In the Jalapão region, State of Tocantins, Brazil, there are large contiguous areas of Neossolos Quartzarênicos (Tropsamments). Golden grass (Syngonanthus nitens) is a native species widely used as material for handicrafts in the region, and it is therefore of great importance for the local economy. The aim of this study was to analyze the relationship between soil physical and chemical properties and golden grass yield in Campos Limpos Úmidos within Jalapão State Park (Parque Estadual do Jalapão). We collected soil samples and golden grass scapes and made slope measurements along the transects perpendicular to the drainage lines. We identified three soil units: Neossolo Quartzarênico órtico típico, Neossolo Quartzarênico hidromórfico típico, and Organossolo Háplico fíbrico típico (Orthpsamment, Hydropsamment, and Hydrofibrist, respectively). Multiple regression models involving 26 variables further portrayed golden grass yield. Six of these variables subsequently proved to be of greatest significance and constituted the final model: pH, Mg2+, Ca2+, fine sand, clay, and soil moisture. The multiple quadratic regression model involving these six variables considered an R2 of 0.85 (adjusted R2of 0.81) of the relationship evaluated. Soil moisture was the most important variable.
Syngonanthus nitens; multivariate regression; sandy soils
