Changes And Prediction Of Water Quality Factors In Litopenaeus Vannamei Aquaculture Ponds

Litopenaeus vannamei is one of the main species of shrimp cultured in our country.During the breeding period,if the water quality deteriorates due to poor management,and then induces bacterial or viral disease outbreaks,it will greatly affect the production of prawns.Therefore,it is necessary to strengthen the monitoring of water quality in aquaculture and predict the water quality in advance so that control measures can be taken in time to effectively reduce risks and improve aquaculture efficiency.In 2021 and 2022,water samples from greenhouse culture ponds,open-air aquaculture ponds and water diversion channels outside the farm were tracked and tested in the greenhouse culture pond,open-air aquaculture pond and water diversion channel outside the farm of Fengxian District Aquaculture Cooperative B,including suspended solids（SS）,water temperature（T）,dissolved oxygen（DO）,p H,total nitrogen（TN）,ammonia nitrogen（TAN）,nitrite nitrogen（NO₂^--N）,nitrate nitrogen(（NO₃^--N）,total phosphorus（TP）,active phosphorus（AP）and chemical oxygen demand（COD）.Firstly,the water quality changes of greenhouse ponds,open-air ponds and diversion channels were analyzed,and the differences between water quality factors of human-intervention ponds and non-human-intervention water-diversion rivers were compared.At the same time,the quality of pond aquaculture tailwater was tested;The correlation between water quality factors in aquaculture ponds was studied.Based on the water quality test data of Litopenaeus vannamei open pond in aquaculture cooperative A in the early stage of 2014-2018,and the test data of cooperative B in 2021 and 2022,establish a single LSTM and PCA-LSTM water quality prediction model,Eight water quality indexes,including water temperature（T）,dissolved oxygen（DO）,chemical oxygen demand（COD）,total phosphorus（TP）,total nitrogen（TN）,ammonia nitrogen（TAN）,nitrite nitrogen（NO₂^--N）and nitrate nitrogen（NO₃^--N）,were selected as inputs,and chemical oxygen demand（COD）and ammonia nitrogen（TAN）were selected as prediction indicators,and a Long short-term memory（LSTM）model was established,the reliability of the model prediction was confirmed by the fitting comparison of the time series and the cross-test of the two aquaculture farms.In order to improve the prediction accuracy of the model,the principal component analysis method is used to extract and reduce the dimensionality of the data,and the LSTM model based on principal components analysis（PCA）is established,and the prediction effect of the two models is compared and analyzed.The main results are as follows:In the process of greenhouse culture in 2021,the open-culture process in 2021 and2022,the water quality indexes of all ponds basically meet the growth needs of Litopenaeus vannamei.Single factor ANOVA test found no significant difference between SS,p H,TN,TP,COD,T,DO,TAN,NO₂^--N,NO₃^--N and AP and no significant difference between the three sampling points of the diversion channel using the principal component analysis,and the results showed the most significant differences in temperature,dissolved oxygen and chemical oxygen demand.Except for suspended matter and total nitrogen,the tail water under different aquaculture modes meets the discharge standards of freshwater aquaculture.Pearson The results of correlation analysis and multiple stepwise regression analysis showed a significant correlation between TAN,COD and T,between SS,TN and NO₂^--N;and between TN,TAN and NO₂^--N in open-air aquaculture ponds.The two analysis methods were verified with each other,and the results were relatively consistent.Based on the water quality data of the No.1 pond in cooperative A from 2014 to2017,A single LSTM model and PCA-LSTM model were established,and the parameters of the established model were optimized.At the same time,the data of Pond 1 in Cooperative A in 2018,the data of Pond 2 in Cooperative A in 2018,Pond 2in Cooperative B in 2021 and Pond No.3 in Cooperative B in 2022 were used for model verification,and the prediction effect of the model was compared horizontally and vertically.At the same time,the PCA-LSTM model was compared with the single LSTM model to verify the quality of the model.The results show that the evaluation index values of the four ponds of the PCA-LSTM model are less than that of the single LSTM,and the prediction results of the PCA-LSTM model are better than that of the single LSTM model,which can be used to predict the water of COD and TAN in the ponds of Litopenaeus vannamei.