Hunting Electromagnetic Counterparts of Gravitational-wave Events Using the Zwicky Transient Facility

Detections of coalescing binary black holes by LIGO have opened a new window of transient astronomy. With increasing sensitivity of LIGO and participation of the Virgo detector in Cascina, Italy, we expect to soon detect coalescence of compact binary systems with one or more neutron stars. These are the prime targets for electromagnetic follow-up of gravitational wave triggers, which holds enormous promise of rich science. However, hunting for electromagnetic counterparts of gravitational wave events is a non-trivial task due to the sheer size of the error regions, which could span hundreds of square degrees. This may require deep observation with large field-of-view telescopes and/or use of galaxy catalogs. The Zwicky Transient facility (ZTF), scheduled to begin operation in 2017, is designed to cover such large sky-localization areas. In this work, we present the strategies of efficiently tiling the sky to facilitate the observation of the gravitational wave error regions using ZTF. To do this, we used simulations consisting of 475 binary neutron star coalescences detected using a mix of two- and three-detector networks. Our studies reveal that, using two overlapping sets of ZTF tiles and a (modified) ranked-tiling algorithm, we can cover the gravitational-wave sky-localization regions with half as many pointings as a simple contour-covering algorithm. We then incorporated the ranked-tiling strategy to study our ability to observe the counterparts. This requires optimization of observation depth and localization area coverage. Our results show that observation in r-band with similar to 600 seconds of integration time per pointing seems to be optimum for typical assumed brightnesses of electromagnetic counterparts, if we plan to spend equal amount of time per pointing. However, our results also reveal that we can gain by as much as 50% in detection efficiency if we linearly scale our integration time per pointing based on the tile probability.