Operational red tide early warning method and computer readable storage medium

Abstract

The invention discloses a commercial red tide early warning method and a computer-readable storage medium. The method includes: obtaining red tide samples and performing preprocessing; screening the red tide samples to obtain fixed sample data; determining the final input variable through a genetic algorithm; constructing Neural network model; obtain temporary sample data and perform preprocessing; obtain forecast sample data according to the final input variables in the fixed sample data and temporary sample data; generate m+n forecast training data corresponding to each forecast sample data, and respectively m+n neural network models for training; obtain the latest m-day real-time data, input the previous m-1+n neural network models respectively, and combine the m-day real-time data into the m+nth neural network model, Obtain the forecast results corresponding to n days in the future; analyze the red tide occurrence probability level of that day according to the forecast results corresponding to the same day in the future. The invention can realize relatively accurate operational red tide early warning.

Description

technical field [0001] The invention relates to the technical field of red tide forecasting, in particular to a business red tide early warning method and a computer-readable storage medium. Background technique [0002] Red tide is an abnormal phenomenon in marine ecosystems, which is caused by the explosive proliferation of red tide algae in the seaweed family under specific environmental conditions. Seaweed is a huge family, except for some large seaweeds, many are very tiny plants, and some are single-celled plants. Depending on the species and number of organisms that cause red tides, seawater sometimes appears in different colors such as yellow, green, and brown. Red tides not only cause serious harm to the marine environment, marine fisheries and marine aquaculture, but also have certain impacts on human health and life safety. With the rapid development of industrial and agricultural production, marine environmental pollution, trophication of water bodies in river ...

AI Technical Summary

Problems solved by technology

However, no matter which forecasting method is used, it is only a means of red tide forecasting. Most of the red tide forecasting methods cannot be operated operationally. Only two kinds of red tide forecasting by meteorological conditions and remote sensing methods can be used for large-scale macro forecasting applications.

Meteorological conditions are only the inducing conditions of red tides, not the main conditions. Because the red tide disaster itself is limited by many conditions, the accuracy of the red tide weather forecasting method is not high; ability, but also has its own shortcomings, for example, it cannot work around the clock, and cannot monitor red tides in rainy weather and at night. In addition, the spatial resolution is low, and it is very difficult to monitor small-scale red tides

[0005] Sufficient, time-sensitive, and quality-assured data support is the basis for the commercial implementation of red tide forecasting. Even a neural network model with a very good fit is only a statistical means of discussion. In the absence of data It is impossible to really start the red tide forecasting work

Images