Discussion
Seasonal flooding and associated changes in the wetness-dryness
conditions of wetlands is a major disturbance for species inhabiting
wetlands, requiring them either to survive during periods of high water
or to disperse and recolonize once the water subsides. The ability to
survive these conditions may depend on the availability of dry refuges
nearby. This study found differences among arthropod groups in their
relationship to dry refuges. For instance, genera within Staphylinidae
and Linyphiidae were more abundant in sites with a higher proportion of
refuges, whereas groups such as Carabidae and Lycosidae appeared less
sensitive. Interestingly, one group, Salticidae, showed the opposite
response, being more abundant in sites that experienced longer
inundation periods. We also observed that the spatial scale of refuges
played an important role. The highest predictive power was found when
the proportion of dry refuges was estimated for a 10x10
m2 area surrounding the sampling site. In contrast,
the proportion of dry refuges immediately surrounding the sampling sites
(i.e., within 2x2 m2) was less significant.
To interpret these results, it is important to first consider the
composition of these dry refuges. Most wetlands are heterogeneous,
consisting of a mixture of wet and dry areas (Diamond, Epstein, Cohen,
McLaughlin, Hsueh, Keim and Duberstein 2021). This particular wetland is
a relatively flat fen, primarily covered by grasses and sedges, but it
also features microtopographic variation. Notably, solid and relatively
high hummocks (height: 30-50 cm, diameter: 20-30 cm), composed mainly ofCarex species, dominate parts of the area. These hummocky areas
represent some of the key dry refuges, although other dry refuges
include areas with higher but more evenly elevated ground. In this
study, the pixel size used to estimate local wetland wetness-dryness
(expressed in terms of inundation frequencies) was 2x2
m2. As a result, each pixel may contain multiple
hummocks as well as intervening hollows, which tend to remain flooded
for longer time periods. Please note that the digital elevation model
(DEM) and inundation frequency estimates should be interpreted relative
and not as absolute measures. While differences between sites are likely
accurate, the absolute levels may be influenced by both measurement
errors and annual variations. For the purpose of this study, the wet
sites are consistently wetter than dry sites whereas the absolute
wetness-dryness estimates or inundation frequencies may be somewhat
biased.