Department of Zoology, University of Vienna, 1010 Wien, AUSTRIA.
In this thesis I investigated moth communities in relation to flood regime across three riparian regions of lowland Eastern Austria (viz. Danube, Morava and Leitha rivers) using light traps once a month over a period of two consecutive years.
Although light trapping is the most widely used method to survey nocturnal moths, little is still known about the distances at which moths respond to an artificial light source. Two community-wide mark-release-recapture experiments were carried out in order to investigate the attraction radius of a weak artificial light source (2 × 15 W UV-light tubes). Altogether 2,331 moths belonging to 167 species were caught at light traps, individually marked, and released again at distances of 2–100 m. Of these only 313 moths returned to the light trap within 5 min of release. Percentage recapture was generally low (gross rate 13.4%) and strongly decreased with increasing the distance at which they had been released. The data confirm that the attraction radius of low-power light traps for moths is very small, often even below 10 m. Therefore, moth samples assembled with such light traps reflect the communities from which they are drawn at a sufficiently high spatial resolution (in the range of tens of meters) to allow for comparisons in a finely grained forest landscape.
As one major question of this thesis was the impact of flooding on moth species diversity, it was important to select an appropriate measure of local diversity which is sensitive at precisely the ecological scales under study. I used a large data set of 448 moth species and 32,181 individuals, collected in the three floodplain forests mentioned above, to empirically explore the performance of a range of alpha-diversity measures. Earlier comparisons of diversity measures have mostly been made using modelled data sets. The studied forests comprised regularly flooded and non-flooded habitats, thus, I expected that local moth diversity should be shaped by both, regional differences and local flood effects. Surprisingly, observed species numbers as well as eight methods to extrapolate species totals completely failed to reflect differences between the three study regions or between flooded and non-flooded habitats. Rarefied species numbers and Fisher’s α of the log-series distribution did capture differences in moth diversity between the regions, but failed to mirror flooding impact. Only Shannon’s diversity captured all expected diversity differences, at high significance levels. Whether using Shannon’s diversity in its original formulation, or a recently developed bias-correction for small sample sizes, did not affect conclusions about species diversity patterns, but the original formulation tended to underestimate species diversity in smaller samples. I therefore decided to adopt the bias-corrected Shannon diversity as the most meaningful species diversity measure for my subsequent analyses.
I then proceeded to compare moth species diversity and species composition between the three floodplain forest regions and between differentially flood-impacted forest stretches. Today’s floodplain forests in Austria consist of small stretches embedded into non-forested cultivated landscape. Accordingly, and in view of the high mobility of these insects, moth samples taken inside forests always contain a fraction of non-breeding individuals that have immigrated from this landscape matrix. To test the impact of these stray species on diversity patterns, moths were segregated into resident and strays according to their larval resource and habitat requirements. Resident moths were further partitioned into arboreal and ground-layer species based on their larval habitat, to find out if flooding affects these groups differently. Stray species were quite numerous, accounting for 17 % of observed species and 6 % of sampled individuals, but they only marginally influenced diversity and species composition patterns. Contrary to expectation, total moth diversity and ground-layer moth diversity were generally not reduced in flooded habitats relative to non-flooded habitats. In two of three riverine regions species diversity of these terrestrial insects was even higher in flood-impacted habitat fractions. I attribute these patterns to the higher heterogeneity and naturalness of flood-impacted areas plus the strong re-colonisation potential of mobile moths after disturbances through floods. Species diversity of arboreal moths did not show any significant differences between flood regimes at all. With regard to species composition, there was a strong differentiation of moth communities between the three floodplain regions and to a lesser degree between flooded and non-flooded forests. Moth ensembles from flooded habitats in different riverine regions did not group together in ordination diagrams. This contradicts to the hypothesis that flooding would result in a characteristic moth community tolerant to frequent inundation. Differences in species composition were mostly caused by changes in abundance relations of eurytopic moths, and could not be attributed to specialist species bound to wetland habitats.
I further investigated if subsamples of moth assemblages differ in their potential to reveal ecological patterns, i.e. such subsamples can serve as surrogates for overall beta-diversity. Concomitantly, I analysed the extent of structural redundancy in the dataset. Various taxonomically or ecologically defined moth subsamples mirrored total beta-diversity patterns to quite different degrees. For these analyses, I compared the three largest superfamilies (Noctuoidea, Geometroidea, and Pyraloidea) as well as 10 functional groups defined by their larval habitats and resource affiliations. Even tough the Noctuoidea showed the highest concordance with all moths, the Geometroidea provide a better surrogate for beta-diversity, because they scored almost as well as the Noctuoidea, but working effort is much lower since they are not that rich in species and less numerous in individuals (i.e. 31.25% of total species and 21.22% of total individuals).
Regarding to structural redundancy I was able to reduce the dataset down to only 8–15 species (i.e. only 1.5–3.35 % of all recorded moth species) that were fully sufficient to reflect the species composition patterns in the overall moth community. The most abundant species did not necessarily carry the greatest weight in that regard. Rather, the results suggest that representation of all (common) functional types which may be expected within an ecosystem is more important to define surrogate groups to monitor species turnover. These observations also lead to hypothesize that floodplain forest moth assemblages likely show considerable functional redundancy.
Overall, the results assembled in this thesis indicate that for moths, as a representative and species-rich group of terrestrial herbivorous insects, floodplain forests cannot be characterised as ‘hotspots’ of biodiversity. Moth species diversity and species composition were more strongly modulated by regional factors than by local habitat conditions.