Apportionment of Narragansett

    The steps in the apportionment of Narragansett are summarized in the following table. Links are provided for those who want to want to move quickly through the sequence.

    The apportionment for Narragansett illustrates how we examine the various solutions, or "threads," made available at step 2. Three possibilities presented themselves here: a North American solution, a European solution, and a Japan/Korea solution. Our operating philosophy is to examine each in turn, beginning with the most likely one (the area with the highest F-ratios), and restrict each solution to that particular continent (because aerosols don't more freely from continent to continent the way longer-lived gases such as CO₂ and the Freons do. Thus the solution to a continental receptor site will generally be one continent or another, not a combination of continents. Exceptions are the margins of certain continents and remote locations such as polar regions and the middle of oceans.
    The North American thread is shown in red italic in the table and in the figures below. The European and Japan/Korea threads are shown separately.

    Step 0 for Narragansett, shown below, displays all the F-ratios before any signatures are "chosen" (entered into the fit). Note how the fits for eastern North America (NAm) are markedly superior to those for the Atlantic (Atl), Europe and Russia (EurRus), East Asia (EAs), Japan and Korea (JK), the West Pacific Islands (WPac), the Central Pacific (CPac), the Arctic (Arc), and the type-signatures (Type). The order within each group is the same as in the table of signatures.

    Step 2 is after sea and crust have been entered. (Their color is changed to yellow to indicate that they have been placed into the fit.) The F-ratios of the sites are now responding mostly to the pollution elements in their signatures. Note how West Point (in northeastern North America) and to a lesser extent Prince George (also NE US) stand out above all others. In conjunction with the other sites in the Northeast, they make it the region to begin with. Note also that other groups of signatures, most notably Europe and Russia and to a lesser extent Japan/Korea have clusters of F-ratios approaching those in the Northeast. This is why the fits from these two other regions have to be examined as well.

    Step 3 is after West Point (the highest F-ratio above) has been entered into the fit. The result is shown below. Choosing West Point collapses nearly all the other North American signatures except for Prince George, which remains the highest unchosen one.

    Step 4 is after Prince George has been chosen and entered. No high F-ratios remain. The fit has become very good (SSE = 0.32, down from the original 22), to the point that we could justifiably stop the apportionment here and explain the Narragansett aerosol as a combination of the nearby West Point and Prince George ( in addition to the sea and crust).

    But the signature for Sudbury, Ontario, the rightmost of the North American group, remains high. Since we know from long experience with Narragansett that a nonferrous signal from three Canadian smelters in Ontario and Quebec, including two in Sudbury, is easily detected when winds come from the NW behind a cold front, we try Sudbury as step 5. It is accepted, and improves the fit to 0.26.

    We then explored whether any other of the North American sources could enter the fit. The next-highest, Elms, did, but improved the fit only marginally ( from 0.26 to 0.23). We thus stopped the North America thread there.

    The final North American solution was then West Point, Prince George, Elms, and Sudbury (in addition to crust and sea). The fit was an excellent 0.23.

Ahead to European Solution
Back to Tracer System