Landsat Data Based Prediction of Loblolly Pine Plantation Attributes in Western Gulf Region, USA

The suitability of using Landsat sensor variables to predict key stand attributes, including average dominant/codominant tree height (HT), mean diameter at breast (DBH), the number trees per hectare (NT), basal area (BA), and density index (SDI), intensively managed loblolly pine plantations in Western Gulf Region plot/stand level was assessed. In total, thirty six original bands, three vegetation indices, Tasseled Cap transformed eighteen texture measure were used as predictors. Field data 125 permanent plots located across east Texas western Louisiana reference data. Individual those measured plot establishment (referred first cycle measurement; about 4.5 years old) remeasured three-year intervals (the second measurement approximately seven old third 10 old). Thus, field represent development from open- (first cycle) closed-canopy (third cycle). Models HT, DBH, NT, BA, SDI developed by multiple linear regression (MLR) also random forests (RF) methods. Results indicated that stands well predicted with R2 > 0.7 low RMSEs. These relationships weakened age, although still moderate being around 0.45 for became practically useless (R2 < 0.30) measurement. For no meaningful models achieved regardless cycle. MLR RF comparable accuracy had similar Overall, shortwave infrared red band, wetness most important predictors, but their dominance declined Texture relatively less a trend increasing importance noted. show promise operationally predicting young plantations, an age class typically presents significant challenges conventional forest methodologies. Potential methods further improve model how use results within context plantation management planning region discussed.