Empirical Modeling of Zodiacal Backgrounds to Improve JWST NIRISS/SOSS Data Reduction

Abstract

Zodiacal light -- arising from both the thermal emission and scattered sunlight by interplanetary dust -- is the dominant component of the sky background in NIRISS Single Object Slitless Spectroscopy (SOSS) observations. The GR700XD grism disperses the zodiacal background across multiple diffraction orders, producing a characteristic two-dimensional, spectral order-dependent background pattern on the detector that reflects the combined contribution of overlapping orders. Observations reveal significant variability in background intensity driven by JWST's sky pointing and seasonal changes, underscoring the need for precise background subtraction during SOSS data reduction. Current methods rely on a generic background template derived during commissioning, scaled to match individual exposures. However, mis-scaled templates can leave behind structured residuals that may compromise the precision of exoplanet transit depth measurements. To improve background modeling, we conducted two calibration programs (PID 4479 and 6658) using the FULL frame readout mode and a 5-row by 2-column mosaic pattern to sample a range of sky positions. These observations enable empirical reconstruction of the sky background and provide detailed insights into its spatial and spectral characteristics. We present a library of empirically derived background templates and evaluate their performance, alongside the current template, by benchmarking against the PID 2113 dataset, which includes contemporaneous background exposures. This work supports aims to enhance background subtraction for SOSS time-series to achieve higher-precision exoplanet spectroscopy with JWST.