Vegetation and Landscape Dynamics in Eastern Taranaki Hill Country

Abstract

An ecological study of hill country landscapes in eastern Taranaki, New Zealand, was undertaken as part of a project concerning the implications of long-term soil mantle changes for sustainable land use. The study was undertaken in a 417 km2 area comprising uplifted and steeply dissected soft Tertiary sediments with a predominantly sandstone lithology. Rapid European settlement in the 1890s modified the natural vegetation cover greatly, so that most remaining forest in the study area occurs in patches surrounded by a matrix of pastoral land. Vegetational and successional patterns and environmental variation : The pattern of woody vegetation was investigated by extensive reconnaissance sampling incorporating semi-quantitative analysis of canopy cover, followed by intensive, environmentally stratified sampling. The vegetation was classified on a structural and floristic basis into 19 units of forest, treeland, scrub and shrubland. The effect of environmental variation on vegetation composition was investigated by reciprocal averaging ordination (DECORANA). The first ordination axis was correlated to vegetation structure and canopy height and was interpreted as a complex disturbance gradient relating to time since disturbance. The second and third axes were related to soil fertility and topographical gradients. Forest plots were dominated by Beilschmiedia tawa and Weinmannia racemosa and had basal area values of up to >250 m2/ha. Basal area, stem and seedling density varied greatly between vegetation structural groups. Regeneration of woody vegetation following various types of disturbance: The disturbance regime was comprehensively documented. Main factors of natural disturbance are landslide erosion and windthrow; main factors of cultural disturbance are direct clearance by felling and burning, and introduced animals. A chronology is presented of successional pathways for about 400 years following major disturbance. Succession proceeds through shrubland and scrub stages dominated by treeferns, Leptospermwn scoparium or other broadleaved woody shrubs, through treeland, to broadleaved forest dominated firstly by W. racemosa or Knightia excelsa, then by B. tawa. Podocarp trees are generally only prominent after a long period of uninterrupted succession. Seedling recruitment, mortality and growth were monitored for 2 years. Seedling dynamics varied considerably between and within sampling plots, some of which contained small exclosures that excluded possums and goats. The effects of introduced animals on seedling recruitment and vegetation growth is strongly modified by microtopography. Most dominant species showed continuous regeneration at the scale of the whole study area, despite local discontinuities. This pattern was consistent with a model of interrpted gap-phase regeneration, which may be widely applicable to New Zealand lowland forests. The vegetation turnover time is in the order of 150-250 years, a period consistent with comparable temperate forest ecosystems. The successional pathway is primarily dependent on topography, previous site history and location and area of disturbance. The existence of residual-soils on landslide scars, variations in plant propagule supply, and rapid loss of soil from steep slopes cleared for agriculture, all suggest that a rigid distinction between primary and secondary succession in the study area is not appropriate. Hillslope processes underlying vegetation and landscape change: Hillslope processes were studied in five 0.1 ha plots in which slope profiles were measured, vegetation and microtopography mapped in detail, vegetation age assessed and soil properties investigated. Ground surface age was assessed as an interpretation of the above data. Mean surface age was c. 450 years, but some swales had a surface age of several thousand years. There was a significant correlation between surface age and soil depth, soil depth increase being faster and continuing for much longer under forest than under pasture. Observations were made of near-surface erosion processes such as soil creep. A model of hillslope erosion is outlined, involving periodic evacuation of swales by landslides and refilling of swales by near-surface erosion. Evidence of past environments supports a fluvial origin for swales in an early Ohakean (glacial maximum) or pre-Ohakean period of high erosion. A concluding synthesis of vegetation, topography and soils emphasises the importance of selecting appropriate temporal and spatial scales at which to study landscape processes.