Abstract
In studies carried out at the Krasnodar Rice Institute, as well as according to data from world rice cultivation centers, it has been shown that evaporation has the greatest effect on rice yield from climatic factors. Evaporation without any significant errors can be taken as evaporation from the waters of water bodies, while the value of rice transpiration most often has a fairly significant correlation or functional relationship with evaporation from water bodies. From the short-term weather forecasts issued by regional subjects of Roshydromet and other organizations, it is possible to obtain forecast values of temperature, wind speed and precipitation with a five-day (pentad) advance. It was on their basis that a method for predicting evaporation for the future pentad was developed. In this paper, the hypothesis of the possibility of predicting the values of evaporation from the water surface of rice checks by the main predicted meteorological elements using neuro-machine learning technologies is tested. In the course of the work, a machine learning algorithm was compiled aimed at approximating the amount of evaporation from the water surface based on forecast data of meteorological parameters. This technology can be used as a new technique for medium-term forecasting of evaporation for the pentad period. In order to obtain the amount of evaporation, it is quite possible to obtain a forecast dependence on the values of air temperature and wind speed, provided there is no precipitation, when such a forecast is most important.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call