Fusion of remotely sensed 3DVegetation structure with a dynamic global terrestrial ecosystem model- Ent for improved estimates of carbon stocks and land-atmosphere exchanges

Background/Question/Methods

Lidar remote sensing provides measurements of horizontal and vertical vegetation structure of ecosystems which will be critical for estimating global carbon storage and assessing ecosystem response to climate change and natural and anthropogenic disturbances.  However, no consistent approach currently exists to derive the lidar based vegetation structure information required by terrestrial ecosystem models. Can vegetation structure information measured by lidar improve terrestrial ecosystem carbon flux estimates? How can lidar data be integrated in the terrestrial ecosystem models?

To answer those questions, we implemented a 3D-vegetation structure based canopy radiative transfer model into the Ent Terrestrial Biosphere Model (TBM) (Ni-Meister et al., 2010). We developed a scheme to derive vegetation structure inputs required by ENT from lidar full waveform measurements. Vegetation height and vertical foliage profiles from lidar are used as inputs to drive the canopy radiative transfer and photosynthesis and conductance schemes in Ent Terrestrial Biosphere Model (TBM) to calculate Photosynthetic Active Radiation (APAR) absorbed by green vegetation and gross primary production (GPP).

We test if using vertical vegetation structure inputs improves fraction of Absorbed Photosynthetic Radiation (FAPAR) and GPP estimates using the 3D-vegetation structure-based and non-structure based radiative transfer schemed implemented in the Ent-TBM.  We also investigate if the vegetation height and vertical foliage profile derived from lidar data can be implemented in the Ent model and if using lidar vegetation structure inputs improves the FAPAR and GPP estimates. 

Results/Conclusions

Our model comparison results in deciduous forests of the Morgan-Monroe State Forests (MMSF), IN show that the non-vegetation-structure based canopy radiative transfer scheme underestimates the FAPAR and GPP estimates. The modeled GPP using the vegetation structure based canopy radiative transfer model matches well with the field measurements as the structure-based scheme deals with the sunlit and shaded components.

Our scheme was able to invert the lidar full waveform measurements to derive the clumping index and vertical foliage profile from lidar. The clumping index and foliage profile were compared with ground-based vegetation structure measurements in Harvard Forests, MA and conifer forests in Howland Experimental Forest, ME.

Currently the structure inputs derived from lidar is being used to derive the Ent in these two sites to investigate if the lidar vegetation structure measurements would improve APAR and GPP estimates. This study presents a physical approach to derive vegetation height and vertical foliage profiles which can be used by the Ent TBM. Future work will continue to expand our method to large regions and use these vegetation structure inputs to drive Ent-TBM to evaluate the improvement of carbon stock and fluxes estimation.