Retrieval of forest leaf functional traits from HySpex imagery using radiative transfer models and continuous wavelet analysis

Abstract

Quantification of vegetation properties plays an important role in the assessment of ecosystem functions with leaf dry mater content (LDMC) and specific leaf area (SLA) being two key functional traits. For the first time, these two leaf traits have been estimated from the airborne images (HySpex) using the INFORM radiative transfer model and Continuous Wavelet Analysis (CWA). Ground truth data, were collected for 33 sample plots during a field campaign in July 2013 in the Bavarian Forest National Park, Germany, concurrent with the hyperspectral overflight. The INFORM model was used to simulate the canopy reflectance of the test site and the simulated spectra were transformed to wavelet features by applying CWA. Next, the top 1% strongly correlated wavelet features with the LDMC and SLA were used to develop predictive (regression) models. The two leaf traits were then retrieved using the CWA transformed HySpex imagery and the predictive models. The results were validated using R 2 and the RMSE of the estimated and measured variables. Our results revealed strong correlations between six wavelet features and LDMC, as well as between four wavelet features and SLA. The wavelet features at 1741 nm (scale 5) and 2281 nm (scale 4) were the two most strongly correlated with LDMC and SLA respectively. The combination of all the identified wavelet features for LDMC yielded the most accurate prediction (R 2 = 0.59 and RMSE = 4.39%). However, for SLA the most accurate prediction was obtained from the single most correlated feature: 2281 nm, scale 4 (R 2 = 0.85 and RMSE = 4.90). Our results demonstrate the applicability of Continuous Wavelet Analysis (CWA) when inverting radiative transfer models, for accurate mapping of forest leaf functional traits.

Monitoring LAI, Chlorophylls, and Carotenoids Content of a Woodland Savanna Using Hyperspectral Imagery and 3D Radiative Transfer Modeling

“…Canopy LAI and biochemistry estimations using hyperspectral imagery have mostly been done over canopies with high LAI (for instance, see Banskota et al [16], le Maire et al [25], Ali et al [26], Darvishzadeh et al [27], Malenovský et al [28]). While various estimations have also specifically been done over open-canopy ecosystems (e.g., the works of Zarco-Tejada et al [29], Hernández-Clemente et al [30], Zarco-Tejada et al [31]), research concerning acceptable modeling methods within RTM is still ongoing [32,33].…”

Detection of Pine Shoot Beetle (PSB) Stress on Pine Forests at Individual Tree Level using UAV-Based Hyperspectral Imagery and Lidar

“…The range of leaf chlorophyll (Cab) and leaf area index (LAI) were defined from field data, while soil reflectance, viewing geometry and solar zenith angle were extracted from the hyperspectral image metadata. Others parameters (shown in Table 2) include, leaf mass per area (Cm), equivalent water thickness (Cw), leaf structure parameter (N), carotenoid content (Car), anthocyanin content (Canth), average leaf angle (ALA), and hot spot size were set according to the similar literature [45,52,62]. The parametrization of the LUT was based on the input parameters and range described in Table 2.…”

Remotely sensing the species of individual trees

“…The ranges of Cw, Cm and Cab were set based on the field data, while the solar zenith angle, observation angle and azimuth angle were determined based on the HySpex acquisition metadata. The ranges of stem density (SD, n/ha), stand height (SH, m), crown diameter (CD, m) and average leaf angle (ALA, degree) were decided based on the field measurement (Ali et al 2016b;Wang et al 2018). The measured background reflectance was introduced as ρsoil (Atzberger 2000).…”

“…Other leaf, canopy, and external input parameters were selected similarly in agreement with the existing literature (e.g. Ali et al 2016b;Casas et al 2014;Clevers et al 2010;Schlerf and Atzberger 2006;Verhoef and Bach 2007). Input parameters and ranges are shown in Table 4.3.…”

“…Although some studies have successfully retrieved plant functional traits in natural forests using physical or empirical approaches (e.g. Ali et al 2016a;Ali et al 2016b;Essery et al 2008;Kötz et al 2004;Wang et al 2015), the role of plant functional traits retrieved from hyperspectral data for tree species classification still remains underexplored.…”