Determination of the physical drivers of Zostera seagrass distribution using a spatial autoregressive lag model
Version 2 2024-06-04, 14:50Version 2 2024-06-04, 14:50
Version 1 2017-09-16, 08:13Version 1 2017-09-16, 08:13
journal contribution
posted on 2017-01-01, 00:00authored byAlastair Hirst, K Giri, D Ball, R S Lee
Seagrass mapping has become a key tool in understanding the causes of change in seagrass habitats. The present study demonstrates a method for examining relationships between seagrass habitat polygons and environmental data generated by hydrodynamic, wave, catchment and dispersion models. Seagrass abundance data are highly spatially autocorrelated and this effect was corrected using a spatially simultaneous autoregressive lag model (SSARLM). The physical processes that determine the spatial distribution of seagrass in Port Phillip Bay, Australia, were investigated by examining the links between seagrass distribution and abundance and broadscale hydrodynamic (waves, currents), physical (light, depth, salinity and temperature) and catchment (nutrient and suspended sediment concentrations) processes. The SSARLM indicated that the distribution of Zostera spp. meadows is principally constrained by two physical thresholds, namely, wave height or exposure and light. The former excludes seagrasses from colonising wave-exposed coastlines, whereas the latter directly determines the depth profile of seagrasses through its influence on light availability. In total, 95% of all seagrass occurred within grid cells with a mean significant wave height of <0.38 m and a mean percentage irradiance of >33% surface levels. By comparison, variation in water quality, represented by variables such as modelled total nitrogen, suspended solids or salinity, had little influence on seagrass distribution.