River water temperature prediction method based on LSTM deep learning

Abstract

The invention discloses a river water temperature prediction method based on LSTM deep learning, and belongs to the technical field of hydrological models, and the method comprises the steps: collecting the historical day-by-day temperature, flow and water temperature data of a certain river; dividing the time sequence into a training set and a test set, and performing regularization processing onthe data; constructing a water temperature prediction model based on a long short-term memory (LSTM) neural network by taking the air temperature and the flow as input factors and taking the water temperature as an output factor based on the training set data; inputting the test set air temperature and flow into the trained model, and obtaining a test set water temperature prediction result through inverse regularization processing; and comparing a water temperature prediction value and an observation value of the test set, and verifying the rationality of the model by taking a mean absoluteerror MAE, a root-mean-square error RMSE, a Nash-Sutcliffe efficiency coefficient NSE and a determination coefficient R2 as test standards. The river water temperature prediction method is high in precision and reliable in result, has superiority compared with an existing machine learning river water temperature prediction method, and provides scientific support for river ecosystem management andprotection.

Description

technical field [0001] The invention belongs to the technical field of hydrological models, and in particular relates to a river water temperature prediction method based on LSTM deep learning. Background technique [0002] Water temperature is an important water quality factor and river habitat variable in river ecosystems, affecting aquatic organisms and biogeochemical processes. Reliable water temperature prediction is of great significance for environmental impact assessment and aquatic ecosystem protection. [0003] Water temperature models are mainly divided into physical models and statistical models. The physical model focuses on describing the physical mechanism of the river heat exchange process, which has high accuracy and interpretability, but requires the collection of a large amount of basic data. The statistical model uses statistical analysis, data mining, etc. to construct the relationship between water temperature and air temperature, flow and other varia...

AI Technical Summary

Problems solved by technology

[0004] Because water temperature is affected by many factors such as climate change and human activities, the accuracy of current traditional machine learning water temperature prediction methods is limited

[0005] In recent ye

Images