| Title | Detecting and adjusting for preferential monitoring site location: A case study of air pollution with novel R packages |
| Publication Type | Manuscript |
| Year of Publication | Submitted |
| Collection Title | Evironmetrics |
| Abstract | The paper concerns preferential siting, i.e. the sampling of sites in networks of monitors designed to measure the levels of atmospheric air pollution over a specified geographical region. More precisely, a case study demonstrates the use of an R software package that detects such preferentiality in the Southern California Air Quality Basin, i.e. SOCAB, in the siting of monitors very fine particulate matter, i.e. PM10 monitors. It then demonstrates newly published R software that simultaneously maps, over time and space, both the P M10 field and the binary selection process. It thus calculates the impact of such preferential sampling on estimates of the annual level of PM10. Furthermore, that new software enables the spatial prediction of those levels at unmonitored locations in California. P M10 was chosen for the case study because of its harmful impact on human health and because it has been monitored over a substantial period. The SOCAB was selected, because of its historical importance in the develop- ment of the US Air Quality Standards (AQS). The paper thus demonstrates the value of the two software packages while showing the impact of preferential sampling on estimates of PM10 over an important geographical domain. |
