Power spectrum to probe turbulence in Interstellar medium using 21cm line

Turbulent dynamics generate random fluctuations in density and velocity in the Interstellar Medium (ISM) of spiral and dwarf galaxies. Observationally, H~{sc i} 21-cm radiation provides a good probe of these stochastic processes and helps us to know more about their nature and generating mechanism. Structure function, auto-correlation function, power spectrum are some of statistical estimators of these fluctuations which have been used in literature, whose studies helps in understanding the generating mechanisms of these structures. In this thesis we are interested in estimating the column density and velocity power spectrum at scales as large as 10 Kpc, to understand the energy associated with these structures and the type of forcing of the turbulence. Estimating power spectrum from observations can be done either by using the directly measured quantity visibility or by reconstructing the sky brightness distribution (or image) from the visibilities. By using simulated observations of a model galaxy with known power spectrum we find that the visibility based estimator reproduce the true power spectrum. The image based estimator has a scale dependent bias that is highly correlated to the incompleteness in the baseline coverage of the interferometer. Interestingly, we found that the noise in the reconstructed image has zero mean and hence the locally averaged quantities estimated from the reconstructed image are unbiased. We implement the visibility moment estimator as discussed in Dutta et al. (2015) and estimate the line of sight velocity fluctuation of the spiral galaxy NGC6946. Our result strongly indicate presence of large scale turbulence in the spiral galaxies. We conclude that we would need to have higher signal to noise H~{sc i}observations to definitively comment on the generating mechanisms of the large scale turbulence.