Understanding the relationship between vegetation greenness and productivity across dryland ecosystems through the integration of PhenoCam, satellite, and eddy covariance data

Authors: YAN, D. - University Of Arizona; Scott, Russell - Russ; MOORE, D.J.P. - University Of Arizona; Biederman, Joel; SMITH, W.K. - University Of Arizona

Submitted to: Remote Sensing of Environment
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 12/21/2018
Publication Date: 3/15/2019
Publication URL: https://handle.nal.usda.gov/10113/6281732
Citation: Yan, D., Scott, R.L., Moore, D., Biederman, J.A., Smith, W. 2019. Understanding the relationship between vegetation greenness and productivity across dryland ecosystems through the integration of PhenoCam, satellite, and eddy covariance data. Remote Sensing of Environment. 223:50-62. https://doi.org/10.1016/j.rse.2018.12.029.
DOI: https://doi.org/10.1016/j.rse.2018.12.029

Interpretive Summary: Arid lands account for approximately 40% of the global land surface and play a dominant role in the net global uptake of carbon dioxide by the land surface. Total ecosystem photosynthesis – termed gross primary productivity (GPP) – represents the total gross carbon uptake but it cannot be observed directly. Currently, vegetation indices that largely capture changes in greenness are the most commonly used datasets in satellite-based GPP estimates. However, there remains significant uncertainty in the relationship between greenness indices and GPP over different spatial areas and over different time periods. We compared vegetation greenness indices from both ground and satellite observations against GPP estimated from ground-based measurements across three representative ecosystem types of the southwestern United States. We found that greenness-GPP relationships were independent of spatial scales as long as land cover type and composition remained relatively constant. We also found that the greenness-GPP relationships became stronger as the time interval increased, with the strongest relationships observed at the monthly resolution. We posit that the greenness-GPP relationship breaks down at short timescales because greenness changes more slowly than plant photosynthesis, which responds rapidly to changes in key meteorological drivers. These findings provide insight into the potential for and limitations of GPP estimates from widely-available greenness indices across dryland ecosystem types.

Technical Abstract: Drylands account for approximately 40% of the global land surface and play a dominant role in the trend and variability of terrestrial carbon uptake and storage. Gross ecosystem photosynthesis – termed gross primary productivity (GPP) – is a critical driver of terrestrial carbon uptake that cannot be observed directly. Currently, vegetation indices that largely capture changes in greenness are the most commonly used datasets in satellite-based GPP modeling. However, there remains significant uncertainty in the spatiotemporal relationship between greenness indices and GPP, especially for relatively heterogeneous dryland ecosystems. In this paper, we compared vegetation greenness indices from PhenoCam and satellite (Landsat and MODIS) observations against GPP estimates from the eddy covariance (EC) technique, across three representative ecosystem types of the southwestern United States. We systematically evaluated the changes in the relationship between vegetation greenness indices and GPP: i) across spatial scales of canopy-level, 30-meter, and 500-meter resolution; and ii) across temporal scale of daily, 8-day, 16-day, and monthly resolution. We found that greenness-GPP relationships were independent of spatial scales as long as land cover type and composition remained relatively constant. We also found that the greenness-GPP relationships became stronger as the time interval increased, with the strongest relationships observed at the monthly resolution. We posit that the greenness-GPP relationship breaks down at short timescales because greenness changes more slowly than plant physiological function, which responds rapidly to changes in key biophysical drivers. These findings provide insight into the potential for and limitations of modeling GPP using remotely sensed greenness indices across dryland ecosystem types.