Sampling variability in nutrient cycling studies in some forested ecosystems of Westland, New Zealand

摘要

Field collectors were used to collect monthly samples of litterfall, rainfall, and throughfall for a two-year period, and soil drainage beneath the soil organic horizon and at the base of the rooting zone for 12 months, in two indigenous and four pine forest plots in Westland. CoeffiCients of variation (CVs) for estimates of components of litterfall, rainfall, through fall, and soil drainage below the organic soil and rooting zone and their associated nutrient fluxes 10 kg/ha/yr are given. Litterfall variability was greatest on plots where canopy closure was incomplete or where wood-fall was a significant component of litterfall. CVs for nutrient fluxes in litterfall was generally similar to, or lower than those for dry weight of litterfall. Incident rainfall was very uniform (CV 2#x2013;3). Variability increased markedly (CV 16#x2013;31) as a result of rainfall passing through the forest canopy. Nutrient fluxes in bulk precipitation and throughfall were considerably more variable than a!l10unts of precipitation and throughfall respectively. For Soil drainage, samples collected by tension-free (tray) lysimeters gave results approximately 2#x2013;5 times more variable than those for tension lysimeters. For the latter variability was generally similar to that for throughfall nutrient fluxes. Choice of the number of field collectors to be used should be based on the degree of inherent variability, which depends on the ecosystem. In practice, it will also depend on the availability of resources. Generally, 3#x2013;5 rain gauges, 20#x2013;25 throughfall gauges and littertraps, and 10#x2013;20 lysimeters are considered to be a suitable compromise for some indigenous and pine forests in Westland, New Zealand. Some form of stratified random sampling is desirable for collection of throughfall and litterfall in forests with open canopies. Wood-fall, other than for small twigs, was inadequately sampled by 12 littertraps. It is suggested that for the major components of wood-fall larger areas of the forest floor be regularly sampled.