South America's research impact is underestimated in the main citation databases, as you suggest (Nature 510, 202–203; 2014). However, incorporating the Scientific Electronic Library Online (SciELO) Index into Thomson Reuter's Science Citation Index will not rectify the situation.
SciELO captures only a fraction of the continent's peer-reviewed publications. For example, it indexes papers from just 267 of Brazil's 1,909 journals. Coverage by Elsevier's Scopus database is also inadequate.
According to Latin America's most comprehensive database on scholarly journals, the Latindex Catalog, 4,882 journals in South America meet specified editorial criteria (see go.nature.com/lbrng2). Scopus includes only 726 of these journals (15%); regional journals are not eligible (see also go.nature.com/laywal).
That leaves 4,156 journals whose impact is hidden from Scopus. If we assume conservatively that each of these publishes only 20 articles per year by South American authors, then at least 83,120 articles are being overlooked annually.
The commercial databases should close this coverage gap to properly reflect the impact of research in South America.