Several mitigation strategies have been proposed to deflect a NEO from its route to the Earth. The efficiency of a given strategy requires a specific knowledge on the physical properties of the NEO to be deflected. For instance, a strategy involving a kinetic impactor needs some knowledge of the surface and internal properties: the momentum transfer efficiency depends on the amount of ejecta generated by the impactor, which in turn depends on the surface conditions at the impact point, the material properties and in some extent the internal properties of the object. A gravity tractor requires an estimate of the asteroid's size, shape and more importantly its mass. The knowledge of thermal and spin properties is also important to better assess the Yarkovsky effect that can play a critical role for the accurate prediction of the impact of a small NEO.

I will review our current knowledge on the physical properties of NEOs obtained by ground-based and space-based observations, as well as numerical modeling, that are relevant for mitigation strategy designs. Ground-based telescopic and radar observations provide information essentially on the magnitude, spin period, shape as well as albedo (size) and thermal inertia when they are made in the thermal infrared. Spectral observations tell us something about the composition of the exposed few micrometers of the surface, although their interpretation can in some cases be ambiguous. These observations provided a statistical knowledge of NEO properties that is the minimum information required for designing mitigation strategies and that will become even more robust thanks to on-going and future programs (e.g. Pan-Starrs, LSST). Space missions are obviously the best method to obtain accurate information on an object, especially when they are dedicated to a NEO rendezvous or sample return. They allow us to obtain information on its internal properties by providing an estimate of its mass, volume and consequently its bulk density, and detailed information on surface properties and composition. Such knowledge is crucial for mitigation strategies, such as the ones based on a kinetic impactor or a gravity tractor. Unfortunately, because they are very limited in number, they cannot give us representative information for the whole population, but they are crucial to provide ground-truth for ground-based observations and check our expectations from numerical models. For instance numerical models allowed us to identify the main source regions of NEOs, and the possible outcomes of catastrophic disruptions. They suggest that most NEOs should be rubble piles or at least heavily shattered objects. This is consistent with the low bulk density generally estimated by space missions, such as the NASA mission NEAR that provided the bulk density of the 23 km-size NEO Eros (about 2.7 g/cc), and the JAXA mission Hayabusa that returned a sample from the 350m-size NEO Itokawa and estimated its bulk density (about 2 g/cc). Moreover, bulk densities of dark-type bodies seem to be systematically lower (e.g. about 1.3 g/cc for the C-type asteroid Mathilde as estimated by NEAR) than those of S-types (> 2 g/cc), suggesting that dark-types are highly porous. Such a high porosity greatly influences the momentum transfer from a kinetic impactor; numerical simulations of impact, as well as laboratory experiments, suggest that porosity decreases the amount of ejecta (and therefore the momentum transfer efficiency) resulting from a cratering impact.

Given the wide diversity of NEO physical properties, it is important to improve our knowledge on these properties and to have a good understanding of which are the most relevant and their influence on a given mitigation method. This is especially important for the most threatening objects, so that the most appropriate choice and design can be decided in case of necessity. Meantime, it would be important to test our ability to deflect an object by one or more methods.

The Wide-Field Infrared Survey Explorer (WISE) has imaged the entire sky two times between January, 2010 and January, 2011 at four wavelengths spanning the near through mid-IR at sensitivities hundreds of times greater than previous surveys [1]. The WISE band-passes (3.4, 4.6, 12 and 22mm) sample the flux from most inner-solar-system bodies near the peak of their thermal emission. Overlapping sky regions were sampled repeatedly at 3 hour intervals. The same region of sky was observed a minimum of 8 times.

While the primary WISE science objectives focus on astrophysical topics, additions to the baseline WISE pipeline (collectively known as "NEOWISE") have enabled the detection and catagorization of formerly undiscovered moving objects, as well as previously known bodies [2]. As of January, 2011, NEOWISE has detected more than 156,000 minor planets, including more than 500 near-Earth objects (NEOs), ~2000 Jupiter Trojans, 120 comets, and ~20 outer Solar System objects such as Centaurs. The survey has discovered ~34,000 new minor planets, including 133 new NEOs and 20 new comets. NEOWISE was the most prolific observer of minor planets in 2010, having detected nearly 100x more objects than its predecessor, the Infrared Astronomical Satellite (IRAS) [3]. The NEOWISE data will drive a panoply of new small body discoveries and investigations.

NEOWISE allows determination of IR-derived diameters and albedos for minor planets throughout the Solar System, with increased sensitivity for the darkest members of these populations. By using infrared observations to determine diameters and albedos (which serve as a proxy for composition and density), a more precise calculation of mass and impact hazard can be made. The survey also provides measurements of:

Radar is an extremely powerful astronomical technique for characterizing near-Earth objects and for refining their orbits. The Arecibo and Goldstone radars can image near-Earth asteroids (NEAs) with resolutions as fine as several meters, which greatly exceeds the finest resolution available from any ground- or space-based optical telescope. Radar images reveal an object's size, shape, spin state, and features on its surface such as craters, valleys, and even large boulders. Among NEAs larger than 200 m in diameter, radar imaging has revealed that ~16% are double and triple systems, that ~10% are contact binaries, and that oblate shapes with pronounced equatorial ridges are relatively common. Two-thirds of binary and ternary systems have been discovered by radar. Radar can determine the masses of binary NEAs and, in some cases, solitary NEAs through detection of the Yarkovsky effect. Multiple opportunities for high-resolution radar imaging occur annually that yield images that are exceeded in resolution only by spacecraft.

Radar echoes from NEAs have revealed stony and metallic objects, featureless spheroids, shapes that are elongated and irregular, monolithic objects, unconsolidated rubble piles, rotation periods ranging from a few minutes to several weeks, objects whose rotation is accelerating, non-principal axis rotators, and surfaces that are extraordinarily rough.

Radar is invaluable for refining orbits of potentially hazardous NEAs. Range-Doppler measurements provide astrometry with precision as fine as 4 m in range and 1 mm/s in velocity, with a fractional precision that is orders of magnitude finer than with optical measurements. Radar astrometry can add centuries to the interval over which we can predict close Earth approaches and dramatically refines collision probability estimates based on optical astrometry alone.

A sequence of radar images can be inverted to estimate an asteroid's 3D shape and spin state, which constrains its formation and evolution and which opens up a number of important geophysical investigations into its dynamical environment, which can be very useful for spacecraft planning and navigation. Radar observations have supported multiple spacecraft missions to asteroids and comets, such as NEAR-Shoemaker, Hayabusa, Rosetta, EPOXI, Dawn, and Clementine; contributed to past mission proposals such as OSIRIS, Galahad, and Marco Polo, and to multiple proposals in current NASA and ESA competitions.

Several significant developments have occurred since the 2009 Planetary Defense Conference. Arecibo replaced its generators and klystrons and has resumed operating at 900 kW for the first time in several years. In 2010 NASA began funding Arecibo, and as a result, the number of asteroids scheduled for radar observations has approximately tripled. Goldstone's imaging resolution has improved by a factor of five to 3.75 m, which is twice as fine as the highest resolution at Arecibo. This provides the ability to see considerably more detail on asteroid surfaces, to spatially resolve NEAs as small as ~20 meters, and it greatly increases the precision of radar ranging astrometry, which has significant ramifications for long-term orbit prediction of sub-100-meter-diameter NEAs.

We analyse the orbital distribution of near-Earth objects (NEOs) using present-day data for more than seven thousand discovered objects. The distribution of large NEOs with absolute magnitude H<18 is in agreement with the earlier predictions (Bottke, Morbidelli, Jedicke, et al., 2002; Stuart, Binzel, 2004), although a maximum found in the distribution of perihelion distances q near q=0.5 AU was not mentioned earlier. The investigation of all observed NEOs reveals new features in the orbital distribution of these objects. In particular, neither arguments of perihelion nor longitudes of perihelion are uniformly distributed. The Kolmogorov-Smirnov statistical test confirms this with a probability exceeding 0.9999. A dynamical explanation of these features associated with secular perturbations is given by using the numerical integration of the orbital evolution of selected NEOs. The improved description of the NEO distribution imposes additional constraints on space and ground-based strategies to find and study the remaining undiscovered NEOs. For this purpose, we create maps describing where in the sky NEOs can be found and comment on their applications for searches of potentially hazardous objects. It is shown that variations of the NEO distribution density function with heliocentric ecliptic longitudes are almost twenty percent of the mean value. We stress large uncertainties in the number and distribution of small (H >>18) objects. In this range of sizes low-albedo, asteroid-like remnants of comets can give a large contribution to the terrestrial impact hazard. The solution to this problem must lie in more observations and a deeper understanding of the physical evolution of comets. This work was supported by the Federal Targeted Programme "Scientific and Educational Human Resources of Innovation-Driven Russia" for 2009-2013.

Predictions of possible impacts are limited in time by a predictability horizon. This results from two effects: 1. chaos introduced by close approaches to major planets, implying exponential divergence of orbits with nearby initial conditions; 2. imperfect knowledge of the dynamic model due to non-gravitational perturbations, implying quadratic divergence of orbits computed with different models.

The relative importance of these two effects changes with time, with non-gravitational effects dominating in the short run and chaos in the long run. This transition is controlled by the presence of intermediate very close approaches in the time span between the observations and the impact. This very complex interaction of different effects has only recently begun to be studied, and in very few cases, including (99942) Apophis (see Chesley, this meeting). For asteroid (101955) 1999 RQ36 (Milani et al. Icarus, vol. 203, 2009) it has been shown that impact possibilites can be identified as far in time as the year 2182, and that the impact probabilities can be significant (of the order of 1 in 1000) because of refocusing effects of resonant returns.

(101955) is again observable optically from August 2011 and may be a radar target in September. If radar observations are succesful, the knowledge of the Yarkovsky non-gravitational effect can be improved by more than an order of magnitude. Then it would become possible to greatly restrict the orbital possibilities for this object, with a very significant change in the impact predictions.

To generalize the kind of analysis performed on (101955) to most Potentially Hazardous Asteroids with very well known orbit, with the goal of extending the predictability horizon beyond the year 2100, would be extremely valuable: this could one day allow the identification of the next real impactor on our planet. However, there are several stumbling blocks preventing this upgrade of the current Impact Monitoring computational engines. We discuss the development path we are following with this long term goal.

Asteroid impact hazard monitoring systems have so far focused primarily on warning of potential impact threats within the next century. This 100-year search horizon is generally appropriate because mitigation efforts over longer time horizons are often precluded because of the difficulty of establishing high impact probabilities for longer term long term predictions. Therefore the time for deflection efforts is generally much less than several decades prior to impact. Moreover, searching farther into the future for potential impacts is problematic, first because the fractal nature of keyholes means that most recent asteroid discoveries harbor a cascade of many very low probability events beyond several decades in the future, leading to a significant increase in computational cost for longer searches. But more importantly, the standard theories used in the current impact monitoring systems become inadequate for longer search intervals.

Well-observed asteroids with high-precision orbits represent a different class of object for impact monitoring. In these cases the future trajectory may be well known for a long period, a century or more, until the object has a close planetary encounter that scatters the range of possible trajectories, thereby injecting a large amount of uncertainty. At that future point the problem can become similar to that of a newly discovered asteroid, with a wide range of future trajectories.

Nongravitational accelerations, namely the Yarkovsky Effect, can also become an important consideration. Assessing the uncertainty due to the Yarkovsky effect can be difficult, especially when the spin state of the object is not known. Even so, reasonable assumptions can be followed to estimate the amount of trajectory spreading due to this effect. In many high precision cases the Yarkovsky spreading remains small relative to the dispersions due to planetary encounters and so it can be neglected. But when the uncertainty due to the Yarkovsky Effect is large enough to significantly alter the circumstances of a planetary encounter it must be incorporated into the impact assessment. We will consider examples that illustrate these properties, such as 99942 Apophis and 101955 (1999 RQ36).

Panoramic Survey Telescope And Rapid Response System (Pan-STARRS) prototype (PS1) is currently in full operation and brings many asteroid discoveries. We present simulation results for the detection efficiency of the telescope in a search for Earth impacting asteroids and discuss the behavior of the impacting population. The synthetic population of impactors derived from Bottke et al. (2002) was processed through the Pan-STARRS Moving Object Processing System (MOPS) which performs detection linking, orbit determination, precovery, attribution and orbit identification. Our simulation emulated a realistic survey cadence, limiting magnitude, field size, weather, and MOPS performance. Our results provide the survey efficiency for finding Earth impacting bodies as a function of their diameter. We also determine the evolution of the minimum orbit intersection distance (MOID) as we obtain more observations and as a function of time to impact. We find that the sky-plane distribution of impactors evolves strongly with time before impact. We confirmed the existence and position of the "sweet spot" regions for identifying potentially hazardous objects at small solar elongations. Using current estimates of the size-impact-frequency distributions of asteroids and fireballs we investigated if PS1 will discover small impactors before they enter the Earth's atmosphere. There is about a 20% chance that PS1 will find a 1m size meteoroid before it enters the atmosphere during a 4 year survey.

If a threatening Near-Earth Object is discovered to be on a collision trajectory, there is a good chance that it will make one or more close approaches to the Earth during the century before impact. The possible 2036 impacting trajectory for Apophis is a good example: it makes a very close approach to Earth seven years earlier, when it passes through a 600m-wide keyhole in the 2029 target plane. In general, a keyhole is a narrow slice of the uncertainty ellipse in the target plane at a position where the encounter perturbs the object onto a trajectory which impacts at a later time. Passage through a keyhole implies impact, and conversely, deflection away from the keyhole implies avoidance of the impact. Since keyholes are typically narrow, it is easier to deflect the object before keyhole passage rather than after; the keyhole thus provides deflection leverage. Because the 2029 Apophis encounter is so close, the 2036 keyhole provides a huge leverage factor of over 10,000. But what is the chance that a random impactor will have keyhole passages with such large deflection leverage? We show statistically that these large leverage scenarios are extremely unlikely. In fact, the chance that an impactor will have a keyhole passage providing the leverage factor of 10 or larger within 25 years of impact is only about 8 percent. If we require a leverage factor of 100 or larger within 25 years of impact, the chances shrink to 3 percent. In our study, we have also looked at the difficult converse problem of dealing with impactors with large and nasty "jabba" keyholes. These have the undesirable property of subverting deflection attempts by reducing the effect of prior deflections, not amplifying them. For these impactors, it may be easier to wait until after passing through the keyhole before deflecting the object, rather than deflecting ahead of the keyhole passage.