Response of individual sizes and spatial patterns of Deyeuxia angustifolia to increasing water level gradient in a freshwater wetland

The wetland plants are very sensitive to hydrological regimes. In this study, the individual sizes of a widely distributed species (i.e., Deyeuxia angustifolia) at three typical marshes with different water table depths (i.e., wet meadow (WM) marsh; seasonal inundated (SI) marsh; perennial inundated (PI) marsh) were investigated in the Sanjiang Plain of Northeast China. Concurrently, three primary point pattern processes (homogeneous Poisson (HP) process, homogeneous Thomas (HT) process, and inhomogeneous Thomas (IT) process) were used to model spatial patterns in the distribution at 0–50 cm scale for this tillering-cloning species. The plant height, diameter at breast height (d.b.h), internode number, branches number, and individual aboveground biomass of D. angustifolia decreased sharply with rising water level; however, its density and coverage increased first and then decreased as water level increases. The distribution of D. angustifolia totally diverged from the complete spatial randomness (CSR) model (i.e., HP process) suggesting strong aggregation at 0–50 cm scale in all marshes, and aggregated intensity enhanced with increasing water level. Interestingly, the spatial distribution of D. angustifolia fits better with the nested double-cluster model (i.e., IT process) at all scales in WM and SI marshes, indicating that there is a series of clustered patterns under the slight flood stress. However, the spatial pattern fits well with the Poisson cluster model (i.e., HT process) at all scales in PI marsh, implying the small-scale clustering disappeared with the intensification of flooding stress. Our results highlight that the D. angustifolia population could adapt to flooding stress in a certain degree via individual miniaturization strategies and multi-aggregation mechanisms in the freshwater wetlands.