Chlorophyll content retrieval of boreal conifers using hyperspectral remote sensing (Pinus banksiana).

摘要

This investigation quantitatively links chlorophyll a+b (chl a+b) concentration, a physiological indicator of forest health condition, to hyperspectral observations of Jack Pine (Pings banksiana), a dominant Boreal forest species. Between June and September of 2001, four field campaigns of data collection were conducted over forested areas near Sudbury, Ontario, Canada. Using the Compact Airborne Spectrographic Imager (CASI), optical image data were collected, in the visible-near infrared domain, over eight selected Jack Pine sites. Supplementing the airborne campaigns was simultaneous on-site collection of foliage samples for laboratory spectral and chemical measurements. The study first linked needle-level optical properties and pigment concentration through the inversion of radiative transfer models, PROSPECT and LIBERTY. Next, a chlorophyll sensitive optical index (R750/R710), was "scaled-up" to the canopy level using turbid medium canopy models and infinite reflectance formulae, which approximate the canopy as an optically thick vegetative medium. However, due to the potential confounding effects of open canopy structure and foliage clumping, such a simple approach requires careful validation. Accordingly, the analysis focused on high spatial resolution CASI imagery (1 meter) to visually target tree crowns, while accounting for shadows, and open (sun-lit) understory areas. Estimating needle-level chl a+b concentration using coupled leaf and canopy models were shown to be feasible with root mean square errors ranging from 4.5 to 6.5 mug/cm2, 16% of typical needle pigment values. Effectively eliminating the confounding variables will permit the prediction of needle pigment content at large spatial scales to highlight regions of potential stress for prospective management strategies.