Data from: Diversity in tropical ecosystems: the species richness and turnover of moths in Malaysian rainforests

As a contribution to accurate estimation of arthropod alpha and beta diversities in tropical forests, we present results of some of the largest moth samples ever collected in the Malaysian region. To estimate alpha diversity, light traps were run at three geographically distinct locations. We generated individual-based and coverage-based rarefaction curves to estimate sampling sufficiency and alpha diversity of the locations. Despite a large number of moths collected (67 282 individuals, from three locations), none of the rarefaction curves reached asymptote. The species accumulation curves based on the Chao1 richness estimator at each location suggested that, even when sampling yielded over 30 000 individuals, Chao1 could not reliably estimate the observed number of species. In one of the three locations, moths were collected systematically by light traps in 1979–1980 and 2000–2001. Despite over 160 trapping nights and 16 500 individuals collected in total, the estimated total number of species (2262) was well below the general collection conducted at the same location over a 35-year period [3921 species (1975–2013)]. Beta diversity was investigated using the samples collected at one location on two occasions over a 20-year period. The faunal composition has changed over the 20-year period, possibly as a result of extensive land-use change around the study location. We estimated the minimum sampling effort required to detect such changes by calculating type II errors. Unlike alpha diversity estimation, we found that only four replicate samples, each with only two trapping nights, would be sufficient to reliably detect changes in assemblage composition