Analysis of Spatial and Temporal Gross Primary Productivity and Evapotranspiration in Korea by Using BESS

Abstract

Analysis on spatial and temporal patterns of gross primary productivity (GPP) and evapotranspiration (ET) is critical in understanding and predicting terrestrial carbon and water cycles. However, there have been a few studies which provide spatially and temporally continuous estimates of GPP and ET in the Korean Peninsula. In this study, spatial and temporal patterns of GPP and ET in the Korean Peninsula were studied by using a bio-physical model, Breathing Earth System Simulator (BESS). GPP and ET from BESS are sensitive to Leaf Area Index (LAI) which is one of important input variables of BESS. LAI refinement was conducted by using re-modified temporal and spatial filter (rTSF) and LAI refinement improved accuracy compared to LAI from BESS and raw MODIS. The simulations of 11-year (2001-2011) GPP and ET in the Korean Peninsula were performed by using refined LAI. BESS derived GPP and ET were verified with eddy covariance measurements at 4 sites and 5 sites, respectively. Also, BESS derived basin ET was evaluated against water balance derived ET at a basin scale. Both GPP and ET from BESS show high positive bias at GDK and GCK sites while BESS derived GPP and ET were comparable to eddy covariance measurement at other sites and water balance derived ET. The mean annual land GPP and ET over the eleven years (2001-2011) were 1,183 gC m-2 year-1and 491 mm year-1. Land cover change was one of major causes for interannual variation of GPP and ET. In dry year, ET at highland became more sensitive to canopy conductance than that in normal year. Spring and autumn droughts caused spring and autumn ET to become most sensitive to canopy conductance. The sensitivity of ET to available energy increased in summer due to clouds during the summer monsoon.