Remote sensing of net ecosystem productivity based on component spectrum and soil respiration observation in a boreal forest, interior Alaska

Abstract

We built a remote‐sensing method for determining leaf area index (LAI) and ground cover mosses/lichens in boreal forests by field component spectral observation and radiative transfer modeling based on the spectrum. The method was applied to evaluate annual net ecosystem productivity (NEP) distribution in a boreal forest, interior Alaska, by accounting for net primary productivity (NPP) of the vegetation compositions and soil respiration observation synchronized to the spectral observation. Spectral reflectance and soil respiration were observed in two 30‐m × 30‐m plots in black spruce stands, central Alaska. Spectral characteristics of the forest floor and the needle leaves were used as input parameters of a radiative transfer model to evaluate nadir reflectances of spruce communities in relation to varying upper layer LAI, forest floor bryophyte types, and leaf spectral characteristics. Using the relationship, we obtained LAI and bryophyte area ratios for each pixel that corresponds to spruce forest on Landsat ETM+ imagery. The LAI‐NPP relationship of spruce forest was estimated from Plonski's [1981] normal yield table data and specific leaf area, and NPP was calculated from LAI. Observations of daily respiration were extrapolated to annual timescales using soil temperature. On the basis of the annual soil respiration and NPP of the upper layer and forest floor, annual NEP geographical distribution in a recent normal year was estimated from remotely sensed LAI and forest floor bryophyte area ratios. The annually estimated NEP was 51 g C/m2/yr, which corresponds to the value (55 g C/m2/yr) for 150‐year‐old black spruce forest in the Boreal Ecosystem‐Atmosphere Study (BOREAS) region, Canada.