Abstract
The paper examines the development of forecasting and modeling technologies for environmental processes using classical and quantum data analysis methods. The main focus is on the integration of deep neural networks and classical algorithms, such as AutoARIMA and BATS, with quantum approaches to improve the accuracy of forecasting environmental parameters. The research is aimed at solving key problems in environmental monitoring, particularly insufficient forecast accuracy and the complexity of processing small data with high discretization. We developed the concept of an adaptive system for predicting environmental conditions in urban agglomerations. Hybrid forecasting methods were proposed, which include the integration of quantum layers in LSTM, Transformer, ARIMA, and other models. Approaches to spatial interpolation of environmental data and the creation of an interactive air pollution simulator based on the A* algorithm and the Gaussian kernel were considered. Experimental results confirmed the effectiveness of the proposed methods. The practical significance lies in the possibility of using the developed models for operational monitoring and forecasting of environmental threats. The results of the work can be applied in environmental information systems to increase the accuracy of forecasts and adaptability to changing environmental conditions.