Cosmic temperature fluctuations from two years of COBE DMR observations

The first two years of COBE DMR observations of the CMB anisotropy are analyzed and compared with our previously published first year results. The results are consistent, but the addition of the second year of data increases the precision and accuracy of the detected CMB temperature fluctuations. The two-year 53 GHz data are characterized by RMS temperature fluctuations of DT=44+/-7 uK at 7 degrees and DT=30.5+/-2.7 uK at 10 degrees angular resolution. The 53X90 GHz cross-correlation amplitude at zero lag is C(0)^{1/2}=36+/-5 uK (68%CL) for the unsmoothed 7 degree DMR data. A likelihood analysis of the cross correlation function, including the quadrupole anisotropy, gives a most likely quadrupole-normalized amplitude Q_{rms-PS}=12.4^{+5.2}_{-3.3} uK (68% CL) and a spectral index n=1.59^{+0.49}_{-0.55} for a power law model of initial density fluctuations, P(k)~k^n. With n fixed to 1.0 the most likely amplitude is 17.4 +/-1.5 uK (68% CL). Excluding the quadrupole anisotropy we find Q_{rms-PS}= 16.0^{+7.5}_{-5.2} uK (68% CL), n=1.21^{+0.60}_{-0.55}, and, with n fixed to 1.0 the most likely amplitude is 18.2+/-1.6 uK (68% CL). Monte Carlo simulations indicate that these derived estimates of n may be biased by ~+0.3 (with the observed low value of the quadrupole included in the analysis) and {}~+0.1 (with the quadrupole excluded). Thus the most likely bias-corrected estimate of n is between 1.1 and 1.3. Our best estimate of the dipole from the two-year DMR data is 3.363+/-0.024 mK towards Galactic coordinates (l,b)= (264.4+/-0.2 degrees, +48.1+/-0.4 degrees), and our best estimate of the RMS quadrupole amplitude in our sky is 6+/-3 uK.