Performance of Four Optical Methods in Estimating Leaf Area Index at Elementary Sampling Unit of <i>Larix principis-rupprechtii</i> Forests

• Main Authors: Jie Zou, Yong Zuo, Peihong Zhong, Wei Hou, Peng Leng, Bin Chen
• Format: Article
• Language: English
• Published: MDPI AG 2019-12-01
• Series: Forests
• Subjects: leaf area index; clumping effects; inversion model; woody components correction method; <i>larix</i>-dominated forest plots; optical method; elementary sampling unit
• Online Access: https://www.mdpi.com/1999-4907/11/1/30

Optical methods are frequently used as a routine method to obtain the elementary sampling unit (ESU) leaf area index (LAI) of forests. However, few studies have attempted to evaluate whether the ESU LAI obtained from optical methods matches the accuracy required by the LAI map product validation community. In this study, four commonly used optical methods, including digital hemispherical photography (DHP), digital cover photography (DCP), tracing radiation of canopy and architecture (TRAC) and multispectral canopy imager (MCI), were adopted to estimate the ESU (25 m &#215; 25 m) LAI of five <i>Larix principis-rupprechtii</i> forests with contrasting structural characteristics. The impacts of three factors, namely, inversion model, canopy element or woody components clumping index (<inline-formula> <math display="inline"> <semantics> <mrow> <msub> <mi>&#937;</mi> <mi>e</mi> </msub> </mrow> </semantics> </math> </inline-formula> or <inline-formula> <math display="inline"> <semantics> <mrow> <msub> <mi>&#937;</mi> <mi>w</mi> </msub> </mrow> </semantics> </math> </inline-formula>) algorithm, and the woody components correction method, on the ESU LAI estimation of the four optical methods were analyzed. Then, the LAI derived from the four optical methods was evaluated using the LAI obtained from litter collection measurements. Results show that the performance of the four optical methods in estimating the ESU LAI of the five forests was largely affected by the three factors. The accuracy of the LAI obtained from the DHP and MCI strongly relied on the inversion model, the <inline-formula> <math display="inline"> <semantics> <mrow> <msub> <mi>&#937;</mi> <mi>e</mi> </msub> </mrow> </semantics> </math> </inline-formula> or <inline-formula> <math display="inline"> <semantics> <mrow> <msub> <mi>&#937;</mi> <mi>w</mi> </msub> </mrow> </semantics> </math> </inline-formula> algorithm, and the woody components correction method adopted in the estimation. Then the best <inline-formula> <math display="inline"> <semantics> <mrow> <msub> <mi>&#937;</mi> <mi>e</mi> </msub> </mrow> </semantics> </math> </inline-formula> or <inline-formula> <math display="inline"> <semantics> <mrow> <msub> <mi>&#937;</mi> <mi>w</mi> </msub> </mrow> </semantics> </math> </inline-formula> algorithm, inversion model and woody components correction method to be used to obtain the ESU LAI of <i>L. principis-rupprechtii</i> forests with the smallest root mean square error (RMSE) and mean absolute error (MAE) were identified. Amongst the three typical woody components correction methods evaluated in this study, the woody-to-total area ratio obtained from the destructive measurements is the most effective method for DHP to derive the ESU LAI with the smallest RMSE and MAE. In contrast, using the woody area index obtained from the leaf-off DHP or DCP images as the woody components correction method would result in a large LAI underestimation. TRAC and MCI outperformed DHP and DCP in the ESU LAI estimation of the five forests, with the smallest RMSE and MAE. All the optical methods, except DCP, are qualified to obtain the ESU LAI of <i>L. principis-rupprechtii</i> forests with an MAE of &lt;20% that is required by the global climate observation system. None of the optical methods, except TRAC, show the potential to obtain the ESU LAI of <i>L. principis-rupprechtii</i> forests with an MAE of &lt;5%.