Using the Modified Nearest Neighbor Method to Correct Fiber-collision Effects on Galaxy Clustering

Fiber collision is a persistent problem faced by modern spectroscopic galaxy surveys. In this work, we propose a new method to correct for this undesired effect, focusing on the clustering from the fiber-collision scale up to $\lesssim 10 \rm Mpc\,h^{-1}$. We assume that the fiber-collided galaxies are in association with their nearest three angular neighbors. Compared with the conventional nearest-neighbor method, we have properly accounted for the foreground (background) galaxies that are associated with the foreground (background) cosmic webs relative to the nearest neighbor. We have tested the new method with mock catalogs of the Sloan Digital Sky Survey (SDSS) Data Release 7 (DR7). The test demonstrates that our new method can recover the projected two-point correlation functions at an accuracy better than 1\% on small (below the fiber-collision scale) to intermediate (i.e., $10 \rm Mpc\,h^{-1}$) scales, where the fiber collision takes effect and the SDSS main sample can probe. The new method also gives a better recovery of the redshift-space correlation functions almost on all scales that we are interested in.