Other formats

    TEI XML file   ePub eBook file  

Connect

    mail icontwitter iconBlogspot iconrss icon

Tuatara: Volume 30, Issue 1, December 1988

The use of increment cores for the analysis of tree ring chronologies for fijian kauri (Agathis Macrophylla)

page 51

The use of increment cores for the analysis of tree ring chronologies for fijian kauri (Agathis Macrophylla)

Abstract

Tree ring data was collected from Fijian Kauri (Agathis macrophylla) trees in Northern Viti Levu, Fiji, during November and December 1986. The study was to test the validity of using single increment cores instead of entire trunk sections in growth rate and tree age studies for this species. One increment core per tree was sampled. The nature of the wood formation in this species led to problems in the analysis of the cores and poor statistical confidence in the results. The paper discusses the problems encountered during the study and suggests that alternative methodology should be used for this species in tree ring analysis.

Key words: Dendrochronology, Fiji, Agathis.

Introduction

The study area lies on the north western slopes of Mt Lomalagi, near the Nadarivatu Government Station at the northern edge of the Nadarivatu plaeau in northern central Viti Levu (17° 35′ S, 177° 55′ E). The elevation of the study area lies between 900m and 990m.

The climate of this area lies in zone D of the Fiji Forest Inventory, corresponding to a 2540-3300mm mean annual rainfall, with a moderate 3-5 month dry season having less than 150m rainfall in those months. There appears to be sufficient seasonality in this region for the formation of annual rings in the wood of forest trees.

The basaltic parent materials have produced soils classified by Berry and Howard (1973) as Manosavu bouldery red brown clay, humic latosols, having a low base status and high acidity. These soils have a dark organic rich surface horizon that overlies brown to reddish brown sub-surface horizons which are moderately weathered. A more detailed description of the soils can be seen in the following paper (Weaver, 1988).

The vegetation has been classified as sub-tropical lower montane rain forest by MacDonald (1987). Two adjacent study sites were chosen. One consisted of lush dense secondary forest dominated in the canopy by Homalium vitiense, Calophyllum vitiense, Podocarpus nerifolius, Syzygium sp., and Agathis macrophylla. This site maintained an uneven 4-6m canopy and had a dense undergrowth of seedlings, grasses and herbaceous species. The A. macrophylla population on this site consisted of seedlings, saplings and pole-sized individuals having a basal trunk diameter of approximately 5-30cm. The other site was an old growth forest dominated by emergent A. macrophylla over a 12-16m canopy. The canopy dominants included P. nerifolius, C. vitiense, Syzygium sp., Arytera brackenridgeii and Dysoxylum richii. The Agathis population on this site consisted mostly of trees having basal trunk diameters of 30-130cm.

Methods

In the field an increment corer was used to extract one 5mm diameter wood core per tree from a sample of forty five Agathis macrophylla trees over a range of trunk diameters from both sites. Each core was taken perpendicular to the main axis of the tree and as low to the ground as possible. Each core was then labelled with the diameter at 1.4m height from the tree it was taken from. In the page 52 laboratory each core was mounted and sanded. Distances between rings were measured where possible and the total ring count was calculated for those cores that had intercepted with the growth centre of the tree. Areas of indistinct ring formation on each core were measured and the average length of uncountable core sections was calculated.

Results

Ring expression was variable in most cases with some cores only showing distinct rings for part of their length (see Fig. 1). Only 1/3 of the cores sampled showed completely distinct ring formation along the entire length of the core. 40% of the cores sampled had zones where rings were indistinct. On average 35% of the length of these cores was uncountable. The remaining 26% of the cores sampled showed practically no distinguishable ring formation, or visible rings were so widely separated by zones of indistinct rings that ring counts and measurements would have been meaningless. Therefore, this data was considered unreliable and was not used for statistical analysis.

Fig. 1. An example of variation and inconsistencies in ring expression in sanded cores traken from Fijian kauri trees at Nadarivatu. Each core has a 5mm diameter. Core ‘a’ shows a zone of indistinct ring formation at left. To the right large variations in ring distances are seen causing difficulties in determining whether they are annually formed. Some rings are likely to be false (arrow). Core ‘b’ shows no visible ring formation for this section of the core. Core ‘c’ shows variation in the visibility of rings. Core ‘d’ shows evidence of ring wedging and false ring formation (arrowed).

Fig. 1. An example of variation and inconsistencies in ring expression in sanded cores traken from Fijian kauri trees at Nadarivatu. Each core has a 5mm diameter. Core ‘a’ shows a zone of indistinct ring formation at left. To the right large variations in ring distances are seen causing difficulties in determining whether they are annually formed. Some rings are likely to be false (arrow). Core ‘b’ shows no visible ring formation for this section of the core. Core ‘c’ shows variation in the visibility of rings. Core ‘d’ shows evidence of ring wedging and false ring formation (arrowed).

page 53

Discussion

Anomalous ring development can frequently cause problems for the dendrochro-nologist if they are not taken into account when constructing a data set from tree rings. False or double rings and the wedging out of rings can lead to very broad error variances in statistical analysis. Irregularities in ring formation can vary from one species to another and between different locations. For example, in New Zealand, studies of Libocedrus bidwillii have shown that false rings are virtually non-existent for this species. Dendrochronological studies of Agathis australis however, have revealed problems with ring wedging as well as a lack of ring pattern consistency around the trunk. Lobate growth in this species is also common as with Phyllocladus aspleniifolius var. alpinus. Dating of these two species may only be possible at some sites (Dunwiddie, 1979). Ash (1985) undertook a study on the tree rings of Fijian Kauri on the Mt Koravaturu plateau in western Viti Levu (alt. c. 800m) and at a site near Suva (alt. c. 150m). By examining trunk discs he found the seasonality sufficient for the formation of annual rings. Here the late wood consisted of smaller compressed xylem elements formed during the dry season in the months between April and October. Ash noted that late wood anatomy was variable with some distinct clearly differentiated bands, but other rings were represented by a gradual transition to late wood anatomy. On most of Ash's sections 5-10% of rings were not distinct. Some trees that were growing on rapidly draining sites produced false rings, and in general, compression wood development necessitated the examination of whole trunk sections to locate all the late wood rings.

In the present study the core samples proved extremely variable and anamalous in ring expression. Because only one core was taken per tree it was impossible to establish the validity of each ring as truly annual. Understanding the nature of wood formation in Fijian kauri described by Ash (1985), growth rate data from a single core are not likely to be representative of the entire trunk even if ring clarity is good for those cores samples. Whitmore (1975) suggested that ring counts are inapplicable or unreliable in the tropics when attempting to determine tree ages.

Conclusion

The technique of sampling one core per tree to study growth rates and tree ages in Agathis macrophylla in lower montane rain forest proved to be inadequate for confident statistical analysis of the data. Entire trunk discs should be used if possible. If entire trunk discs are not available then at least 3 cores per tree should be sampled to allow for anaomalies in ring formation such as eccentricity, ring wedging, lobate growth and locally absent rings.

Acknowledgements

This paper forms part of a B.Sc. Honours project at Victoria University, Wellington. I am therefore grateful to the following people in the Botany Department: Dr. Ross McQueen for supervision of the project; Stephen Fuller for useful comments on the draft; Ian MacDonald for assistance in the field. The Royal Forest and Bird Protection Society and the Wellington Botanical Society provided financial assistance.

page 54

References

Ash, J. 1985, Growth rings and longevity of Agathis vitiensis (Seeman) in Fiji. Australian Journal of Botany 33: 81-88.

Berry, M. J., and Howard, W.J. 1973. Fiji Forest Inventory. Vols. 1, 2 and 3. Overseas Development Authority: Surbiton, U.K.

Dunwiddie, P.W. 1979. Dendrochronological studies of indigenous New Zealand trees. New Zealand Journal of Botany 17 (3): 251-66.

MacDonald, I. 1987. Observations on the forest dynamics and regeneration of Fijian Kauri (Agathis vitiensis) within old growth forest, Viti Levu, Fiji. Unpublished B.Sc. Honours project. Victoria University, Wellington, New Zealand.

Weaver, S.A. 1987. An introduction to the regeneration of Fijian Kauri following logging. Unpublished B.Sc. Honours dissertation. Victoria University, Wellington, New Zealand.

Weaver, S.A. 1988. Soil differences between secondary and old growth Agathis macrophylla forest at Hadarivatu, Fiji. Tuatara 30: 55-61.

Whitmore, T.C. 1975: Tropical Rainforests of the Far East. 1st ed. Oxford University Press.