A case for multi-model and multi-approach based event attribution: The 2015 European drought