Extreme temperatures detection and attribution related to external forcing in Madagascar

Abstract

In this article, we use standard extreme temperature indices to detect and attribute external forcing in Madagascar. These indices are calculated from observations and multi-model ensemble mean responses on anthropogenic-plus-natural (ALL), greenhouse gases (GHG), natural (NAT) and anthropogenic (ANT) which are subtracted from ALL and NAT forcings over 1950–2018. Correlation analysis emphasizes that the observed changes are more influenced by ENSO events, especially in minimum temperature. The observed changes are regressed or combined with model simulations from the sixth phase of the Coupled Model Inter-comparison Project (CMIP6) to assess human impacts in indices. CMIP6 models with ALL, GHG and ANT forcings correspond well with the observations for the frequency indices than the intensity indices. Moreover, decadal trends indicate the existence of anthropogenic warming according to observations and multi-model ensembles with ALL, GHG and ANT forcings. Detection and attribution parties identify and justify the causes of the observed changes. We do this by performing the single-signal and two-signal analysis using the Regularized Optimal Fingerprinting (ROF) method with Total Least Square (TLS) regression. We estimate internal climate variability by means of control model simulations. As a result, we note an inconsistency in the warming trend with the NAT forcing. The influence of ALL, GHG and ANT forcings is detectable for standard extreme temperature indices between 1950 and 2018. Nearly, observed changes are attributed to GHG and ANT forcings except for coldest night and warm nights in Madagascar.