Hemispherical photography simulations with an architectural model to assess retrieval of leaf area index

Three-dimensional simulations of forest scenes rendered with the freeware POV-Ray (www.povray.org) were used to assess retrievals of plant and leaf area index (PAI and LAI) from hemispherical photography (HP). Ten HPs were simulated in 75 scenes with random, erectophile and planophile leaf angle distributions. PAI retrieved using Beer's law near 57.3° (55-60°) and with the truncated Miller integration over zenith angles from 10° to 65° gave similar results. The importance of clumping was also shown to obtain accurate PAI. Foliage density inside the crown was a major factor causing underestimates of the PAI because the clumping effect was underestimated. Moreover, best PAI retrievals were found with the clumping index method using the Chen and Cihlar method combined with the Lang and Xiang method (relative RMSE of 23-24% and relative median absolute error of 12-14%) and was slightly less affected by the averaging length when compared with only using the Lang and Xiang. Previous error measurements of 29% were reduced to 16-23% for a black spruce stand using updated input variables and the method presented here, and they were shown to fall within the errors of the simulation framework.