Above-ground biomass estimation of tuberous bulrush ( Bolboschoenus planiculmis) in mudflats using remotely sensed multispectral image.

We present a multivariate regression approach for mapping the spatial distribution of above-ground biomass (AGB) of B. planiculmis using field data and coincident moderate spatial resolution satellite imagery. A total of 232 ground sample plots were used to estimate the biomass distribution in the Nakdong River estuary. Field data were overlain and correlated with digital values from an atmospherically corrected multispectral image (Landsat 8). The AGB distribution was derived using empirical models trained with field-measured AGB data. The final regression model for AGB estimation was composed using the OLI3, OLI4, and OLI7 spectral bands. The Pearson correlation between the observed and predicted biomass was significant (R = 0.84, p < 0.0001). OLI3 made the largest contribution to the final model (relative coefficient value: 53.4%) and revealed a negative relationship with the AGB biomass. The total distribution area of B. planiculmis was 1,922,979 m. Based on the model estimation, the total AGB had a dry weight (DW) of approximately 298.2 tons. The distribution of high biomass stands (> 200 kg DW/900 m) constituted approximately 23.91% of the total vegetated area. Our findings suggest the expandability of remotely sensed products to understand the distribution pattern of estuarine plant productivity at the landscape level. [ABSTRACT FROM AUTHOR]