Plant ecology above timber line on Mt Ruapehu, North Island, New Zealand

Abstract

Height frequency of 67 spccies was measured on 100 sites together with environmental variables. Data were analysed by correlation coefficients, association analysis of vegetation data using agglomerative “d squared centroid” method, mapping of variables using two-factor principal component analysis, and relationships between individual spccies and environmental variables using multiple and path analysis regression methods. The 21 more-abundant species for which the latter analysis was made were: Anisotome aromatica, Celmisia spectabiiis, Chionochloa rubra, Coprosma cheesemanii, C. pumita, Dacrydium laxifolium, Dracophyllum recurvum, Euphrasia cuneata, Forstera bidwillii, Gaultheria colensoi, Gentiana bellidifolia, Hebe tetragona, Notodanthonia setifolia, Ourisia vulcanica, Pentachondra pumila, Poa colensoi, Podocarpus nivalis, Rhacomitrium lanuginosum, Schoenus pauciflorus, Senecio bidwiili, and Wahlenbergia pygmaea.

Environmental factors measured were: altitude, aspect, slope, depth to rock, depth A₁ soil horizon, soil colour, total plant frequency, ground cover, vegetation height distribution, solar radiation, catchment area, minimum soil moisture, soil temperature, snow depth, wind exposure, aspcct attenuation of temperature, soil bulk density, colloids, clay, aeration, organic matter, pH, cation exchange capacity, available calcium, potassium, magnesium, nitrate, ammonium, sulphur, and phosphorus. The expected relationship between variables is given in multi-level, multi-factor diagrams. Statistical tests were based on data from 80 sites after the null hypothesis had been determined from the preliminary analysis of 20 sites.

The most frequent significant relationships between species and environmental factors were with: depth to rock, soil acration, solar radiation, available potassium, wind exposure, and available phosphorus. Rescreening using data from all sites indicated that other important relationships requiring testing were with: midsummer soil temperature, catchment area, and soil moisture. Quantitative equations are given for each of the relationships.