A comparison of alternative modelling approaches to evaluate the European forest carbon fluxes

The European forest carbon balance studied by various methods shows different results We compared the regional and national net primary production (NPP) estimated by the forest inventory-based model EFISCEN and the climate-based terrestrial ecosystem models (TEMs.BIOME-BGC, ORCHIDEE. and JULES), and single forests NPP derived from the international network of eddy-covariance towers (FLUXNET) In addition, the paper presents the net ecosystem production (NEP) and the net biome production (NBP) calculated with EFISCEN and discusses the influence of forest management onto carbon fluxes We aimed to better understand the variance between EFISCEN and TEMs NPP estimates, and to improve the assessment of European forest mitigation potential for the year 2005 The NPP comparison between the EFISCEN inventory method and the TEMs process-based method showed similar average values for Europe and its countries The European NIT average 508 +/- 183 (+/- standard deviation) gC/m(2)/year of EFISCEN was close to 487 +/- 126 gC/m(2)/year of TEMs The country level average EFISCEN-TEMs difference was just 57 +/- 153 gC/m(2)/year Larger differences were apparent at the regional level for the species groups Especially for coniferous forests, EFISCEN projected higher values (NPP maximum 1480 gC/m(2)/year) than TEMs (NIT reaching saturation below 700 gC/m(2)/year) Compared to regional TEMs NPP across Europe, the range of regional EFISCEN NIT was consistently larger and with larger variance. Regionally EFISCEN and TEMs NIT averages were close to the individual FLUXNET data Similar to broadleaves of TEMs, the FLUXNET broadleaves NPP were more productive compared to coniferous forests We conclude that the two methods produce similar results, except for higher regional EFISCEN NPP of coniferous forests The NPP difference between modelling methods was presumably result of TEMs assuming mature steady state forests, and lacking the distribution of highly productive and abundant intermediate age classes (integrated into EFISCEN) Both approaches have their advantages. TEMs include climate and environmental change, whereas EFISCEN includes past and current management Combining the two approaches will allow more accurate assessment of the forest carbon balance, including direct and indirect human effects. (C) 2010 Elsevier B V All rights reserved