In situ visualization, i.e., computing images while the simulation is running, requires automatically choosing visualization parameters, such as isovalues for isosurface extraction. This image compares the common approach of selecting equally spaced values across the value range (left) to selecting isovalues based on the contour tree right. Both panels visualize a simulation of an ion accelerator using the WARP code (warpx.readthedocs.io/). Topological analysis makes it possible to choose isovalues that capture the important structure of the simulation data much better compared to the simple approach of using equally spaced isovalues.