CIBER 4th flight fluctuation analysis: Pseudo-power spectrum formalism, improved source masking and validation on mocks

TL;DR

This paper improves the measurement of extragalactic background anisotropies by refining flat field correction and source masking techniques for CIBER, validated through simulations to ensure unbiased fluctuation recovery.

Contribution

It introduces a pseudo-power spectrum formalism for flat field correction and a machine learning-based source masking method, enhancing sensitivity and accuracy in CIBER measurements.

Findings

Residual flat field errors contribute less than 20% to power spectrum uncertainty.

The new masking technique achieves over two magnitudes deeper masking than 2MASS.

Simulations confirm unbiased recovery of sky fluctuations.

Abstract

Precise, unbiased measurements of extragalactic background anisotropies require careful treatment of systematic effects in fluctuation-based, broad-band intensity mapping measurements. In this paper we detail improvements in methodology for the Cosmic Infrared Background ExpeRiment (CIBER), concentrating on flat field errors and source masking errors. In order to bypass the use of field differences, which mitigate flat field errors but reduce sensitivity, we characterize and correct for the flat field on pseudo-power spectra, which includes both additive and multiplicative biases. To more effectively mask point sources at 1.1 $\mu$m and 1.8 $\mu$m, we develop a technique for predicting masking catalogs that utilizes optical and NIR photometry through random forest regression. This allows us to mask over two Vega magnitudes deeper than the completeness limits of 2MASS alone, with errors in the shot noise power remaining below $<10\%$ at all masking depths considered. Through detailed simulations of CIBER observations, we validate our formalism and demonstrate unbiased recovery of the sky fluctuations on realistic mocks. We demonstrate that residual flat field errors comprise $<20\%$ of the final CIBER power spectrum uncertainty with this methodology.