The meaning and applicability of kernel methods in environmental research

The methods of primary quantitative processing of data series for the targeted identification of significant trends, including smoothing, filling in gaps, and detecting fluctuations in the level of values, are considered. The emphasis is placed on the ideological similarity of processing methods from different fields of knowledge - the use of sliding window technology, in which local processing of initial values and the formation of a number of values with new properties takes place. Such methods include filtering, approximation, kernel methods, etc., which help to get rid of excessive variability and identify stable relationships and dependencies. Examples of processing real data using special functions of the R language environment are given.