Multi-color characterization of optically invisible FU Orionis-type outbursts: Demonstration and prospects for the WINTER survey

TL;DR

This study demonstrates that a multi-color infrared time-domain approach combining WINTER with NEOWISE enables the discovery and detailed characterization of heavily embedded FU Orionis-type outbursts, which are largely missed by optical surveys. By confirming two infrared-bright FUor events—one previously known and one newly identified—the work maps their environments, progenitors, spectral signatures, and luminosity-driven accretion rates, using near-infrared spectroscopy and broad SED modeling. The results place the outbursts within the canonical FUor accretion framework and show that distances and extinction critically shape derived luminosities and $\dot{M}$ values, with implications for the census of FUors in the Galactic plane. Overall, WINTER’s real-time, three-color infrared monitoring promises to advance understanding of episodic accretion and early disk evolution in deeply embedded YSOs, enabling large-sample statistics in future surveys.

Abstract

Episodic mass accretion is the dominant mechanism for mass assembly in the proto-stellar phase. Although prior optical time-domain searches have allowed detailed studies of individual outbursts, these searches remain insensitive to the earliest stages of star formation. In this paper, we present the characterization of two FU Orionis (FUor) outbursts identified using the combination of the ground-based, near-infrared Wide-field Infrared Transient Explorer (WINTER) and the space-based, mid-infrared NEOWISE survey. Supplemented with near-infrared spectroscopic follow-up, we show that both objects are bona fide FUor type outbursts based on i) their proximity to star-forming regions, ii) large amplitude (2-4 magnitudes) infrared brightening over the last decade, iii) progenitor colors consistent with embedded (Class I) protostars, and iv) "mixed-temperature" infrared spectra exhibiting characteristic signatures of cool outer envelopes and a hot inner disk with a wind. While one source, WNTR24-cua, is a known FUor which we independently recover; the second source, WNTR24-egv, is a newly confirmed object. Neither source is detected in contemporaneous ground-based optical imaging, despite flux limits $\gtrsim 100\times$ fainter than their infrared brightness, demonstrating the capabilities of WINTER to identify heavily obscured young stellar object (YSO) outbursts. We highlight the capabilities of the Galactic Plane survey of the recently commissioned WINTER observatory in addressing the poorly understood FUor population with its unique combination of real-time detection capabilities, multi-color sensitivity, weekly cadence, and wide area coverage.

Figures (6)

• Figure 1: WINTER $J$-band detections of WNTR24-cua (left) and WNTR24-egv (right) overlaid on images of the events' environments. The larger image shows the area surrounding the outburst taken with the Spitzer Space Telescope Spitzer with IRAC channels 1 (3.6 $\mu$m), 2 (4.5 $\mu$m), and 4 (8.0 $\mu$m) used as blue, green, and red, respectively, in the composite images. Inset: details from WINTER science images showing the initial detection of the event (left), a corresponding archival reference image from UKIRT (middle), and the difference image between WINTER and UKIRT (right).
• Figure 2: Multi-band light curves for WNTR24-cua (top) and WNTR24-egv (bottom) combining data from WISE/NEOWISE ($W1$, $W2$), WINTER ($Y$, $J$, $H_s$), VVV ($J$, $H$, $K_s$), and Gaia ($G$). Inverted open triangles mark 5$\sigma$ upper limits corresponding to non-detections. For WNTR24-cua, the progenitor is detected in VVV before the outburst but not in WISE/NEOWISE. The outburst begins in 2016, visible in VVV $K_s$ and WISE $W2$. The initial rise is not detected in WISE $W1$ or Gaia $G$, which provide upper limits at that time. The peak and decline are captured from Gaia $G$ through WISE $W2$, with WINTER observations starting as Gaia and WISE coverage end. In contrast, WNTR24-egv shows a slower outburst beginning in 2014 that continues through 2025, covered by WISE and WINTER. Neither source is detected in the optical by ZTF to $\sim$20 mag, confirming that both are heavily obscured at optical wavelengths. For clarity, Gaia $G$ has been shifted upward by 2 magnitudes, highlighting how much fainter the optical detections are compared to the infrared bands.
• Figure 3: Spectra for the FUors compared with Gaia 17bpi Gaia17bpi in purple. The inset highlights the He I $\lambda$10830 line demonstrating blue-shifted absorption, indicative of a strong wind. Blue: WNTR24-cua spectrum taken with FIRE. Black: WNTR24-egv spectrum taken with SpeX with an inset of the He I line from a FIRE spectrum. High noise regions are masked. All three spectra show mixed temperatures typical of FUors, with deep molecular absorption at longer wavelengths, notably the CO bandhead near 2.3 $\mu$m, which becomes less prominent at shorter wavelengths.
• Figure 4: Classification scheme and data adapted from Connelley:2018 with the two WINTER sources over plotted. The WINTER sources fall near other bona fide FUors and show notably high CO equivalent widths.
• Figure 5: The spectral energy distributions (SEDs) for WNTR24-cua (left) and WNTR24-egv (right) showing pre- (reds) and post-outburst (blues) energy. The pre-outburst systems are shown with reddened photospheric models with a range of temperatures and extinctions in purple lines. The same models with additional blackbody components due to dust are shown in orange lines to explain the IR excess. The dereddened ($A_V = 0$) pre-outburst data are shown, along with a representative dereddened photosphere model, using extinction corrections of $A_V = 5.0$ mag and $A_V = 19.0$ mag, for WNTR24-cua and WNTR24-egv, respectively, derived from the photospheric model fits.