Identification of precipitation and maximum temperature patterns in Spain during 1951 to 2021 using clustering based on multiscale analysis of time series

Abstract

Spain is a territory with spatial variations highly influenced by its distance from the sea and its complex orography, where it is possible to note an uneven distribution of both temperature and precipitation. This study presents an analysis of trends in maximum temperature and precipitation by zone over the period 1951-2021 using monthly data. The database used includes 16156 multivariate time series (maximum temperatures and precipitation) corresponding to different areas of the Spanish territory, distributed over a grid of 5x5km2. The methodology used starts by reducing the dimensionality of the time series and with this version are clustered using an approach based on multiscale analysis using a clustering algorithm. In the following, the prototypes of each group are defined, which allows to identify and analyse patterns of change in maximum temperatures and precipitation by zones. An increase in average maximum temperature has been identified in eight zones distributed in Spain from 1951 to 2021. The rate of change of maximum temperature was between 0.060ºC and 0.2155ºC per decade. Areas further south showed a higher rate of increase than areas found in the north. It has been observed that May was the month with the highest variation for all areas in maximum temperature, nevertheless, differences in seasonal variation are evident when passing from one zone to other, as in some there is greater variation in spring months and in others in winter months. An analysis of trends and seasonal variations of precipitation in the identified zones will be carried out and the correlation between patterns of maximum temperature and precipitation will be studied in each of the eight zones.