Earthquake tree-ring impacts in the middle and upper Buller River catchment

Abstract

Earthquakes produce strong ground shaking and numerous secondary effects such as landslides, rockfalls and liquefaction. In steep mountainous terrain the abrupt increase in debris following the earthquake also leads to major sediment build up in the valley and river systems (aggradation). Where earthquakes occur in densely forested regions the trees can suffer severe damage. Some of the damage occurs simply from the earthquake shaking, in which the tree acts as a relatively top heavy inverted “pendulum”, and can suffer breakage of the main trunk, side branches or root system.

A lot of damage also arises from secondary damage to the ground on which the trees grow by landslides or liquefaction and the impact of debris at rock fall margins and on alluvial fans. Finally aggradation can bury the trunks of trees in alluvium sufficient to cause their decay and eventual death. While all these factors cause damage, an earthquake is also beneficial for some of the surviving trees. Frequently there is an increase in the available light and sometimes the less severe incursions of debris and alluvium bring new beneficial nutrients.

In all these ways earthquakes impact on tree growth and the impacts are ultimately recorded in the tree ring widths of those trees that survive. Potentially tree rings provide a reliable dating method and where trees are old enough they may provide an important tool to help date pre-historic large earthquakes in forested areas. The Buller region of the West Coast of the South Island is a forested area that has experienced two large earthquakes in the 20th century, the 1929 Buller earthquake and the 1968 Inangahua earthquake. A study of growth rings of trees in this area provides an excellent opportunity to better understand the way New Zealand trees react and record the effects of a strong earthquake. The extraction of a thin core of wood with an appropriate boring implement does not harm the tree but provides a record of the growth rings over the life of the tree. By counting the rings, and measuring the relative width of the rings, the rate of tree growth before and after the earthquakes can be determined.

We have found that both the 1929 and 1968 earthquakes have left a clear impact in the tree growth patterns. There are some trees in which the negative impacts of damage have dominated but most of these trees regain their former growth patterns as they “heal” following the earthquake. Those that do not recover, or progressively decline, tend to die and so now there are not many living examples of the most severely affected trees remaining for us to sample. However, in many cases we have sampled trees that have experienced the benefits of extra light and nutrients so that they have accelerated their growth and never slowed back down. By adjusting our method of analysis to focus on those trees that have undergone life long changes in growth, we have been able to make a very clear distinction between the earthquake impacts and other normal growth fluctuations caused by climatic factors such as drought, severe frost and wind.

We have also sampled a range of tree species, and the landforms on which they grow, so that we can determine the species and landform that most clearly and consistently record the earthquake signal. Knowing this will enable future studies in forested areas to better target the most useful trees and the areas in which they grow. We have found that swampy hollows and pakihis (swamps on slightly elevated older terraces with relatively impermeable iron pans) are generally the best sites. They appear to provide the most consistent earthquake record and are also often the areas that are least affected by drought and strong wind.

Our study has included a total of 12 long-lived indigenous tree species commonly found growing on the landforms, but in particular has focussed on the three most common beech species (red beech, silver beech, mountain beech) along with mountain cedar, pink pine and rimu. The results show that while virtually all species have the potential to record growth changes from earthquakes, silver beech is the best. Fortuitously silver beech is one of the most widely distributed indigenous tree species in New Zealand, and can be found growing in a wide range of elevations and settings.

The study areas that have been selected include areas that have experienced a reasonable range in the degree of strong earthquake shaking in the two historical earthquakes. The degree of earthquake shaking is measured in units of an intensity scale (Modified Mercalli Intensity ranging from 1 to 10). We have found that the earthquake record starts to clearly stand out amongst the normal growth fluctuations of the trees once the earthquake shaking approached Modified Mercalli Intensity 8. Recognition of this threshold level provides a possible way to delineate the main epicentral areas of pre-historic earthquake shaking.

We conclude that tree ring analysis provides a very good tool in helping to better understand the timing and extent of prehistoric earthquakes in New Zealand, particularly where there is reliable collaborating evidence from geological investigations of the faults which have been responsible. Two important faults that have forests in close proximity, and which would repay further study, is the Alpine Fault of the South Island, and the Wellington Fault in the lower North Island.

---

Technical Abstract

Earthquakes can be a major disturbance agent in forests because strong earthquake shaking breaks branches and tree crowns, damages root systems, and causes trees to fall amongst closely growing neighbours. Earthquakes also affect trees by secondary effects such as the generation of landslides and debris movement on slopes and fans, as well as aggradation and liquefaction in valley and swamp areas. Forests not only preserve an earthquake signal in their forest age structure (reflecting tree mortality and subsequent regeneration) but also in the tree ring patterns of the trees that survive the event. Potentially the analysis of tree ring patterns provides a possible method to narrow down the dates of inferred prehistoric earthquake events where paleoseismic trenching and related geological investigations demonstrate earthquake occurrence.

The Buller area of the West Coast of the South Island is a densely forested region that has experienced two large earthquakes in the twentieth century. The largest was the 1929 M 7.8 Buller earthquake, followed almost 40 years later by the 1968 M 7.4 Inangahua earthquake. Potentially the area offers an opportunity to carefully assess the nature and variability of tree ring impacts from large earthquakes in New Zealand terrain. We have selected two forested areas in the Buller River catchment within the epicentral area of each of these earthquake events. At each location we have systematically sampled tree rings from a representative variety of indigenous New Zealand tree species, and across a range of landforms, so that we can determine both the most useful tree species, and the landform that best records the earthquake signal.

At both the study sites the earthquakes of 1929 and 1968 resulted in clearly, relatively easily distinguishable pulses of impact on tree growth that stand out amongst the normal fluctuations from climatic factors. We have also developed an improved analysis method to better distinguish between tree ring impacts caused by earthquakes, and more normal fluctuations from wind and drought. It appears that by focussing on the subset of trees that record extremely severe and long-lasting growth changes it is possible to most clearly distinguish earthquakes from other factors. For some trees earthquakes are life changing events from which they either never recover their former growth rate, or else are able to exploit new nutrient and gaps in the forest canopy to abruptly accelerate and maintain their growth. Using this long-lasting analysis method we are able to clearly distinguish the two historical twentieth century earthquakes from background fluctuations.

We conclude that swamps and pakihis record earthquake events most clearly and consistently because tree response is not dependent on spatially variable secondary processes such as debris movement and rockfall, and the potential complications in tree ring patterns from severe climatic events (wind and drought) are generally minimised. The most extreme tree ring response is recorded in some alluvial fan and rockfall margins, but the record at such sites is much less consistent, and is frequently spatially restricted within each landform.

Our results show that while virtually all tree species have the potential to record growth changes resulting from earthquake impacts, some species record the impacts much more frequently and consistently than others, and in this sense can be considered “sensitive” to earthquake impacts. The most promising species are silver beech and mountain beech. Silver beech has the added advantage of being one of the most widely distributed indigenous tree species in New Zealand, and can be found growing in most areas from valley floors to sub-alpine timberlines, and on well drained sites through to wet infertile bogs.

The two study areas included a range in Modified Mercalli shaking intensity from MM7 to MM10 and thus provide an opportunity to investigate whether tree ring patterns vary in a systematic way with shaking intensity. We conclude that is possible to at least map the outer limits of MM8 shaking, on the basis that MM7 does not always show up significantly above background levels of normal growth variation. This has important potential in allowing an assessment of the spatial distribution of prehistoric earthquake shaking.

The results of the study will allow future research to target the most suitable tree species and landforms and will provide a foundation for better interpretation of the paleoseismic record in forested areas over the last approximately 800 years (the typical maximum age of the key tree species).