Modeling And Prediction Of Chlorophyll A In Lake Taihu Based On Bayesian Network

As the global climate continues to warm,the problem of algal blooms in lakes has become an important environmental issue worldwide.In the field of lake ecology,Bayesian network models use probability to describe the correlation between factors,which can solve the problem of inconsistency or even opposition among ecological data.With this unique advantage,Bayesian network has become an effective research tool to predict algal bloom outbreak.As a typical eutrophic shallow lake in China,the problem of algal bloom is particularly prominent in Lake Taihu.In this thesis,a Bayesian network chlorophyll a simulation model was constructed to simulate and study the chlorophyll a in Lake Taihu under different climatic and nutrient conditions based on the water quality and meteorological actual measurement data from Taihu Laboratory from 1992 to 2020,using Bayesian network model to simulate lake biological indicators.In this thesis,we first analyzed the long-term trends of water quality and meteorological indicators in Lake Taihu by the generalized additive model,and found that the water quality and climate of Lake Taihu have strong seasonal differences.Combining with the seasonal characteristics of algal growth in Lake Taihu,Bayesian network models were built for each of the four seasons.From the analysis of the model’s own structure,the simulation found that the contribution of climate indicators to chlorophyll a was greater than that of total phosphorus concentration in spring and winter;in summer and autumn,the contribution of total phosphorus concentration to chlorophyll a was greater than that of climate indicators.In spring and winter,the climate is the main factor limiting the growth of algae;in summer and autumn,the climate is suitable for algae growth and total phosphorus concentration is the main factor limiting the reproduction of algae.The thesis then used the model to quantify the contribution of environmental elements to historical chlorophyll a in Lake Taihu,and found that Lake Taihu chlorophyll a was most sensitive to changes in the external environment when total phosphorus concentration was 0.135 mg/L and climate indicator was 6.35 ℃·s/m in spring,phosphorus concentration was 0.231 mg/L and climate indicator was11.48 ℃·s/m in summer,and phosphorus concentration was 0.136 mg/L and climate indicator was 7.00 ℃·s/m in autumn.The algae growth and reproduction are the most efficiency under this condition.Finally,based on the climate projections in the IPCC,the model was used to simulate the nutrient control targets in the context of future climate change,and it was concluded that to control the chlorophyll a in Lake Taihu in the next 20 years at a low risk level of water bloom outbreak(the probability of chlorophyll a concentration below the historical 75% quantile value is greater than 75%),the total phosphorus concentration in Lake Taihu should be controlled below 0.024 mg/L in spring,below0.118 mg/L in summer and below 0.129 mg/L in autumn,which needs to be reduced by78.57%,39.80% and 26.29% based on the average total phosphorus concentrations of0.112 mg/L,0.196 mg/L and 0.175 mg/L in spring,summer and autumn of Lake Taihu in the past two years,respectively.This thesis has 32 figures,4 tables and 101 references.