Abstract

Estimates of biomass integrated over forest management areas such as selective logging coupes, can be used to assess available timber stocks, variation in ecological status and allow extrapolation of local measurements of carbon stocks. This study uses fifty 0.1 ha plots to quantify mean tree biomass of eight logging coupes (each 450–2500 ha) and two similarly sized areas in un-logged forest. These data were then correlated with the spectral radiance of individual Landsat-5 TM bands over the 15 km × 15 km study area. Explanation of the differences in radiance between the ten forest sites was aided by measurements of the relative reflectance of selected leaves and canopies from ground and helicopter platforms. The analysis showed a marked variation in the stand biomass from 172 t ha −1 in coupe C88 that was disturbed by high-lead logging to 506 t ha −1 in a similarly sized area of protection forest. A two-parameter linear model of Landsat TM radiance in the near-infrared (NIR) band was able to explain 76% of the variation in the biomass at this coupe-scale. The local-scale measurements indicated that the differences in the mean radiance of each coupe (in cloud-free areas) may relate to a change in the proportion of climax tree canopy relative to a cover of either pioneer trees or ginger/shrubs; the canopies of climax trees have the lowest NIR radiance of the vegetation characteristic of selectively logged forest. The coupe harvested following ‘Reduced Impact Logging’ guidelines had a residual biomass and NIR radiance more like that of undisturbed lowland dipterocarp forest than coupes disturbed by ‘conventional’ selection felling. The predictability of tree biomass (at the coupe-scale) by such a parsimonious model makes remote sensing a valuable tool in the management of tropical natural forests.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call