Estimating the Seasonal Dynamics of the Leaf Area Index Using Piecewise LAI-VI Relationships Based on Phenophases

Kun Qiao,Wenquan Zhu,Zhiying Xie,Peixian Li

doi:10.3390/rs11060689

Abstract

The leaf area index (LAI) is not only an important parameter used to describe the geometry of vegetation canopy but also a key input variable for ecological models. One of the most commonly used methods for LAI estimation is to establish an empirical relationship between the LAI and the vegetation index (VI). However, the LAI-VI relationships had high seasonal variability, and they differed among phenophases and VIs. In this study, the LAI-VI relationships in different phenophases and for different VIs (i.e., the normalized difference vegetation index (NDVI), enhanced vegetation index (EVI) and near-infrared reflectance of vegetation (NIRv)) were investigated based on 82 site-years of LAI observed data and the Moderate Resolution Imaging Spectroradiometer (MODIS) VI products. Significant LAI-VI relationships were observed during the vegetation growing and declining periods. There were weak LAI-VI relationships (p > 0.05) during the flourishing period. The accuracies for the LAIs estimated with the piecewise LAI-VI relationships based on different phenophases were significantly higher than those estimated based on a single LAI-VI relationship for the entire vegetation active period. The average root mean square error (RMSE) ± standard deviation (SD) value for the LAIs estimated with the piecewise LAI-VI relationships was 0.38 ± 0.13 (based on the NDVI), 0.41 ± 0.13 (based on the EVI) and 0.41 ± 0.14 (based on the NIRv), respectively. In comparison, it was 0.46 ± 0.13 (based on the NDVI), 0.55 ± 0.15 (based on the EVI) and 0.55 ± 0.15 (based on the NIRv) for those estimated with a single LAI-VI relationship. The performance of the three VIs in estimating the LAI also varied among phenophases. During the growing period, the mean RMSE ± SD value for the estimated LAIs was 0.30 ± 0.11 (LAI-NDVI relationships), 0.37 ± 0.11 (LAI-EVI relationships) and 0.36 ± 0.13 (LAI-NIRv relationships), respectively, indicating the NDVI produced significantly better LAI estimations than those from the other two VIs. In contrast, the EVI produced slightly better LAI estimations than those from the other two VIs during the declining period (p > 0.05), and the mean RMSE ± SD value for the estimated LAIs was 0.45 ± 0.16 (LAI-NDVI relationships), 0.43 ± 0.23 (LAI-EVI relationships) and 0.45 ± 0.25 (LAI-NIRv relationships), respectively. Hence, the piecewise LAI-VI relationships based on different phenophases were recommended for the estimations of the LAI instead of a single LAI-VI relationship for the entire vegetation active period. Furthermore, the optimal VI in each phenophase should be selected for the estimations of the LAI according to the characteristics of vegetation growth.

Highlights

The leaf area index (LAI), as a dimensionless quantity, is usually defined as the ratio of the leaf surface area to the unit ground surface area [1]
LAI Estimation with the Piecewise LAI-vegetation indices (VI) Relationships Based on Phenophases Versus That with a Single L4A.1I.-VLAI RI eElsattiimonasthioipn fworitthhethEenPtiierceeVweigseetLatAioIn-VAI cRtievlaetPioenrsiohdips Based on Phenophases Versus That with a SingTlehLeALIA-VIIaRndelaVtiIosngsehniperfaorllythsehEonwtierde VsiemgeiltaartisoenaAsocntiavlevPaerriiaotdions throughout the year, corresponding to the cycle of the vegetation phenology
Significant LAI-VI relationships were observed during the growing period and declining period, and they showed an almost linear trend

Summary

Introduction

The leaf area index (LAI), as a dimensionless quantity, is usually defined as the ratio of the leaf surface area to the unit ground surface area [1]. It is an important parameter used to describe the geometry of the vegetation canopy and a key indicator for understanding the ecological processes at the global and regional scale. Grayson et al proposed a new VI, the near-infrared reflectance of vegetation (NIRv) [20] It consistently untangles the confounding effects of background brightness, leaf area and the distribution of photosynthetic capacity with depth in canopies using existing moderate spatial and spectral resolution satellite sensors [20]. The three VIs have their own advantages in estimating the LAI for various types of vegetation in different regions

Methods

Results

Discussion

Conclusion