Aboveground Forest Biomass Estimation with Landsat and LiDAR Data and Uncertainty Analysis of the Estimates

Table 1

A summary of regression analysis results in the western Brazilian Amazon.

Variables	Models	𝑅^2	Beta value
Spectral signature	AGBSF=667.72−13.92×𝑓𝑏4	0.75
	AGBMF=1024.14−54.96×𝑓𝑏5	0.16
Texture	AGBSF=164.62−2.27×𝑓var	0.23
	AGBMF=134.57+19.29×𝑓con	0.39
Combination	AGBSF=480.82−8.06×𝑓𝑏4−0.98×𝑓var	0.76	−0.71 (sp), −0.24 (txt)
	AGBMF=753.31−43.21×𝑓𝑏5+17.89×𝑓con	0.50	−0.31 (sp), 0.65 (txt)

(1) AGB_SF and AGB_MF represent biomass estimates for successional forest and mature forest.
(2) 𝑓_𝑏4 and 𝑓_𝑏5 represent reflectance values of TM bands 4 and 5.
(3) 𝑓_var and 𝑓_con represent the textural image developed with the variance texture measure with TM band 4 and a window size of 15 × 15 pixels, and the textural image developed with the contrast texture measure with TM band 5 and a window size of 19 × 19 pixels.
(4) sp and txt represent the variables of spectral signatures and textures.
(5)  𝑅² is the coefficient of determination.