JC gave me a simulated 1 deg square SZ image (from Martin white?). I processed it as follows to try and make predictions of what experiments might actually be able to see.
An illustration of the impact of angular resolution requirements for an SZ survey instrument. All the images are 1 degree square (axis in arc minutes), and the color scales are in uK. The top three frames show an image of the SZ effect at 140 GHz from a state-of-the-art N-body simulation, a realization of the CMB from a currently popular cosmological model, and the sum of the two effects. The lower nine frames are arranged column-wise by the assumed instrument angular resolution. The upper row shows the SZ plus CMB filtered with three assumed FWHM beam sizes. The middle row shows what happens when we take these images, add pixel noise equivalent to 10 uK per beam, and then apply an optimal filter designed to re-extract the SZ effect. For this row the color scale has an additional set of labels on the left side showing the detection significance in standard deviations. The final row of frames are identical to the upper-left, with the addition of crosses marking the location of each cluster detection with significance greter than 4 sigma.
Top three frames are image power spectra illustrating the construction of the optimal filter used above. Note y axis is log. The red curve is the beam smeared SZ signal, the blue curves are the noise (exponentially falling CMB, and flat instrumental noise), and the green curve is the spectrum of the realized image. Note that the SZ has comparable power to the noise in the 1 arc minute case, for a narrow range of spactial frequencies.
The lower three frames show histograms of pixel values for the filtered images. Y axis is again log. A Gaussian fit has been made in the red region (+/- 1.5 sigma) - the green curve is an extrapolation into the un-fit region. The low side tail is caused by SZ detections. Note that in the 4 arc minute case there is no evidence of SZ signal.
Finally here is the first figure again, but this time for 0.75, 1.0 and 1.5 arc minute beams.