Using Digital Repeat Photography to Link Vegetative Phenology and Carbon Fluxes to Biotic and Abiotic Drivers in Three Semi-arid Systems (New Mexico, USA)

Abstract:

Arid and semi-arid ecosystems account for over 45% of global land cover. While mean annual carbon uptake in these ecosystems is relatively low, aridlands collectively store a significant amount of carbon. There is high inter- and intra-annual variability of plant growth in aridlands, depending largely on the timing and size of rainfall events. This variation is also of great significance, as the variation in annual semi-aridland carbon uptake accounts for ~39% of the inter-annual variability of the global terrestrial carbon sink, the largest percentage of any land cover type. Although arid and semi-arid ecosystems are of global importance, they are understudied. To better understand the drivers and variability of carbon uptake in these critical ecosystems, we utilize a six-year record of digital images (45,000+ images), carbon flux and meteorological data, soil water content, and associated ecological measurements from three eddy covariance tower sites in central New Mexico. These sites include a Chihuahuan Desert/short-grass Plains grassland site, and post-fire successional grassland site, and a creosote-encroached shrubland site, each of which have unique species compositions, carbon fluxes, and reactions to disturbance and resource addition. All images used are co-registered and corrected for radial lens distortions (when necessary) and greenness indices (2GRBi, gcc, and/or NDVI) are calculated for each scene’s overall “canopy” and for individual species and plant functional types therein. At all three sites, camera-derived greenness is correlated to measured carbon uptake with fine resolution (R² up to 0.8), capturing temporal and spatial variation usually not seen in satellite-based imagery. At sites with lower LAI, species-specific ROI’s were more correlated to the site’s measured carbon flux across shorter time scales. Understanding the biota comprising each image and its contribution to changing scene greenness at different times of year can lead to more accurate carbon flux predictions in semi-arid systems, with species-specific biotic constraints (maximum growth rate, lifespan, and seasonality), growth parameters (light availability, VPD, soil water content, and temperature) as well as community-wide abiotic drivers considered.