TEMPERATURE AND PRECIPITATION FORECAST FOR WESTERN TRANSYLVANIAN PLAINS USING ARTIFICIAL INTELLIGENCE

Abstract

Climate change represents a growing risk factor for agriculture, climate predictions being very important for adaptation measures. The ecological frame from the Transylvanian Plain is given by the existence of interaction among a great number of factors, of which two are very important for the agro-ecosystem: the first is the thermal background with high temperature variations and low rainfall, characteristics which impose significant restrictions for crop plants and secondly it is the hill orography, with slope land. The present paper analyze the weather evolution of the farm from Agricultural Research and Development Station Turda located on Western Transylvanian Plains, using artificial intelligence. Study developed mathematical models based on data measured in the last 63 years (1957-2019). The average annual temperature for the period 1957-2019 is 9.2°C and multiannual average precipitation is 531.4 mm. From the simulation of the evolution of the main climatic factors, temperature and precipitation, in the next decade it can be observed the sharp increase of the average annual temperature and a lower fluctuation of the precipitation regime. It could be noticed that the predicted average temperature values for the summer of the next 10 years are: 17.3°C for June, 23.8°C for July and 25.9°C for August, resulting an average of 22.33°C for summer. Agriculture must adapt to the new conditions, taking some measures, the modification or improvement of current plant cultivation technologies, the use of varieties/hybrids that are more resistant or at least tolerant to prolonged drought and during periods with high temperatures, the adoption of new technologies to accumulate/preserve water in the soil for a longer period.