Table 4. Validation accuracy assessment of the biophysical forest attributes (BFA) estimates.
S BFA NEW approach REUSE approach FORECAST approach
MPD
 
RMSE RMSE% λ0 λ1 σεem MPD RMSE RMSE% λ0 λ1 σεem MPD RMSE RMSE% λ0 λ1 σεem
S1 Ho 0.5ns 0.9 5 –1.1 1.03 0.8 0.4ns 0.8 4 4.6 0.73 0.7 1.4 1.5 8 4.4a 0.68b 0.4
HL 0.9 1.0 6 1.6 0.84 0.4 0.5 0.7 4 1.0 0.90 0.4 1.7 1.8 12 2.8 0.71 0.5
N –439.3 471.4 38 431.8 1.01 179.2 –769.3 793.6 63 851.3a 0.93 203.6 –865.6 913.6 73 389.5 1.38 281.8
G –2.0 2.9 11 5.0 0.88 2.2 –1.4ns 3.9 15 1.7 0.99 3.9 –6.2 6.9 27 7.7 0.94 3.1
V –10.8ns 26.4 14 1.3 1.05 25.2 15.6ns 36.1 19 17.3 0.83 33.3 –38.3 49.0 25 42.8 0.98 32.0
AGB –6.7ns 14.7 12 6.3 1.00 13.7 17.4 23.7 19 –1.6 0.87 16.5 –17.5 21.7 17 20.0 0.98 13.4
S2 Ho 0.3ns 0.8 4 3.1a 0.81b 0.5 –0.1ns 1.0 5 4.7a 0.75b 0.6 –0.1ns 1.4 8 7.6a 0.58b 0.6
HL –0.2ns 0.7 4 2.2 0.87 0.6 –0.4 0.7 4 2.3a 0.88 0.5 –0.1ns 0.6 4 3.0a 0.82b 0.4
N 7.2ns 116.6 18 121.8 0.80b 101.8 76.1ns 164.4 26 150.1a 0.64b 99.4 50.2ns 171.9 27 254.1a 0.52b 74.8
G –1.4ns 3.9 17 4.3 0.87 3.6 1.4ns 3.6 16 5.8a 0.69b 1.7 –2.4 4.2 19 8.5a 0.73b 2.6
V –19.9 32.6 18 38.1a 0.90 25.1 25.9 40.5 23 29.7a 0.69b 14.8 –20.4 33.3 19 53.2a 0.82b 21.4
AGB –12.8 23.2 21 22.5 0.91 19.4 25.5 33.4 31 12.6a 0.65b 9.0 –16.2 22.9 21 21.6 0.95 16.5
S3 Ho 0.1ns 0.8 4 3.3a 0.85b 0.7 –1.7 1.9 8 7.6a 0.74b 0.6 0.9 1.2 5 2.7a 0.84b 0.6
HL 0.7 0.9 5 2.3a 0.86b 0.5 0.0ns 0.8 4 4.5a 0.78b 0.4 0.9 1.2 6 –0.1 0.96 0.7
N –52.0ns 171.1 22 158.6 0.86 163.2 39.1ns 163.2 21 362.2a 0.47b 104.5 –184.5 261.4 34 329.2a 0.81 183.9
G –1.7ns 4.5 15 4.6 0.91 4.2 4.3 6.3 20 3.8 0.73 4.2 –5.5 7.3 24 8.8 0.89 4.9
V –13.5ns 38.5 13 12.5 1.00 36.8 33.0 56.2 19 57.5a 0.70b 36.3 –42.6 73.5 25 35.0 1.03 61.1
AGB –5.0ns 21.3 13 20.5 0.91 20.6 20.9 30.9 18 23.6 0.74b 18.9 –24.5 41.0 24 42.2 0.90 33.1
S = Stratum. S1 = Young forest, S2 = Mature forest with Site index ≤ 11 m, S3 = Mature forest with Site index > 11 m. Biophysical forest attributes; Ho = Dominant height (m), HL = Lorey’s mean height (m), N = Stem number (ha–1), G = Basal area (m2 ha–1), V = Volume (m3 ha1), AGB = Aboveground biomass (Mg ha–1). MPD = Mean predictor difference.­ RMSE = Root mean square error. RMSE% = relative RMSE. λ0, λ1, σεem = parameters from the error model equation. ns = Not statistically significant different from zero (p-value > 0.05). a rejected null hypothesis λ0 = 0. b rejected null hypothesis λ1 = 1 (based on t-tests Persson and Ståhl (2020))