Representation results for absolutely continuous curves $μ:[0,T]\to \mathcal{P}_p(\mathbb{R}^d)$, $p>1$, with values in the Wasserstein space $(\mathcal{P}_p(\mathbb{R}^d),W_p)$ of Borel probability measures in $\mathbb{R}^d$ with finite $p$-moment, provide a crucial tool to study evolutionary PDEs in a measure-theoretic setting. They are strictly related to the superposition principle for measure-valued solutions to the continuity equation. This paper addresses the extension of these results to the case $p=1$, and to curves $μ:[0,+\infty)\to\mathcal{P}_1(\mathbb{R}^d)$ that are only of bounded variation in time: in the corresponding continuity equation, the flux measure $ν\in\mathcal{M}_{loc}([0,+\infty)\times\mathbb{R}^{d};\mathbb{R}^{d})$ thus possesses a non-trivial singular part w.r.t. $μ$ in addition to the absolutely continuous part featuring the velocity field. Firstly, we carefully address the relation between curves in ${\rm BV}_{loc}([0,+\infty);\mathcal{P}_1(\mathbb{R}^d))$ and solutions to the associated continuity equation, among which we select those with minimal singular (contribution to the) flux $ν$. We show that, with those distinguished solutions it is possible to associate an `auxiliary' continuity equation, in an augmented phase space, solely driven by its velocity field. For that continuity equation, a standard version of the superposition principle can be thus obtained. In this way, we derive a first probabilistic representation of the pair $(μ,ν)$ solutions by projection over the time and space marginals. This representation involves Lipschitz trajectories in the augmented phase space, reparametrized in time and solving the characteristic system of ODEs. Finally, for the same pair $(μ,ν)$ we also prove a superposition principle in terms of BV curves on the actual time interval, providing a fine description of their behaviour at jump points.