We explore the initial conditions for fragments in the extended regions (r 50 AU) of gravitationally unstable disks. We combine analytic estimates for the fragmentation of spiral arms with 3D SPH simulations to show that initial fragment masses are in the gas giant regime. These initial fragments will have substantial angular momentum, and should form disks with radii of a few AU.We show that clumps will survive for multiple orbits before they undergo a second, rapid collapse due to H 2 dissociation and that it is possible to destroy bound clumps by transporting them into the inner disk. The consequences of disrupted clumps for planet formation, dust processing, and disk evolution are discussed. We argue that it is possible to produce Earth-mass cores in the outer disk during the earliest phases of disk evolution.

I argue for two modes of gas giant planet formation and discuss the conditions under which each mode operates. Gas giant planets at disk radii r > 100 AU are likely to form in situ by disk instability, while core accretion plus gas capture remains the dominant formation mechanism for r < 100 AU. During the mass accretion phase, mass loading can push disks toward fragmentation conditions at large r. Massive, extended disks can fragment into clumps of a few to tens of Jupiter masses. This is confirmed by radiation hydrodynamics simulations. The two modes of gas giant formation should lead to a bimodal distribution of gas giant semi-major axes. Because core accretion is expected to be less efficient in low-metallicity systems, the ratio of gas giants at large r to planets at small r should increase with decreasing metallicity.

This paper presents a fully three-dimensional radiative hydrodymanics simulation with realistic opacities for a gravitationally unstable 0.07 M disk around a 0.5 M star. We address the following aspects of disk evolution: the strength of gravitational instabilities under realistic cooling, mass transport in the disk that arises from GIs, comparisons between the gravitational and Reynolds stresses measured in the disk and those expected in an -disk, and comparisons between the SED derived for the disk and SEDs derived from observationally determined parameters. The mass transport in this disk is dominated by global modes, and the cooling times are too long to permit fragmentation for all radii. Moreover, our results suggest a plausible explanation for the FU Ori outburst phenomenon.

The dynamical evolution of planetary systems leaves observable signatures in debris disks. Optical images trace micron-sized grains, which are strongly affected by stellar radiation and need not coincide with their parent body population. Observations of mm-size grains accurately trace parent bodies, but previous images lack the resolution and sensitivity needed to characterize the ring's morphology.Here we present ALMA 350 GHz observations of the Fomalhaut debris ring. These observations demonstrate that the parent body population is 13-19 AU wide with a sharp inner and outer boundary. We discuss three possible origins for the ring, and suggest that debris confined by shepherd planets is the most consistent with the ring's morphology.

The large number of detected giant exoplanets offers the opportunity to improve our understanding of the formation mechanism, evolution, and interior structure of gas giant planets. The two main models for giant planet formation are core accretion and disk instability. There are substantial differences between these formation models, including formation timescale, favorable formation location, ideal disk properties for planetary formation, early evolution, planetary composition, etc. First, we summarize the two models including their substantial differences, advantages, and disadvantages, and suggest how theoretical models should be connected to available (and future) data. We next summarize current knowledge of the internal structures of solar- and extrasolar- giant planets. Finally, we suggest the next steps to be taken in giant planet exploration.Comment: Accepted for publication as a chapter in Protostars and Planets VI, to be published in 2014 by University of Arizona Pres