© The Authors 2025

1. Introduction

Galaxy groups are the smallest aggregates of galaxies, typically composed of fewer than one hundred galaxies, and represent the smallest structures that collapse to form galaxy clusters (Press & Schechter 1974). Groups are common in the nearby Universe, containing approximately half of the galaxies in this volume (Eke et al. 2004). They represent an intermediate environment between galaxies in the field and those in dense clusters.

Groups are associated with a range of galaxy properties that reflect the diversity of processes that act on these scales (Wetzel et al. 2012). Comparisons between galaxies in groups, clusters, and the field have revealed systematic differences in star formation rates (SFRs), gas content, metallicity, and structural properties (e.g., Blanton & Moustakas 2009).

Recently, miniJPAS (Bonoli et al. 2021) has illustrated the power of the Javalambre Physics of the Accelerating Universe Astrophysical Survey (J-PAS; Benitez et al. 2014) to detect groups with masses of up to 10¹³ M_⊙ and to characterize their galaxy populations up to a redshift of z ∼ 1 (Maturi et al. 2023; Doubrawa et al. 2024). Its multiwavelength photometry system, with 56 narrowband filters covering the entire optical range, allows the identification and characterization of galaxy populations and their evolution since z = 1 (González Delgado et al. 2021; Díaz-García et al. 2024), as well as the stellar population properties of group members (Rodríguez-Martín et al. 2022; González Delgado et al. 2022).

Similarly, the Physics of the Accelerating Universe (PAU; Csizi et al. 2024) and Advanced Large, Homogeneous Area Medium-Band Redshift Astronomical (ALHAMBRA; Díaz-García et al. 2019, 2015) surveys have also demonstrated the capability of consecutive narrowband filters to determine the stellar population properties of galaxies. Naturally, these methods do not achieve the accuracy available with spectroscopic data (see, for example, Nersesian et al. 2024, 2025), but the results are nonetheless useful.

These recent studies, as well as most previous ones (e.g., Boselli & Gavazzi 2006), used integrated light to examine galaxy properties. Judging from the progress achieved when integrated-light surveys such as the Sloan Digital Sky Surveys (SDSS; York et al. 2000) were complemented with Integral Field Spectroscopy (IFS) surveys like Calar Alto Legacy Integral Field Area Survey (CALIFA; Sánchez et al. 2012) and Mapping Nearby Galaxies at Apache Point Observatory (MaNGA; Bundy et al. 2015), examining the spatially resolved stellar and nebular properties of galaxies in groups is a worthwhile endeavor. Observationally, the challenge is to cover galaxies over their full spatial extent, as well as wide areas of the sky, to map all group members. Additionally, statistics over many groups are needed to firmly establish whether galaxies in groups at different stages of evolution differ in their internal stellar and nebular properties. These requirements are not easy to satisfy with IFS.

This paper is the first in a series aimed at characterizing the spatially resolved stellar population and emission line (EL) properties of galaxies in groups and clusters with Javalambre Photometric Local Universe Survey (J-PLUS) (Cenarro et al. 2019) and Southern Photometric Local Universe Survey (S-PLUS) (Mendes de Oliveira et al. 2019; Herpich et al. 2024) data, with the methodology developed by Thainá-Batista et al. (2023) (hereinafter TB23). The large field of view of these surveys makes them ideally suited to study groups and clusters, while their unique combination of narrow- and broadband filters provides information on both stellar and nebular properties.

This work focuses primarily on describing the data, the methodology, and an initial analysis of the results for the six galaxies in the M101 group (Garcia 1993). Subsequent papers will target other nearby groups and clusters, such as the M51 group and Fornax (Smith Castelli et al. 2024), gradually building a sample suitable for comparative analyses of galaxy properties across different environments. Section 2 describes the sample and the preprocessing procedures applied to the data. Section 3 reviews and illustrates the fitting technique used. Our primary 2D results are presented in Sect. 4, where we examine maps of the stellar population and the EL properties, as well as the relations between them. Section 5 compares the galaxies in the group with each other in terms of their scaling relations. Finally, our main results are summarized in Sect. 6.

2. Data, sample, and preprocessing

This work explores the properties of galaxies in the M101 group, which, in addition to M101, includes NGC 5474, NGC 5585, NGC 5204, UGC 8837, and NGC 5477. After a brief description of the data (Sect. 2.1) and the individual galaxies in the sample (Sect. 2.2), we explain the series of preprocessing steps applied to the data (Sect. 2.3) prior to the stellar population and EL analysis.

2.1. Data

All six galaxies were observed as part of J-PLUS with an 80 cm robotic telescope at the Observatorio Astrofísico de Javalambre (Cenarro et al. 2014), in Teruel, Spain. The Javalambre Auxiliary Survey Telescope (JAST80; Marín-Franch et al. 2015) is equipped with a 9216 × 9232 pixel camera (T80Cam), which provides a 1.4 × 1.4 deg² field of view and a pixel scale of 0.55 arcsec pix⁻¹. In this work, we use the third data release (DR3; López-Sanjuan et al. 2024) of J-PLUS. The distinctive feature of J-PLUS is its set of 12 filters, with five broad and seven narrow bands, covering the ∼3500–9000 Å range. The filters were designed for a multipurpose survey, spanning applications ranging from the Solar System to extragalactic scales. A full description of J-PLUS, including details on the observations and data reduction, can be found in Cenarro et al. (2019), with example studies based on its data available in Logroño-García et al. (2019), San Roman et al. (2019), Lumbreras-Calle et al. (2022), López-Sanjuan et al. (2022), and Rahna et al. (2025).

Figure 1 shows J-PLUS images of all six galaxies in the group. These composite images are constructed using the J0660 flux in the R channel, the g band in the G channel, and the sum of the fluxes in the five bluest bands (u, J0378, J0395, J0410, J0430) in the B channel. As these images reveal, the group consists of spiral and irregular galaxies, with sizes ranging from ∼1 to 18 arcmin. The figure also reveals that all galaxies are asymmetric to some degree, and, as evidenced by the red regions in these images, they all exhibit Hα emission, which falls well within the J0660 filter¹ for the redshifts of the galaxies in the group.

Fig. 1. Composites of the original data for galaxies in the M101 group, constructed using the J0660, g, and the sum of the five bluest filters in the R, G, and B channels, respectively. The white bars indicate a scale of 1 arcmin.

2.2. Individual objects

Table 1 lists basic information on our sample galaxies. We adopted a single distance of 7.24 Mpc to all galaxies in the group, as determined by Lee & Jang (2012) for M101. Information about morphologies was extracted from LEDA (Makarov et al. 2014), NED (Helou 1990), and SIMBAD (Wenger et al. 2000).

M101: Widely known as the Pinwheel Galaxy, M101 (also named NGC 5457, PGC 50063, and UGC 8981) is the dominant galaxy of the group. With a redshift of 0.000804, this nearly face-on, lopsided spiral with an optical diameter of nearly 30 arcmin exhibits tidal features as well as numerous star-forming regions, including giant H II regions of nearly 1 kpc in size (García-Benito et al. 2011). Due to these characteristics, M101 is the subject of many studies in the literature (see, e.g., Bresolin (2007), Hu et al. (2018), Garner et al. (2024) and references therein).

NGC 5474: This dwarf spiral (also known as PGC 50216) also exhibits a lopsided structure, a feature often attributed to its gravitational interaction with M101 (Linden & Mihos 2022). NGC 5474 is the most massive satellite of M101 and has a redshift of 0.000874. Its peculiar structure and the presence of regions with ongoing star formation make it an interesting example of tidal effects within galaxy groups.

NGC 5585: Also known as PGC 51210, this is a late-type galaxy of morphological type SAB(s)d with a redshift of 0.001011. It is particularly notable for being a dark-halo-dominated galaxy (Cote et al. 1991).

NGC 5204: This low-surface-brightness galaxy (also known as PGC 47368) has a redshift of 0.000670. Like the other galaxies in the group, NGC 5204 exhibits an asymmetric structure with numerous regions of star formation.

UGC 8837: This irregular galaxy, also known as PGC 49448, has a redshift of 0.00048. It is the only edge-on galaxy in the group.

NGC 5477: With a redshift of 0.00105, this irregular dwarf galaxy (also known as PGC 50262) is the smallest in the M101 group, with a diameter of approximately 1 arcmin in the R band. Its spectrum is characterized by strong emission lines that indicate ongoing star formation.

2.3. Preprocessing

The original flux-calibrated J-PLUS images underwent a series of preprocessing steps before being analyzed for their stellar population and EL information content. The general goals of these preprocessing steps are to clean the data of unwanted or spurious sources (such as foreground stars and background galaxies), reduce the noise level, homogenize the spatial resolution, and ensure photo-spectra with a minimal quality to warrant an analysis with the tools detailed in Sect. 3. The main steps are summarized as follows.

1. Spatial masks – The first step is to define spatial masks mapping (a) the object of interest, (b) the foreground stars and background galaxies, and (c) the sky. The “stars” mask (which includes foreground stars, background galaxies, and artifacts) was built using functions from the PHOTUTILS package (Bradley et al. 2024) combined with our own code to detect sources, followed by an interactive review to confirm source identification and to add saturated stars. The galaxy limits were defined using isophotes at r-band surface brightness μ_r = 24 mag/arcsec², allowing for adjustments when necessary.

2. Butterworth filter – Inspired by the work of Ricci et al. (2014), we apply a low-pass Butterworth filter that reduces noise and instrumental artifacts, particularly in the outer regions, where the signal is weaker.

3. PSF homogenization – The full width at half maximum (FWHM) for the different bands is then used to convolve all images to the worst point spread function (PSF), resulting in a common spatial resolution.

4. Binning – We rebin the data in 2 × 2 pixels, changing the original 0.55 arcsec/pix scale to 1.1 arcsec/pix. This step increases the signal-to-noise ratio (S/N) and optimizes computational operations while maintaining the spatial resolution close to that of the original data.

5. Pixel tagging – Next, we assign each pixel to one of three categories: (a) H II region, (b) dusty region, and (c) unlabeled region. Pixels tagged as H II region are selected as those with a combined Hα + [N II] equivalent width of > 40 Å and a flux > 5.5σ above the average over the whole galaxy. The Hα + [N II] values used in this stage are obtained using the method of Vilella-Rojo et al. (2015) (see also Rahna et al. 2025). Dusty pixels are defined as those with a z-band/g-band flux ratio greater than 1.2, a condition that does not occur for any galaxy in this work. These tags are used in the next step (Voronoi binning), which is carried out separately for each tag, thus mitigating the mixing of physically different pixel types within the same bin. We reserve a graphical illustration of this tagging strategy for the case of M51 (Thainá-Batista et al., in prep.), which has many dusty and H II regions.

6. Voronoi Binning – Finally, we apply Voronoi binning (Cappellari & Copin 2003) to ensure a minimum S/N across bins. After testing different target S/N and filter combinations, we selected the u-band as the reference band with a specific S/N for each galaxy: S/N_u = 8 for NGC 5477 and UGC 8837, S/N_u = 10 for M101, NGC 5474, and NGC 5204, and S/N_u = 20 for NGC 5585.

7. Local sky re-subtraction – If the sky regions present a systematically net positive spectrum, we refine the original sky subtraction by subtracting the average sky spectrum in the datacube. This optional extra step affects only the fainter regions of the galaxy.

Figure A.1 illustrates the progressive effects of these steps, while Fig. A.2 shows how Fig. 1 appears after preprocessing. The final datacubes are divided into 11706, 752, 1512, 2753, 120, and 133 Voronoi zones for M101, NGC 5474, NGC 5585, NGC 5204, UGC 8837, and NGC 5477, respectively, covering areas corresponding to equivalent circular radii ranging from 28 kpc (M101) to 2 kpc (NGC 5477).

The errors in the photometric fluxes (ϵ_λ) are important both in step six (Voronoi binning) and in the photo-spectral fits discussed below. A simple propagation of errors produces S/N ratios in excess of 100 in the r band, a level of precision that cannot realistically be achieved given the simplifications and limitations in the spectral modeling of galaxies, including our own. We therefore limit the errors in the r band such that S/N_r does not exceed 25. Errors in the other bands are adjusted in order to preserve the shape of the error spectrum of the original data, as computed from the median ϵ_λ/ϵ_r spectrum within the galaxy mask. This spectrum is computed for each galaxy separately, although it is very similar in all cases, with errors in the five bluest bands averaging 7.7 × ϵ_r, while those in the four reddest bands are 1.6 × ϵ_r.

3. Data analysis

The preprocessed photo-spectra are fitted using the ALSTAR code, a spectral energy distribution (SED) fitting tool similar to others such as CIGALE (Boquien et al. 2019), PROSPECTOR (Johnson et al. 2021), and PROSPECT (Robotham et al. 2020). Like MUFFIT (Díaz-García et al. 2015) and BAYSEAGAL (González Delgado et al. 2021), ALSTAR was developed in the context of J-PLUS and J-PAS and has already been tested and tailored to handle data from these surveys. A key advantage of ALSTAR is its ability to simultaneously derive stellar population properties and emission line characteristics.

A detailed description of the code is given in TB23, including applications to integrated and spatially resolved S-PLUS data that are essentially identical to those analyzed here. This section briefly reviews the code, its ingredients, and updates. Examples of fits to different regions of M101 are also presented.

3.1. ALSTAR

The ALSTAR code performs a non-parametric decomposition of a photo-spectrum in terms of a spectral base containing both stellar population and EL elements. The stellar base used in this work comprises composite stellar populations corresponding to periods of constant star formation rate, with 16 approximately logarithmically spaced age intervals spanning from 0 to 14 Gyr, and three different metallicities (0.2, 0.5, and 1 Z_⊙). We note that the adopted metallicity range does not reach the over-solar values allowed for in TB23. Our choice of a more restricted Z-range is justified by the fact that previous spectroscopic estimates of the stellar metallicity across the body M101 point to values ≲Z_⊙ (Lin et al. 2013; Simanton-Coogan et al. 2017). Imposing this external constraint helps to reduce the inherent degeneracies of spectral synthesis of stellar populations.

Spectra of these 48 components are built from an updated version of the Bruzual & Charlot (2003) models (see Martínez-Paredes et al. 2023). The Chabrier (2003) initial mass function (IMF) is adopted. This stellar base is expanded to include 94 EL base elements of unitary Hα flux and relative line intensities mimicking those observed in real galaxies, including H II regions, diffuse ionized gas (DIG), and active galactic nuclei.

The empirical EL base of TB23 was updated to include the corresponding nebular continuum emission, which can be significant in spaxels dominated by H II regions (see, e.g., Corbin et al. 2006; Miranda et al. 2025). The PyNeb tool (Luridiana et al. 2015) was used for this purpose. The (relatively minor) dependence of the nebular continuum on the electron temperature is accounted for in an approximate way by following the relation between T_e and gas metallicity in the models by Byler et al. (2017), and by using the O3N2 $\equiv log([\mathrm{O}\mathrm{I}\mathrm{I}\mathrm{I}]/\text{H}\beta )/([\mathrm{N}\mathrm{I}\mathrm{I}]\text{H}\alpha )$ calibration of Marino et al. (2013) to estimate the gas metallicity.

The Calzetti et al. (2000) law is used to model dust attenuation. Two dust components are included: one applied to all base components (parameterized by the V-band optical depth τ^ISM), and an extra component applied only to populations younger than 10 Myr to represent the dust in the birth clouds (τ^BC). The ratio τ^BC/τ^ISM is kept fixed at 1.27, as found in the original Calzetti et al. (1994) paper on differential extinction (see also Charlot & Fall 2000). The ELs associated with components compatible with star-formation are attenuated by τ^BC + τ^ISM (as are the ≤10 Myr components), whereas other components of the EL base are attenuated by only τ^ISM.

This general scheme is complemented with an empirical prior which (1) uses a first ALSTAR fit to estimate the combined equivalent width of the [N II]λλ6548,6584 and Hα lines ($W_{\mathrm{H}\alpha \mathrm{N}\mathrm{I}\mathrm{I}}$), (2) reduces the EL base to elements where such values of $W_{\mathrm{H}\alpha \mathrm{N}\mathrm{I}\mathrm{I}}$ are found in real galaxies, and (3) reruns ALSTAR. The major effect of this prior is to break the degeneracy between [N II] and Hα fluxes, which are both covered by the same J0660 narrow band.

The fits are repeated 100 times with fluxes perturbed with Gaussian noise of amplitude ϵ_λ. Averages over these Monte Carlo (MC) runs are used to define our estimates of physical properties such as masses, mean ages, EL fluxes, and others, as well as their uncertainties.

3.2. Example fits to M101

Figure 2 illustrates the quality of our fits. The top panels show three examples of “J-spectra”, corresponding to the nucleus (A), an interarm zone (B), and an H II region (C), depicted in the stamps on the right. The observed (O_λ) and model (M_λ) photo-spectra are shown in the left panels, where the solid black line represents the model photometry, while the dashed line and error bars represent the data. In all cases, the fluxes shown correspond to those inside a single 1.1″ × 1.1″ (rebinned) spaxel, although region B falls inside a Voronoi zone comprising 333 pixels.

Fig. 2. Top: Example ALSTAR fits for individual spaxels in three different regions of M101. Colored lines with error bars show the data (Oλ), while black lines show the model photometric fluxes (Mλ), and the gray lines show the corresponding high-resolution model spectrum. Images on the right show 278″ × 278″ composites built with the J0660, r, and g fluxes in the R, G, and B channels, respectively. Bottom: Distributions of the (Oλ − Mλ)/Mλ relative residuals of the fits for the 12 J-PLUS bands for 11 705 zones in M101. Median residuals are marked by solid black circles, while the horizontal bars mark the 16 and 84 percentiles.

As expected from our previous work with S-PLUS (TB23), the ALSTAR fits are very good in all three examples, which span both different spectral shapes and over an order of magnitude in surface brightness. The mean relative absolute deviation, $\overline{\mathrm{\Delta }}\equiv ⟨|O_{\lambda }-M_{\lambda }|/M_{\lambda }⟩$, between data and model fluxes is just 2.26, 4.61, and 1.96% for regions A, B, and C, respectively. For the M101 datacube as a whole, the median $\overline{\mathrm{\Delta }}$ is 3.6%, and only 323 of the 11 706 (2.7%) fits have $\overline{\mathrm{\Delta }}>10\%$. These results are also typical of the quality of the fits for NGC 5474, NGC 5585, and NGC 5204, whose median $\overline{\mathrm{\Delta }}$ ranges from 2.3 to 3.2%. For NGC 5477 and UGC 8837, the fainter galaxies in the M101 group, the fits are somewhat worse, with median $\overline{\mathrm{\Delta }}$ values of 6.0 and 4.0%, respectively. Figure A.3 shows example fits for these other galaxies, as well as $\overline{\mathrm{\Delta }}$ maps for the whole sample.

The violin plots in the bottom panel of Fig. 2 show the distributions of the relative residuals, (O_λ − M_λ)/M_λ, for the 12 bands. As expected, residuals are statistically larger towards the bluer (noisier) bands.

Overall, the ALSTAR fits are very good. From TB23, we further know that the stellar population and EL properties derived from these fits are reliable, in the sense that they agree with those derived from much more detailed (“Å by Å”) full spectral fits.

4. Results

We examine the properties derived from our analysis of J-PLUS data for galaxies in the M101 group. We begin with maps of stellar populations (Sect. 4.1) and EL properties (Sect. 4.2). We then investigate the relations between these properties within each galaxy (Sect. 4.3).

4.1. Stellar population maps

Figure 3 presents maps of stellar population properties obtained with our method for each of the galaxies in the M101 group. The maps are sorted by stellar mass, with larger M_⋆ at the top.

Fig. 3. Maps of stellar population properties for galaxies in the M101 group. From left to right: surface density, mean age, star formation rate (SFR) surface density, and an RGB with the fluxes at 5635 Å of old, intermediate-age, and young populations. See text for details.

The first column panels show the stellar mass surface density, Σ_⋆. In M101 (top row), Σ_⋆ decreases from ∼3000 M_⊙ pc⁻² in the nucleus to < 10 M_⊙ pc⁻² in the outer parts. As is usual in spirals, the high contrast between arms and interarm regions seen in optical images is not as strong when viewed in mass density. The peak Σ_⋆ values reached by the other galaxies decrease from ∼460 M_⊙ pc⁻² for NGC 5474 to ∼40 M_⊙ pc⁻² for NGC 5477, following a sequence in M_⋆.

The second column in Fig. 3 shows the luminosity-weighted mean log stellar age, defined as ⟨log t⟩_L ≡ ∑_jx_jlog t_j, where x_j is the fraction of the flux at 5635 Å (our reference normalization wavelength) associated with the population j at the base. In M101 there is a gradual decrease of ⟨log t⟩_L toward the outer regions, as well as many “islands” of very low age along the arms, corresponding to the many star-forming regions in this galaxy (also detected in our EL maps, as discussed below). Although a direct quantitative comparison is not warranted because of methodological differences, our negative age gradient is consistent with the trend reported by Hu et al. (2018) on the basis of full spectral fitting (their Figure 6), as well as with the map produced by Lin et al. (2013), whose use broadband photometry from the UV to the mid-IR and a very different modeling strategy based on parametric star formation histories (their Figure 11). Both works further identify a young central component, in line with our Fig. 3.

This same global description applies to NGC 5574, NGC 5585, and NGC 5204, but this pattern is no longer recognizable in UGC 8837 and NGC 5477, the two less massive galaxies in the group. At the same time, low ⟨log t⟩_L values become progressively more prevalent as one moves down the M_⋆ scale. These results are in line with those obtained from IFS surveys, such as the CALIFA-based study by González Delgado et al. (2014a), who found flatter age gradients for lower-mass galaxies, or the MaNGA-based study by Sánchez (2020), who confirm this trend for a larger sample.

A detailed study of star formation rates (SFRs) will be presented in an upcoming paper, but for completeness, the third column of Fig. 3 shows maps of the SFR surface density derived from the ALSTAR fits, Σ_SFR^⋆. These are computed by adding the mass in stars formed in the last t_SF years and dividing it by t_SF (as in Asari et al. 2007). We adopt t_SF = 100 Myr as a compromise between the desire to map the recent SFR and the need to group the base ages to produce a more robust estimator (see, e.g., the experiments in Cid Fernandes et al. 2004). The maps combine the patterns of the first and second panels, showing higher Σ_SFR^⋆ in regions of higher Σ_⋆, as expected from the star-forming main sequence (MS, e.g., Enia et al. 2020), but also in regions of low ⟨log t⟩_L. We also note the strong asymmetry in the Σ_SFR^⋆ map of M101.

Finally, the rightmost images in Fig. 3 combine the fluxes at 5635 Å of young (t ≤ 10 Myr), intermediate-age (10 Myr < t < 1 Gyr), and old (t ≥ 1 Gyr) populations into an RGB composite. The overall inside-out R → G → B run of colors in the M101 to NGC 5204 panels reflects the negative mean age gradient previously seen in the ⟨log t⟩_L maps, while for UGC 8837 and NGC 5477 no organized pattern is discernible. In M101, where the spiral arms are well defined, old populations are also ubiquitous in the inter-arm regions, while the arms themselves are dominated by intermediate-age and young populations.

Maps of the dust optical depth, along with dereddened color composites (similar to those presented in TB23 for NGC1365), are presented in Fig. A.4. Unlike in the M51 group (Thainá-Batista et al., in prep.), dust attenuation is generally small throughout all galaxies in the M101 group (typically around 0.2 mag in terms of A_V), with the most notable exceptions occurring in star-forming regions, which are dustier. The typical value of A_V for Sd galaxies, as reported by González Delgado et al. 2015, is approximately 0.2 mag, which is also representative of the galaxies in the M101 group (Table 1).

The uncertainties in the properties discussed above are estimated from the dispersion among the MC realizations performed by ALSTAR. Using the robust σ_NMAD² statistic to quantify these noise-related uncertainties, we find median uncertainties of 0.22–0.25 dex in logΣ_⋆ over the six M101 group galaxies, rising to ∼0.4 dex in regions of strong star formation. Mean light-weighted ages have median σ_NMAD(⟨log t⟩_L) values of 0.3–0.5 dex, with greater uncertainty for ages near 10⁸ yr and in the fainter outskirts. Stellar metallicities exhibit smaller formal errors (∼0.2 dex), largely reflecting the restricted metallicity grid used, while SFR surface densities have σ_NMAD(logΣ_SFR)∼0.2 dex when considering only zones where young populations contribute at least 30% of the 5635 Å flux. Appendix A.5 discusses these and other uncertainties in more detail.

4.2. Emission line maps

We now examine the EL properties derived from our fits. Figure 4, illustrates this, with the first column showing the Hα surface brightness (Σ_Hα) maps for galaxies in the M101 group, ordered as in Fig. 3. Star-forming regions stand out clearly in all cases. Diffuse emission is also detected, particularly in the central regions of the top four galaxies and between the arms of M101. Although expected from a physical perspective, it is interesting to note the strong correspondence between these Hα maps and the young stellar populations mapped in Fig. 3, which shows that our methodology is able to identify young populations by their contributions to both ELs and stellar continuum.

Fig. 4. Maps of emission-line properties. From left to right: Hα surface brightness; Hα equivalent width; RGB with the ([N II], Hα, [O III]) fluxes; BPT diagram of the spaxels of each galaxy, with points color-coded as in the RGB panel.

The second column of Fig. 4 shows maps of the Hα equivalent width (W_Hα). The maps span values from as low as 1 Å to more than 1000 Å. Again, there is an excellent correspondence between regions of large W_Hα and those of small ⟨log t⟩_L in Fig. 3. Regions depicted in red-orange colors, where W_Hα values are relatively small, are visibly more diffuse than those in blue (W_Hα > 100 Å), which appear to be knotty. This diffuse emission matches the characteristics of the mixed diffuse ionized gas (DIG) component defined by Lacerda et al. (2018), with W_Hα in the ∼3–14 Å range. Given the widespread star formation in these galaxies, it is likely that this diffuse emission is mainly powered by leakage of ionizing photons from H II regions (Watkins et al. 2024).

Regions B and C of Fig. 2 are examples of regions whose ELs are dominated by DIG and H II regions, respectively. Their W_Hα values are 9.4 and 700 Å, respectively, while $[\mathrm{N}\mathrm{I}\mathrm{I}]/\text{H}\alpha =0.5$ and 0.10. The large W_Hα in region C is in fact evident from its J-spectrum in Fig. 2, which also shows a strong [O II] ($W_{[\mathrm{O}\mathrm{I}\mathrm{I}]}=231$ Å according to our fits). The nucleus (region A) has intermediate values, with W_Hα = 15.9 Å and $[\mathrm{N}\mathrm{I}\mathrm{I}]/\text{H}\alpha =0.37$.

The third column of Fig. 4 combines the [N II], Hα, and [O III] maps into an RGB composite, which synthesizes our main results regarding ELs and their spatial variations. The fourth column shows the corresponding BPT diagrams (Baldwin et al. 1981), $log[\mathrm{O}\mathrm{I}\mathrm{I}\mathrm{I}]/\text{H}\beta$ versus $log[\mathrm{N}\mathrm{I}\mathrm{I}]/\text{H}\alpha$, where the points are colored as they appear in the RGB panel. The rather schematic appearance of our BPT diagrams stems from our methodology to estimate ELs from J-PLUS photometry, which involves a discrete EL base constructed on the basis of the BPT diagram itself (see TB23).

The redder hue in the central regions of the top four galaxies, as well as in inter-arm regions of M101, trace regions of elevated $[\mathrm{N}\mathrm{I}\mathrm{I}]/\text{H}\alpha$, characteristic of DIG-like emission. Star-forming regions gradually change their colors in this image from green in the inner disk to blue in the outer parts due to a systematic increase of the ratio between [O III] (plotted in the B channel) to Hα (G). In the case of M101 we associate this behavior with its well-documented negative nebular metallicity gradient (e.g., Bresolin 2007, Croxall et al. 2016). This is confirmed in its BPT diagram, where spaxels with bluer points from the outer H II regions in the RGB composite populate broadband in the upper-left, low oxygen-to-hydrogen abundance ratio (O/H) regions of the BPT. The diagram also shows how reddish regions in the ([N II], Hα, [O III]) composite map onto the right wing of the BPT, as one would expect for DIG-like ELs. We also note how dark (low flux, mostly inter-arm) regions in the EL RGB map onto the right wing of the BPT. The same trend also appears present in the ([N II], Hα, [O III]) images of NGC 5474, NGC 5585, and NGC 5204, although it is not as pronounced as in M101.

Our map W_Hα, the BPT diagram and the radial trends in the line ratios for M101 agree well with the results of Hu et al. (2018). However, this conclusion is based on a coarse visual analysis, since their “maps” are not continuous, but focus on individual H II regions, which also explains why their study barely samples the right wing DIG-dominated regions of the BPT in M101.

Typical (MC-based) uncertainties in Σ_Hα range from 0.1 to 0.3 dex across our galaxies, with the larger relative uncertainties occurring in regions of weak line emission. Restricting the analysis to the regions where $W_{\text{H}\alpha }>30\mathrm{\AA }$, we obtain σ_NMAD(logΣ_Hα) that ranges from 0.04 to 0.13 dex. For W_Hα, we obtain σ_NMAD(log W_Hα)∼0.1 dex for most galaxies (up to 0.3 dex in the noisiest cases). Ratios such as $[\mathrm{N}\mathrm{I}\mathrm{I}]/\text{H}\alpha$ show median uncertainties near 0.2 dex. For a more detailed discussion of uncertainties in EL properties, see Appendix A.5.

4.3. Age-metallicity-mass relations

We now examine our results in terms of the relations among the derived properties. Some of the most studied relations in galaxies are the mass-age, mass-metallicity, and mass-SFR relations. These relations are mapped both for galaxies as a whole (e.g., Gallazzi et al. 2005; Renzini & Peng 2015), and with spatially resolved observations (e.g., González Delgado et al. 2016; Sánchez 2020).

Figure 5 plots the mean stellar age and metallicity, as well as the nebular metallicity and the SFR surface density, against the local stellar mass surface density. As in previous figures, galaxies are sorted by mass. Points are colored by log W_Hα, which facilitates the identification of spaxels associated with star-forming regions. Bins along the Σ_⋆-axis were defined to draw median curves (thick line with circles) and 16th and 84th percentiles (gray lines). Given the systematic decrease in Σ_⋆ with distance to the center, these plots can be seen as equivalent to radial profiles, with the advantage of bypassing the need to define a meaningful radial coordinate in these distorted systems.

Fig. 5. Scaling relations for galaxies in the M101 group. Individual zones are plotted as circles colored by log WHα. The thick black lines with circle markers indicate the median curves for bins in logΣ⋆; thin lines show the corresponding 16th and 84th percentile curves. In the third column, the black dashed line is the median curve from Barrera-Ballesteros et al. (2016) for 653 MaNGA galaxies. The gray line shows the 80% contour of Barrera-Ballesteros et al. (2016) for MaNGA galaxies with log M⋆/M⊙ < 9.2. The dotted contour indicates the typical loci of Sd galaxies from Sánchez 2020. Dashed lines in the fourth column panels mark lines of specific SFR = 0.01 and 1 Gyr−1. The solid gray line is the relation obtained by Enia et al. (2020) – see text for details.

4.3.1. The mass-age relation

The tendency for more massive galaxies to have older stellar populations (also known as “downsizing”; Pérez-González et al. 2008), has a local counterpart within galaxies (González Delgado et al. 2014b). The left panels in Fig. 5 show how our ⟨log t⟩_L age indicator varies as a function of Σ_⋆ for galaxies in the M101 group.

Except for the dwarfs, all other galaxies in the group show a clear mass-age relation, with relatively tight ranges in ⟨log t⟩_L for fixed Σ_⋆. The increase in age with mass depicted in these plots reflects the overall negative age gradients seen in Fig. 3 and is well documented in the literature, particularly in studies based on IFS data (Sánchez 2020 and references therein). The two galaxies that deviate from this pattern are UGC 8837 and NGC 5477, the smallest and least massive galaxies of the group, which exhibit a flat mass-age relation, in line with the IFS-based results of González Delgado et al. (2015), which indicate flatter age profiles at lower galactic masses.

In particular, regions that deviate from the median trend tend primarily toward younger ages and have larger W_Hα (more than 50 Å), as indicated by their green-blue symbols. Despite the abundance and prominence of these star-forming regions, they occupy only a fraction of the area of the galaxies, and thus have relatively little influence on the median relation. We note in passing that the systematic gradient in symbol colors along the vertical direction in these panels reinforces the connection between ⟨log t⟩_L and W_Hα noted while comparing their maps in Figs. 3 and 4.

4.3.2. The stellar mass-metallicity relation

To investigate the stellar mass-metallicity relation (MZR), we use the mass weighted mean (log) stellar metallicity, ⟨log Z⟩_M ≡ ∑_jμ_jlog Z_j, where μ_j denotes the mass fraction associated with the component j. Since most of the mass in a stellar population lies in low-mass, long-lived stars, this index naturally gives a larger weight to old populations. This is desirable, given the well-known difficulty in estimating stellar metallicities in young populations (e.g., Cid Fernandes et al. 2005). Still, given the profusion of star-forming regions in our galaxies, many spaxels have their light dominated by young stars to the point of affecting the derived ⟨log Z⟩_M values. To mitigate this effect, we focus on regions where populations older than 1 Gyr account for more than x_O = 50% of the flux at 5635 Å.

The second column of Fig. 5 shows the resulting stellar MZRs. Again, points are colored by W_Hα, and median, 16th, and 84th percentile relations are indicated. Due to its much larger statistics, M101 exhibits a better-defined stellar MZR, with a clear increase of ⟨log Z⟩_M for increasing Σ_⋆. Ultimately, this reflects a negative stellar metallicity gradient, with inner (denser) regions being more metal-rich (∼1/2 solar), than the outer (less dense) ones (∼1/4 solar at the median point). The color scheme shows that the scatter in the MZR is related to W_Hα, and hence (indirectly) to ⟨log t⟩_L, in the sense that, for a fixed Σ_⋆, younger zones have larger metallicities. This is consistent with chemical evolution, although one should keep in mind the caveats associated with estimating stellar metallicities in the presence of young populations.

Unlike M101, NGC 5474 exhibits a flat MZR. The values of ⟨log Z⟩_M are generally lower than those in M101, oscillating around 0.3 solar. NGC 5585 and NGC 5204 exhibit, at best, weak MZRs. No meaningful statement can be made about the stellar MZR in the two smallest galaxies in the sample, as they have too few regions satisfying our x_O > 50% criterion to obtain a statistically meaningful estimate of ⟨log Z⟩_M.

4.3.3. The nebular mass-metallicity relation

The nebular version of the local MZR is shown in the third column of Fig. 5. We estimate 12 + logO/H using the O3N2 index, following the calibration by Marino et al. (2013). Since this strong-line method is only applicable to star-forming regions, we exclude all zones whose ELs place them above the Kauffmann et al. (2003) line in the BPT diagram. We further exclude zones where W_Hα < 14 Å to mitigate effects of the mixed-DIG component discussed by Lacerda et al. (2018), although this additional constraint has negligible effects on the resulting MZRs. Finally, we disregard zones larger than 50 pixels to avoid their disproportionate weight in the overall statistics and to minimize the mixing of very different spatial scales.

The correlation between Σ_⋆ and O/H is evident in M101, as reinforced by the thick black solid lines in Fig. 5, which trace the median relation. This nebular MZR is consistent with the well-documented O/H gradient in M101 (e.g., Kennicutt et al. 2003; Berg et al. 2020), as well as with MZRs derived from IFS studies of other galaxies. The points overlap well with the dotted contour, which outlines the 50% contour for Sd galaxies from the review by Sánchez (2020) (shifted by −0.26 dex to correct for differences in the IMF). The median curve is also close to the MZR obtained by Barrera-Ballesteros et al. (2016) in their analysis of 653 MaNGA galaxies, marked with a dashed line (also adjusted for IMF), a difference that we attribute to the contribution of more massive, metal-rich systems in their sample, particularly at large Σ_⋆.

Nebular MZRs are also found for NGC 5474, NGC 5585, and NGC 5204, although they are not as well defined as in M101. In the cases of UGC 8837 and NGC 5477, MRZ is only weakly indicated, consistent with IFS-based studies that generally report weak or absent systematic nebular metallicity gradients in low-mass galaxies (e.g., Belfiore et al. 2017). Our results for these five less massive galaxies in the M101 group align well with the region occupied by spaxels in MaNGA galaxies with log M_⋆/M_⊙ < 9.2 from Barrera-Ballesteros et al. (2016), shown as a solid gray line (also adjusted for IMF). These contours also overlap with the outer regions of M101.

Evidently, these results are subject to the usual caveats and limitations of strong-line methods, further compounded by the uncertainties inherent in our approach for estimating EL fluxes from purely photometric data.

4.3.4. The spatially resolved star-forming main sequence

The rightmost panels of Fig. 5 show the relation between the SFR surface density estimated from the fits ALSTAR, Σ_SFR^⋆, and Σ_⋆. In addition to the median and percentile curves, these panels include lines for specific SFRs of 0.01 and 1 Gyr⁻¹, as well as the relation obtained by Enia et al. (2020), based on multiband photometry from UV to far IR of eight M_⋆ = 1–5 × 10¹⁰ M_⊙ grand design spirals (shown as a thick gray line).

The effects of spatial sampling (which varies from 39 pc spaxels to 2 kpc Voronoi zones across our galaxies) on the scatter in this relation are explored in detail in a separate study. We anticipate that these effects are significant. However, median curves should be immune to this issue, so we focus on them here.

All galaxies exhibit Σ_SFR^⋆-Σ_⋆ relations that are either approximately ∼ parallel to the Enia et al. (2020) line or, in the case of M101, very close to it, at least for Σ_⋆ ≳ 10 M_⊙/pc². In all cases, regions with high W_Hα lie systematically above the median relations, consistent with W_Hα being a tracer of a specific SFR. The flattening or upturn seen at low Σ_⋆ is not unexpected, as most zones contributing to the median curve in these regions have large W_Hα, as indicated by their green-blue colors. This is likely a selection effect: low Σ_⋆ regions only enter the galaxy mask if they are sufficiently bright, a criterion that naturally favors star-forming regions over more passive ones. More interestingly, the right panels in Fig. 5 show a systematic upward shift in the median relation with decreasing M_⋆ or towards later morphological types, as observed in CALIFA (González Delgado et al. 2016, 2017).

Our total stellar mass (2.2 × 10¹⁰ M_⊙) and SFR (3.2 M_⊙/yr) derived from the spatially resolved M101 cube agree within 6% and 24%, respectively, with those reported by Enia et al. (2020) based on UV-to-far-IR data. This reinforces the conclusion of TB23 that, despite the limitations of our data and somewhat unconventional aspects of our analysis (mainly the EL-base), our methodology yields physical properties consistent with those derived from more conventional methods.

5. Discussion

In this section, we compare the scaling relations and other derived properties of the galaxies in the M101 group. We explore similarities and differences between the galaxies and compare our findings with results from the literature.

To facilitate intercomparison of the scaling relations among the different galaxies in the group, Fig. 6 overplots the median relations in Fig. 5 for all six galaxies. The local mass-age relation (Σ_⋆-⟨log t⟩_L) shows that all galaxies approximately converge at ${\mathrm{\Sigma }}_{\star }\lesssim 10M_{\odot }/{\text{pc}}^2$ and diverge at higher densities, with more massive galaxies generally exhibiting older stellar populations for the same ${\mathrm{\Sigma }}_{\star }$ (Pérez-González et al. 2008; González Delgado et al. 2015). This trend is consistent with the expected inside-out growth of galaxies, where central regions are more evolved. Only UGC 8837 deviates from this sequence.

Fig. 6. Median scaling relations for all galaxies. The dotted lines in the right panel represent lines of specific SFR at 0.01 and 1 Gyr−1. The thick gray line shows the star-forming MS relation obtained by Enia et al. (2020) for nearby spirals.

The stellar and nebular MZRs exhibit distinct behaviors. The stellar MZR in the second panel of Fig. 6 shows that our most massive galaxy, M101, reaches the highest ⟨log Z/Z_⊙⟩_M values, with a clear gradient towards denser regions. In contrast, NGC 5474, NGC 5585 and NGC 5204 exhibit flatter relations, suggesting a less efficient metal enrichment process, possibly due to the shallower gravitational potential wells in these low-mass systems. This is in line with results from IFS-based studies (Neumann et al. 2021). Beyond this, the values of ⟨log Z/Z_⊙⟩_M are around −0.5, which is consistent with Sd galaxies (González Delgado et al. 2015).

The nebular MZR reveals systematic trends with ${\mathrm{\Sigma }}_{\star }$ across all galaxies. In M101, O/H increases steadily with ${\mathrm{\Sigma }}_{\star }$. This gradient becomes less pronounced in less massive galaxies, such as NGC 5474, reflecting lower overall enrichment levels. This is consistent with Belfiore et al. (2017), who observe flatter radial gradients of O/H as M_⋆ decreases.

The spatially resolved Σ_SFR^⋆-Σ_⋆ relations for all galaxies align well with the MS of star formation. For M101, the relation matches the Enia et al. (2020) results at Σ_⋆ ≳ 10 M_⊙ pc², while less massive galaxies show systematic M_⋆-dependent offsets, possibly reflecting differences in star formation efficiency and evolutionary states. The upturn toward a large specific SFR at low Σ_⋆ occurs because of the star-forming regions sampled in the outskirts of our galaxies, whose W_Hα values exceed 200 Å.

Finally, we note that the widespread star formation across all galaxies in the group indicates that they have not undergone any substantial environmental quenching. The efficacy of many environmental processes is known to increase with the mass of the group or cluster (e.g., Alonso et al. 2012; Raj et al. 2019). In terms of its total stellar mass, 2.5 × 10¹⁰ M_⊙ according to our analysis, M101 is a relatively low mass group, which explains why its galaxies show no sign of suppressed star formation. This is in line with the results of González Delgado et al. (2022) on the effect of the stellar mass on the excess of quenched galaxies with respect to the field. They find that, for a fixed M_⋆, the quenched fraction excess in low-mass groups (defined as those with < 5 × 10¹¹ M_⊙ in stars, the regime of the M101 group) is significantly lower than in more massive structures.

At this stage, these interpretations remain speculative and are based on results from previous integrated-light studies. As we expand this work on the 2D characterization of galaxy properties to include other groups and clusters, we will be better equipped to evaluate environmental effects on the properties of galaxies at local scales, as highlighted in other studies (see, e.g., Coenda et al. 2019, Bluck et al. 2020, Epinat et al. 2024 ).

6. Summary

This paper marks the beginning of a series of studies focused on galaxies in nearby groups and clusters, utilizing 12-band optical imaging photometry data from the J-PLUS and S-PLUS surveys. In this first article, we have described our data-handling and photo-spectral synthesis methodologies, along with results for the M101 group derived from J-PLUS observations. We have presented (Figs. 3–5):

1. Maps of stellar mass surface density (Σ_⋆), mean stellar age (⟨log t⟩_L), SFR surface density derived from the synthesis (Σ_SFR^⋆), and maps of the stellar populations divided into young, intermediate, and old age bins.

2. Maps of emission line properties including Hα fluxes, equivalent widths, RGB composites of ([N II], Hα, [O III]), as well as the corresponding BPT diagram.

3. Spatially resolved relations between the mean stellar age, stellar and nebular metallicities, and Σ_SFR^⋆ with Σ_⋆.

We also compared the local scaling relations for the different galaxies in the group, obtaining: (a) Σ_⋆-age relations shifted towards younger ages as M_⋆ decreases; (b) a well-defined stellar MZR in M101, while less massive galaxies exhibit flatter or less distinct trends; (c) a nebular MZR that agrees with trends found in IFS studies of galaxies in the same mass range; (d) Σ_SFR^⋆-Σ_⋆ MS relations offset toward higher specific SFR values as M_⋆ decreases.

We have therefore achieved our goal of characterizing the spatially resolved stellar population and EL properties of galaxies in the M101 group. Our results demonstrate the capability of J-PLUS to perform spatially resolved analyses of galaxy properties, providing insights comparable to those from traditional IFS surveys.

However, this study alone is insufficient to fully assess the nature and intensity of environmental effects acting on galaxies in groups. To do this, comparable characterization of galaxies in different environments must be compiled, and this study represents a first step in this direction. We plan to apply the same overall methodology employed here to other nearby groups and clusters in future studies, with the ultimate goal of building a large, homogeneous dataset of IFS-like observations and analyses. This will enable meaningful comparisons across galaxies with varying masses, dynamical states, and evolutionary histories in different environments.

Acknowledgments

This work was supported by CAPES under grant 88881.892595/2023-01 and FAPESC (CP 48/2021). RCF acknowledges support from CNPq (grants 302270/2018-3 and 404238/2021-1). JTB, RGD, RGB, JRM, GMS and LADG acknowledge financial support from the Severo Ochoa grant CEX2021-001131-S funded by MCIN/AEI/ 10.13039/501100011033 and to grant PID2022-141755NB-I00. Based on observations made with the JAST80 telescope and T80Cam camera at the Observatorio Astrofísico de Javalambre (OAJ), in Teruel, owned, managed, and operated by the Centro de Estudios de Física del Cosmos de Aragón. We acknowledge the OAJ Data Processing and Archiving Department (DPAD) for reducing and calibrating the OAJ data used in this work, as well as the distribution of the data products through a dedicated web portal. Funding for OAJ, UPAD, and CEFCA has been provided by the Governments of Spain and Aragón through the Fondo de Inversiones de Teruel and their general budgets; the Aragonese Government through the Research Groups E96, E103, E16_17R, E16_20R and E16_23R; the Spanish Ministry of Science and Innovation (MCIN/AEI/10.13039/501100011033 y FEDER, Una manera de hacer Europa) with grants PID2021-124918NB-C41, PID2021-124918NB-C42, PID2021-124918NA-C43, and PID2021-124918NB-C44; the Spanish Ministry of Science, Innovation and Universities (MCIU/AEI/FEDER, UE) with grant PGC2018-097585-B-C21; the Spanish Ministry of Economy and Competitiveness (MINECO) under AYA2015-66211-C2-1-P, AYA2015-66211-C2-2, AYA2012-30789, and ICTS-2009-14; and European FEDER funding (FCDD10-4E-867, FCDD13-4E-2685). The Brazilian agencies FINEP, FAPESP, and the National Observatory of Brazil have also contributed to J-PLUS. We also thank the comments from J. A. Fernández-Ontiveros, P.T. Rahna, and E. Telles.

References

Appendix A: Complementary material

A.1. Preprocessing

Figure A.1 uses the u and r band images of M101 to illustrate the sequence of preprocessing operations described in Sect. 2. It is notable in Fig. A.1 that the original data in the u-band are quite faint, exhibiting a significant number of negative flux pixels (in white in the image), approximately 36% of those within the galaxy mask, which implies an even higher count of pixels with bad signal. In contrast, the r-band, which has a superior S/N ratio, shows 6% of negative flux pixels. The figure demonstrates the sequential reduction in negative flux pixels after each preprocessing step, which can be read as a measure of improvement in images. Figure A.2 shows how the composites in Fig. 1 change after preprocessing.

Fig. A.1. Step-by-step of the preprocessing applied to the J-PLUS data cube of M101. The columns show, respectively: the original data, Butterworth filter, PSF homogenization, 2 × 2 binning, and Voronoi binning (defined by the u-band). The blue line represents the contour of the galaxy mask. The first and second rows are the surface brightness maps (in AB mag/arcsec2) in the u and r-band respectively. Pixels with < 0 flux appear in white. The legends fu− and fr− show the number of pixels with negative fluxes inside the galaxy mask in the u and r bands, while the f12− numbers under the plots are the sum of the fλ < 0 pixels in all the 12 bands for each of the preprocessing steps. Masked stars are marked in green.

Fig. A.2. Same as Fig. 1 but for the data after the preprocessing.

Fig. A.3. Maps of the mean relative absolute deviation between observed and model fluxes, $\overline{\mathrm{\Delta }}\equiv ⟨|O_{\lambda }-M_{\lambda }|/M_{\lambda }⟩$. Fits are presented as Fig.2, but for the other galaxies. The last panel shows the fit positions (A, B, and C) in the galaxy.

Fig. A.4. Left: Maps of the effective V-band dust optical depth ($\stackrel{\sim }{\tau }$). The central and right panels show RGB composites with the i, r, and g bands before and after correction for dust, respectively.

Fig. A.5. From left to right: Maps of the [O II]3727, [O III]5007, and [N II]6584 surface brightness, and the [N II]/Hα ratio.

A.2. Photo-spectral fits and the mean relative absolute deviation maps

Like Fig. 2 in the main text, Fig. A.3 shows three example fits for zones in galaxies other than M101. Region A in all cases shows the fit for the central spaxel. The resulting $\overline{\mathrm{\Delta }}$ range from 0.79 to 1.93%. Region B represents fainter regions, near the boundary of our galaxy mask (as seen in the last panel). As expected, these B-fits have the worst $\overline{\mathrm{\Delta }}$, though it is not too bad for the NGC 5474, NGC 5585, and NGC 5204 (the most massive in the figure), varying from 2.05 to 3.75%. However, for UGC 8837, the noisiest galaxy in the sample, region B has residuals of $\overline{\mathrm{\Delta }}=6.58$%, a good example of a bad fit. Lastly, the regions labeled C show examples of H II regions, which show significant Hα and [N II] fluxes in the J0660 filter, and, in some cases, a prominent [O II] contribution to the J0378 filter.

Figure A.3 shows maps of $\overline{\mathrm{\Delta }}$, the mean relative absolute deviation between data and model fluxes, for all galaxies in the sample. Red colors in these maps indicate relatively bad fits ($\overline{\mathrm{\Delta }}>6\%$). Properties derived from these zones should be treated with care. It is notable that, despite the generally good fits of M101, the inter-arm areas exhibit the poorest $\overline{\mathrm{\Delta }}$, which is expected given that these areas have less signal. NGC 5474 is the galaxy with the smallest $\overline{\mathrm{\Delta }}$ values, with few regions where $\overline{\mathrm{\Delta }}\ge 6$%. Both NGC 5585 and NGC 5204 demonstrate good fits, with $\overline{\mathrm{\Delta }}\sim 2\%$ at the center, increasing towards the edges as the signal wanes. The dwarfs UGC 8837 and NGC 5477 exhibit a similar pattern.

A.3. Dust maps and de-reddened images

The left panels in Fig. A.4 show maps of the effective V-band dust optical depth $\stackrel{\sim }{\tau }$, defined as

$$\begin{array}{c}\hfill \stackrel{\sim }{\tau }\equiv x_{\mathrm{BC}}({\tau }^{\mathrm{ISM}}+{\tau }^{\mathrm{BC}})+(1-x_{\mathrm{BC}}){\tau }^{\mathrm{ISM}}=x_{\mathrm{BC}}{\tau }^{\mathrm{BC}}+{\tau }^{\mathrm{ISM}},\end{array}$$(A.1)

where x_BC is the sum of all x_j fractions associated with populations with age t_j < t_BC = 10 Myr . In our two-τ’s model, these young stars are attenuated by τ^ISM + τ^BC, while the older ones are attenuated by only τ^ISM, so that $\stackrel{\sim }{\tau }$ is a convenient weighted average that condenses these two components of dust onto a single number. By construction, $\stackrel{\sim }{\tau }\to {\tau }^{\mathrm{ISM}}$ when young populations are not present, and to τ^ISM + τ^BC when they dominate.

In M101 our dust map follows the spiral arms, with peaks reaching $\stackrel{\sim }{\tau }\sim 1$ aligning with its star-forming regions and a slight general decrease towards large radii. An inner, more diffuse component is also seen in the central parts, as well as a horizontal inner bar-like feature like that seen in CO (Kenney et al. 1991). Except for these regions, overall our fits indicate that there is relatively little dust attenuation in M101, consistent with its ∼ face-on orientation. The equivalent A_V values compare relatively well with those derived by Lin et al. (2013) and Watkins et al. (2017) with very different methodologies.

Other galaxies in the group also show relatively little dust. The $\stackrel{\sim }{\tau }$ maps are patchy, with no clear structure, with peaks in the star-forming regions (compare with the ⟨log t⟩_L and W_Hα maps). Note that we find very little dust in UGC 8837, despite its edge-on orientation.

TB23 used their ALSTAR fits to correct S-PLUS images of NGC 1365 for dust. Our galaxies are not as dusty as NGC 1365, whose inner regions are heavily obscured, but it is still worth repeating this exercise for our J-PLUS data on galaxies in the M101 group. Figure A.4 shows the results. Its middle panels show composites made with the i, r, and g images in the R, G, and B channels, respectively, while the panels on the right show the same composite after correcting for dust attenuation.

For M101, the dusty filaments along the inner arms, visible as red stripes in the original image, are absent in the corrected one, indicating that, despite its simplicity, our modeling is able to account for dust attenuation adequately. Moreover, the star-forming regions, where the largest values of $\stackrel{\sim }{\tau }$ are found, now appear bluer, brighter, and with a stronger contrast to the underlying disk. For the other galaxies, where the dust is less structured, the middle and right panels differ in their overall blueness and in the brightness of the star-forming regions.

A.4. Emission lines

Figure A.5 shows complementary maps related to our EL results. The Σ_Hα maps were previously shown in Fig. 4, while Fig. A.5 shows the surface brightness maps for the forbidden lines [O II], [O III] and [N II], as well as the [N II]/Hα flux ratio. Note that, for the redshifts of galaxies in the M101 group, only [O II] and [N II] + Hα fall under narrow bands (J0378 and J0660). Yet, as demonstrated in TB23, because of the empirical nature of the EL base used in our fits, even lines covered only by broad bands are well recovered by our method.

As in the Hα maps, the [O III] maps in Fig. A.5 also show the star-forming regions standing out clearly and in sharp contrast with their surroundings. These same regions stand out in the [O II] and [N II] maps, although these are visibly more diffuse than the [O III] and Hα ones. These low ionization lines are characteristic of DIG emission.

The [N II]/Hα map for M101 shows this EL flux ratio decreasing from ∼0.5 in the inner star-forming regions to less than 0.1 in the outer ones. As suspected from Fig. A.5, in the central parts of M101, as well as in between its spiral arms, [N II]/Hα is larger, reaching values typical of DIG. As discussed in Sect. 4.2, the W_Hα map in the second column panel of Fig. 4 confirms this interpretation. A gradient in [N II]/Hα is also present in NGC 5474, NGC 5585, and NGC 5204, though not as strong as in M101, while no clear trend is identifiable in the two dwarf galaxies.

A.5. Uncertainties

This final section addresses the uncertainties in the main properties derived from our methodology. As one can infer from the overall consistency of the results, as well as from the reported agreement with previous studies, uncertainties are not a limiting factor in our analysis. It is, nonetheless, relevant to discuss them, especially given some of the novelties in our approach (like the modeling of ELs). Though based on the results for galaxies in the M101 group, the discussion below serves as a guide for future studies where the same kind of analysis is applied to other targets.

A.5.1. Uncertainties in individual fits

The dispersion among the MC runs in ALSTAR offers a way to estimate the uncertainties in the derived properties. The experiments in TB23, where these dispersions were compared to the dispersion of the difference (δ) between properties derived from STARLIGHT full spectral fits of SDSS galaxies and those obtained from ALSTAR fits of the corresponding synthetic photometry in the S-PLUS bands, confirm that this is a reasonable method to estimate uncertainties, even though a tendency was noted for the ALSTAR-based MC dispersions to overestimate the more heuristic δ-based ones by factors of up to 2, depending on the property. Let us thus examine the MC-based dispersions in the properties derived for individual fits, be they single pixels or Voronoi zones. We will use the σ_NMAD statistic as a measure of dispersion.

The median values of σ_NMAD for logΣ_⋆ span the 0.22–0.25 dex range for the six galaxies in the M101 group. In star-forming regions, however, σ_NMAD can reach values of 0.4 dex or larger. These regions have their light completely dominated by young populations, but even a 1% light fraction due to old stars would be enough to make them dominant in mass. Variations in this very poorly constrained old component among the MC fits explain the large spread in Σ_⋆ in these regions. As an illustration, the median σ_NMAD(logΣ_⋆) in M101 changes from 0.17 dex for regions where W_Hα < 30 Å to 0.26 dex for those where Hα has a larger equivalent width.

Figure A.6a shows the map of σ_NMAD(logΣ_⋆) for M101. Comparing this map with the ones for ⟨log t⟩_L or W_Hα (Figs. 3 and 4) one sees low σ_NMAD in regions dominated by old populations, and larger uncertainties in regions of ongoing star formation, as summarized above.

Regarding mean ages, we find median values of σ_NMAD(⟨log t⟩_L) in the 0.34–0.52 dex range for our galaxies. σ_NMAD(⟨log t⟩_L) tends to be larger for regions with ages close to the middle of the age grid (at ∼10⁸ yr), where the larger number of combinations of different components leading to approximately the same total spectrum inevitably translates to a larger dispersion in ⟨log t⟩_L. The map of σ_NMAD(⟨log t⟩_L) in Fig. A.6b shows a general progression towards larger uncertainties towards the outer regions. The red regions, where σ_NMAD ≥ 0.5 dex, are mostly located in regions of intermediate age and W_Hα, often around H II regions (see Figs. 3 and 4). In the H II regions themselves, uncertainties are smaller, though not as small as in zones dominated by older populations. Overall, though acceptable for photometric work, our uncertainties in ⟨log t⟩_L are substantially larger than those attainable with spectroscopy, which are typically about 0.1 dex (e.g., Cid Fernandes et al. 2014).

Fig. A.6. Maps of the σNMAD dispersion among the Monte Carlo runs for logΣ⋆, ⟨log t⟩L, $log{W}_{\text{H}\alpha \mathrm{N}\mathrm{I}\mathrm{I}}$, and log WHα for M101.

Uncertainties in dust attenuation are also larger than the typical ±0.1 mag in A_V attainable with full spectral synthesis over a similar spectral range. We obtain median σ_NMAD(τ^ISM) values of 0.15–0.22 for our galaxies, corresponding to 0.16–0.24 mag in A_V. Stellar metallicities, on the other hand, turn out to have small uncertainties (∼0.2 dex), but this is mainly due to the restricted range in Z allowed for in the fits. Regarding SFRs, we obtain median σ_NMAD(logΣ_SFR^⋆) values of 0.15–0.22 dex across our sample. These dispersions are obtained after discarding zones where < 100 Myr stars contribute less than 30% to the flux at 5635 Å, i.e., by focusing on regions where these populations are reliably detected.

Turning to EL properties, let us first examine the combined equivalent widths of the Hα and $[\mathrm{N}\mathrm{I}\mathrm{I}]\lambda \lambda 6548,6584$ lines, $W_{\text{H}\alpha \mathrm{N}\mathrm{I}\mathrm{I}}$. We find median ${\sigma }_{\mathrm{NMAD}}(log{W}_{\text{H}\alpha \mathrm{N}\mathrm{I}\mathrm{I}})$ values of 0.05–0.25 dex over our sample (see Fig. A.6c). The highest value of 0.25 dex is for NGC 5585, due to its unusual noisy data in the J0660 filter, with photometric uncertainties about ∼3.3× larger than in the other galaxies.

Considering only zones where $W_{\text{H}\alpha \mathrm{N}\mathrm{I}\mathrm{I}}>30\mathrm{\AA }$, the median values of ${\sigma }_{\mathrm{NMAD}}(log{W}_{\text{H}\alpha \mathrm{N}\mathrm{I}\mathrm{I}})$ for all galaxies is just ∼0.05 dex. Such small uncertainties are not surprising, given that, at the redshift of our sources, these lines all fall within the J0660 filter, so that their combined equivalent width can be reliably estimated. This is in fact the reason why in TB23 we have devised a whole empirical $W_{\text{H}\alpha \mathrm{N}\mathrm{I}\mathrm{I}}$-based scheme to disentangle Hα from [N II] which greatly improves the EL diagnostic power of S-PLUS and J-PLUS data.

In itself, however, $W_{\text{H}\alpha \mathrm{N}\mathrm{I}\mathrm{I}}$ is not a particularly interesting index. W_Hα, on the other hand, is a very useful tracer of specific SFR (or age of an H II region, depending on the scale one is looking at). We obtain median σ_NMAD(log W_Hα) values of ∼0.1 dex for M101, NGC 5204, UGC 8837, and NGC 5477, while for NGC 5474 the value is 0.2 dex and for NGC 5585 it is 0.3 dex. In all cases, the relative uncertainty (i.e., σ_W/W) decreases as W_Hα increases. Even when W_Hα is small (and thus uncertain) we are able to identify is as small, which is good enough for basic diagnostics, like distinguishing DIG from H II regions.

Figure A.6d shows the map of σ_NMAD(log W_Hα) for M101. Comparing to Fig. 4, one sees that log W_Hα is most uncertain in regions where W_Hα is just a few Å. In fact, up to W_Hα ∼ 10 Å the MC dispersion is of the order of W_Hα itself. Conversely, uncertainties in log W_Hα are very low in star-forming regions.

Another EL property of interest is the [N II]/Hα ratio, which, besides being a nebular metallicity indicator by itself, has a major diagnostic role in separating star-forming regions from those where other sources of ionization are present. The median ${\sigma }_{\mathrm{NMAD}}(log[\mathrm{N}\mathrm{I}\mathrm{I}]/\text{H}\alpha )$ values for our galaxies are all very close to 0.2 dex, and in very few regions it reaches 0.3 dex. This apparently good precision (for photometric data) is, however, in most part built into our method, which uses an initial estimate of $W_{\text{H}\alpha \mathrm{N}\mathrm{I}\mathrm{I}}$ to restrict the possible EL-space available to the fits.

Finally, the MC uncertainties on 12 + logO/H run from 0.05 to 0.2 dex over the body of our galaxies, increasing systematically as O/H increases. This happens because our empirical EL-base is built to span the observed BPT diagram, which is narrower at its top-left, low O/H regime, than at its center-bottom, where O/H is larger. Consequently, the MC fits mix fewer base elements at low O/H than at high O/H. Notice that these uncertainties are of the same order of the overall variations in log O/H within our galaxies. This, however, does not challenge the significance of the nebular MZRs seen in Fig. , since they were obtained not on the basis of individual fits, but from median curves over many such fits, and thus much less uncertain them each of them. Statistics, in fact, plays a major role in the assessment of uncertainties, as discussed next.

A.5.2. Uncertainties in practice

The uncertainties in the individual fits to J-PLUS photometry discussed just above are clearly larger than those ones grew accustomed to in the era of spectroscopic surveys, as expected from the disparity in information content between, say, the thousands of flux entries in an SDSS or CALIFA spectrum and the 12 points “summary of a spectrum” offered by J-PLUS.

In practice, however, both in bona fide IFS and IFS-like studies, one never focuses on results for an individual spaxel (or Voronoi zone). Instead, spatial averages are always considered in one way or another. This averaging can be either explicit, like when one examines radial profiles (including alternative representations like those in Fig. 5, where the x-axis is only indirectly related to radius), or implicit, like when one inspects maps like those in Figs. 3 and 4. This averaging process reduces the formal uncertainties of individual fits by factors of $\sqrt{N}$, which in most circumstances make them negligible.

There are, of course, other less formal sources of uncertainty inherent in this kind of analysis. For instance, we adopted a fixed dust attenuation law and a pre-defined recipe for differential extinction, which is at best an approximation to reality. Our choice of spectral models for stellar populations, with its built-in assumptions and approximations, is also a source of uncertainty, as is the way we define our EL base as well as our implementation of an empirical EL prior to disentangle [N II] from Hα. In summary, each of these modeling choices carries its own set of assumptions, contributing to the overall uncertainty in the analysis.

While we have not quantified these modeling-related uncertainties in this work, we can borrow from the thorough analysis of uncertainties in full spectral fitting of CALIFA datacubes by Cid Fernandes et al. (2014). They find that different choices of stellar population models impact the results more than uncertainties related to the method itself, a conclusion that most likely applies to this work too. Substantial changes in the results reported here are thus more likely to come from changes in the ingredients of the analysis or eventual recalibrations of the data.

All Tables

All Figures

	Fig. 1. Composites of the original data for galaxies in the M101 group, constructed using the J0660, g, and the sum of the five bluest filters in the R, G, and B channels, respectively. The white bars indicate a scale of 1 arcmin.
In the text

Fig. 2. Top: Example ALSTAR fits for individual spaxels in three different regions of M101. Colored lines with error bars show the data (Oλ), while black lines show the model photometric fluxes (Mλ), and the gray lines show the corresponding high-resolution model spectrum. Images on the right show 278″ × 278″ composites built with the J0660, r, and g fluxes in the R, G, and B channels, respectively. Bottom: Distributions of the (Oλ − Mλ)/Mλ relative residuals of the fits for the 12 J-PLUS bands for 11 705 zones in M101. Median residuals are marked by solid black circles, while the horizontal bars mark the 16 and 84 percentiles.

In the text

	Fig. 3. Maps of stellar population properties for galaxies in the M101 group. From left to right: surface density, mean age, star formation rate (SFR) surface density, and an RGB with the fluxes at 5635 Å of old, intermediate-age, and young populations. See text for details.
In the text

	Fig. 4. Maps of emission-line properties. From left to right: Hα surface brightness; Hα equivalent width; RGB with the ([N II], Hα, [O III]) fluxes; BPT diagram of the spaxels of each galaxy, with points color-coded as in the RGB panel.
In the text

Fig. 5. Scaling relations for galaxies in the M101 group. Individual zones are plotted as circles colored by log WHα. The thick black lines with circle markers indicate the median curves for bins in logΣ⋆; thin lines show the corresponding 16th and 84th percentile curves. In the third column, the black dashed line is the median curve from Barrera-Ballesteros et al. (2016) for 653 MaNGA galaxies. The gray line shows the 80% contour of Barrera-Ballesteros et al. (2016) for MaNGA galaxies with log M⋆/M⊙ < 9.2. The dotted contour indicates the typical loci of Sd galaxies from Sánchez 2020. Dashed lines in the fourth column panels mark lines of specific SFR = 0.01 and 1 Gyr−1. The solid gray line is the relation obtained by Enia et al. (2020) – see text for details.

In the text

	Fig. 6. Median scaling relations for all galaxies. The dotted lines in the right panel represent lines of specific SFR at 0.01 and 1 Gyr−1. The thick gray line shows the star-forming MS relation obtained by Enia et al. (2020) for nearby spirals.
In the text

Fig. A.1. Step-by-step of the preprocessing applied to the J-PLUS data cube of M101. The columns show, respectively: the original data, Butterworth filter, PSF homogenization, 2 × 2 binning, and Voronoi binning (defined by the u-band). The blue line represents the contour of the galaxy mask. The first and second rows are the surface brightness maps (in AB mag/arcsec2) in the u and r-band respectively. Pixels with < 0 flux appear in white. The legends fu− and fr− show the number of pixels with negative fluxes inside the galaxy mask in the u and r bands, while the f12− numbers under the plots are the sum of the fλ < 0 pixels in all the 12 bands for each of the preprocessing steps. Masked stars are marked in green.

In the text

	Fig. A.2. Same as Fig. 1 but for the data after the preprocessing.
In the text

	Fig. A.3. Maps of the mean relative absolute deviation between observed and model fluxes, $\overline{\mathrm{\Delta }}\equiv ⟨|O_{\lambda }-M_{\lambda }|/M_{\lambda }⟩$. Fits are presented as Fig.2, but for the other galaxies. The last panel shows the fit positions (A, B, and C) in the galaxy.
In the text

	Fig. A.4. Left: Maps of the effective V-band dust optical depth ($\stackrel{\sim }{\tau }$). The central and right panels show RGB composites with the i, r, and g bands before and after correction for dust, respectively.
In the text

	Fig. A.5. From left to right: Maps of the [O II]3727, [O III]5007, and [N II]6584 surface brightness, and the [N II]/Hα ratio.
In the text

	Fig. A.6. Maps of the σNMAD dispersion among the Monte Carlo runs for logΣ⋆, ⟨log t⟩L, $log{W}_{\text{H}\alpha \mathrm{N}\mathrm{I}\mathrm{I}}$, and log WHα for M101.
In the text