Work

Decoding climate-induced phytoplankton dynamics in a tropical macrotidal estuary using explainable machine learning

Year: 2025

Type: article

Abstract: Phytoplankton communities in tropical macrotidal estuaries are highly sensitive to hydroclimatic variability, particularly during extreme El Niño-Southern Oscillation (ENSO) events. This study assessed ENSO effects on phytoplankton dynamics in the São Marcos Estuarine Complex (SMEC), Brazilian equatorial margin, using data from 2011 to 2020. Multivariate analyses indicated that variability was primarily driven by seasonality, interannual changes, and ENSO forcing, with significant impacts on salinity, dissolved oxygen, nutrients, and suspended particulate matter. During El Niño phases, reduced river discharge and higher salinity coincided with sharp declines in density and chlorophyll-a. Following the 2015-2016 extreme El Niño, cell density increased while biomass and size fractions declined, indicating a decoupling between abundance and chlorophyll-a and a shift toward smaller, less efficient species. La Niña phases enhanced nutrient inputs, increased DIN:DIP ratios, and supported only partial biomass and diversity recovery, without returning to pre-El Niño conditions. Temporal β-diversity showed significant turnover centred on the 2015-2016 event, revealing incomplete recovery and persistent ecological change. Across the time series, 321 taxa were identified, with Bacillariophyta dominant but less diverse after the extreme event, accompanied by reductions in dinoflagellates and microphytoplankton. SHapley Additive exPlanations (SHAP) identified water temperature, dissolved inorganic nitrogen, and rainfall as the most influential predictors, acting through nonlinear and context-dependent effects. Sensitivity analyses (2021-2050) indicated density declines under La Niña-like regimes and increases during El Niño-like phases. These findings demonstrate that extreme ENSO events restructure phytoplankton communities and underscore the importance of integrating climate extremes into monitoring and adaptive management.

Source: The Science of The Total Environment

DOI: 10.1016/j.scitotenv.2025.180837