Modifying geographically weighted regression for estimating aboveground biomass in tropical rainforests by multispectral remote sensing data

► Aboveground biomass in a tropical rainforest was estimated using satellite data. ► A local regression approach was used to capture relationships nonstationarity. ► Altitudinal effects were incorporated into geographically weighted regression (GWR). ► A new geographically and altitudinally weighted regression (GAWR) was introduced. ► GAWR provided more accurate estimations of aboveground biomass than GWR.