Abstract/Summary

Addressing the impact of vegetation architecture on the treatment of shortwave radiation in land surface models (LSMs) is important for accurate weather forecast and climate predictions. The study of the carbon budget is also impacted by vegetation architecture because shortwave radiation is used by plants to photosynthesise. Three pieces of research are presented in this thesis: the implementation and evaluation of different parameterisations of vegetation architecture in a commonly used radiative transfer scheme; analysis of the impact of Sun zenith angular variability on vegetation structural parameters including the effect that these parameters have on Gross Primary Productivity (GPP) at site level; and a study on how the simulation of global carbon assimilation is impacted when considering vegetation architecture with satellite derived data sets. Neglecting canopy heterogeneity in radiative transfer schemes leads to significant uncertainties in shortwave radiation absorption and reflectance. The best agreement between detailed 3D radiative transfer schemes and a parameterised 1D version that accounts for vegetation architecture heterogeneity is given when considering zenith angular variability of the parameters. The major impacts on shortwave radiation distribution along the vertical axis are found at the bottom layers of the canopy, which absorbs more radiation when structure is considered. Further impacts on photosynthesis are evaluated at site level with digital hemispherical photography and eddy covariance measurements, and at global level with satellite data and global modelling. Impacts on GPP are dependent on the vertical distribution of the photosynthesis limiting regimes and the variation of the structural parameters with Sun zenith angle is more important over sites with denser foliage than sites with sparser foliage. At global level, prediction of GPP increases by 5.53 ± 1.02 PgC.yr−1 when considering canopy structure, with a strong signal in the tropics. This work establishes the importance of considering vegetation canopy architecture in land surface modelling and predicts that current values of global GPP might be underestimated by LSMs