It remains a major challenge to derive a theory of cloud-scale ($\lesssim100$ pc) star formation and feedback, describing how galaxies convert gas into stars as a function of the galactic environment. Progress has been hampered by a lack of robust empirical constraints on the giant molecular cloud (GMC) lifecycle. We address this problem by systematically applying a new statistical method for measuring the evolutionary timeline of the GMC lifecycle, star formation, and feedback to a sample of nine nearby disc galaxies, observed as part of the PHANGS-ALMA survey. We measure the spatially-resolved ($\sim100$ pc) CO-to-H$\alpha$ flux ratio and find a universal de-correlation between molecular gas and young stars on GMC scales, allowing us to quantify the underlying evolutionary timeline. GMC lifetimes are short, typically 10-30 Myr, and exhibit environmental variation, between and within galaxies. At kpc-scale molecular gas surface densities $\Sigma_{\rm H_2}\geqslant8$M$_{\odot}$pc$^{-2}$, the GMC lifetime correlates with time-scales for galactic dynamical processes, whereas at $\Sigma_{\rm H_2}\leqslant8$M$_{\odot}$pc$^{-2}$ GMCs decouple from galactic dynamics and live for an internal dynamical time-scale. After a long inert phase without massive star formation traced by H$\alpha$ (75-90% of the cloud lifetime), GMCs disperse within just 1-5 Myr once massive stars emerge. The dispersal is most likely due to early stellar feedback, causing GMCs to achieve integrated star formation efficiencies of 4-10% These results show that galactic star formation is governed by cloud-scale, environmentally-dependent, dynamical processes driving rapid evolutionary cycling. GMCs and HII regions are the fundamental units undergoing these lifecycles, with mean separations of 100-300 pc in star-forming discs. Future work should characterise the multi-scale physics and mass flows driving these lifecycles.

The physics of star formation and the deposition of mass, momentum, and energy into the interstellar medium by massive stars (`feedback') are the main uncertainties in modern cosmological simulations of galaxy formation and evolution. These processes determine the properties of galaxies, but are poorly understood on the $\lesssim$100 pc scale of individual giant molecular clouds (GMCs) resolved in modern galaxy formation simulations. The key question is why the timescale for depleting molecular gas through star formation in galaxies ($t_{\rm dep}\approx 2$ Gyr) exceeds the dynamical timescale of GMCs by two orders of magnitude. Either most of a GMC's mass is converted into stars over many dynamical times, or only a small fraction turns into stars before the GMC is dispersed on a dynamical timescale. Here we report our observation that molecular gas and star formation are spatially de-correlated on GMC scales in the nearby flocculent spiral galaxy NGC300, contrary to their tight correlation on galactic scales. We demonstrate that this de-correlation implies rapid evolutionary cycling between GMCs, star formation, and feedback. We apply a novel statistical method to quantify the evolutionary timeline and find that star formation is regulated by efficient stellar feedback, driving GMC dispersal on short timescales (~1.5 Myr) due to radiation and stellar winds, prior to supernova explosions. This feedback limits GMC lifetimes to about one dynamical timescale (~10 Myr), with integrated star formation efficiencies of only 2-3%. Our findings reveal that galaxies consist of building blocks undergoing vigorous, feedback-driven lifecycles, that vary with the galactic environment and collectively define how galaxies form stars. Systematic applications of this multi-scale analysis to large galaxy samples will provide key input for a predictive, bottom-up theory of galaxy formation and evolution.

Direct comparisons between galaxy simulations and observations that both reach scales < 100 pc are strong tools to investigate the cloud-scale physics of star formation and feedback in nearby galaxies. Here we carry out such a comparison for hydrodynamical simulations of a Milky Way-like galaxy, including stochastic star formation, HII region and supernova feedback, and chemical post-processing at 8 pc resolution. Our simulation shows excellent agreement with almost all kpc-scale and larger observables, including total star formation rates, radial profiles of CO, HI, and star formation through the galactic disc, mass ratios of the ISM components, both whole-galaxy and resolved Kennicutt-Schmidt relations, and giant molecular cloud properties. However, we find that our simulation does not reproduce the observed de-correlation between tracers of gas and star formation on < 100 pc scales, known as the star formation 'uncertainty principle', which indicates that observed clouds undergo rapid evolutionary lifecycles. We conclude that the discrepancy is driven by insufficiently-strong pre-supernova feedback in our simulation, which does not disperse the surrounding gas completely, leaving star formation tracer emission too strongly associated with molecular gas tracer emission, inconsistent with observations. This result implies that the cloud-scale de-correlation of gas and star formation is a fundamental test for feedback prescriptions in galaxy simulations, one that can fail even in simulations that reproduce all other macroscopic properties of star-forming galaxies.

Diffuse emission is observed in galaxies in many tracers across the electromagnetic spectrum, including tracers of star formation, such as H$\alpha$ and ultraviolet (UV), and tracers of gas mass, such as carbon monoxide (CO) transition lines and the 21-cm line of atomic hydrogen (HI). Its treatment is key to extracting meaningful information from observations such as cloud-scale star formation rates. Finally, studying diffuse emission can reveal information about the physical processes taking place in the ISM, such as chemical transitions and the nature of stellar feedback (through the photon escape fraction). We present a physically-motivated method for decomposing astronomical images containing both diffuse emission and compact regions of interest, such as HII regions or molecular clouds, into diffuse and compact component images through filtering in Fourier space. We have previously presented a statistical method for constraining the evolutionary timeline of star formation and mean separation length between compact star forming regions with galaxy-scale observations. We demonstrate how these measurements are biased by the presence of diffuse emission in tracer maps and that by using the mean separation length as a critical length scale to separate diffuse emission from compact emission, we are able to filter out this diffuse emission, thus removing its biasing effect. Furthermore, this method provides, without the need for interferometry or ancillary spectral data, a measurement of the diffuse emission fraction in input tracer maps and decomposed diffuse and compact emission maps for further analysis.

We recently presented a new statistical method to constrain the physics of star formation and feedback on the cloud scale. Fundamentally, this new method is only able to recover the relative durations of different evolutionary phases. It therefore requires a "reference time-scale" to convert the relative time-scales into absolute values. The phase over which the star formation rate (SFR) tracer is visible is the logical choice to assign this reference time-scale to, since it can be characterised using stellar population synthesis models. In this paper, we use hydrodynamical disc galaxy simulations to produce synthetic emission maps of several SFR tracers and apply our statistical method to measure the associated characteristic time-scale of each tracer. These cover 12 ultraviolet (UV) filters (from GALEX, Swift, and HST) covering a wavelength range 150-350 nm, as well as H$\alpha$. For solar-metallicity environments, we find the characteristic time-scales for H$\alpha$ with (without) continuum subtraction to be ~4.3 (6-16) Myr and for the UV filters to be in the range 17-33 Myr, monotonically increasing with wavelength. We find that the characteristic time-scale decreases towards higher metallicities, as well as to lower star formation rate surface densities, if the stellar initial mass function is not well-sampled. We provide fitting functions to the resulting reference time-scales to facilitate observational applications of our statistical method across a wide range of galactic environments. More generally, our results predict the time-scales over which photoionisation and UV heating take place around star-forming regions.

The cloud-scale physics of star formation and feedback represent the main uncertainty in galaxy formation studies. Progress is hampered by the limited empirical constraints outside the restricted environment of the Local Group. In particular, the poorly-quantified time evolution of the molecular cloud lifecycle, star formation, and feedback obstructs robust predictions on the scales smaller than the disc scale height that are resolved in modern galaxy formation simulations. We present a new statistical method to derive the evolutionary timeline of molecular clouds and star-forming regions. By quantifying the excess or deficit of the gas-to-stellar flux ratio around peaks of gas or star formation tracer emission, we directly measure the relative rarity of these peaks, which allows us to derive their lifetimes. We present a step-by-step, quantitative description of the method and demonstrate its practical application. The method's accuracy is tested in nearly 300 experiments using simulated galaxy maps, showing that it is capable of constraining the molecular cloud lifetime and feedback time-scale to $<0.1$ dex precision. Access to the evolutionary timeline provides a variety of additional physical quantities, such as the cloud-scale star formation efficiency, the feedback outflow velocity, the mass loading factor, and the feedback energy or momentum coupling efficiencies to the ambient medium. We show that the results are robust for a wide variety of gas and star formation tracers, spatial resolutions, galaxy inclinations, and galaxy sizes. Finally, we demonstrate that our method can be applied out to high redshift ($z\lesssim4$) with a feasible time investment on current large-scale observatories. This is a major shift from previous studies that constrained the physics of star formation and feedback in the immediate vicinity of the Sun.