Home‎ > ‎

Science Requirements LRIFU


June 27th 2010:
My goal is to determine the optimum spectral resolution to observe supernovae in order to "classify" them. Classification is somewhat ill defined at this point, but parameters we are interested in recovering are redshift, phase, type, and "weirdness". Right now, I do not have an answer, instead I'm working on an objective approach to determine the "power" of a spectral resolution. My initial approach shows some promise but I'm a bit confused by the results.

My approach:
1. Load all the SN Ias with phase of minus one day or plus ten days from the set of spectra provided by Quimby.( Need to find out again what the name of this database is.)
2. Convolve the spectra with the spectral response of a lenslet IFU. The lensets produce circular images that in turn are converted into one dimensional spectra. As a result, the kernel has a hard edge (in velocity space) which can be seen in the attached file "velocity_profiles.pdf". In said file I plot the velocity profile of the lenslet array in red (total width of 3000 km/s or R=100) and the equivalent gaussian with FWHM=1500 km/s. Note that the gaussian distributes significant light to the wings.
3. Normalize all the spectra (there are six) using a broad median filter.
4. Measure the standard deviation of the six spectra.
5. Plot the standard deviation of the spectra as a function of wavelength.

Results are attached at the end of this web page. Plots throw three types of spectra:
    1. The top labeled curves are SN Ia spectra with some arbitrary constant added to enhance readability.
    2. The median filtered versions of the same spectra are shown normalized to 0.7 or so.
    3. A line with hexagonal point showing the standard deviation of the set of median spectra.
    4. Note that the low resolution IFU (LRIFU) has a S/N spec of 30 per resolution element, thus the line of a five sigma detection is 5/30 (the black horizontal line).
    5. You'll note that there seem to be several critical features for classification
        The MgII Triplet Red wing?
        Fe and Si features from 5500 - 6000
        The SiII 6150 feature
        The OIII triplet at 7770
        The CaII IR triplet at 8500
 

Conclusions:
    The negative of this approach is the limited set of spectra used in it: only Ia SN. Yet, I'm confused as to why even very low resolution (R=25) seems to carry a lot of weight?


List of Mountaintop Observatories with primary diameter 1 < D < 7 m
Anderson Mesa Station (Lowell?): 1.1, 1.8
Anglo Australian Observatory: 1.2, 3.9
Apache Point: 1, 2.5 (SDSS), 3.5
Asiago (Italy): 1.22
Boyden Observatory (South Africa): 1.5
Byurakan Observatory (Armenia): 1, 2.6
Calar Alto Observatory (MPIA): 1.2, 1.5, 2.2, 3.5
Canopus (Australia): 1 
CTIO: 1, 1.3 (2mass), 1.5
David Dunlap Observatory: 1.9
Dominion (Canada): 1.2, 1.8
ESO: 1, 2.2, 3.6, 3.5 (NTT) --- AT Least
Fan Mountain (Virgina): 1
Guillerm Haro (spain): 2.2
Haleakala: 2, 2 (Magnum), 3.1 (AEOST) 
Haute-Provence: 1.2, 1.5, 1.9
Hoher (Bonn): 1
Indian Astronomical: 2 (IR)
La Palma (Canary Islands): 1, 1.2, 2, 2.5, 2.6, 3.5, 4.2
Karl Schwarzschild: 2 m (Schmidt!)
Kitt Peak: 1.3, 1.8, 2.1, 2.3, 2.4, 3.5, 4
Klet' (Czech Republic): 1
LCO: 1, 1.3, 2.5, 6.5
Lenocito (Argentina): 2.2
Lick: 1, 2.4, 3
Llando del Hato (Venezual): 1 ??
Lowell: 1.1, 1.8, 4.2
Lulin: 1
Magdelna ridge: 2.4
McDonald: 2.1, 2.7
MDM: 1.3, 2.4
Mont Megnatic (Canada): 1.6
Morgan-Monroe (Indiana): 1.3
Mount Graham: 1.8 (VATT)
Mount John (New Zealand): 1, 1.8
Mount Laguna (SD): 1
Mt Lemmon: 1, 1.5, 1.5
NAOJ (Japan): 1.9
Turkey National: 1.5
Nordic Optical: 2.6
Palomar: 1.2, 1.5, 5.1
Pic du Midi de Biogrre: 1, 2
Pico dos Dias (Brazil): 1.6
Rozhen (Bulgaria): 2
Shamakhi (Ajerbijan): 2
Shanghi: 1.5
Siding Spring (Australia): 1.2, 1.3, 2, 2.3, 3.9
Sierr Nevada (Spain): 1.5
South African: 1, 1.2, 1.4, 1.9
Teide: 1, 1.2, 1.5
Tuolra (Finland): 1
Whipple: 1.2, 1.3, 2.5, 6.5
Yunnan: 1, 2.4
Zimmerwarld (Swiss): 1

Ċ
Nick Konidaris,
Jun 27, 2010, 3:08 PM
Ċ
Nick Konidaris,
Jun 27, 2010, 3:08 PM
Ċ
Nick Konidaris,
Jun 27, 2010, 3:08 PM
Ċ
Nick Konidaris,
Jun 27, 2010, 3:08 PM
Ċ
Nick Konidaris,
Jun 27, 2010, 2:46 PM
Comments