The Structure of Molecular Clouds: I - All Sky Near Infrared Extinction Maps

We are studying the column density distribution of all nearby giant molecular clouds. As part of this project we generated several all sky extinction maps. They are calculated using the median near infrared colour excess technique applied to data from the Two Micron All-Sky Survey (2MASS). Our large scale approach allows us to fit spline functions to extinction free regions in order to accurately determine the colour excess values. Two types of maps are presented: i) Maps with a constant noise and variable spatial resolution; ii) Maps with a constant spatial resolution and variable noise. Our standard Av map uses the nearest 49 stars to the centre of each pixel for the determination of the extinction. The one sigma variance is constant at 0.28mag Av in the entire map. The distance to the 49th nearest star varies from below 1arcmin near the Galactic Plane to about 10arcmin at the poles, but is below 5arcmin for all giant molecular clouds (|b|< 30degr). A comparison with existing large scale maps shows that our extinction values are systematically larger by 20% compared to Dobashi et al. and 40% smaller compared to Schlegel et al.. This is most likely caused by the applied star counting technique in Dobashi et al. and systematic uncertainties in the dust temperature and emissivity in Schlegel et al.. Our superior resolution allows us to detect more small scale high extinction cores compared to the other two maps.

💡 Research Summary

The paper presents a new set of all‑sky near‑infrared (NIR) extinction maps aimed at characterizing the column‑density distribution of nearby giant molecular clouds (GMCs). Using the Two Micron All‑Sky Survey (2MASS) point‑source catalog, the authors apply a median colour‑excess technique to the J, H, and Kₛ bands. By taking the median of the colour differences rather than the mean, the method is robust against outliers such as variable stars, binaries, or photometric errors.

A crucial innovation is the spline‑fitting of extinction‑free regions. The authors first identify sky patches where the intrinsic stellar colours follow the expected Galactic background, fit a smooth spline surface to these reference colours, and then subtract this surface from the observed median colours. This step removes large‑scale systematic biases that can arise from spatial variations in the underlying stellar population, especially near the Galactic plane where colour gradients are steep.

Two families of maps are produced. The first maintains a constant statistical uncertainty (σ₍Av₎ ≈ 0.28 mag) while allowing the spatial resolution to vary according to the local stellar density. For each map pixel the extinction is derived from the 49 nearest stars; the distance to the 49th star ranges from <1′ in the crowded plane to ~10′ at the poles, but stays below 5′ for essentially all GMCs (|b| < 30°). This “fixed‑noise, variable‑resolution” approach yields high fidelity in dense regions while preserving reliable statistics in sparse areas. The second family fixes the angular resolution (e.g., 3′) and permits the noise level to vary, providing a complementary product useful for direct comparison with other fixed‑resolution datasets.

When compared with existing large‑scale extinction maps, the new NIR maps show systematic differences. Relative to the Dobashi et al. star‑count map, the authors find their Av values are on average 20 % higher. They attribute this to the star‑count method’s tendency to underestimate extinction in high‑density regions where many background stars are completely obscured. Conversely, the Schlegel et al. far‑infrared (FIR) dust emission map yields Av values about 40 % larger than the NIR results, a discrepancy the authors ascribe to uncertainties in dust temperature and emissivity assumptions inherent in the FIR method. The NIR median‑colour technique therefore occupies an intermediate, arguably more reliable, position between the two extremes.

A notable advantage of the high spatial resolution (≤5′) is the detection of numerous small, high‑extinction cores that are invisible in the lower‑resolution Dobashi and Schlegel maps. These compact structures are likely sites of early star formation and provide valuable constraints on the initial mass function and star‑formation efficiency within GMCs.

The paper also discusses limitations. In high‑latitude regions where the stellar density is low, the variable‑resolution map expands to larger pixel sizes, potentially smoothing out the smallest features. The fixed‑resolution map can mitigate this but at the cost of increased noise. Nonetheless, the dual‑product strategy offers flexibility for a wide range of scientific applications.

In summary, the authors deliver a robust, all‑sky NIR extinction dataset that improves upon previous methods by combining median colour‑excess with spline‑based background correction, delivering uniform statistical uncertainties, and achieving sub‑5′ resolution across most of the Galactic plane. The maps are poised to become a standard reference for studies of molecular cloud structure, Galactic dust distribution, and star‑formation processes, and they provide a critical benchmark for testing theoretical models of interstellar medium evolution.