[yt-dev] bug in SFRD calculation

[Geoffrey So]

Hi,

I've issued a pull request to yt_analysis, the change diverting from the
norm for yt to output in proper units, but as Greg pointed out in the case
of star formation rate density everyone uses comoving Mpc, so I think the
change will be useful.  The original quantity doesn't make much sense, it
was dividing Msol/yr at all redshifts by the proper volume at the last
redshift.

The script should show identical curves offset by a tiny bit.

From
G.S.

#------------
from yt.mods import *
from yt.analysis_modules.star_analysis.api import *
import pylab

pf=load('HD1200/HD1200')
dd=pf.h.all_data()
sfr = StarFormationRate(pf, data_source=dd, bins=10)
sfr.write_out(name="StarFormationRate.out")
time, lookback, z, Msolyr, MsolyrMpc3, Msol,
cumMsol=na.loadtxt('StarFormationRate.out',usecols=(0,1,2,3,4,5,6),skiprows=2,unpack=True)

cMpc3,=pylab.plot(z,Msolyr/(pf['CosmologyComovingBoxSize']/pf['CosmologyHubbleConstantNow'])**3)
New,=pylab.plot(z,MsolyrMpc3+3e-5)

pylab.legend([cMpc3,NewMpc3],
             ['[s$^{-1}$cMpc$^{-3}$]',
              'New'])

pylab.savefig('SFRD.eps')
#------------


On Fri, Apr 4, 2014 at 9:36 AM, Geoffrey So <gsiisg at gmail.com> wrote:

> Hi Greg,
>
> My original script has them converted from proper to comoving and that's
> when I noticed the bug :)
>
> Nathan,
>
> I was trying to convey the idea, wasn't sure if what I described is the
> proper fix, I'll issue the PR if it tests out correctly.
>
> From
> G.S.
>
> Sent from my iPhone
>
> On Apr 4, 2014, at 6:15 AM, Greg Bryan <gbryan at astro.columbia.edu> wrote:
>
> Yes, although the star formation rate in cosmology is usually calculated
> in a comoving volume, not proper (see, e.g.,
> http://arxiv.org/pdf/1403.0007v2.pdf).
>
> Greg
>
> On Apr 3, 2014, at 11:15 PM, Nathan Goldbaum <nathan12343 at gmail.com>
> wrote:
>
> Hi Geoffrey,
>
> I'm not very familiar with the star formation rate calculator, but if you
> think there is a bug I'd encourage you to submit a pull request with a fix.
>
> -Nathan
>
>
> On Thu, Apr 3, 2014 at 6:37 PM, Geoffrey So <gsiisg at gmail.com> wrote:
>
>> There seems to be a bug in the star formation rate density calculation
>> routine.  The volume is only the real 'proper volume' at the current
>> redshift and not at any other redshifts, because it uses the current
>> redshift in the calculation of proper volume in all redshifts.
>>
>> #---------------------------------
>> from yt.mods import *
>> from yt.analysis_modules.star_analysis.api import *
>> import pylab
>>
>> pf=load('HD1200/HD1200')
>> dd=pf.h.all_data()
>> sfr = StarFormationRate(pf, data_source=dd, bins=10)
>> sfr.write_out(name="StarFormationRate.out")
>> time, lookback, z, Msolyr, MsolyrMpc3, Msol,
>> cumMsol=na.loadtxt('StarFormationRate.out',usecols=(0,1,2,3,4,5,6),skiprows=2,unpack=True)
>>
>>
>> cMpc3,=pylab.plot(z,Msolyr/(pf['CosmologyComovingBoxSize']/pf['CosmologyHubbleConstantNow'])**3)
>> currentz,=pylab.plot(z,MsolyrMpc3/(1+pf.current_redshift)**3+1e-5)
>> Expect,=pylab.plot(z,MsolyrMpc3/(1+z)**3)
>>
>> pylab.legend([cMpc3,currentz,Expect],
>>              ['[s$^{-1}$cMpc$^{-3}$]',
>>               'use only current z to fit',
>>               'Expected to fit'])
>>
>> pylab.savefig('SFRD.eps')
>> #---------------------------------
>>
>>
>> in the file:
>> yt/analysis_modules/star_analysis/sfr_spectrum.py
>> line 151:
>>             self.Msol_yr_vol.append(self.mass_bins[i] / \
>>                 (self.time_bins_dt[i] * tc / YEAR) / vol)
>>
>> The vol is a fixed number instead of varying by the redshift, even though
>> the numerator in solar masses per year is from different redshifts.
>>
>> So I think a couple lines above it should create an empty array of vol
>> vol = []
>> In the loop
>> for i, time...
>> calculate the right volume vol[i]
>> Then instead of dividing by the constant vol, divide by vol[i]
>> corresponding to the right redshift.
>>             self.Msol_yr_vol.append(self.mass_bins[i] / \
>>                 (self.time_bins_dt[i] * tc / YEAR) / vol[i])
>>
>> From
>> G.S.
>>
>>
>>
>> _______________________________________________
>> yt-dev mailing list
>> yt-dev at lists.spacepope.org
>> http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
>>
>>
> _______________________________________________
> yt-dev mailing list
> yt-dev at lists.spacepope.org
> http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
>
>
> ----------------------------------
> Greg Bryan
> Associate Professor of Astronomy
> Department of Astronomy, Columbia University
> 1325 Pupin Physics Laboratories, Mail Code 5246
> 550 West 120th Street
> New York, New York 10027 (USA)
>
> email: gbryan at astro.columbia.edu
> phone: +1(212)854-6837
> fax  : +1(212)854-8121
>
>
>
>
>
> _______________________________________________
> yt-dev mailing list
> yt-dev at lists.spacepope.org
> http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.spacepope.org/pipermail/yt-dev-spacepope.org/attachments/20140404/2604cee3/attachment.htm>