I just got back from the “nIFTy” Cosmology workshop, which took place at the IFT (Instituto de Física Teórica) of the Universidad Autonoma de Madrid. It was organized primarily by Alexander Knebe, Frazer Pearce, Gustavo Yepes, and Francisco Prada. As usual, it was a very international workshop, which could’ve been interesting in the context of the World Cup, except that most of the participants’ teams had already been eliminated before the workshop began! In spite of Spain’s early exit, the stadium of Real Madrid (which I visited on a day of sightseeing) was nonetheless a popular tourist spot. I also visited the Prado museum, which had an interesting painting by Rubens involving the Milky Way.

This was one of a series of workshops and comparison projects, and I was involved in some of the previous ones as well. For example, following a conference in 2009, some colleagues and I compared measures of galaxy environment—which are supposed to quantify to what extent galaxy properties are affected by whether they’re in clustered or less dense regions—using a galaxy catalog produced by my model. (The overview paper is here.) I also participated in a project comparing the clustering properties of dark matter substructures identified with different methods (here is the paper). Then last year, colleagues and I participated in a workshop in Nottingham, in which we modeled galaxy cluster catalogs that were then analyzed by different methods for estimating masses, richnesses and membership in these clusters. (See this paper for details.)

This time, we had an ambitious three week workshop in which each week’s program is sort of related to the other weeks. During the first week, we compared codes of different hydrodynamical simulations, including the code used by the popular Illustris simulation, while focusing on simulated galaxy clusters. In week #2, we compared a variety of models of galaxy formation as well as models of the spatial distributions and dynamics of dark matter haloes. Then in week #3, we’re continuing the work from that Nottingham workshop I mentioned above. (All of these topics are also related to those of the conference in Xi’an that I attended a couple months ago, and a couple other attendees were here as well.)

The motivation of these workshops and comparison workshops is to compare popular models, simulations, and observational methods in order to better understand our points of agreement and disagreement and to investigate our systematic uncertainties and assumptions that are often ignored or not taken sufficiently seriously. (This is also relevant to my posts on scientific consensus and so-called paradigm shifts.)

Last week, I would say that we had surprisingly strong disagreement and interesting debates about dark matter halo masses, which are the primary drivers of environmental effects on galaxies; about the treatment of tidally stripped substructures and ‘orphan’ satellite galaxies in models; and various assumptions about ‘merger trees’ (see also this previous workshop.) These debates highlight the importance of such comparisons: they’re very useful for the scientific community and for science in general. I’ve found that the scatter among different models and methods often turns out to be far larger than assumed, with important implications. For example, before we can learn about how a galaxy’s environment affects its evolution, we need to figure out how to properly characterize its environment, but it turns out that this is difficult to do precisely. Before we can learn about the physical mechanisms involved in galaxy formation, we need to better understand how accurate our models’ assumptions might be, especially assumptions about how galaxy formation processes are associated with evolving dark matter haloes. Considering the many systematic uncertainties involved, it seems that these models can’t be used reliably for “precision cosmology” either.