Interannual variation in net ecosystem productivity of Canadian forests as affected by regional weather patterns – A Fluxnet-Canada synthesis

Large-scale weather events such as the El Niño Southern Oscillation (ENSO), Pacific Decadal Oscillation (PDO), and droughts are known to cause substantial interannual variation in the net ecosystem productivity (NEP) of tropical, temperate and boreal forests. Hypotheses for the impacts on NEP of changes in air temperature (T a) and precipitation associated with these events were tested at diurnal, seasonal and annual time scales using the terrestrial ecosystem model ecosys with measurements of CO2 and energy exchange from 1998 to 2006 at eddy covariance (EC) flux towers along a transcontinental transect of forest stands in the Fluxnet-Canada Research Network (FCRN). 1 1 Now the Canada Carbon Project. These tests were supported at seasonal time scales by remotely-sensed vegetation indices, and at decadal time scales by wood growth increments from tree-ring and inventory studies. Collectively, results from this testing indicate that large-scale weather events during the study period caused spatially coherent changes in NEP, although these changes may vary with climate zone, species and topography. High T a episodes, such as occurred with greater frequency during ENSO/PDO events, adversely affected diurnal CO2 exchange of temperate and boreal conifers, but had little effect on that of a boreal deciduous forest. These contrasting responses of CO2 exchange to T a were attributed in the model to greater xylem resistance to water uptake in coniferous vs. deciduous trees. Sustained warming such as occurred during ENSO/PDO events extended the period of net C uptake and thus raised annual NEP at boreal coniferous and deciduous sites, but did not do so at a temperate coniferous site where annual NEP was reduced. However the rise in NEP of boreal conifers with warming was partially offset by the adverse effects of high T a on diurnal CO2 exchange, so that the rise in NEP with warming remained smaller than that at a boreal deciduous site. A 3-year drought during the study period adversely affected annual NEP of well-drained boreal deciduous forests but did not affect that of poorly-drained boreal conifers. This lack of effect was attributed in the model to low coniferous evapotranspiration rates and to subsurface water recharge. Drought effects on NEP were therefore largely determined by topography. These contrasting responses of different forest stands to warming and drought indicate divergent changes in forest growth with interannual changes in weather. Such divergent changes are consistent with the complex changes in forest NDVI and net C uptake observed over time in several large-scale remote-sensing studies.