four Rivers Plain area in 1929.Reliquefaction occurred during at least one 1968 aftershock at Walkers Flat.The most distant cases £rom the Inangahua earthquake occurred at Westport, 34 km from the epicentre.Here, sand was ejected at a few isolated places in the Buller and Orawaiti river floodplains and in a much more pervasive manner in the dune sands in the North Beach-Kilkenny Park area of Westport.It is worth noting that the greatest concentration of sand boils occurred in the very recent beach sands that have built up since the harbour breakwater extensions in the 1890 1 s.would expect liquefaction to be possible up to about 70 km from the epicentre given sufficiently loose soils.Enquiries of residents south of Reefton and Westport uncovered no further cases.The extensive liquefaction in 1968 at Three Channel Flat and at Kilkenny Park and at Four Rivers Plain in 1929 suggest that these would be interesting sites to investigate further, as would the Grey Lagoon site if it can be located more precisely.
FIGURE 17: MAGNITUDE-DISTANCE PLOT OF THE 1929 AND 1968 SITES,