As Near-Earth Object (NEO) surveys continue, we improve our knowledge of the population of NEOs and become better able to estimate both the total population versus size and the fraction that remains undiscovered. As of January 19, 2009, the present surveys had discovered 765 Near-Earth Asteroids (NEAs) larger than 1 km in diameter out of an estimated total population of 940, or about 81 % of that population. Since most of the risk resides in the largest impactors, finding that fraction of the largest bodies, and that none of them has a significant chance of impacting in the next century or so, "retires" more than 90 % of the impact risk, including most of the risk of a globally catastrophic impact event. It appears that ground damage extends to considerably smaller impactor sizes than was previously inferred by modeling them as equivalent to nuclear airbursts. This increases the expected frequency of damaging events, although it only modestly increases the "fatality rate", since the smallest events are not very damaging. In the mid-size range, from ~150 m to ~1 km, the main risk is from tsunami generated by an ocean impact. The detailed analyses of the 2003 NASA SDT report estimated a "persons affected" rate of ~182 per year associated with impact tsunami. They did mention that for earthquake-generated tsunami, the actual death rate is typically only 10 % or less of the population in the inundation zone, but did not take full account of that in their risk analysis. Here we re-evaluate the impact hazard, using our new population and completion estimates, and revised "kill curves" including the airburst damage down to smaller size and lower tsunami fatality rate. We estimate that the a priori impact risk (not allowing for any discovered NEAs) is (was) ~40/year for local/regional land impacts, ~6/year from impact tsunami (the dramatic decrease in this number is due to a reduction of a factor of 3 in the estimated population in this size range, times the factor of 10 reduction due to expected actual deaths in the inundation zone), ~1100/year due to globally catastrophic events, and ~10/year from comet nucleus impacts. With the current level of survey completion, the remaining risk from the undiscovered population is ~20/year from local/regional land impacts, ~4/year from impact tsunami, ~54/year from globally catastrophic events, and still ~10/year from comet nuclei. It is noteworthy that even though the risk from globally catastrophic events has been 95 % retired for the short term, it is still the largest component of the remaining risk. Looking to the future, the "next generation survey", aimed to find 90 % of NEAs larger than 140 m diameter, will further reduce the impact risk, using our models of population, completion, and impact damage, to ~6/year from local/regional land impacts, ~0.3/year from impact tsunami, ~11/year from globally catastrophic events, and ~10/year from comet nuclei. In addition to providing long-term (decades) warning of an impact, optical surveys have the capability to spot an impending impactor days or weeks before an impact, if it is coming from a direction being covered by the survey (currently ~35 % of the sky area), providing short term warning long enough for evacuation of affected area or other "civil defense" measures*.

---