Application of remote sensing, an artificial neural network leaf area model, and a process-based simulation model to estimate carbon storage in Florida slash pine plantations

Abstract

Carbon sequestration in forests is of great interest due to concerns about global climate change. Carbon storage rates depend on ecosystem fluxes (photosynthesis and ecosystem respiration), typically quantified as net ecosystem exchange (NEE). Methods to estimate forest NEE without intensive Site sampling are needed to accurately assess rates of carbon sequestration at stand-level and larger scales. We produced spatially-explicit estimates of NEE for 9770ha of slash pine (Pinus elliottii) plantations in North-Central Florida for a single year by coupling remote sensing-based estimates of leaf area index (LAI) with a process-based growth simulation model. LAI estimates produced from a neural-network modeling of ground plot and Landsat TM satellite data had a mean of 1.06 (range 0-3.93, including forest edges). Using the neural network LAI values as inputs, the slash pine simulation model (SPM2) estimates of NEE ranged from -5.52 to 11.06Mg•ha^(-1)-a^(-1) with a mean of 3.47Mg•ha^(-1)-a^(-1) Total carbon storage for the year was 33920t, or about 3.5 tons per hectare. Both estimated LAI and NEE were highly sensitive to fertilization.