The formation of the solar system's giant planets predated the ultimate epoch of massive impacts that concluded the process of terrestrial planet formation. Following their formation, the giant planets' orbits evolved through an episode of dynamical instability. Several qualifies of the solar system have recently been interpreted as evidence of this event transpiring within the first similar to 100 Myr after the Sun's birth; around the same time as the final assembly of the inner planets. In a series of recent papers we argued that such an early instability could resolve several problems revealed in classic numerical studies of terrestrial planet formation; namely the small masses of Mars and the asteroid belt. In this paper, we revisit the early instability scenario with a large suite of simulations specifically designed to understand the degree to which Earth and Mars' formation are sensitive to the specific evolution of Jupiter and Saturn's orbits. By deriving our initial terrestrial disks directly from recent high-resolution simulations of planetesimal accretion, our results largely confirm our previous findings regarding the instability's efficiency of truncating the terrestrial disk outside of the Earth-forming region in simulations that best replicate the outer solar system. Moreover, our work validates the primordial 2:1 Jupiter-Saturn resonance within the early instability framework as a viable evolutionary path for the solar system. While our simulations elucidate the fragility of the terrestrial system during the epoch of giant planet migration, many realizations yield outstanding solar system analogs when scrutinized against a number of observational constraints. Finally, we highlight the inability of models to form adequate Mercury-analogs and the low eccentricities of Earth and Venus as the most significant outstanding problems for future numerical studies to resolve.

In a recent paper we proposed that the giant planets' primordial orbits may have been eccentric (e(J) similar to e(S) similar to 0.05), and used a suite of dynamical simulations to show outcomes of the giant planet instability that are consistent with their present-day orbits. In this follow-up investigation, we present more comprehensive simulations incorporating superior particle resolution, longer integration times, and eliminating our prior means of artificially forcing instabilities to occur at specified times by shifting a planets' position in its orbit. While we find that the residual phase of planetary migration only minimally alters the planets' ultimate eccentricities, our work uncovers several intriguing outcomes in realizations where Jupiter and Saturn are born with extremely large eccentricities (e(J) similar or equal to 0.10; e(S) similar or equal to 0.25). In successful simulations, the planets' orbits damp through interactions with the planetesimal disk prior to the instability, thus loosely replicating the initial conditions considered in our previous work. Our results therefore suggest an even wider range of plausible evolutionary pathways are capable of replicating Jupiter and Saturn's modern orbital architecture.

In spite of substantial advancements in simulating planet formation, the planet Mercury's diminutive mass and isolated orbit and the absence of planets with shorter orbital periods in the solar system continue to befuddle numerical accretion models. Recent studies have shown that if massive embryos (or even giant planet cores) formed early in the innermost parts of the Sun's gaseous disk, they would have migrated outward. This migration may have reshaped the surface density profile of terrestrial planet-forming material and generated conditions favorable to the formation of Mercury-like planets. Here we continue to develop this model with an updated suite of numerical simulations. We favor a scenario where Earth's and Venus's progenitor nuclei form closer to the Sun and subsequently sculpt the Mercury-forming region by migrating toward their modern orbits. This rapid formation of similar to 0.5 M (circle plus) cores at similar to 0.1-0.5 au is consistent with modern high-resolution simulations of planetesimal accretion. In successful realizations, Earth and Venus accrete mostly dry, enstatite chondrite-like material as they migrate, thus providing a simple explanation for the masses of all four terrestrial planets, the inferred isotopic differences between Earth and Mars, and Mercury's isolated orbit. Furthermore, our models predict that Venus's composition should be similar to the Earth's and possibly derived from a larger fraction of dry material. Conversely, Mercury analogs in our simulations attain a range of final compositions.