clumpy home :: register :: catalog :: run models :: contact

Welcome to Clumpy

Welcome to the home of CLUMPY, a code modeling AGN clumpy dust torus emission, developed at the University of Kentucky.

Latest news

The CLUMPY models have been updated.

The emission of CLUMPY clouds is the angle-averaged emission of all slab orientations. To "label" such "synthetic" clouds we previously used the optical depth along the normal of the slab.

We recently compared the synthetic clouds to spherical clouds (3-d radiative transfer). For a sphere, the optical depth along a central ray is the highest possible of all rays. For a slab the optical depth along the normal is the smallest of all rays. The calculations show that the effective optical depth through a synthetic cloud is two times higher than the optical depth along a slab normal.

The absorption caused by clouds is not affected by this (the spectral shape doesn't change, only the "optical depth label" assigned to the clouds is now two times higher). However, the cloud emission does change, because the cloud source function is a wavelength-dependent function τ(λ). Thus, the overall spectral shape of the model SEDs does change (up to 20 percent at any given wavelength, in our experience).

These new results will be published in Heymann, Nikutta, and Elitzur (in prep.). Meanwhile, we make available the newly calculated CLUMPY model database as an HDF5 file (1.2 GB). The old online model database is still active, but we discourage you from using it in favor of the new models.

Note that the optical depth values in the new models are now "effective optical depths", rather than the optical depth along the slab normal used before.

Usage of the HDF5 file (here with Python and numpy, h5py, matplotlib modules installed):

import h5py
from numpy import exp
import pylab as p

h = h5py.File('clumpy_models_201410.hdf5')

# see what's there

	[(u'N0', <HDF5 dataset "N0": shape (1247400,), type "<i4">),
	 (u'Y', <HDF5 dataset "Y": shape (1247400,), type "<i4">),
	 (u'f2', <HDF5 dataset "f2": shape (1247400,), type "<f4">),
	 (u'flux_tor', <HDF5 dataset "flux_tor": shape (1247400, 119), type "<f4">),
	 (u'flux_toragn', <HDF5 dataset "flux_toragn": shape (1247400, 119), type "<f4">),
	 (u'i', <HDF5 dataset "i": shape (1247400,), type "<i4">),
	 (u'lam10', <HDF5 dataset "lam10": shape (1247400,), type "<f8">),
	 (u'lam18', <HDF5 dataset "lam18": shape (1247400,), type "<f8">),
	 (u'ptype1', <HDF5 dataset "ptype1": shape (1247400,), type "<f4">),
	 (u'q', <HDF5 dataset "q": shape (1247400,), type "<f4">),
	 (u's10', <HDF5 dataset "s10": shape (1247400,), type "<f4">),
	 (u's18', <HDF5 dataset "s18": shape (1247400,), type "<f4">),
	 (u'sig', <HDF5 dataset "sig": shape (1247400,), type "<i4">),
	 (u'tv', <HDF5 dataset "tv": shape (1247400,), type "<f4">),
	 (u'wave', <HDF5 dataset "wave": shape (119,), type "<f8">)]

# As you can see there are 1247400 models in the database, with
# 119 wavelength per SED. The CLUMPY parameters are
# N0,Y,i,q,sig,tv. The wavelength array is wave. Some derived
# properties are also stored: f2 (dust covering fraction), ptype1
# (the probability to see the AGN unobscured), s10 & s18 (the
# 10 & 18-micron silicate feature strengths), lam10 & lam18
# (the wavelengths where the extrema of the 10 and 18-micron
# features occur). Finally, the model SEDs are stored in flux_tor
# (torus-only SED) and flux_toragn (torus with AGN input spectrum
# added in, regardless of the actual probability to see the AGN
# in that model).

# load some parameter arrays into RAM
N0 = h['N0'][:]
sig = h['sig'][:]
i = h['i'][:]
pesc = h['ptype1'][:]
s10 = h['s10'][:]
s18 = h['s18'][:]

# make selection mask for all type1 models
type1 = (pesc > 0.5)

# plot an s10 vs. s18 diagram
p.scatter(s10[~type1],s18[~type1],marker='.',s=0.5,linewidths=0,c='red',alpha=0.2) # type2
p.scatter(s10[type1],s18[type1],marker='.',s=0.5,linewidths=0,c='blue',alpha=0.2) # type1
CLUMPY description
# get the torus-only SEDs for all models with i=0 and N0<5
mask = (i==0) & (N0<5)
seds = flux_tor[mask,:]

# close the HDF5 file

# And so on... If you have questions, email me: robert dot nikutta at gmail dot com

2010-September-24 3:00pm EST: We have found a bug in the old code that caused the torus models to have incorrect normalization. The spectral shapes of the SEDs were not affected by the bug, only the absolute flux levels were. The bug has been fixed, and the models re-generated. The catalog and the online creation of models are now re-enabled. An erratum to Nenkova et al. 2008b (ApJ, 685, 160) has been published in ApJ, 723, 1827.

The torus covering factor 'f2' -- see Eq. 9 in Nenkova et al. 2008a (ApJ, 685, 147) -- is now called 'f2' in the CLUMPY output files (it used to be called 'Pnorm').

Together with the catalog update we have increased the number of Y values that can be selected (multiple selections OK now). The maximum number of models that can be calculated online in one run is now 1000 (up from 100). Finally, the models will now be generated faster (more CPUs).

What would you like to do?

Search for already existing models (please notice our disclaimer):

Get interactive with CLUMPY:

CLUMPY description

Click here for a description of our CLUMPY model


All data available on our website are public, and reflect the latest state of our scientific research and development. Please feel free to use them for the benefit of your research, provided that acknowledgement is made in each publication. The bibliographic references are:

Nenkova, M., Sirocky, M.M., Ivezic, Z. & Elitzur, M.,
"AGN Dusty Tori: I. Handling of Clumpy Media", 2008,
ApJ, 685, 147, and

Nenkova, M., Sirocky, M.M., Nikutta, R., Ivezic, Z. & Elitzur, M.,
"AGN Dusty Tori: II. Observational Implications of Clumpiness", 2008,
ApJ, 685, 160

Please note that models may be subject to change or update at any time. Therefore, please make sure to use the latest data available (the catalog provides the date of creation of each model file).

The CLUMPY group, University of Kentucky

Last updated: October 15, 2014 by Robert Nikutta (robert at pa dot uky dot edu)