Topology-controlled volume rendering has proven to be a useful tool for exploration of volumetric data by highlighting the global, high-level structure of data sets. However, topological analysis is difficult to parallelize on distributed memory systems – and thus to utilize for in situ visualization – due to the global nature of topological descriptors. This chapter presents and evaluates a task-parallel formulation of topology-controlled volume rendering applicable to visualization of large scalar field data. It evaluates previous efforts towards parallel topology extraction and introduces a distributed computation schema for augmented contour trees. Through data partitioning into rectilinear blocks, the algorithm is designed to be in-situ suitable. The use of a task-parallel framework aims at latency hiding and dataflow-specific scheduling. It thereby also allows for combining contour tree computation and subsequent volume rendering. The technique divides the scalar field with separate transfer functions according to the branch decomposition of the full data set while each local block only has to keep track of its own vertex augmentation. Beyond describing the approach and its implementation in the task-parallel framework HPX, initial experiments on scaling behaviour are presented.