A considerable literature on the horizontal variability in plankton distribution has grown up, seemingly aimed, in part, towards invalidating any preconceived assumption of homogeneity. It is difficult to determine just where this assumption might have arisen, as investigations readily demonstrate that distributions are often far from homogeneous. It may be that a predominance of papers in the mid-twentieth century focused on the population dynamics and vertical distributions of phytoplankton in small lakes paying insufficient attention to simultaneous horizontal heterogeneity. Such assumptions, real or supposed, have no place in modern plankton science. However, even now, it is important to present a perspective on just how much variation might be expected, over what sort of horizontal distances and how it might reflect the contributory physical processes.

Omitting the very smallest scales (see preamble to Section 2.7), phytoplankton is generally well-randomised within freshly collected water samples (typical volumes in the range 0.5-5 litres, roughly corresponding to a linear scale of 50-200 mm). Thus, there is normally a low coefficient of variation between the concentrations of plankton in successive samples taken at the same place. The first focus of this section is the horizontal distance separating similar samples show ing significant or systematic differences in concentration.

There have been few systematic attempts to resolve this question directly. In a rarely cited study, Nasev et al. (1978) analysed the confidence interval about phytoplankton counting by partitioning the variance attaching to each step in the estimation - from sampling through to counting. Provided adequate steps were taken to suppress the errors of subsampling and counting (Javornicky, 1958; Lund et al, 1958; WillĂ©n, 1976), systematic differences in the numbers present in the original samples could be detected at scales of a few tens of metres but, on other occasions, not for hundreds. Irish and Clarke (1984) analysed the estimates of specific algal populations of algae in similar samples collected from within the confines of a single Blelham enclosure (area 1641 m2, diameter, 45.7 m) at locations nominated on a stratified-random grid. They found that the coefficients of variation varied among different species of plankton, from about 5%, in the case of non-motile, neutrally buoyant algae, to up to 22% for some larger, buoyancy-regulating Cyanobacte-ria. In another, unrelated study, Stephenson et al. (1984), showed that spatial variability increased with increasing enclosure size.

A general conclusion is that sampling designs underpinning in-situ studies of phytoplankton population dynamics must not fail to take notice of the horizontal dimension. However, the size of the basin under investigation is also important. For instance, a coefficient of variation of even 22% is small compared with the outcome of growth and cell division, where a population doubling represents a variation of 100% per

Was this article helpful?

## Post a comment