At 20 sites, incorporating mixtures of black, red, and white mangroves, canopy reflectance spectra were derived from high resolution spectral data taken from a helicopter platform. Canopy characteristics were predicted from the canopy reflectance spectra by using measured and estimated data as inputs into a light-canopy interaction model within a optimization routine. Pertinent to average conditions typifying the area and time of the study, the light-canopy interaction model accomplished two goals. Using the model as a predictor, a sensitivity analysis suggested that little error in modelling the near nadir view canopy reflectance (Rcv) would result from assuming an average soil reflectance of about 0.1, at leaf area index (LAI) values above 2, at near infrared (NIR) leaf reflectances higher than about 0.45. and at sun elevation angles >40o. Moderate errors could result from assuming a spherical leaf angel distribution (LAD), and relatively high errors could result from errors in estimating visible leaf reflectances (and NIR leaf reflectances <0.45) and percent skylight. Differences between canopy hemispherical reflectance (Rc) and Rcv were dominated by percent skylight variation, while differences between Rc and Rcv were moderate to slight at a sun elevation above 20o to 30o, a near spherical LAD, a soil reflectance near 0.1, a LAI up to 4, and a NIR leaf reflectance less than 0.7.
Simulated canopy reflectance spectra were close predictors of obtained spectra, with R2 values >0.97. Mean predicted LAI values were 2.6±0.86 (mean ±1 standard deviation) and were highly related to LAI values derived from field measurements. Seventy-eight percent of the modelled LAI variance was predicted by a normalized difference vegetation index transform of the field canopy spectra data. Predicted LAD values had a near spherical mean value, while the mean difference between input (estimated from laboratory measurements) and predicted leaf reflectances was nearly zero.
|Title||Modeling mangrove canopy reflectance using a light interaction model and an optimization technique|
|Authors||Elijah Ramsey III, John R. Jensen|
|Publication Type||Book Chapter|
|Publication Subtype||Book Chapter|
|Record Source||USGS Publications Warehouse|
|USGS Organization||National Wetlands Research Center|