Spatial and Temporal Variability of Precipitation and Temperature: Analysis of Recent Changes and Future Development with Use of the Weather Generator and RCM-Based Climate Change Scenarios

While much effort has been devoted to analyzing long-term changes of temperature and precipitation in mean values and extremes, studies on changes in variability have been rather scarce. Trends in variability are, however, important, among others because their interaction with trends in mean values determines the degree with which extremes would change. The knowledge of long-term changes in temporal variability is essential for assessments of climate change impacts on various sectors, including hydrology (floods and droughts), agriculture, health, and energy demand and production.

SPAGETTA is a stochastic spatial daily weather generator (WG), which uses first-order multivariate (dimension = number of variables X number of gridpoints) autoregressive model to represent the spatial and temporal variability of surface weather variables (including precipitation and temperature). We consider the generator to be a suitable tool for assessing changes in the spatial and temporal variability of the weather series because of following reasons: (A) The inter-gridpoint lag-0 and lag-1(day) correlations included in a set of WG parameters may serve as representatives for spatial and temporal variability of input weather variables. (B) Statistical significance of changes in the lag-0 and lag-1 correlations derived from the input series may be easily assessed by comparing the changes with a variability of the lag-0 and lag-1 correlations related to the stochasticity in input weather series (the variability is assessed across a set of multiple realisations of the synthetic series). (C) Separate effects of changes in various statistical characteristics on any climatic characteristic may be easily assessed. Specifically, having analysed changes in the means, variability and inter-gridpoint correlations (e.g. based on RCM simulations of the future climate), we may modify only a selected (possibly only a single one) WG parameter(s) before producing the synthetic series and analysing effect of climate change on the climatic characteristics.

In the first part of the contribution, we employ SPAGETTA generator to analyse changes in interdiurnal variability of precipitation and temperature in 8 European regions (defined in Dubrovsky et al 2020, Theor Appl Climatol) using (a) gridded observational (last N years vs. first N years in available E-OBS times series) and (b) RCM-simulated surface weather series (2070-2099 vs 1971-2000; outputs from 19 RCMs available from the CORDEX database are analysed). In doing this, we assess the statistical significance of the detected changes. In the second part, we assess separate effects of changes in the means, variability and lag-0 & lag-1 correlations of temperature and precipitation (the changes based on a set of 19 RCM simulations are used to modify the corresponding WG parameters) on a set of climatic indices - including a set of compound precipitation-temperature characteristics representing spells of days with spatially significant extent of significantly non-normal weather (e.g. hot-dry spells).