Spatiotemporal Dynamics of Ammonia Nitrogen and Its Response to Inflow River Based on Random Forest Model in Northwest Taihu Lake, China

Abstract

Inflow rivers have crucial impact on lake eutrophication. Ammonia nitrogen (NH4 + -N) in inflow rivers have tightly coupled relationships with water quality in Lake Taihu, China. Based on the monthly concentrations from 2009 to 2015, this study focused on the spatiotemporal dynamics of NH4 + -N in northwest Taihu Lake and utilised random forest (RF) model to simulate its response to inflow rivers. Results indicated that (1) in northwest Lake Taihu, the spatiotemporal distribution patterns of NH4 + -N concentrations are distinct, and river inputs were the major source of NH4 + -N loadings. (2) Scenario simulation results in RF models indicated that the inflow loads should be controlled under 6427.38, 3248.01, 2206.92 and 1107.58 ta -1 for the protective targets. (3) In 2015, NH4 + -N concentrations in four lake regions and NH4 + -N loads from inflow rivers were consistent with the simulation results in RF models. The NH4 + -N concentrations decreased approximately 6%, 34%, 44.44% and 6.67% in North Zone and Zhushan, Meiliang and Gonghu Bays, respectively, despite reduction in external NH4 + -N loads by 8.99%, 11.41%, 51.38% and 62.87%. This study provides further understand on quantifying the reduction of nutrient loading from inflow rivers and governing Lake Taihu or other typical eutrophic lakes.