A Critical Assessment of Photometric Redshift Methods: A CANDELS Investigation

Abstract: We present results from the Cosmic Assembly Near-infrared Deep Extragalactic Legacy Survey (CANDELS) photometric redshift methods investigation. In this investigation, the results from eleven participants, each using a different combination of photometric redshift code, template spectral energy distributions (SEDs) and priors, are used to examine the properties of photometric redshifts applied to deep fields with broad-band multi-wavelength coverage. The photometry used includes U-band through mid-infrared filters and was derived using the TFIT method. Comparing the results, we find that there is no particular code or set of template SEDs that results in significantly better photometric redshifts compared to others. However, we find codes producing the lowest scatter and outlier fraction utilize a training sample to optimize photometric redshifts by adding zero-point offsets, template adjusting or adding extra smoothing errors. These results therefore stress the importance of the training procedure. We find a strong dependence of the photometric redshift accuracy on the signal-to-noise ratio of the photometry. On the other hand, we find a weak dependence of the photometric redshift scatter with redshift and galaxy color. We find that most photometric redshift codes quote redshift errors (e.g., 68% confidence intervals) that are too small compared to that expected from the spectroscopic control sample. We find that all codes show a statistically significant bias in the photometric redshifts. However, the bias is in all cases smaller than the scatter, the latter therefore dominates the errors. Finally, we find that combining results from multiple codes significantly decreases the photometric redshift scatter and outlier fraction. We discuss different ways of combining data to produce accurate photometric redshifts and error estimates.

• The paper finds that no single photometric redshift method significantly outperforms others, with training optimizations reducing scatter and errors.
• The paper demonstrates that photometric redshift accuracy strongly depends on signal-to-noise ratios, with weaker correlations to redshift and galaxy color.
• The paper shows that aggregating results from multiple codes diminishes scatter and outlier fractions, underscoring the advantage of combined methodologies.

A Critical Assessment of Photometric Redshift Methods: A CANDELS Investigation

This paper presents a thorough examination of photometric redshift techniques within the framework of the Cosmic Assembly Near-infrared Deep Extragalactic Legacy Survey (CANDELS). The study leverages eleven different approaches, each with unique combinations of photometric redshift codes, template spectral energy distributions (SEDs), and priors, to ascertain the effectiveness and variability of photometric redshift applications in deep fields with expansive multi-wavelength coverage. The photometry employed, spanning from $U$-band to mid-infrared filters, was derived using the TFIT method.

The primary outcome of this investigation is that no single photometric redshift determination method or SED template set significantly outperforms others. However, methods that produce minimal scatter and outliers frequently employ training samples for optimization, such as photometry zero-point offsets, template adjustments, or additional smoothing errors, underscoring the pivotal role of the training procedure. Importantly, photometric redshift precision shows a robust dependency on photometric signal-to-noise ratios while exhibiting a weaker correlation with redshift and galaxy color.

The study consistently reveals that all codes predispose slightly biased photometric redshifts. Despite this, the bias remains consistently smaller than the overall scatter, rendering the scatter the dominant error factor. Intriguingly, amalgamating results from multiple codes demonstrably mitigates photometric redshift scatter and outlier fractions. This approach suggests the potential benefits of combined methodologies in photometric redshift determinations, aligning with the "wisdom of crowds" phenomenon where aggregated results outperform individual contributions.

Implications of the research are extensive, affecting both theoretical frameworks and practical applications in astronomical survey data processing. The demonstration that combined photometric redshift determinations yield superior precision suggests avenues for improved accuracy in galaxy distance estimates, which are foundational to cosmological studies. The analysis intimates that further refinements in training processes and error assessments could enhance the reliability of vast cosmic surveys.

The study also informs the potential evolution of photometric redshift methodologies by emphasizing the necessity for rigorous optimization and training strategies. Future investigations might focus on integrating more sophisticated machine learning approaches for template adjustment, as well as refining error estimation mechanisms through hierarchical Bayesian frameworks as explored in the study.

In summary, this comprehensive assessment underscores the nuanced complexity and significant potential of photometric redshift methods in contemporary astrophysics, whilst providing a benchmark for future enhancements in this critical domain.