Моделирование интенсивности засухи в горных районах с использованием метеорологических параметров

Abstract

Listen

Translate article

Studying the intensity of drought in mountainous areas is critical for water resource management, disaster forecasting and the development of sustainable development strategies. Meteorological parameters play a key role in the formation of arid conditions in the mountains, so their analysis is necessary to understand and predict drought. Research methods and materials. The study used data on meteorological parameters (temperature, precipitation, humidity, wind speed) from the NASA POWER project and data on drought intensity from the US Drought Monitoring (USDM). Machine learning algorithms “random forest” and “decision tree” were used to build models of drought intensity classification. The results of the research. The constructed models have demonstrated high accuracy in predicting the intensity of drought. The analysis of the importance of the signs revealed the importance of surface pressure (PS), specific humidity (QV2M) and wind speed (WS10M, WS10M_MAX, WS10M_MIN, WS10M_RANGE) for drought forecasting. Experiments with a reduced set of factors have confirmed the possibility of constructing accurate models based on a limited number of parameters. Discussion of research results. The high accuracy of the models indicates the effectiveness of using meteorological data and machine learning algorithms to predict the intensity of drought in mountainous areas. The resulting models can serve as a tool for assessing drought risks, developing mitigation measures and managing water resources in mountainous areas. Conclusion. The study confirmed the possibility of effective modeling of drought intensity in mountainous areas based on meteorological parameters. The key factors influencing the intensity of drought have been identified and the possibility of constructing accurate models using a limited data set has been demonstrated. Conclusions of the article. The use of meteorological data and machine learning algorithms makes it possible to create effective tools for predicting the intensity of drought in mountainous areas. Suggestions for practical application and direction of future research. The obtained models can be used in drought monitoring and early warning systems, as well as to optimize water resources management in mountainous areas. Further research may be aimed at taking into account climatic zones in modeling the intensity of drought. Dividing the data into separate climatic zones can improve the accuracy of forecasting and improve understanding of the impact of meteorological factors on drought in different climatic conditions.