Threshold air pollutant

Consider a city of one million people trying to cope with a pollution problem. The pollutant is produced by dozens of sources, each of which has a different marginal cost of abatement. Moreover, for most of the sources the marginal cost of abatement is unknown. Public health studies suggest that when more than 100 tons of the pollutant are emitted, 5 people will die from diseases caused by it. However, when less than 100 tons are emitted, the pollutant causes no health problems. At all levels of emission the pollutant causes no problems other than its effects on health. At the moment, 150 tons of pollution are being produced. In addition, an economist has estimated that individuals in the city are willing to pay $1 for each 1/1,000,000 reduction in the risk of death.

1. Propose a specific policy to deal with this problem. By "specific" I mean that you must indicate the specific numerical value of any key part of the policy. Explain in detail why you chose the policy you did.
2. Now suppose that the legislature put the policy you recommended above into place and a couple of years have passed. The firms have had a chance to react to the policy. Suppose that you, as the economist, are still unable to observe the marginal abatement cost curves of the individual firms. (That is, you can't get a MC of abatement curve for each of the firms, either because there are too many firms or because the information is a trade secret.) How would you decide whether the policy was actually efficient? Explain.