(Proposal ID) S15A-047 (PI) Kodama, Tadayuki (Proposal Title) MAHALO-DEEP: Ultra-deep NB Halpha imaging of two protoclusters at z>2 (Abstract) Our Mahalo-Subaru project has successfully mapped out Halpha emitters associated to the two protoclusters (USS1558 and PKS1138) at z>2 using dedicated narrow-band filters, and they turn out to be the densest/richest protoclusters ever identified at z>2. As such, they offer us the unique laboratories to test the effect of galaxy environments on galaxy formation. We now propose to go much deeper to yet unexplored regime of the main sequence (MS) of star forming galaxies (SFGs) in protoclusters with Halpha, down to dust corrected star formation rate of 10M_{sun}/yr and stellar mass of 2.5x10^9M_{sun}. This is motivated to better balance the depth of our current MOIRCS data (NB, K_s) to our recently obtained HST data (WFC3, ACS). The current shallower data already show a hint of environmental dependence below 5x10^9M_{sun}, in the sense that SF activity may be more often boosted in high density regions. The proposed deeper observations will make this situation much clearer. Together with the morphological information by HST we will interpret the physical origin of the possible SF boosting. Moreover, deep Halpha imaging will probe diffuse, spatially extended Halpha emission in bright sources. They will hold the key to understanding gaseous outflows driven by starburst and/or AGN feedback.