Sparse Bayesian mass-mapping using trans-dimensional MCMC

Uncertainty quantification is a crucial step of cosmological mass-mapping that often ignored. Suggested methods are typically only approximate or make strong assumptions Gaussianity the shear field. Probabilistic sampling methods, such as Markov chain Monte Carlo (MCMC), draw samples form probability distribution, allowing for full and flexible uncertainty quantification, however these notoriously slow struggle in high-dimensional parameter spaces imaging problems. In this work we use, first time, trans-dimensional MCMC sampler mass-mapping, promoting sparsity wavelet basis. This gradually grows space required by data, exploiting extremely sparse nature mass maps space. The coefficients arranged tree-like structure, which adds finer scale detail grows. We demonstrate on galaxy cluster-scale images where planar modelling approximation valid. high-resolution experiments, method produces naturally parsimonious solutions, requiring less than 1% potential maximum number still producing good fit to observed data. presence noisy better reconstruction mass-maps standard smoothed Kaiser-Squires method, with addition uncertainties fully quantified. opens up possibility new inferences about dark matter using data from upcoming weak lensing surveys Euclid.