Hindcasting estuary ecological states using sediment cores, modelled historic nutrient loads, and a Bayesian network

Abstract

Estuaries are impacted by catchment land use changes, driving degradation associated with eutrophication and alterations in sediment dynamics. Estuarine ecological monitoring has typically covered only recent periods, so magnitudes and trajectories of degradation are often poorly described. Here, we develop a multi-method approach to hindcast historic estuary condition. We determined geochemical properties in dated sediment cores from two sites in the eutrophic New River Estuary (NRE), Aotearoa New Zealand. Mud, organic matter, carbon, and nitrogen (N) contents increased from ~1847 to 2019 (~20 to 100%, ~1 to ~7%, ~0.2 to ~3%, and ~0.03 to ~0.37% respectively). Historic land use modelling shows increasing catchment loads and water column total N rising from 165 to 663 mg m-3, with more rapid change in the late 20th and early 21st centuries. Strong correlations (R2 = 0.93-0.95) were found between sediment N content and water column N concentrations. Trophic indicators hindcasted by a Bayesian Belief Network worsened from the mid-20th to 21st centuries. This degradation is associated with increased sheep, beef and dairy farming in the catchment. Our approach identifies nutrient and sediment loadings associated with ‘tipping points’ for indicators and can inform policy by identification of hazardous levels of contaminant loading for other estuaries.