Diagnose and correct unexplained dark-current fluctuations

Determine the cause of the WINTER detectors’ dark-current fluctuations between two possible levels, separated by approximately 50–100 counts at constant temperature, and develop a method to correct for this effect in the WINTER data-processing pipeline.

Background

The WINTER InGaAs detectors exhibit dark-current variations that depend on ambient temperature, time of day, and other environmental conditions. In addition to these understood or partially mitigated variations, the authors observe that the dark current can randomly switch between two distinct levels even when the temperature remains constant.

This unexplained fluctuation contributes to the systematic uncertainty floor of WINTER photometry, at approximately 5% relative to the sky counts. The pipeline does not currently correct for it, making diagnosis and mitigation an unresolved technical problem for improving photometric accuracy.

References

Additionally, we observe a third effect, where even at constant temperature the dark current fluctuates between two possible levels, separated by $\approx50-100$ counts (see Figure \ref{fig:dark_counts_variation}). This effect is currently undiagnosed and we do not have a way of correcting for it.

— A data processing pipeline for the WINTER near-infrared surveyor using the $\texttt{mirar}$ framework  (2609.10675 - Karambelkar et al., 9 Sep 2026) in Section 3.2, subsection “Dark subtraction”