We will deliver upscaled global GPP using SIF-GPP relationships derived at global eddy covariance towers from FLUXNET to scale Global SIF, taking advantage of our team members’ expertise and close involvement with the FLUXNET community. Our method is as follows: (1) Predict Network SIF at FLUXNET GPP towers every 16-days; (2) Estimate SIFyield and LUE from MERRA-2 PAR, Network SIF, and Fluxnet GPP at each location; (3) Quantify linear relationships between SIF and GPP as the ratio of SIF-yield to LUE as a function of biome and light conditions (sunny vs cloudy); (4) Use biome and light specific relationships to scale Global SIF over the entire record. Our method uses LUE model logic to compute GPP (see Fig. 3), which captures the efficiency of plants to use light in photosynthesis (Monteith, 1972):
GPP = LUE * APAR (1)
where LUE (g C MJ-1) is the conversion efficiency of photosynthetic active radiation (PAR) to vegetation biomass, and APAR (MJ m-2) is the amount of PAR absorbed by the canopy for photosynthesis. Traditional practices that use this approach parameterize LUE as a spatially and temporally dynamic parameter sensitive to biome and reduced for suboptimal environmental conditions due to temperature and water stress, but there has been no effective global record to clarify LUE or associated limits to GPP or ET. Recent work shows this model can lead to overestimates of growing season length and CO2 uptake if structural leaf changes vs physiological stress effects are not accounted for correctly (Luus et al., 2017; Commane et al., 2017). However, SIF is closely related to LUE and GPP (Grace 2007). The following equation defines the relationship between SIF and APAR:
SIF = SIFyield * APAR (2)
where SIF is the radiation re-emission (fluorescence) of a portion of APAR by the plant canopy. Equations (1) and (2) can be rearranged to estimate GPP:
GPP = (LUE / SIFyield) * SIF (3)
The SIF-GPP relationship as a function of LUE and SIFyield will be validated against tower SIF and GPP