Comparing in-situ leaf observations in early spring with flux tower CO2 exchange, MODIS EVI and modeled LAI in a northern mixed forest

Abstract

Changes in the timing and duration of spring leaf development have implications for the start of the carbon uptake period and are therefore fundamental to the accurate calculation of carbon budgets, and in determining the potential for forests to sequester CO2. Here, we examined trends in CO2 exchange (Net Ecosystem Exchange (NEE), Gross Primary Production (GPP) and Ecosystem Respiration (ER)) (1997–2016) and satellite derived measures (Enhanced Vegetation Index (EVI) and modeled Leaf Area Index (LAI)) of the start of spring from the MODIS product MOD13Q1 (2001–2016) for a mixed forest landscape in northern Wisconsin, USA. We then explored the relationship between these indirect determinants of spring phenology and the timing and duration of spring phenophases (bud-burst, leaf-out, full-leaf unfolded) of trees over a 5-year period (2006–2010). Contrary to earlier studies focus’, our analysis did not find a consistent link between the early transition, of the forest stand from C source to sink with increased annual productivity. Interestingly, while annual regional NEE trended from a source to a sink over the study period, there were no significant concomitant trends in the timing of the start of the season, peak season or the duration of the season derived from (i) satellite data (2001–2016), (ii) flux data (1997–2016) nor from in situ observations (2006–2010). The range of time periods used and difference in phenological determinants examined likely contributed to a lack of expected relationships. The results highlight the need for in situ observations of different forest layers, in particular shrubs, which could help explain current discrepancies between direct and indirect determinants of spring phenology. Furthermore, characterization of abiotic influences on C flux measurements may further explain some of the observed discrepancies.