Skip to main content
ARS Home » Research » Publications at this Location » Publication #220151

Title: Utility of an image-based canopy reflectance modeling tool for remote estimation of LAI and leaf chlorophyll content at the field scale

item Houborg, Rasmus
item Anderson, Martha
item Daughtry, Craig

Submitted to: Remote Sensing of Environment
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 9/26/2008
Publication Date: 1/15/2009
Citation: Houborg, R., Anderson, M.C., Daughtry, C.S. 2009. Utility of an image-based canopy reflectance modeling tool for remote estimation of LAI and leaf chlorophyll content at the field scale. Remote Sensing of Environment. 113:259-274.

Interpretive Summary: Remotely sensed data in the reflective optical domain function as a unique cost-effective source for providing spatially and temporally distributed information on key biophysical and biochemical parameters of land surface vegetation. This paper describes a tool (REGFLEC) for entirely image-based retrievals of leaf area index and leaf chlorophyll content. REGFLEC links atmospheric radiative transfer and inverse and forward canopy reflectance modeling and computes key biophysical properties by considering wide variations in leaf structure, vegetation clumping, leaf inclination angle, fraction of senescent vegetation and soil reflectance. Image-based regularization techniques are introduced to constrain the parameter space and improve the reliability of the retrievals. The model is applied to a stressed corn field in Maryland, USA using aerial imagery and satellite data, and the accuracy of LAI and leaf chlorophyll retrievals are evaluated using ground truth data.

Technical Abstract: A REGularized canopy reFLECtance (REGFLEC) modeling tool that couples leaf optics (PROSPECT), canopy reflectance (ACRM), and atmospheric radiative transfer (6SV1) models is described and the model output of leaf chlorophyll (Cab) and total leaf area index (LAI) is validated against ground measurements. Leaf structure, vegetation clumping, and leaf inclination angle are assumed to possess limited variability within a given land cover class and are estimated by iteratively inverting the canopy reflectance model with the spectral radiometric information of pixels belonging to the same land cover type. Only pixels originating from medium to high density vegetation areas are included in the inversion to maximize the sensitivity of the reflectance signal to the class-specific model parameters and reduce the influence of background effects. Image-based techniques are introduced to constrain the parameter space and reduce the number of the computationally demanding inversions. Subsequently, a family of pre-computed spectral reflectance relationships (look-up tables), each a function of canopy characteristics, soil background effects and external conditions, are employed for fast pixel-wise biophysical parameter retrievals. Using 1 m resolution aircraft and 10 m resolution SPOT-5 imagery, REGFLEC effectuated robust biophysical parameter retrievals for a stressed corn field characterized by a wide range in leaf chlorophyll levels and intermixed green and senescent leaf material. Relative root-mean-square deviations (RMSD) were on the order of 10 % for the 1 m resolution LAI (RMSD = 0.25) and Cab (RMSD = 4.4 µg cm-2) estimates, due in part to an efficient correction for background influences. The use of 10 m resolution radiance data resulted in comparable retrieval accuracies and the overall intra-field pattern in LAI and Cab was well established at this resolution. The developed method is entirely image-based, does not rely on impractical in-situ measurements, can be applied at a range of scale using radiometric information from only 3 spectral bands (green, red, and near-infrared), and can be used for agricultural fields characterized by widely varying distributions of model variables. REGFLEC holds promise as a valuable tool for predicting LAI and leaf chlorophyll in the context of precision farming, for determining vegetation stress and photosynthetic productivity, and for reliable biophysical parameter estimations at a range of scales.