Riqueza de especies arbóreas en bosques de montaña de Chiapas: estimaciones a partir de datos de herbarios e inventarios florísticos

Estimates of tree species richness (S) have usually been approached with systematic surveys. More recently, with the increasing on-line availability of digital herbaria data, alternative S estimates are possible. How comparable are these two estimates? We fitted generalized linear models (GLM) to obtain estimates of S with both data types from three different mountain regions in Chiapas (>1,500 elevation) for which substantial numbers of both herbarium vouchers and survey plots are available. A total number of 792 tree species were detected with both types of data. Clench's model estimated 862 and 425 tree species with herbarium and survey data, respectively. The selected GLM models rendered S estimates that were positively correlated in two out of the three studied regions (not in the most species-rich region). Systematic surveys provide a snap-shot of species equitability among real samples of forest communities, and therefore, represent a useful set of data to approach the understanding of tree diversity structure. On the other hand, herbarium vouchers allow estimates of how many rare or restricted species may still be absent in the survey-based inventory. It seems convenient to make complementary use of both types of data in order to estimate S in poorly known tropical forests.