Estimating biomass for boreal forests using ASTER satellite data combined with standwise forest inventory data

Abstract

In the present study, the suitability of optical ASTER satellite data (with 9 spectral bands) for estimating the biomass of boreal forest stands in mineral soils was tested. The remote sensing data were analysed and tested together with standwise forest inventory data. Stand volume estimates were converted to aboveground tree biomass using biomass expansion factors, and the aboveground biomass of understory vegetation was predicted according to the stand age. Non-linear regression analysis and neural networks were applied to develop models for predicting biomass according to standwise ASTER reflectance. All ASTER bands appeared to be sensitive to tree biomass, in particular the green band 1. The relative estimation errors (RMSE r) of the total aboveground biomass of the forest stands were 44.7% and 41.0% using multiple regression analysis and neural networks, respectively. Although the estimation errors remained large, the predictions were relatively accurate in comparison to previous studies. Furthermore, the predictions obtained here were significantly close to the municipality-level mean values provided by the National Forest Inventory of Finland.