The Global ecosystem Production in Space and Time (GePiSaT) Model of the Terrestrial Biosphere

Abstract:

This work is a continuation of the development and analysis of the Global ecosystem Production in Space and Time (GePiSaT) model. The development of the GePiSaT model was prompted due to inconsistencies found amongst the more complex global vegetation and biogeochemical models that are focused more on prediction rather than explanation. GePiSaT takes a simplistic approach to modelling terrestrial gross primary production (GPP) by making the best use of in situ observations while defensibly representing the principal ecophysiological processes that govern GPP, including (1) the eddy-covariance method of partitioning net CO₂ and solar radiative fluxes into monthly quantities of GPP and respiration and (2) the optimality principle of vegetation minimizing the summed costs associated with maintaining carbon fixation and water transport capabilities. The free and fair-use FLUXNET archive provided the half-hourly in situ observations and were screened for outliers using Peirce's criterion. The limitations of the observational data (i.e., missing or invalid data points) were overcome by means of gap-filling based on modelled half-hourly extraterrestrial solar radiation with daily magnitudes scaled to match WATCH global modelled data products of shortwave downwelling solar radiation and converted to photosynthetic photon flux density (PPFD). Monthly totals of GPP and PPFD, together with MODIS-based estimates of fractionally absorbed photosynthetically active radiation (fPAR), are used in a new theoretically-based light-use efficiency model. Estimates have been made for the intrinsic quantum efficiency and carbon to water cost ratio at individual flux sites to the global terrestrial scale. Analyses of meteorological variables and soil moisture conditions have revealed that vapor pressure deficit and moisture availability (i.e., the Cramer-Prentice bioclimatic moisture index) are important for explaining the deviation from the expected in the light-use efficiency. The new insights on modelling GPP from first principles are useful in developing next-generation vegetation and land-surface models for the characterization and prediction of future climate change impacts.