Modeling the incidence of temporary pond microcrustacea: The importance of dry phase and linkage between ponds

The incidence of seven species of microcrustacea (3 Ostracoda, 2 Cladocera, and 2 Copepoda) was recorded from 30 small experimental pools in late winter and early summer throughout 1995-1998. All ponds were created in autumn 1994 and were the same size, shape, and age, distributed over a freshwater wetland in Northumberland, UK. The length of time during which each pond dried out in summer or linked to other ponds during floods was recorded. Logistic regression was used to test the hypotheses that summer dry phase and flood linkage provided significant explanation for the presence or absence of each species throughout the ponds. Late winter incidence was regressed against the length of the previous summer's dry phase and the length of time each pond was linked to other ponds by subsequent re-flooding. Summer incidence was regressed against the length of the previous summer's dry phase, length of time linked by subsequent re-flooding, macrophyte density, and incidence at the preceding late winter sample. Individual ponds were also identified as separate factors in regressions to quantify the effect of repeat-measures from the same ponds. For January samples, the length of previous summer dry phase was a significant predictor of incidence for four species, and linkage was significant for one species. In summer, the length of previous summer dry phase was a significant predictor of incidence for four species, linkage was significant for three species, and macrophyte density was significant for four species. In summer, two species showed significant relationships with specific ponds. The relationship between length of dry period and subsequent presence varied, and was negative or positive for different species. The results suggest that variation of a few major systematic environmental factors, such as dry phase, explains species presence and absence, but that individual species vary in response and that factors contingent on individual pond characteristics also affect distribution.