Weak lensing reconstruction through cosmic magnification. II. Improved power spectrum determination and map-making
The existence of galaxy intrinsic clustering severely hampers the weak lensing reconstruction from cosmic magnification. In paper I \citep{Yang2011}, we proposed a minimal variance estimator to overcome this problem. By utilizing the different dependences of cosmic magnification and galaxy intrinsic clustering on galaxy flux, we demonstrated that the otherwise overwhelming galaxy intrinsic clustering can be significantly suppressed such that lensing maps can be reconstructed with promising accuracy. This procedure relies heavily on the accuracy of determining the galaxy bias from the same data. Paper I adopts an iterative approach, which degrades toward high redshift. The current paper presents an alternative method, improving over paper I. We prove that the measured galaxy clustering between flux bins allows for simultaneous determination of the lensing power spectrum and the flux dependence of galaxy bias, at this redshift bin. Comparing to paper I, the new approach is not only more straightforward, but also more robust. It identifies an ambiguity in determining the galaxy bias and further discovers a mathematically robust way to suppress this ambiguity to non-negligible level ($\sim 0.1%$). The accurately determined galaxy bias can then be applied to the minimal variance estimator proposed in paper I to improve the lensing map-making. The gain at high redshift is significant. These maps can be used to measure other statistics, such as cluster finding and peak statistics. Furthermore, by including galaxy clustering measurement between different redshift bins, we can also determine the lensing cross power spectrum between these bins, up to a small and correctable multiplicative factor.