Perturbed parameter ensembles (PPEs) are based on a single climate model where each ensemble member samples a different set of values for some key parameters according to a plausible range elicited from the parameterisation experts. PPEs have been used in climate projections like those that underpinned the 2009 UK Climate Projections (UKCP09), to allow impact studies and adaptation planners to account for parametric modelling uncertainty. But the PPEs also provide a rich source of information about parameter sensitivities and for identifying model biases that are systematic regardless of the parameter values. This information is really only starting to be exploited; in particular, it can help model developers prioritise their work using evidence of those systematic biases that have a link to important projection variables across the PPE. With the next set of UK Climate Projections, which are based on a PPE of the Met Office Hadley Centre’s latest climate model, due out in 2018, this talk will show that this PPE provides a complementary database for research to that provided by the CMIP multimodel ensembles.