The Diurnal Dynamics of Gross Primary Productivity Using Observations From the Advanced Baseline Imager on the Geostationary Operational Environmental Satellite-R Series at an Oak Savanna Ecosystem

Gross primary productivity (GPP) is the largest flux in the global carbon cycle and satellite-based GPP estimates have long been used to study the trends and interannual variability of GPP. With recent updates to geostationary satellites, we can now explore the diurnal variability of GPP at a comparable spatial resolution to polar-orbiting satellites and at temporal frequencies comparable to eddy covariance (EC) tower sites. We used observations from the Advanced Baseline Imager on the Geostationary Operational Environmental Satellite-R series (GOES-R) to test the ability of subdaily satellite data to capture the shifts in the diurnal course of GPP at an oak savanna EC site in California, USA that is subject to seasonal soil moisture declines. We compared three methods to estimate GPP: (a) a light-use efficiency model, (b) a linear relationship between the product of near-infrared reflectance of vegetation and photosynthetically active radiation (LIN-NIRvP) and EC tower GPP, and (c) a light response curve (LRC-NIRvP) between NIRvP and EC GPP. The LRC-NIRvP achieved the lowest mean absolute error for winter (2 mu mol CO2 m(-2) s(-1)), spring (2.51 mu mol CO2 m(-2) s(-1)), summer (1.43 mu mol CO2 m(-2) s(-1)), and fall (1.35 mu mol CO2 m(-2) s(-1)). The ecosystem experienced the largest shift in daily peak GPP in relation to the peak of incoming solar radiation toward the morning hours during the dry summers. The LRC-NIRvP and the light-use efficiency model were in agreement with these patterns of a shift in peak daily GPP toward the morning hours during summer. Our results can help develop diurnal estimates of GPP from geostationary satellites that are sensitive to fluctuating environmental conditions during the day.

Automated summary

This study used satellite data to explore the diurnal variability of GPP and its relationship with environmental conditions. By comparing three methods, it was found that the LRC-NIRvP model provided relatively accurate estimates of GPP and revealed the characteristic of the peak of GPP shifting towards the morning hours during summer in relation to incoming solar radiation.