Toward a mathematical theory of environmental monitoring: the infrequent sampling problem

Optimal monitoring of pollutants in diffusive environmental media was studied in the contexts of the subproblems of the optimal design and management of environmental monitors for bounds on maximum allowable errors in the estimate of the monitor state or output variables. Concise problem statements were made. Continuous-time finite-dimensional normal mode models for distributed stochastic diffusive pollutant transport were developed. The resultant set of state equations was discretized in time for implementation in the Kalman Filter in the problem of optimal state estimation. The main results of this thesis concern the special class of optimal monitoring problem called the infrequent sampling problem. Extensions to systems including pollutant scavenging and systems with emission or radiation boundary conditions were made. (U.S.)