No abstract

[1] Over 1 year we followed the seasonal variations of N 2 O and CH 4 fluxes at a tropical rain forest site in Australia. In addition, meteorological parameters, litter fall and decomposition, plant species composition, and concentrations of NH 4 + /NO 3 À in the soil and N 2 O and CH 4 in the soil atmosphere were measured. N 2 O emissions showed a pronounced seasonal pattern with highest rates in the wet season (108.6 mg N m , and, thus, approximately 7 times lower than a previous estimate for the year 2000. The marked differences in N 2 O emissions between different years indicate that the interannual variability of N 2 O emissions from rain forest soils cannot be neglected. With regard to CH 4 the soil functioned throughout the entire year as a significant sink. Rates of CH 4 uptake during the dry period (35-68 mg CH 4 m À2 h À1 ) were higher as compared to the wet period (4-45 mg CH 4 m À2 h À1 ). A close linear correlation between soil moisture and magnitude of CH 4 uptake was found. The calculated annual CH 4 uptake (N = 6090) is 3.2 kg CH 4 ha À1 yr À1. This implies that tropical rain forest soils function as significant sinks for atmospheric CH 4 on a global scale.

Summary1 Mathematical proofs show that rate estimates, for example of mortality and recruitment, will decrease with increasing census interval when obtained from censuses of non-homogeneous populations. This census interval effect could be confounding or perhaps even driving conclusions from comparative studies involving such rate estimates. 2 We quantify this artefact for tropical forest trees, develop correction methods and re-assess some previously published conclusions about forest dynamics. 3 Mortality rates of > 50 species at each of seven sites in Africa, Latin America, Asia and Australia were used as subpopulations to simulate stand-level mortality rates in a heterogeneous population when census intervals varied: all sites showed decreasing stand mortality rates with increasing census interval length. 4 Stand-level mortality rates from 14 multicensus long-term forest plots from Africa, Latin America, Asia and Australia also showed that, on average, mortality rates decreased with increasing census interval length. 5 Mortality, recruitment or turnover rates with differing census interval lengths can be compared using the mean rate of decline from the 14 long-term plots to standardize estimates to a common census length using λ corr = λ × t 0.08 , where λ is the rate and t is time between censuses in years. This simple general correction should reduce the bias associated with census interval variation, where it is unavoidable. 6 Re-analysis of published results shows that the pan-tropical increase in stem turnover rates over the late 20th century cannot be attributed to combining data with differing census intervals. In addition, after correction, Old World tropical forests do not have significantly lower turnover rates than New World sites, as previously reported. Our pan-tropical best estimate adjusted stem turnover rate is 1.81 ± 0.16% a − 1 (mean ± 95% CI, n = 65). 7 As differing census intervals affect comparisons of mortality, recruitment and turnover rates, and can lead to erroneous conclusions, standardized field methods, the calculation of local correction factors at sites where adequate data are available, or the use of our general standardizing formula to take account of sample intervals, are to be recommended.

Summary1. Despite great concern about the eects of fragmentation on biodiversity, quantitative studies are still scarce with respect to many major groups and important environments. Well-studied natural reference sites are few. 2. Extensive light trapping surveys for moths were thus carried out in both dry and wet seasons in nine remnants of complex notophyll vine forest on basalt on the Atherton Tablelands in tropical north Queensland, Australia. Three sites had never been cleared, three secondary sites had substantial regrowth, and three sites were newly cleared. 3. A total of 15 632 moths of 835 species was collected, counted and identi®ed. These represent more than 17% of the named Australian fauna of our target families. 4. A principal components analysis (PCA) indicated clear discrimination among assemblages based on forest type. This discrimination did not dier qualitatively between seasons (although abundance levels of moths did) but the pattern was most evident in the smaller dry season samples. 5. Taxa, the relative abundance of which increased signi®cantly with disturbance, were the Arctiinae, Amphipyrinae, Catocalinae, Hadeninae, Heliothinae, Hypeninae, Noctuinae, Plusiinae, Hermeniidae and Phycitinae. In contrast, a number of subfamilies showed a marked decrease in relative abundance with increased disturbance, namely Ennominae, Geometrinae, Larentiinae, Oenochrominae, Epipaschiinae, Lymantriidae and Anthelidae. 6. A weighted sum with importance values based on the eigenvalues associated with each of these taxa derived from the PCA is a powerful predictor set of forest quality. 7. These dierential responses may be explained on the basis of broad known and expected host-plant patterns. The results show how moth assemblages are powerful indicators of forest disturbance, and should prompt parallel studies elsewhere in the world.