Conclusions for the conservation of arthropods in floodplain grassland

As outlined in the introduction, conservation schemes of wet grassland often include management by mowing and changes of the flooding regime. The present study revealed that in temporarily flooded grassland mowing twice a year, i.e. in late spring and late summer, led to a decrease of species numbers and favoured pioneer species which are well adapted to disturbed habitats. Flooding in summer affected the

long time during winter led to plant- and leafhopper and spider communities not very rich in species but consisting of many specialists. In contrast, in both areas of the floodplain a lot of carabid species occurred, many of which were specialists.

The present results suggest that mowing should be stopped if it is the aim to preserve typical floodplain arthropod communities. The question remains how to maintain open grassland habitats, i.e. how to prevent the growth of shrubs and trees. For plant- and leafhoppers Achtziger et al. (1999) suggest a more extensive mowing regime. In moist grassland mowing parts of the area in autumn resulted in an even more species rich Auchenorrhyncha community as compared to fallows (Achtziger et al. 1999).

Furthermore, many of these species were specialists (Achtziger et al. 1999). For moist grassland it was found that a decrease in mowing intensity favoured hygrophilous carabids, i.e. typical wetland species (Klieber et al. 1995). Decleer (1990) states that cutting in summer was more severe for spider communities than cutting in winter. But can we apply these results to temporarily flooded habitats?

In chapter 4 it was outlined that many of the specialist plant- and leafhoppers and two typical wetland spiders overwinter and tolerate winter submersion in the egg stage. But currently we have no knowledge about mortality rates during the time of submersion. It is known from ground beetles that a species being capable of surviving submersion, needs immigrating individuals to maintain a viable population (Fuellhaas 1997). This might also be the case for a number of other arthropods. Mowing in autumn might further increase mortality rates in the egg stage and hence lower the populations’

chances of survival during winter submersion.

Grazing is a second common management to maintain open grassland. The impact of grazing on arthropods was out of the scope of the present study but was investigated in a number of earlier studies: Such as for mown sites, Morris (1973) identified a number of Auchenorrhyncha species responding differently to the impact of grazing. Nickel &

Hildebrandt (2003) found that moderate grazing had less severe effects than mowing twice a year when regarding species numbers and specialists. Spiders might benefit from a mosaic of short and longer swards which results from grazing, but at present there is no clear evidence whether grazing is preferable to cutting in relation to spiders (Bell et al. 2001). Reducing grazing intensity seems to have no effect on numbers of carabid species, however changes in species assemblage were found (Irmler &

Heydemann 1985).

Mowing and grazing are both common management measures in a variety of grassland habitats. However, in temporarily flooded grassland one might ask the question whether mowing or grazing is really necessary to keep shrubs and trees out. In the Polish parts of the Lower Oder Valley there are areas that belonged to the polder system until World War II. During the war, dykes were damaged and not rebuilt afterwards and the area has since been left for succession. The climax vegetation in the area mainly consists of alder and to a lesser extent of willow forests, but at present large areas are still devoid of trees. This might be caused by the frequent flooding (Jasnowska et al. 1999). Especially

(Jasnowska et al. 1999). These findings suggest that restoring natural flooding dynamics in floodplains might as well contribute to preserve open grassland habitats.

The results of the present study help to understand in what respect natural flooding will affect arthropod communities. Natural flooding dynamics of the lower river Oder, as an example of a typical lowland river of Central Europe, are characterised by regular and long lasting winter floods. Summer floods are more rare events and high water levels that are comparable to winter floods occur only once in ten years in average (Vössing 1998).

At first sight, study sites subject to high flooding impact, i.e. to winter and summer floods, seem to be subject to a near-natural flooding regime. As outlined above, this flooding influence favours arthropod communities which are similar to those found in mown meadows. But in contrast to the relatively wide floodplain of lowland rivers, these study sites were situated on a very narrow stretch of the river bank. Here, the water level rises and falls rapidly and the grassland as well as the arthropods are subject to very variable water levels and high currents. The former natural floodplain was much wider, consequently summer floods had a less severe impact on the arthropod communities. The water flooded a larger area thus inundation was slower and less severe or changes in the water level were even just detectable by a rise in the ground water level.

Restoring natural flooding dynamics seems to favour areas which are comparable to some of our study sites subject to medium flooding impact. Here, long lasting floods occur regularly in winter and even during summer ground water levels remain relatively high. In these study sites we found very specialised communities which, however, were not rich in Auchenorrhyncha and Araneida species. Conversely, in the latter two groups significantly more species were only collected in sites not subject to flooding at all.

Thus, it seems that there is a trade-off between diversity and naturalness when it is the aim to conserve plant- and leafhoppers and spiders in floodplain grassland. Ground beetles, on the other hand, occur in high species numbers especially in sites subject to long winter floods and in those subject to summer and winter floods. Differences in the relief and in the distance to the river will - under natural flooding dynamics - lead to a number of habitat patches differing in flooding influence. Some might be comparable to those of our study sites subject to high, medium and low flooding impact. The present results (see chapter 3) and those of Krüger (2004) and Sellner (2004) demonstrate that species similarity between these sites was 20-60%. Furthermore, in sites subject to medium flooding impact already at present a variety of patches can be found that are dominated by different plant species such as sedges (Carex spec.), sweet-grass (Glyceria spec.), reed grass (Phalaris arundinacea) and common reed (Phragmites australis). These differences might be due to different moisture conditions. For the present study we selected only sites dominated by sedges (Carex spec.) and reed grass (Phalaris arundinacea). In habitat patches dominated by other plant species one can expect to find a number of additional Auchenorrhyncha species due to the availability of different food plants. Furthermore, additional species of spiders and ground beetles

structure. Consequently, in a floodplain which is characterised by a lot of different habitat patches due to natural flooding dynamics one does not only find very specialised communities but also a high β-diversity.

Additionally, the findings of chapter 4 revealed, that both submersion tolerant and immigrating arthropods were abundant in temporarily flooded grassland. Many of the submersion tolerant species seem to have a low dispersal capacity. It is important to investigate the dispersal range of these species to assess whether they are capable to colonise new suitable habitats such as those created by restoration of wetlands within a specific conservation scheme. Furthermore, the assemblage of immigrating species is dependent of the regional species pool. They need suitable non-flooded overwintering sites inside as well as in the surroundings of the floodplain, from which they can recolonise it with receding water levels each spring. Thus, in conservation projects that have the aim to preserve typical floodplain arthropod communities not only land use should be stopped and a more natural flooding regime should be restored, but also the surrounding landscape of the sites in question should be taken in account.

6.3 Evaluation of the plans for future national park development