I have the TOA solar spectral irradiance data produced by Chance& Kurucz 2010 (can be downloaded in the link). The full-width-half-maximum (FWHM) of this data is 0.04 nm and the series is sampled every 0.01 nm. I want to convolve the data to 1 nm resolution (1 nm FWHM). But how can I do this considering that it is already convolved at 0.04?
Assuming a filter that has a Gaussian response, using R, I started by creating a "Guassian weights" by:
sigma
The requirement is basically to obtain more data points among a given set of x-y data. The estimated data should follow the same y-x variation as that of the given set. This can be tried with Lagrange interpolation technique. That needs a functional form for y depending on x. Gaussian form can be used. The given algorithm seems to be correct.
Thank you Barun. I will test the result integrating the curves (original and convolved), that should give me the same value.
I don't Know your code, but I know that if you have two Gaussian profiles and you convolve one with width a and another with width b, then the result is a Gaussian of width c where
c= sqrt ( a-squared + b-squared )
So in you Case you want c = 1 and you have a=0.04
So you should convolve your spectrum with a value b that solves the above equation, That will be very close to 1. Is this what you were asking?
Thank you Philip, I believe it is. I want to know the "solar constant" (TOA irradiance) in 380 nm, one of the bands of the TOMS (Total Ozone Mapping Spectrometer) sensor. TOMS bands have all 1nm FWHM. The data of Chance & Kurucz 2010 is convolved in 0.04 nm, so I want to go from the original data to the TOMS resolution.
Alexandre
It seems that the problem is one in interpolation from one grid spacing to another. I'll take a second look at this problem in a week or so.
Did you make any comparison plots? If so, please send them to me. I can start there.
Rick
If what you want is the TOA solar spectral irradiance at 1 nm, your source is not a very good one, at least at short wavelengths. There is a published observed solar spectrum from space Thuillier et al. 2003), and also better models that take into account non-LTE (my spectra).
- Badges
- Science topic