Workshop "Towards Accurate Lightcones for Cosmology"

18 Oct 2016, 08:00 → 20 Oct 2016, 18:00 Europe/Berlin

MIAPP Building, Seminar Room (Excellence Cluster Universe, Garching)

Stefan Hilbert (LMU)

The Excellence Cluster Universe organises a workshop "Towards accurate lightcones for cosmology", which will take place during 18-20 October 2016 in Garching.

Rationale
Large surveys of galaxies and gravitational lensing provide us with excellent data on the large scale structure (e.g. galaxy and gravitational shear correlations) with very small statistical uncertainties, which is essential for a better understanding of the nature and properties of dark energy and dark matter, the evolution galaxies, and the origin of cosmic structure. To properly interpret these data, model predictions for these quantities must be very accurate (e.g. stage-IV dark energy experiments require sub-percent accurate predictions for galaxy and lensing correlations). Currently, lightcone and gravitational lensing simulations are crucial to obtain these predictions. Thus, a systematic approach for developing, testing, and validating lightcone and lensing simulation software and their outputs is required.

Aims
The workshop brings together leading experts in the field of generating and analyzing galaxy and gravitational lensing lightcones from cosmic structure formation simulations to address the following questions:

• Which tests are required to quantify the accuracy of simulated lightcones and lightcone generation codes for their application in cosmology?
• What is the level of accuracy of existing lightcone simulation implementations?
• For a given target accuracy, what is the accuracy required for the input from structure formation simulations,
• which physical and observational effects need to be taken into account, and
• what numerical schemes are feasible?

Format:
Before the worksop, participants are asked to engage in preparatory discussions of, e.g., the first steps for testing and validating lightcones and lightcone simulation software, as well as matters of data input and output formats. Participants are then provided with common data from structure formation simulations, are encouraged to produce lightcones from these data, and to analyse the outputs. At the workshop, participants will discuss their lightcone simulations. Participants will also draft the outline of a joint publication, which will summarize these results, and which will be jointly written by the participants after the workshop.

If you are interested to attend, please register. We encourage you to submit a title and abstract if you like to present a talk.

There will be no fee.

Scientific Organising Committee:
Pablo Fosalba,
Stefan Hilbert,
Jens Jasche,
Alexander Knebe,
Frazier Pearce

Slides

• Bose_lightcones2016.pdf
• Carbone_Euclid_CSWG_LightCone_Garching_Oct2016.pdf
• Castander_MICE_short161018.pdf
• Giocoli_hmWeakLensingMunch.pdf
• Jasche_talk.pdf
• Lightcone_Workshop_2016_Introduction.pdf
• Munari.pdf
• Nelson_talk.pdf
• Rodriguez-Torres_Garching_workshop.pdf
• Santiago_Avila_short.pdf
• Smith_Alex_lightcone_talk.pdf
• Stadel_Lightcone.pdf

Tuesday 18 October

09:00 → 09:30 Registration

09:30 → 10:30 day 1 morning session 1

09:30 Welcome

Welcome, practical information, and outline of workshop

Speaker: Stefan Hilbert (LMU)

10:00 Lightcones - Introduction

topical introduction to workshop

Speaker: Stefan Hilbert (LMU)

Slides Lightcone_Workshop_2016_Introduction.pdf

10:30 → 11:00 day 1 morning break

11:00 → 12:30 day 1 morning session 2

11:00 An Algorithm for the Generation of Continuous Light Cones in Simulations

I will discuss the method employed in our recent 2 trillion particle simulation for generating a light cone where the look-back time and comoving distance from the observer is an exact linear function for every particle. While such an approach may seem to be extremely costly at face value, it turns out to be quite efficient for reasonably large cosmological volumes. I will also discuss our approach to handling this 240 TB of raw data during analysis and what direction future improvements could take.

Speaker: Dr Joachim Stadel (University of Zurich)

Slides Stadel_Lightcone.pdf

11:30 Ray-Ramses: a code for ray-tracing on the fly with N-body simulations (1601.02012)

We present a ray tracing code to compute integrated cosmological observables on the fly in AMR N-body simulations. Unlike conventional ray tracing techniques, this code takes full advantage of the time and spatial resolution attained by the N-body simulation by computing the integrals along the line of sight on a cell-by-cell basis through the AMR simulation grid. Moreover, since it runs on the fly in the N-body run, our code can produce maps of the desired observables without storing large (or any) amounts of data for post-processing, which is a welcomed feature in light of upcoming large scale structure surveys that will require several mock ray tracing maps to calibrate/validate analysis pipelines. This code, with its less conventional ray-tracing approach, can also be used in cross-checks of the more conventional methods, which can be important in tests of theory systematics in preparation for the upcoming observational missions. The Ray-Ramses routines are currently implemented in the RAMSES N-body code and can and have already been used in cosmological analysis of weak lensing, SZ and ISW effects, in both GR and modified gravity cosmologies. In this talk, we will outline the main operations of the code, present a few validation tests and application to the study of weak-lensing with modified gravity, as well as pinpointing future developments and improvements of the current code implementation.

Speaker: Mr Sownak Bose (Institute for Computational Cosmology)

Slides Bose_lightcones2016.pdf

12:00 An alternative route for high order lensing

While doing predictions for gravitational lensing using first order equations greatly simplifies the calculations, it has been shown that the approximations involved can not reach the accuracy required by the next generation of galaxy surveys. Second order effects must be included in lensing predictions. Since this is new territory, it is important that we do not focus only on one technique, but that explore several techniques until we find the most appropiate. In other words, the only way to fully understand the biases associated to a particular technique is to compare with alternatives. I will present a new method for making lensing predictions in the context of N-body cosmological simulations, which uses only geometrical information of ray bundles. Thus, the method does not rely on an expansion including effects at different orders, but includes all the high orders at once. The implementation of the algorithm in the RAY-RAMSES is in its testing phase.

Speaker: Dr Claudio Llinares (Durham University)

12:30 → 14:00 day 1 lunch break

14:00 → 15:30 day 1 afternoon session 1

14:00 A Lightcone Catalogue from MXXL

With upcoming galaxy surveys, such as DESI and Euclid, it is important to have realistic mock catalogues which can be used, for example, to test methods of measuring statistics from the survey and removing systematic effects. For this purpose, we have created a full sky mock catalogue that is complete to r=20 from the Millennium-XXL (MXXL) simulation. We have generated a MXXL halo lightcone catalogue, which has then been populated with galaxies using a halo occupation distribution (HOD) scheme. The HODs we use are based on the SDSS HODs from Zehavi 2011, and we evolve these with redshift in order to produce the GAMA luminosity function at higher redshifts. Each galaxy in the catalogue is assigned an r-band absolute magnitude and g-r colour, and this is based on the methods described in Skibba 2006 and Skibba 2009. We have extended these methods for a 5 parameter HOD, and to produce a distribution of colours which is in agreement with GAMA at each redshift. The catalogue is able to reproduce the luminosity function and clustering seen in SDSS at low redshifts, and GAMA at high redshifts.

Speaker: Mr Alexander Smith (Durham University)

Slides alex_smith_lightcone_talk.pdf

14:30 MICE simulations

We will summarize the MICE simulations and the way we have built a lightcone in the dark matter distribution and for the lensing observables

Speaker: Dr Francisco Javier Castander (ICE, IEEC-CSIC)

Slides Castander_MICE_short161018.pdf

15:00 Requirements for Euclid galaxy mocks

Speaker: Dr Katarina Markovic (ICG Portsmouth)

15:30 → 16:00 afternoon break

16:00 → 17:30 day 1 afternoon session 2

16:00 Group discussions

Wednesday 19 October

09:00 → 10:30 day 2 morning session 1

09:00 Born and post-Born full-sky lensing ray-tracing in N-Body simulations

I discuss Born and post-Born full-sky lensing ray-tracing in N-Body simulations.

Speaker: Dr Carmelita Carbone (INAF - Milan)

Slides Carbone_Euclid_CSWG_LightCone_Garching_Oct2016.pdf

09:45 Bayesian Inference with physical models of the cosmic Large Scale Structure

Next generation cosmological surveys will provide an avalanche of cosmological observations. This increase in scientific data needs to be accompanied with the development of novel information processing techniques to interpret such observations. Analyses of three dimensional inhomogeneous large scale structures require to jointly account for complex dynamical processes, associated to the gravitational formation history of the dark matter distribution, together with systematic as well as stochastic effects arising from observational procedures. We address this problem of large scale Bayesian inference from cosmological data sets via our recently developed BORG algorithm. This approach incorporates three dimensional physical models of cosmological structure formation immediately into a Bayesian inference framework. Detailed physical modelling of data permits us to increase the signal to noise ratio of inferred quantities by at least one order of magnitude. In conclusion, our approach provides new avenues to study the full four dimensional state of our Universe in cosmological data.

Speaker: Dr Jens Jasche (TUM)

Slides Jasche_talk.pdf

10:30 → 11:00 day 2 morning break

11:00 → 12:30 day 2 morning session 2

11:00 Halo Model Characterisation for Weak Lensing Predictions

Speaker: Dr Carlo Giocoli (Department of Physics and Astronomy - Alma Mater Studiorum UniBO)

Slides Giocoli_hmWeakLensingMunch.pdf

11:45 Fast generation of approximate halo catalogues on PLC with PINOCCHIO

Pinocchio is a tool to generate in an approximate but fast way catalogs of DM halos. The new version comes with on-the-fly production of halo catalog on the past light cone with continuous time sampling.

Speaker: Emiliano Munari (INAF Osservatorio Astronomico Trieste)

Slides Munari.pdf

12:30 → 14:00 day 2 lunch break

14:00 → 15:30 day 2 afternonn session 1

14:00 Lightcones, mock catalogues and photometric surveys

Speaker: Dr Santiago Avila (Institute of Cosmology and Gravitation)

Slides Santiago_Avila_short.pdf

14:30 SUrvey GenerARtor (SUGAR)

Speaker: Mr Rodríguez-Torres Sergio (Instituto de Física Teórica/Departamento de Física Teórica UAM/CSIC)

Slides Rodriguez-Torres_Garching_workshop.pdf

15:00 Fast, approximate NISP simulations

Speaker: Dr Katarina Markovic (ICG Portsmouth)

15:30 → 16:00 day 2 afternoon break

16:00 → 17:30 day 2 afternoon session 2

16:00 group discussions

19:00 → 21:00 Workshop Dinner

Thursday 20 October

09:00 → 10:30 day 3 morning session 1

10:00 The Illustris Simulations

Speaker: Dr Dylan Nelson (MPA)

Slides Nelson_talk.pdf

10:30 → 11:00 day 3 morning break

11:00 → 12:30 day 3 morning session 2

11:00 Group discussions / hack session

12:30 → 14:00 day 3 lunch break

14:00 → 15:30 day 3 afternoon session 1

14:00 Discussion of future plans

Speaker: Stefan Hilbert (LMU)

15:00 Workshop Summary

Speaker: Stefan Hilbert (LMU)