Final Report Summary - DUSTPEDIA (A Definitive Study of Dust in the Local Universe (DustPedia)) Executive Summary:The European Space Agency invested heavily in two cornerstone missions; Herschel and Planck. These space observatories provided us with an unprecedented opportunity to study, at far infrared wavelengths, the cold Universe beyond our galaxy. Although these missions have now ended they have left us with a huge legacy data set that we have been able to exploit. The data has provided us with an opportunity to study cosmic dust in galaxies to answer fundamental questions about: the origin of the chemical elements, physical processes in the interstellar medium (ISM), its effect on stellar radiation, its relation to star formation and the cosmic far infrared background. To achieve our goals we have combined the Herschel/Planck data with that from many other databases that contain observations at other wavelengths (from the ultra-violet to the sub-mm) and developed our own photometric analysis package (CAAPR). We have added to this all available observations of atomic and molecular gas mass and metallicity. From this we have created the publicly available DustPedia multi-wavelength database and archive. To maximise our spatial resolution and sensitivity to cosmic dust the DustPedia archive contains 875 nearby (within about 40 Mpc) galaxies. In the course of our work we have developed both computer and physical dust grain models (THEMIS) that enabled us to relate the measured cosmic dust emission to the physical properties of the dust grains (chemical composition, size distribution, temperature) and to the origins of dust (evolved stars, super novae, growth in the ISM) and the processes that destroy it (high energy collisions and shock heated gas). To help us interpret the data we have developed our own, world leading, Monte Carlo photon tracing radiative transfer code (SKIRT). We have used SKIRT to carry out multi-wavelength studies of galaxies accounting for the prominent role that dust plays in influencing what we observe and our interpretation of physical processes in the inter-stellar medium. Using our multi-wavelength data we have created galaxy spectral energy distributions and interpreted them using a new Bayesian inference technique (HerBIE) enabling us to quantify and compare dust properties of galaxies of many different types. We have used our models to:1. Measure the emissivity of cosmic dust.2. Derive the fraction of stellar radiation absorbed by dust as a function of a galaxy's properties (morphology, stellar mass, etc.).3. Quantify the importance of the varying stellar populations that heat the dust (old and young stars).4. Consider the scaling relations (star, gas, dust mass, star formation rate (SFR), etc.) for galaxies of different morphological types.5. Compare the relative spatial distributions of stars, gas, dust and SFR and how this depends on galaxy morphology.6. Interpret the very different properties of early type galaxies.7. Understand the particular properties of low metallicity dwarf galaxies.8. Consider the relationship between gas-to-dust ratio and metallicity.9. Compare the properties of galaxies in different environments (cluster, field).10. Measure the distribution of cosmic dust above the mid-plane of edge-on galaxies.11. Relate our observational data to that produced by numerical simulations of galaxy formation and evolution.Project Context and Objectives:Four hundred thousand years after the Big Bang the Universe had cooled sufficiently to form a gas of atoms. Predominantly these atoms were of the simplest form possible (Hydrogen and Helium) with just a small fraction of heavier elements, which astronomers collectively refer to as ‘metals’. Subsequently clouds of this gas collapsed and merged together to form stars and galaxies. Since then the stars have been enriching the gas in galaxies with metals via the process of nucleosynthesis. This increasing abundance of metals was a crucial step in the path that led to the formation of the Earth and eventually intelligent life.As a galaxy evolves new stars form from the metal enriched gas, which contains about 60% of its metals directly as atoms or molecules in the gas phase and about 40%, which are contained within larger particles that astronomers refer to as cosmic dust. Cosmic dust forms by nucleation and growth from the vapour phase in the cool atmospheres of low mass stars as they come to the end of their lives and also in the gas ejected from supernovae as more massive stars expire. Once deposited into the interstellar medium the dust grains are subject to various physical processes that allow them to grow via the accretion of atoms and molecules and disintegrate in shock heated gas or via high-energy photon or cosmic ray processing. The dust is composed of a mixture of carbonaceous and amorphous silicate grains with a size distribution governed by the growth and destruction mechanisms (sizes of order 0.01-1.0μm).The cosmic dust reveals itself to us primarily in three ways. Firstly, the dust absorbs and scatters radiation from the stars, which not only causes the extinction of the light, but also causes the stellar spectrum to become ‘redder’, as bluer wavelengths are more efficiently extinguished. Secondly, because the absorbed stellar light heats the dust it subsequently radiates some of its energy away. Dust temperatures range from about 10-100K and so the energy radiated from dust is emitted predominantly in the far-infrared and sub-mm part of the electromagnetic spectrum (wavelengths of about 10μm to 1mm). Imprinted on this spectrum are signatures of dust composition, structure and chemistry. Finally, the alignment of dust gains in the Galactic magnetic field leads to the polarisation of starlight.Dust extinction and polarisation are effects imprinted on stellar radiation, while the only direct measure from the dust itself comes from the radiation it emits. Observation of this radiation had to await the arrival of space telescopes because, by far the majority of the cosmic far-infrared and sub-mm radiation is efficiently absorbed by molecules in the Earth’s atmosphere. The first far-infrared space telescope (IRAS) was launched in 1984 and it revolutionised our ideas about what the physical properties of cosmic dust are, just how much dust there was and how important the dust is in governing physical processes in the interstellar medium. We now know that for a typical galaxy like the Milky Way a little under half of the radiation produced by stars is subsequently re-processed through cosmic dust. There are other galaxies in which 99% of the stellar radiation is reprocessed in this way. Subsequent space missions (ISO, Spitzer) have greatly extended our understanding of cosmic dust. In addition, considerable steps forward have been gained through observations of dust emission from our own Galaxy by telescopes designed to observe the cosmic microwave background (COBE, WMAP, Planck). These observational advances have been greatly enhanced by the recent success of the Herschel Space Observatory (HSO), the availability of HSO data was a prime motivator for this project. The HSO provided a huge step forward because at 3.5m its collecting mirror was 5-6 times larger than any of the previous far-infrared telescopes, giving both improved sensitivity and spatial resolution. In addition, one of the major discoveries from previous far-infrared missions was that the cosmic dust was somewhat colder than expected, so the Herschel instruments were designed to look at a previously un-explored part of the electromagnetic spectrum between the far-infrared and sub-mm (250-500μm) as well as the regions previously explored with smaller telescopes (70-160μm). From the theoretical side previous to our work, we still did not have good models of the physical properties of cosmic dust, its origin and fate and how it influenced processes in the inter-stellar medium, such as star formation and radiation transfer. So, along with the analysis of observational data our intension was to develop a new physical dust grain model, a multi-wavelength photon tracing radiative transfer model of galaxies and a Bayesian spectral energy distribution fitting model to interpret our measured spectral energy distributions. By combining these parallel advances in, observational data, theoretical modelling and advanced computing techniques we have been able to make significant progress in the role cosmic dust plays in our understanding of galaxies. For example; as a depository of metals the dust content of a galaxy is a measure of how far along the evolutionary path a galaxy has progressed. Secondly, cosmic dust plays an important role in many of the physical processes that regulate the evolution of galaxies. For example, it provides opacity so that giant clouds of gas collapsing under gravity can heat up to temperatures sufficient for stars to form and nucleosynthesis to start. It is on the surface of dust grains that molecular hydrogen, which is the crucial gaseous ingredient for star formation, forms. Thirdly, dust traces other physical processes and galaxy constituents that are not as easily measured. For example, the far infrared emission from galaxies is closely related to the rate at which stars form and the relatively easily measured mass of dust relates closely to the difficult to measure mass of molecular hydrogen. Fourthly, dust can greatly affect what you measure at other wavelengths. The ultra-violet emission of hot young stars is, for example greatly attenuated by dust and may lay hidden and the reddening effect can mislead us in our determination of the ages of stellar populations. On cosmological scales the far-infrared background can be used as a measure of the star formation history of the Universe. Dust extinction through the Universe may noticeably influence our observations of the most distant objects. It is problems related to and developed from the above that have provided the main objectives for our project.These final objectives are more clearly described below by the tasks we have successfully completed:1. Extraction from the Herschel archive of data on all galaxies observed by Herschel that reside within 3000 km/s and have angular sizes greater than 1 arcmin (875 galaxies).2. Extraction from other data archives of all available data for the selected galaxies. This provides coverage across the electromagnetic spectrum from the ultra-violet to the mm.3. The production of consistent photometry on all multi-wavelength images of all of the galaxies in the sample.4. The measurement of the complete ultra-violet to mm spectral energy distribution of all the galaxies in the sample. 5. Development of a physical dust model based on observation and laboratory based experiments (grain constituents, sizes, optical properties etc.) and the use of this to interpret galaxy spectral energy distributions (the THEMIS dust model). 6. Use of the THEMIS dust model to make inferences about the origin of cosmic dust (stars, growth in the inter-stellar medium etc.) and its destruction (supernova driven shocked gas, x-ray sputtering etc.).7. The interpretation of the galaxy spectral energy distributions using full Bayesian SED models (the HerBIE and CIGALE fitting packages in conjunction with the THEMIS dust model), to derive stellar, gas and dust properties (masses, densities, temperatures etc.), star formation rates and histories as a function of morphological type. 8. Development of a versatile Monte Carlo photon tracing radiative transfer model that is adaptable to different relative star and dust spatial distributions and viewing angles (SKIRT in conjunction with the THEMIS dust model). The radiative transfer model is able to re-produce observations of galaxies from ultra-violet to mm wavelengths. This provides an excellent basis for comparing observation with the theoretical model, for example checking for consistency between observations of dust emission and dust extinction and the energy balance between the total energy produced by stars and that directly escaping plus that absorbed by cosmic dust.9. Use of the above models to study the mass and spatial distribution of cosmic dust in galaxies of different morphological types and those residing within different environments. Particularly the apparent very different relationships of these quantities for early-type galaxies (elliptical compared to spiral galaxies).10. Use of the above models to study variations in dust emissivity both within and between galaxies - changing physical properties of the dust grains and the connection to metallicity.11. Within the context of galaxy evolution to quantify the relationships between stars, gas and dust and how these change with time through the process of star formation and the subsequent production of metals. We have also compared our data to numerical models of galaxy formation and evolution.Project Results:DustPedia is a collaboration of six research institutions across Europe (Cardiff University (UK),The National Observatory of Athens (Greece), INAF-Osservatorio Astrofisico di Arcetri (Italy), Ghent University (Begium), Service d’Astrophysique CEA (France) and Universite Paris-Sud (France). The intention was to bring together experts who have varied skills related to the study of cosmic dust. These include experts in, multi-wavelength data reduction and photometry, database construction, the physical properties of cosmic dust grains, radiative transfer modelling, the fitting and extraction of information from spectral energy distributions, applications to a better understanding of the way galaxies evolve and processes in the inter-stellar medium. Further details can be found at http://www.dustpedia.com/ and https://en.wikipedia.org/wiki/DustPedia.The DustPedia project has also capitalised on the legacy of the Herschel Space Observatory, using cutting-edge modelling techniques to study dust in 875 galaxies - representing the vast majority of extended galaxies within 3000 km/s that were observed by Herschel. This project requires a database of multi-wavelength imagery and photometry that greatly exceeds the scope (in terms of wavelength coverage and number of galaxies) of any previous local-Universe survey.The 875 selected galaxies have optical sizes (D_25) greater than one arcmin and lie within a redshift velocity of 3000 km/s. All available far-infrared photometric data for this sample has been extracted from the Herschel archive. In addition we have extracted all available multi-wavelength ancillary data for each galaxy from GALEX, SDSS, DSS, 2MASS, WISE, Spitzer, and Planck. Using these data, we have performed consistent aperture-matched photometry, which we have combined with external supplementary photometry from IRAS and Planck. To achieve this objective we have produced the Comprehensive & Adaptable Aperture Photometry Routine (CAAPR). This is our own, but publicly available automated photometry pipeline designed to carry out aperture-matched photometry. CAAPR is designed to produce consistent photometry for the enormous range of galaxy and observation types in our data. In particular, CAAPR is able to determine robust cross-compatible uncertainties, thanks to a novel method for reliably extrapolating the aperture noise for observations that cover a very limited amount of background. Our data consists of multi-wavelength imagery and photometry across 42 ultra-violet to mm bands for the 875 DustPedia galaxies. Our aperture-matched photometry, combined with the external supplementary photometry, represents a total of 21,857 photometric measurements. A typical DustPedia galaxy has multi-wavelength photometry spanning 25 bands. The publicly available DustPedia Archive (http://dustpedia.astro.noa.gr) is hosted on a Virtual Machine located at the Greek Research & Technology Network (GRNET). The Database is structured in such a way so that it can accommodate the full set of observations used in DustPedia as well as the computed parameters of the models. The database website is implemented following the light and well established MVC design pattern (a popular variation of the Front Controller pattern when it comes to websites). The actual implementation has been done under Microsoft Visual Studio development environment using C# 4.0 as the programming language with Razor v2.0 View Engine and Microsoft SQL Server 2014 as the Database backend. The website design implements the Repository pattern in order to fetch/commit data from/to the Database. Dependency Injection has also been utilized (using the Ninject library) allowing thus to easily switch to any other means of Data access (such a web service) may the need arise in the future.The Architecture of the DustPedia Archive in essence is a database driven website, where the Database holds information about all content/text that appears on the user interface as well as all necessary information about Galaxies. In order to keep the Database light and fast FITS files are not stored on the Database itself. The Database only holds information about how to access the FITS files associated with a galaxy. The actual FITS files are stored as real files on the file system and in particular on the dedicated 4TB Network Disk and they are only accessed in a secure way (filenames are not exposed on the URL). The Virtual Machine which hosts the website is running Windows Server 2012 R2 having Internet Information Services (IIS) serving the website to the world.The archive contains multi-wavelength imagery for the 875 DustPedia galaxies. The data hosted in the archive are all the available maps for the DustPedia sample coming from the GALEX, SDSS, DSS, 2MASS, WISE, Spitzer, Herschel, and Planck surveys. In the left-hand side of each entry some basic galaxy properties are provided, as well as, photometry cutouts (in PNG format) for each galaxy with the exact aperture used for the photometry, per band, and the annuli used to calculate the background.The user can retrieve all the available maps (and in some cases its associated error map) in FITS format. All maps are in units of Jy/pixel except for DSS (left as photographic densities). The user can search with Galaxy Name, Hubble Stage (T), Velocity (in km/s), inclination angle (in degrees), and size (D_25 in arcmin). In each search parameter the user can define the range desired but also have the option to define only the lower or the higher limit. A "show all results in a single page" button allows the user to print all results on a single page.Consistent aperture-matched photometry (CAAPR) was performed to all available maps hosted in the data archive, combined with external supplementary photometry for IRAS and Planck surveys. The user can retrieve the derived photometry tables (in .csv format). The "DustPedia Main Photometry" link provides all the available photometric measurements in Jy. Redshift-independent distances for all the DustPedia galaxies are provided in "Homogenized redshift-independent distances".Other ancillary information on the DustPedia galaxies has been collected from the literature. This includes HI and H_2 masses as well as line and metallicity measurements. In addition, a 2D morphological study provides morphology information for every galaxy in the sample. In addition the results from fitting the DustPedia galaxies with a Modified Black-Body model are included. The fit is applied to all available fluxes beyond 70 microns. The model assumes an emissivity consistent with the THEMIS grain model (see below). 682 out of the 875 galaxies had sufficient data to be fitted in this way. A bootstrap analysis was used to calculate the uncertainties in dust temperatures, luminosities and masses. We have also fitted the spectral energy distributions (SEDs) of the DustPedia galaxies using the well-known CIGALE package. We have modified CIGALE accordingly in order to give the dust properties predicted using our own THEMIS dust grain model (see below). 815 out of the 875 galaxies had sufficient data and could be fitted. This provides, in addition to the dust properties, a measure of the current star formation rate, the stellar mass and an indication of star formation history. CIGALE generated template SEDs for galaxies of various Hubble Stages (T) and luminosity are obtainable from the database. The database also contains the results from using our hierarchical Bayesian dust SED model (HerBIE, see below), that provides a more sophisticated, detailed and consistent means of measuring the global dust properties (dust mass, temperature etc.) of galaxies. For the most spatially extended galaxies we have produced resolved maps for the dust surface density, the inter-stellar radiation field, and the small grain fraction. This includes results from our Monte Carlo photon tracing radiative transfer code (SKIRT, see below), which provides resolved and global SEDs and model images in each photometric band. The physical nature of cosmic dust grains has been a major topic in astronomical observational studies for more than 80 years, beginning with studies of interstellar absorption in the early 1930s. Since then many materials have been proposed as viable dust components, beginning with the “dirty ice” model in the 1940s. Since then graphite and amorphous silicates have become firm favourites for most dust modellers, with dust properties derived, in some cases, empirically and circularly, from observations. However, more recent analyses of the observations from the Spitzer, Herschel and Planck satellite missions have underlined the need for a more sophisticated and non-empirical approach, and one that does not involve graphite. Within this framework we developed a new dust modelling approach, which is solidly-founded upon laboratory data, insisting that the data used to model the interstellar dust optical properties are derived as much as is practically possible from direct laboratory measurements. Good quality laboratory data for amorphous silicates are widely available, albeit with fewer data at FIR-mm wavelengths, and so we focussed our efforts on using the optical and physical properties of the family of hydrogenated carbonaceous materials, which are thought to be more physically-realistic analogues for interstellar carbonaceous dust.The result of this is a new core/mantle interstellar dust model that is entirely consistent with the dust extinction and emission in the diffuse interstellar medium. In a comparison with some widely-used models the new model was found to perform significantly better being able to explain the observed variations in the dust properties throughout the diffuse ISM within a physically-reasonable parameter space. This interstellar dust modelling framework was subsequently named THEMIS (The Heterogeneous dust Evolution Model for Interstellar Solids; https:// www.ias.u-psud.fr/themis). The THEMIS approach to dust modelling takes a global view of the dust composition and structure and their evolution in response to the local physical conditions (density, temperature and radiation field) in the interstellar media (ISM). This diffuse ISM model is further extended to encompass the effects of dust evolution (through accretion and coagulation) in the transition from the diffuse ISM towards dense molecular clouds. In the process, this work has resulted in a self-consistent explanation for the observed cloud- and core-shine effects via dust evolution resulting from the combined effects of carbon accretion from the gas phase and grain-grain coagulation.The primary, and most significant, difference between the THEMIS dust model and other existing interstellar dust models is that the dust is more emissive at FIR-mm wavelengths and therefore yields dust masses derived from dust emission studies at long wavelengths that are lower by about a factor of two. Additionally, the carbonaceous dust properties are derived with a size- and composition-dependent methodology that avoids the need for artificially-determined and separate “PAH-like” and “very small grain” dust populations. Further, the “out of the box” THEMIS model has been shown to be consistent with numerous dust observables (continuum absorption, spectroscopic absorption bands, scattering, albedo, spectroscopic emission bands, thermal emission, elemental depletions etc.) without the need for any empirical “fine-tuning”. THEMIS has now also been fully implemented into the CIGALE SED fitting code and also into the HerBIE dust SED modeller, and adopted as the model of choice for DustPedia. The THEMIS approach is under continuous development and currently we are extending the framework to explore the optical properties of large grains and aggregates, nano-carbon dust in discs and the implications of dust evolution in the photon-dominated regions and HII regions associated with star formation.The general approach of previously developed state-of-the-art dust SED fitting models for galaxies and resolved regions does not contain explicit assumptions about the geometry of the source being modelled, and simply assumes a general distribution of radiation fields. The significant advantage of the method we have developed is that we are able to derive values of the physical parameters independent of the geometry (IR power, dust mass, moments of the radiation field distribution, PAH mass fraction, sub-mm excess, etc.) with a better precision than previous radiative transfer (RT) models. Another limitation of previous SED fitting models was that we could not reliably interpret observations with poor signal-to-noise ratio, and that there were potential noise-induced correlations and biases. These limitations originated in the simplicity of the statistical approaches, usually the well-known least-squares minimisation.To this end we have transformed our model into a hierarchical Bayesian fitting method. The principle of this approach consists in modelling simultaneously each SED, and the overall distribution of physical parameters (called a prior). In this way, by making very unrestrictive assumptions about the functional interdependency of each physical parameter within our sample (sample of galaxies, or pixels in an image), we can remove false correlations, some biases, and exploit the information contained in poor-signal-to-noise ratio observations. Our hierarchical Bayesian dust SED fitting model, called HerBIE (HiERarchical Bayesian Inference for dust Emission) performs the complete inference of physical parameters, for an arbitrary large spectral cube with the linear combination of any of the following components: – an arbitrary number of modified blackbodies; – a uniformly-illuminated dust mixture; – a dust mixture illuminated by a distribution of radiation fields; – a stellar continuum; – a free-free continuum; – a synchrotron continuum.The output is the full probability distribution of the physical parameters of each galaxy/pixel. The model also takes into account the correlated calibration uncertainties of each instrument.Since this code is numerically intensive (large images may take several weeks to run), we have optimised it by pre-computing finely sampled templates of dust emission. We have developed a series of programs computing: – the Mie theory for most grain optical properties; – the stochastic heating of each species, in a variety of heating conditions; – integrated dust mixtures; –synthetic photometry (including colour correction) of most of IR-submm observatories.We have built this database, using adaptive grids in wavelength, grain radius, grain temperature and radiation field intensity, in order to obtain precise results. The THEMIS dust model has been fully implemented in HerBIE. The physics in the THEMIS code is different than that of previous dust grain models - the main difference being in the physics of the small carbon grains. Previous dust grain mixtures assumed neutral and charged PAHs to reproduce aromatic features, and small graphite or amorphous carbon, to account for the mid-IR continuum. The THEMIS dust model accounts for this ensemble of observables with hydrogenated amorphous carbons, a-C(:H). We parameterise the size distribution of these particles, separating several bins of sizes, in order to have a model with the same degree of freedom.Generation of the SED model fits of the galaxies in the DustPedia data base, using HerBIE, has now been completed. Of the 875 galaxies we found that 19 had a significant AGN component and thus, would not be successfully described by the dust SED model. These 19 galaxies will be fitted later using AGN radiative transfer models. We have successfully fitted and analysed the remaining 856 galaxies with HerBIE, using the DustPedia homogenised integrated photometry. The analysis of the dust properties of the DustPedia galaxies which emerge from the SED modelling have been delivered to the DustPedia database.One of the advantages of the hierarchical Bayesian approach is its holistic account of the statistical properties of the studied sample. With such an approach, one can study the statistical relations not only between the fitted dust properties, but also with ancillary dependencies, such as the stellar and gas masses, the metallicity, etc. In this way, we are able to characterise the various scaling relations, with optimal precision and accuracy. We have thus included in our hierarchical Bayesian run the most important ancillary data we had at our disposal: the stellar masses and star formation rates derived from CIGALE fits (see below), the HI masses and metallicities and the H_2. These ancillary dependencies provide valuable information to ascertain the fits of the most poorly observed objects.We have also made use of the widely used galaxy SED fitting code CIGALE, and adapted it to include our unifying dust grain model (THEMIS). Using the DustPedia photometry we have determined the physical properties of the galaxies, such as, the dust mass, the stellar mass, the star-formation rate, the bolometric luminosity as well as the unattenuated stellar light and the stellar light absorbed by the dust (for both the old and the young stellar populations). We find that the THEMIS dust grain model predicts dust masses lower, by approximately 42%, compared to the widely adopted Draine model. We show how the masses of stars, dust, and gas, as well as the star-formation rate and the dust temperature vary between galaxies of different morphologies. We have also derived recipes to approximate these parameters given a galaxies Hubble stage (T). Using CIGALE we find that early-type galaxies (ETGs) with Hubble stages T < 0 contain, mainly, old stars with only a small fraction of the bolometric luminosity (< 10%) originating from young stars. For galaxies with Hubble stages 0 < T < 5 the fraction of young stars increases up to ~ 25%, while it stays roughly constant for galaxies with Hubble stages T > 5. Both the old and the young stars are mostly affected by dust in intermediate Hubble stages (1 ≤ T ≤ 7) with a drop of more than 15% in their luminosities. On average, young stars are very efficient in heating the dust, with the absorbed, by the dust, luminosity reaching as high as ~ 77% (at T = 3) of the total, unattenuated luminosity of this population. On the other hand, the maximum luminosity of the old stars used in the dust heating is ~ 24%, again at T = 3. The dust heating in ETGs is mainly due to the old stars, to a level of up to ~ 90%, while the young stars progressively contributing more for galaxies with Hubble stages 0 < T < 5, and they become the dominant source of dust heating for galaxies with Hubble stages T > 5 donating up to ~ 60% of their luminosity to this purpose.We have compared the dust properties derived with HerBIE and CIGALE. The comparison shows that the two are consistent, though the use of HerBIE leads to significantly less scatter in derived scaling relations. We also analyzed the fit residuals, looking for systematic features, but none were found.As one of the main goals in the DustPedia project, we have developed a framework to construct detailed 3D dust radiative transfer models for nearby galaxies with the SKIRT Monte Carlo code. The code is publicly available on github (https://github.com/SKIRT/SKIRT8) and on the SKIRT website (http://www.skirt.ugent.be).For a given galaxy, the modelling uses a set of observed images (specifically the GALEX FUV, Hα, 2MASS/UKIDSS H, IRAC 3.6 μm, MIPS 24 μm, and PACS 70 μm images) as input, and creates a multi-component 3D model galaxy that fits the observed images as well as the observed UV-mm SED as closely as possible. The standard output of the modelling consists of the model SED and a set of model images in the most common broadband filters, covering the entire UV-submm wavelength range.We have developed a general and flexible open-source radiative transfer modelling framework. This framework consists of two major parts. The first component is the SKIRT Monte Carlo radiative transfer code. This pre-existing code has been developed by the UGent node over the last 15 years. As part of the DustPedia project, the code has been improved and optimized substantially, such that it is suitable for the radiative transfer modelling of the DustPedia galaxies. Major DustPedia-critical improvements include the implementation of a hybrid parallelization framework, the incorporation of the THEMIS dust model, the setup of an extensive library of input modes, and new optimisation techniques. The second component is the SKIRT-DustPedia modelling pipeline that is designed to set up 3D radiative transfer models for a given galaxy. This pipeline is written in Python, and is publicly available as part of the PTS package, both on github (http://github.com/SKIRT/PTS) and on the SKIRT website (http://www.skirt.ugent.be/pts).The general modelling approach we have developed is to construct a 3D model for the galaxy, consisting of several stellar and a dust component. The stellar components consist of a central bulge, an evolved stellar disc, a young stellar disc, and a disc of ionising stars. The geometrical distribution of each of these components is based on observed images of the galaxy at different wavelengths, using a two-step procedure. The first step consists of combining different images to physical maps that characterise, for example, the distribution of dust or old stellar populations on the sky. The second step consists of de-projecting these 2D maps on the sky to a 3D distribution. Apart from the geometrical distribution, each stellar component is assigned an intrinsic spectral energy distribution, and a total luminosity that is either fixed or a free parameter in the model. Similarly, the optical properties of the dust component are specified, the total dust mass is a free parameter in the model.For any choice of the free parameters in the galaxy model, we can set up a 3D dust radiative transfer simulation that fully takes into account the emission by the different stellar populations, and the absorption, scattering and thermal emission by the dust. The result of each simulation includes the spectral energy distribution (SED) and a set of broadband images of the galaxy, from UV to sub-mm wavelengths, that can directly be compared to the observed images. The best fitting model is determined through a chi-sq optimisation procedure: using a multi-stage fitting approach, we determine the free parameters of the model that best reproduce the observed SED. As soon as the best fitting model is determined, a wealth of possible information can be extracted. For example, additional images of the galaxy at arbitrary viewing points and arbitrary wavelengths can be calculated, and the intrinsic properties of the interaction between starlight and dust can be studied in 3D.We have taken special care to automate the entire procedure as much as possible. The different steps in the procedure (discussed below) are implemented in python and are publicly available as part of the Python Toolkit for SKIRT (PTS). This automation and open-science approach has the obvious advantage that our results can immediately be reproduced or extended by other researchers, and that the same approach can easily be applied to other galaxies. The main steps involved are the image preparation, the map making, the decomposition, and the fitting steps. To ensure that the modelling pipeline is generally applicable (i.e. that it can be used for many different galaxies) and user-friendly (i.e. that it can be used by different users without major issues) we have taken the following steps:1. The pipeline has been developed with the comprehensively observed early-type spiral galaxy M81 as a direct test case. 2. The different steps in the pipeline have then been tested on another well observed sample of four other DustPedia spiral galaxies (M83, M95, M100, NGC1365). 3. As a stringent test, the pipeline has also been applied to NGC1068. NGC1068 is a Seyfert 2 galaxy and was chosen specifically because it was expected that the radiative transfer modelling would be challenging, and indeed, various challenges were encountered. For example, the mid-infrared WISE and IRAC images are completely dominated by the PSF signature from the central point source. In addition, a new, sub-grid implementation was developed to model the AGN component. Again, feedback from this test was fed back into the main pipeline.4. We have also produced the necessary documentation for the Python code, and tutorials on how to implement the package are available on the SKIRT web page. As a result of the above, we are convinced that we have generated a pipeline that can and will be used beyond the lifetime of the DustPedia project itself, and hence will enhance the DustPedia legacy. In principle, it can be applied to any galaxy (as long as sufficient resolution is available), and hopefully will be further tuned or refined as required. For each galaxy so far modelled we have created the SKIRT input the ski file, together with the all the input images required to run the SKIRT simulation. These are all available on the DustPedia section of the SKIRT website. Given the ski file, any user can immediately re-run SKIRT and regenerate the full radiative transfer model. Apart from the obvious observables, such as the SED and the simulated images, the SKIRT code generates other information about the model:1. Simulated images from other observing angles, e.g. edge-on or face-on, can be generated.2. Information about the 3D structure of the stellar and dust distribution can be extracted. For example, the SKIRT code automatically calculates and outputs the interstellar radiation field at every position of the galaxy model, as well as the mean dust temperature for every dust species and grain size.3. By tuning the ski file, additional information can be extracted and written out, such as the contribution of each of the stellar components to the heating of the dust at every position in the model.An important testbed for models of dust grains is the emissivity (i.e. the surface brightness in the FIR per hydrogen column density) this has typically been measured at high galactic latitude in the Milky Way cirrus. From the emissivity and an estimate of the local interstellar radiation field, the dust absorption cross-section per hydrogen column density can be derived, and ultimately the absorption cross-section per dust mass - after assuming a dust-to-gas mass ratio. The mass absorption cross-section, measured either from simplistic assumptions or derived using complex theoretical distributions of grain sizes and materials, is a fundamental quantity critical to the derivation of the dust mass in galaxies. It is normally assumed that the absorption cross-section in various environments of different galaxies is the same as that of the dust in the Milky Way cirrus.We have used the DustPedia data to measure the emissivity and absorption cross-section directly from observations of external galaxies. To achieve this goal, we collected from the literature information on the atomic and molecular gas (now included in the DustPedia database) and homogenized them in order to have the total hydrogen mass within a galaxy's optical radius. Where possible we have also obtained galaxy metallicities and also included these in the database. We have derived the global (i.e. integrated) emissivity at 250um for all DustPedia galaxies detected in the FIR and with the above information on atomic hydrogen, molecular gas and metallicity. In total the sample amounts to 183 objects. The molecular hydrogen column density was derived from observations of the CO molecule after assuming a metallicity dependent conversion factor (see below). The conversion factor was found to provide consistent trends in dust/gas scaling relations. The emissivity shows a large scatter between galaxies, of about 1 dex. A weak trend with the temperature is seen, reflecting different average intensities of the radiation field in different objects. The temperatures were derived by modified-blackbody fits to FIR/sub-mm data and are made available in the DustPedia Archive. In the derivation, it was assumed that the dust absorption cross-section can be described by a power-law: we used a single value spectral index, fixed to the value derived on the MW cirrus.After removing the temperature dependence, the emissivity can be converted into the absorption cross-section per hydrogen column density. However, the true dust emission results from a variety of grains of different sizes and materials, exposed to different intensities of the interstellar radiation field, and thus attaining different temperatures. Our modelling, instead, assumes a single temperature. We have tested the effects of this simplification using the CIGALE models of DustPedia galaxies, where the full THEMIS grain model and a distribution of ISRF intensities is used. We derived the absorption cross-section with the single-temperature assumption and compared it with the intrinsic, built-in, cross-section. The value we retrieved is consistent with the THEMIS averaged absorption cross-section within 20%, a marginal error when compared to the error and scatter in the measurements. The absorption cross-section per hydrogen column density shows a weak dependence on metallicity, which could result from variation in the dust-to-gas mass ratio within the sample. A common assumption in the literature is that the dust-to-metal mass ratio is constant, and thus the dust-to-gas ratio depends linearly on the metallicity. After normalising this dependence to the metallicity and dust-to-gas ratio derived from the metal depletion in the MW, we determined the absorption cross-section per dust mass. While the scatter is only marginally reduced, the dependence of the absorption cross-section on metallicity disappears. The absorption cross section of the sample at 250um is 2.7 cm^2 g^-1. This is lower than the MW value of 3.5 cm^2 g^-1, yet compatible with it within the scatter. Work is still ongoing and we are currently investigating the influence on the results of the various assumptions and recipes we used, with the final goal of finding dependences of the absorption cross-section on other physical quantities and so reducing the scatter in the relations.We have also used a similar approach to derive the global mass absorption cross-section for a sample of 22 galaxies. The major difference is that the dust-to-gas ratio dependence on metallicity is not directly scaled on MW values, but derived from other quantities such as a (constant) dust-to-metal mass ratio and the metal depletion factor. The derived mass absorption cross-section is smaller at 1 cm^2 g^-1 at 250um, the difference being the result of the different numerical values of the various quantities used in the derivation.The above method has been extended to derive the mass absorption cross-section on a resolved scale. Maps of the absorption cross-section have been produced for two well resolved DustPedia galaxies, NGC 5236 (M83) and NGC628 (M74). Both galaxies have available FIR/sub-mm images from the DustPedia database; atomic and molecular gas maps were obtained from the literature; metallicity maps were reconstructed with Gaussian Process Regression from estimates at sparse positions across the galaxies (mainly from integrated field unit spectra). The resolved metallicity data are again available from the DustPedia database. Preliminary results show both gradients and localised (arm-interarm) variations in the absorption cross-section. The absolute values might depend on the various assumptions made within the model, the variations are robust. They reflect true changes in the dust properties with the environment.We have also carried out a characterisation of the radial distribution of dust, stars, gas, and star-formation rate (SFR) in a sub-sample of 18 face-on spiral galaxies extracted from the DustPedia sample. This study exploits the multi-wavelength DustPedia database, from ultraviolet to sub-mm bands, as well as molecular and atomic gas maps and metallicity abundance information. We have made exponential fits to the surface-brightness profiles of the tracers of dust and stars, the mass surface-density profiles of dust, stars, molecular gas, and total gas, and the SFR surface-density profiles and derived their scale-lengths. We have also developed a method to solve for the CO-to-H2 conversion factor (alpha_CO) per galaxy by using dust and gas mass profiles.Although each galaxy has its own peculiar behaviour, we identified a common trend of the exponential scale-lengths versus wavelength. On average, the scale-lengths normalised to the B-band 25 mag/arcsec^2 radius decrease from the ultraviolet to 70um, from 0.4 to 0.2 and then increase back up to 0.3 at 500 microns. The main result is that, on average, the dust-mass surface-density scale-length is about 1.8 times the stellar one derived from IRAC data and the 3.6 μm surface brightness, and close to that in the ultraviolet. There is a weak dependence of the scale-lengths on the Hubble stage T: the scale-lengths of the Herschel bands and the 3.6 μm scale-length tend to increase from earlier to later types, the scale-length at 70 μm tends to be smaller than that at longer sub-mm wavelength with ratios between longer sub-mm wavelengths and 70 μm that decrease with increasing T. The scale-length ratio of SFR and stars shows a weak increasing trend towards later types. Our alpha_CO determinations are in the range (0.3-9) M_0 pc^-2 (K km/s)^-1, almost invariant by using a fixed dust-to-gas ratio mass (DGR) or a DGR depending on metallicity gradient.Most radiative transfer models assume that dust in spiral galaxies is distributed exponentially. We have reconsidered this assumption by analysing the two-dimensional large-scale distribution of dust in galaxies from the DustPedia sample. For this purpose, we have made use of Herschel imaging in the five bands, from 100 to 500um, in which the cold dust constituent is primarily traced and makes up the bulk of the dust mass in spiral galaxies. For a sub-sample of 320 disc galaxies, we successfully perform a simultaneous fitting with a single Sersic model of the Herschel images in all five bands using the multi-band modelling code GALFITM. The derived Sersic index n, which characterises the shape of the profile, lies systematically below 1 in all Herschel bands and is almost constant with wavelength. The average value at 250um is 0.67+/-0.37 (187 galaxies are fitted with n_250<0.75 87 galaxies have 0.751.25). Most observed profiles exhibit a depletion in the inner region (at r<0.3-0.4 of the optical radius r_25) and are more or less exponential in the outer part. We also find breaks in the dust emission profiles at larger distances (0.5-0.6)r_25 which are associated with the breaks in the optical and near-infrared. We assume that the observed deficit of dust emission in the inner galaxy region is related to the depression in the radial profile of the HI surface density in the same region because the atomic gas reaches high enough surface densities there to be transformed into molecular gas. If a galaxy has a triggered star formation in the inner region (for example, because of a strong bar instability, which transfers the gas inwards to the centre, or due to pseudo-bulge formation), no depletion or even an excess of dust emission in the centre is observed. Observations of evolution in the dust-to-metal ratio allow us to constrain the dominant dust processing mechanisms. To this end we have studied the dust-to-metal and dust-to-gas ratios in a subsample of approximately 500 DustPedia galaxies. Using literature and MUSE emission line fluxes, we derive gas-phase metallicities (oxygen abundances) for over 10000 individual regions and determine characteristic metallicities for each galaxy. We study how the relative dust, gas and metal contents of galaxies evolve by using metallicity and gas fraction as proxies for evolutionary state. The global oxygen abundance and nitrogen-to-oxygen ratio are found to increase monotonically as galaxies evolve. Additionally, unevolved galaxies (gas fraction > 60%, metallicity 12 + log(O/H) < 8.2) have dust-to-metal ratios that are about a factor of 2.1 lower (factor of 6 lower for galaxies with gas fraction > 80%) than the typical dust-to-metal ratio (M_d/M_Z ~ 0.214) for more evolved sources. However, for high gas fractions, the scatter is larger due to larger observational uncertainties as well as a potential dependence of the dust grain growth timescale and supernova dust yield on local conditions and star formation histories. We find chemical evolution models with a strong contribution from dust grain growth describe these observations reasonably well. The dust-to-metal ratio is also found to be lower for low stellar masses and high specific star formation rates (with the exception of some sources undergoing a starburst). We find that the metallicity gradient correlates weakly with the HI-to-stellar mass ratio, the effective radius and the dust-to-stellar mass ratio, but not with stellar mass.Understanding the interplay between the various components of the interstellar medium (ISM: dust, atomic and molecular gas) in galaxies of the Local Universe is of fundamental importance for studies of galactic formation and evolution. In this framework, DustPedia –thanks to Herschel– has been devised aimed at performing a complete characterisation of dust in the Local Universe. However, we need information on all the phases of the ISM, including the gas, to draw definitive conclusions on it.Nearby galaxies are characterised by an intricate system of correlations between their global properties. These correlations form the basis of the so-called `scaling relations’, which are fundamental because they provide a quantitative means to characterise how the physical properties of galaxies relate to each other. Galaxy scaling relations also provide the tool to study the internal physics of galaxies, as well as their formation and evolutionary histories, and different galaxy populations.We have considered the scaling relations between dust mass and molecular, atomic, and total gas. We find that all single gaseous components are well correlated with the dust, and the total gas being the best correlated with dust. Within the optical disk of galaxies, dust and atomic gas are better correlated than dust and molecular gas. This finding is opposite to what happens for the small-scale physics of the ISM. The scaling relations we provide have been tested under several physical assumptions, for instance by adopting a constant and metallicity dependent CO-H_2 conversion factor (X_CO). We also study the dust-to-gas ratio (DGR) as a function of morphology, separating the various gas phases. We find that the assumption of a X_CO depending on metallicity is able to reproduce the expected decreasing of DGR with the morphological stage T. With the large and homogenous DustPedia sample and dataset, we are also able to characterise the DGR at a given T for all gas phases.A sub-sample of DustPedia early-type galaxies (ETGs) has been analysed in order to produce various scaling relations using the parameters derived from detailed spectral energy distribution modelling. It was found that the ETGs deviate from the dust mass-SFR relation and the Schmidt-Kennicutt relation that SDSS star forming galaxies define. Compared to SDSS galaxies, ETGs have more dust at the same SFR, or less SFR at the same dust mass. When placing them in the M_star-SFR plane, ETGs show a much lower specific SFR as compared to normal star-forming galaxies. ETGs show a large scatter compared to the Schmidt-Kennicutt relation found locally within our Galaxy, extending to lower SFRs and gas mass surface densities. Using an ETG's SFR and the Schmidt-Kennicutt law to predict its gas mass leads to an underestimate. ETGs have similar observed gas-to-modelled-dust mass ratios as star forming-galaxies of the same stellar mass.We are also conducting a study of the combined gas and dust properties of a sample of 28 low metallicity dwarf galaxies extracted from the DustPedia sample. We have chosen all DustPedia galaxies with metallicities below 20% solar – thereby focusing on the regime where the dust-to-gas ratio (DGR) has been known to decrease steeply with decreasing metallicity. Given the unbiased selection criteria for DustPedia galaxies, this happens to be the first unbiased dwarf galaxy sample on which such a study is being conducted. Also for the first time, such a study is being conducted by adding resolved maps of the gas distribution to the already existing resolved maps from the DustPedia database. The gas maps come from new interferometric observations of the atomic hydrogen in the DustPedia dwarf galaxies. The resolved gas and dust maps are enabling the study to go beyond the historical limitation of deriving one measurement of any given property per galaxy, and can be done on a region-by-region basis within individual galaxies. The dust properties for the sample galaxies are being determined using the HerBIE SED fitting code. The use of HerBIE is critical for this work because the faint dwarf galaxies in our sample can have multiple non-detections in the far-infrared bands. HerBIE enables us to extract meaningful dust parameters even from galaxies/regions that have SEDs with multiple non-detections. A partial study completed for 15 of the 28 galaxies in our sample reveals many interesting and hitherto unknown facets of dust within dwarf galaxies. We determine the global (one measurement per galaxy) dust and gas properties for dwarf galaxies in a consistent way for the first time by measuring the gas and dust properties within matched apertures - the aperture used for deriving the photometry from the various DustPedia bands. The resolved study (region-by-region within galaxies) reveals the rich variation in dust properties within each dwarf galaxy. We find that even for the galaxies with the most number of individual metallicity measurements, such measurements are limited to the very inner regions of dwarf galaxies. Comparing the average metallicity determined from such measurements to the global dust+gas properties have historically led to oversimplified interpretations of the effect of metallicity on the evolution of the gas to dust chain in dwarf galaxies. We find that dust properties themselves show large variations within dwarf galaxies, especially in outer regions of the galaxies with higher radiation fields compared to inner regions where dust properties are more uniform.We have also been considering the fraction of stellar radiation absorbed by dust (f_abs), in 814 DustPedia galaxies of different morphological types. The targets constitute the vast majority (93%) of the DustPedia sample. For each object, we have modelled the spectral energy distribution from the ultra-violet to the sub-millimetre using the dedicated, aperture-matched DustPedia photometry and the CIGALE SED fitting code. The value of f_abs was obtained from the total luminosity emitted by dust and from the bolometric luminosity, which are estimated by the fit.On average, 19% of the stellar radiation is absorbed by dust in DustPedia galaxies. The fraction rises to 25% if only late-type galaxies are considered. The dependence of f_abs on morphology, showing a peak for Sb-Sc galaxies, is weak; it reflects a stronger, yet broad, positive correlation with the bolometric luminosity, particularly late-type, disk-dominated, high-specific-star-formation rate, gas-rich objects. We find no variation of f_abs with inclination, at odds with radiative transfer models of edge-on galaxies. These results call for a better self-consistent modelling of the evolution of the dust mass and geometry along the build-up of the stellar content. We have also provided template spectral energy distributions in bins of morphology and luminosity and have studied the variation of f_abs with stellar mass and specific star-formation rate. We confirm that the local Universe is missing the high f_abs, luminous and actively star-forming objects necessary to explain the energy budget in observations of the extragalactic background light.We have used a sub-set of the DustPedia galaxy sample (461 galaxies) to investigate the effect the environment has had on galaxies. We separately considered Virgo cluster and field samples and also assign a density contrast parameter to each galaxy, as defined by the local density of SDSS galaxies. We have considered their chemical evolution (using M_Dust/M_Baryon and M_Gas/M_Baryon), their specific star formation rate (SFR/M_Stars), star formation efficiency (SFR/M_Gas), stars-to-dust mass ratio (M_Stars/M_Dust), gas-to-dust mass ratio (M_Gas/M_Dust) and the relationship between star formation rate per unit mass of dust and dust temperature (SFR/M_Dust and T_Dust). Late type galaxies (later than Sc) in all of the environments can be modelled using simple closed box chemical evolution and a simple star formation history (SFR(t) proportional to t*exp^(-t/tau)). For earlier type galaxies the physical mechanisms that give rise to their properties are clearly much more varied and require a more complicated model (mergers, gas in or outflow). However, we find little or no difference in the properties of galaxies of the same morphological type within the cluster, field or with different density contrasts. It appears that it is morphology, how and whenever this is laid down, and consistent internal physical processes that primarily determine the derived properties of galaxies in the DustPedia sample and not processes related to differences in the local environment.Despite the fact that over the last decades our knowledge of galaxy formation theory has improved substantially, we still have only a fragmentary understanding of all the complex and coupled physical phenomena that shape galaxies. Numerical simulations of galaxy formation are a needed and valuable tool to alleviate these difficulties and of course crucially that ultimately agree with observations of the Universe. An important of the DustPedia project is to compare simulations with observations in a proper and consistent way, to map the successes and limitations of our numerical models and ultimately improve our understanding of galaxy formation and evolution. DustPedia, being the largest sample of nearby galaxies with accurate and consistent photometry over the entire UV to sub-mm range, is an ideal sample to compare with.The EAGLE state-of-the-art simulation consists of cosmological hydrodynamical simulations with different resolutions, sub-grid models and a range of box sizes up to 100 Mpc. For our study, we are focused on two EAGLE simulations: The reference Ref-L100N1504 (hereafter called the Ref-100 simulation) and Recal-L025N0752 (hereafter Recal-25) simulations. Recal-25 has higher resolution and, appropriately, a different set of sub-grid recipes than Ref-100. Since dust is not modelled in the EAGLE simulations, post-processing of the input data (stellar and gas particles) is necessary. We model two dust sources: star forming regions which are not resolved in the EAGLE simulations, and diffuse dust. Former are modelled using MAPPINGS-III and latter using SKIRT. The diffuse dust distribution is derived from the distribution of gas. For the radiative transfer simulations it is necessary to have a sufficiently resolved dust distribution. As a consequence, our sample will exclude galaxies that have little or no dust, mainly red, elliptical galaxies with low SFR.We have compared the EAGLE-SKIRT sample with 814 DustPedia galaxies already modelled by CIGALE (see above). We analysed different scaling relations and find remarkable agreement between simulated and observed galaxies. The WISE 3.4 μm band is most sensitive to the evolved stellar populations that dominate the baryonic mass in galaxies. This combined with the fact that in this wavelength range dust extinction is not so important, ensures the quality of WISE 3.4 luminosity as a stellar mass tracer. There is overall agreement between Eagle and the DustPedia observations, although some discrepancies are present. We have confirmed that combining the EAGLE simulations with SKIRT can reproduce a wide range of observables of the local Universe, despite the incompleteness of the sample and optimisation of sub-grid properties based only on global dust scaling relations.Potential Impact:European governments financed the Herschel and Planck projects at a cost of over one billion euro. Although the spin-off technologies have been considerable, ultimately the project will be judged by the progress projects like ours make in understanding the Universe, after all that is why the observatories were constructed. Our project has utilised data from a wide range of disparate projects and unites them into a single large focused research programme. The gross properties of the Universe have been observed to correspond very well with those predicted by cosmological models and now one of the great challenges left is a deeper understanding of the physical processes that make the diversity of galaxies. The major impact made by us within the scientific community has been emphasised in our scientific description of the project and will be further justified by our published results in prestigious well-read scientific journals. This will promote European excellence and competitiveness and make Europe an increasingly attractive place to do research. During the course of DustPedia we have produced a number of world leading products:1. A publicly accessible multi-wavelength database of galaxies containing over 25,000 calibrated observations and associated photometric measurements. In addition we carried out an extensive literature search to also assimilate into the database all available molecular and atomic gas masses and metallicities.2. A new publicly available model for cosmic dust grains that completely specifies their optical properties enabling a new interpretation of observational data.3. A publicly available Bayesian spectral energy fitting method.4. A publicly available and completely updated Monte Carlo photon tracing radiative transfer model applicable to a wide range of astrophysical problems.The DustPedia project has provided a new framework for the extensive study of cosmic dust. It utilises data from a part of the electromagnetic spectrum that has hardly been used before and it has made use of state-of-the-art telescopes that have provide un-paralleled spatial resolution and sensitivity. However, satisfying scientific objectives and convincing the public that their money has been well spent are two completely different things. To make an impact with those who fund our work we must clearly disseminate our results in a way that is both accessible and understandable and to successfully communicate the importance and significance of the discoveries we have made. In addition we hope that our success in meeting our objectives and successfully demonstrating them to the public will encourage new generations to look for careers in science.The success of our project relied heavily on the cooperation of the researchers involved. The success of our project relied heavily on the cooperation of the researchers involved. As we were an existing collaboration with a demonstrable record of successfully working together, we believe that have been very effective – we actually ‘hit the ground running’. To do this we required a varied skill set of people with a range of talents from astronomy, theoretical modelling and computing infrastructure.We expect our results and inferences will be used to investigate new problems and so act as a stimulus to the develop of new instrumentation and associated technology. For example much of our work relies upon far infrared/submm broad band observations using comparatively small format arrays. The next generation of astronomical space missions (e.g. SPICA ) will require large format arrays and also much higher resolution spectroscopic options. Across Europe such technologies are being developed and it is important that the work done in DustPedia acts as a driver for not just new science but also new technology that can produce that science.In addition to the above our project has impacted on the employment opportunities of new science graduates. This is both through the opportunities made available to post doctoral researchers but also to the education and training of postgraduate students who have been involved at every level. These new scientists have learnt many state-of-the-art skills that will hopefully enable them to develop into world-class researchers. The data and models they produce will provide the input into new projects both, observational, theoretical and technological as new problems that challenge our understanding are posed. One important legacy of DustPedia will to publicise and promote research in this area and therefore generate opportunities for new research alliances and strengthen existing collaborations with international partners.As a Europe wide (thankfully pre-Brexit) collaboration we hope we have demonstrated the importance of cooperation in carrying out a productive research programme and how the wide reaching European skills base can be effectively used. An essential part of this cooperation is the free and easy transfer of knowledge, techniques and ideas and importantly the promotion of the excitement of science. There is no single institute, or even a single EU country that could have provided all the necessary expertise and experience to make the progress we have made in this field. Primary addressees of the results of the DustPedia project are:• The scientific community primarily those studying the interstellar medium and galaxies.• The international space agencies (ESA, NASA).• The ground based astronomical community.• National scientific strategy boards and funding agencies.• University and secondary schools and the general public.During the course of the project we have hopefully created awareness outside of the immediate consortium. The dissemination strategy has been as follows:• To publish results in the highest impact journals appropriate to the subject. • To disseminate the knowledge to the scientific community through meetings, workshops and conferences.• To disseminate the knowledge through the publication of popular articles, conference proceedings andrefereed journal articles.• To disseminate our results to those who develop future astronomical and/or space technologies.• To inform the general public and promote the technical and scientific aims of the project through the public part of the website and through educational outreach activities.• Dissemination to the consortium through consortium meetings.• The public website will includes a project description, activities and objectives and a general background on the subject.• Press releases for wider circulation in consortium member countries.• Educational dissemination has occurred via the participation of undergraduate and graduate students in the project institutions. Our objective has been to communicate to the public the activities we have been participating in and to place these in context with our general understanding of the Universe. We have done this through the continued update of our publically available website (dustpedia.com) our entry in Wikipedia (https://en.wikipedia.org/wiki/DustPedia) and the public availability of data via our database (http://dustpedia.astro.noa.gr/). In addition to this we have been actively involved in promoting the DustPedia project by giving public/school talks, participating in summer schools and writing popular articles.List of Websites:Further information about the DustPedia project can be found at:dustpedia.comhttp://dustpedia.astro.noa.gr/https://en.wikipedia.org/wiki/or contact Professor Jon Davies (Jonathan.Davies@astro.cf.ac.uk)
