Estimation of Spatial Distributions from Point Sources with Application to Air Pollution Measurement. Technical Report No. 9.

摘要

A class of weighted estimators of a stationary distribution function is introduced which exploits correlation structure in the data as frequently found in air quality measurements. The nature of these weighted estimates is examined for several simple cases,in particular for spatially clustered measurements where the advantage over the usual empirical distribution function seems most pronounced. An illustration is presented using total suspended particulate concentrations from a 14-station monitoring network. The class of weighted estimators is also proposed for the related problem of estimating the actual realization of a spatial distribution function separately at each fixed time period. (ERA citation 03:025016)