All Figures

$\begin{figure} \par\includegraphics[width=7.5cm, draft=false]{6962f1a.ps}\par\includegraphics[width=7.5cm, draft=false]{6962f1b.ps} \end{figure}$ Figure 1: Top: cumulative histogram of 100 surface brightness over 98% of the sky (solid line): fraction of the sky with a brightness lower or equal to I₁₀₀. The dashed line represent what would be expected from a cylindrical disk (cosecant law). Bottom: histogram of I₁₀₀ brightness over the whole sky with bin scaled logarithmically. For both curves we used the IRIS data projected on the Healpix grid (equal area pixels - http://healpix.jpl.nasa.gov) with a pixel size of 1.7 arcmin (nside = 2048). A background of 0.78 MJy sr^-1 was removed from the data to take into account the cosmic infrared background and any zodiacal light residual (Lagache et al. 2000; Hauser et al. 1998).
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=7.3cm, draft=false]{6962f2.ps} \end{figure}$ Figure 2: Power spectrum of a typical IRIS map (point sources removed) with its associated noise power spectrum and the estimated CIB level (convolved by the IRAS beam). The power law at large scale (small k) is due to the Galactic dust emission. The cutoff at $k\sim 0.1$ arcmin^-1 is due to the IRAS beam.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=7.5cm, draft=false]{6962f3.ps} \end{figure}$ Figure 3: Spectral index $\gamma$ of the power spectrum for our sample. Top. Histogram of the spectral index values. Median is -2.93, average is -2.96 and standard deviation is 0.3. Bottom. Spectral index as a function of average 100 brightness (CIB and zodiacal light subtracted).
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=7.5cm, draft=false]{6962f4.ps} \end{figure}$ Figure 4: Normalisation P(0.01) of the power spectrum at 0.01 arcmin^-1 of each map as a function of its mean 100 brightness. The solid line is our fit to the data (see Eqs. (5) and (6)). The dashed line is the relation given by Gautier et al. (1992).
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=7.5cm, draft=false]{6962f5.ps} \end{figure}$ Figure 5: Standard deviation as a function of the average interstellar brightness at 100 for each map of our sample. The noise level of each map and the CIB fluctuation level (0.09 MJy sr^-1 - see Miville-Deschênes et al. 2002) were removed quadratically from $\sigma$ . The solid line represents the two regimes given by Eqs. (7) and (8).
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=16.5cm, draft=false]{6962f6_col.ps} \end{figure}$ Figure 6: Left: a typical IRIS map at 100 . Center: a classical fBm map with same power spectrum as the IRIS map shown on the left. This fBm has Gaussian brightness fluctuations at all scale and everywhere. Right: a modified version of the classical fBm with only positive values and the same average, standard deviation and skewness values as the IRIS map. The fBm has been modified in order to reproduce the $\sigma(I) \propto \langle I\rangle$ relation which results in stronger brightness fluctuations in brighter regions.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=7.5cm, draft=false]{6962f7.ps} \end{figure}$ Figure 7: Cirrus noise at 100 for a 1 m diffraction limited telescope ( FHWM = 33 arcsec) as a function of I₁₀₀ brightness. Solid line is our estimate and dotted line is the estimate of Helou & Beichman (1990).
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=7.5cm, draft=false]{6962f8.ps} \end{figure}$ Figure 8: Contrast of interstellar emission: the standard deviation $\sigma _L$ is the standard deviation of interstellar fluctuations at the $12^\circ$ scale. Left: contrast values as a function of average brightness. Right: histogram of contrast values. According to Eq. (7), the contrast is close to constant for $\langle I\rangle$ lower than 10 MJy sr^-1. Outliers at low brightness corresponds to regions where there are bright structures on a very low level background, like the Magellanic Clouds.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=16.5cm, draft=false]{6962f9_col2.eps} %\end{figure}$ Figure 9: Wavelet decomposition of the IRIS map shown in Fig. 6-left. Top: wavelet coefficients map at scales 4, 8, 16 and 32 pixels (1 pixel = 1.5'). Bottom: Histogram of the wavelet coefficients in linear-log. A Gaussian fit to the histogram is superposed highlighting the non-Gaussian behavior.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=16.5cm, draft=false]{6962f10_col2.eps} %\end{figure}$ Figure 10: Wavelet decomposition of the classical fBm map shown in Fig. 6-center. See Fig. 9 for details.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=7.5cm, draft=false]{6962f11.ps} \end{figure}$ Figure 11: Skewness ( top) and Kurtosis ( bottom) of the wavelet coefficients from l=4 to 32 pixels for each IRIS map of our sample.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=16.5cm, draft=false]{6962f12_col2.eps} %\end{figure}$ Figure 12: Wavelet decomposition of the modified fBm map shown in Fig. 6-right. See Fig. 9 for details. The histogram of the wavelet coefficients of the IRIS map are over plotted (dots) showing how well the modified fBm reproduces the observed statistics.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=7.5cm, draft=false]{6962f13.ps} \end{figure}$ Figure 13: Top: Histogram of all the exponent $\psi$ used to simulate realistic fBm for each IRIS map of our sample. Bottom: Exponent $\psi$ as a function of the skewness $\chi$ of each IRIS map. The straight line is $\psi = 2.6 ~ \chi^{0.8}$ .
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=7.3cm, draft=false]{6962f2.ps} \end{figure}$	Figure 2: Power spectrum of a typical IRIS map (point sources removed) with its associated noise power spectrum and the estimated CIB level (convolved by the IRAS beam). The power law at large scale (small k) is due to the Galactic dust emission. The cutoff at $k\sim 0.1$ arcmin^-1 is due to the IRAS beam.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=7.5cm, draft=false]{6962f3.ps} \end{figure}$	Figure 3: Spectral index $\gamma$ of the power spectrum for our sample. Top. Histogram of the spectral index values. Median is -2.93, average is -2.96 and standard deviation is 0.3. Bottom. Spectral index as a function of average 100 brightness (CIB and zodiacal light subtracted).
Open with DEXTER

$\begin{figure} \par\includegraphics[width=7.5cm, draft=false]{6962f4.ps} \end{figure}$	Figure 4: Normalisation P(0.01) of the power spectrum at 0.01 arcmin^-1 of each map as a function of its mean 100 brightness. The solid line is our fit to the data (see Eqs. (5) and (6)). The dashed line is the relation given by Gautier et al. (1992).
Open with DEXTER

$\begin{figure} \par\includegraphics[width=7.5cm, draft=false]{6962f5.ps} \end{figure}$	Figure 5: Standard deviation as a function of the average interstellar brightness at 100 for each map of our sample. The noise level of each map and the CIB fluctuation level (0.09 MJy sr^-1 - see Miville-Deschênes et al. 2002) were removed quadratically from $\sigma$ . The solid line represents the two regimes given by Eqs. (7) and (8).
Open with DEXTER

$\begin{figure} \par\includegraphics[width=7.5cm, draft=false]{6962f7.ps} \end{figure}$	Figure 7: Cirrus noise at 100 for a 1 m diffraction limited telescope ( FHWM = 33 arcsec) as a function of I₁₀₀ brightness. Solid line is our estimate and dotted line is the estimate of Helou & Beichman (1990).
Open with DEXTER

$\begin{figure} \par\includegraphics[width=16.5cm, draft=false]{6962f9_col2.eps} %\end{figure}$	Figure 9: Wavelet decomposition of the IRIS map shown in Fig. 6-left. Top: wavelet coefficients map at scales 4, 8, 16 and 32 pixels (1 pixel = 1.5'). Bottom: Histogram of the wavelet coefficients in linear-log. A Gaussian fit to the histogram is superposed highlighting the non-Gaussian behavior.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=16.5cm, draft=false]{6962f10_col2.eps} %\end{figure}$	Figure 10: Wavelet decomposition of the classical fBm map shown in Fig. 6-center. See Fig. 9 for details.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=7.5cm, draft=false]{6962f11.ps} \end{figure}$	Figure 11: Skewness ( top) and Kurtosis ( bottom) of the wavelet coefficients from l=4 to 32 pixels for each IRIS map of our sample.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=16.5cm, draft=false]{6962f12_col2.eps} %\end{figure}$	Figure 12: Wavelet decomposition of the modified fBm map shown in Fig. 6-right. See Fig. 9 for details. The histogram of the wavelet coefficients of the IRIS map are over plotted (dots) showing how well the modified fBm reproduces the observed statistics.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=7.5cm, draft=false]{6962f13.ps} \end{figure}$	Figure 13: Top: Histogram of all the exponent $\psi$ used to simulate realistic fBm for each IRIS map of our sample. Bottom: Exponent $\psi$ as a function of the skewness $\chi$ of each IRIS map. The straight line is $\psi = 2.6 ~ \chi^{0.8}$ .
Open with DEXTER