Technologies for terrestrial ecosystem management in the area of precision agriculture and ecosystem restoration have made significant advances recently for more sustainable practices by optimizing water, nutrients, and fertilizers Many of these technologies include monitoring and imaging of plants, soil, and crop harvest as well as their interactions, using in situ sensors, remote sensing, and geophysics Their development is mostly industry-driven focused on the local-scale information In parallel, there have been significant efforts by the US Department of Energy (DOE) to establish public databases for quantifying ecosystem functions in regional and national scales, including greenhouse gas (GHG) fluxes, evapotranspiration (ET), soil biogeochemistry, and microbial genomics through Ameriflux, KBase, and ESS- Dive, and soon the National Microbiome Data Collective (NMDC) The DOEs databases and user facilities are powerful in such assessments yet, they have been rarely used for ecosystem management In this project, we will develop an open, scalable software system for multi-scale, multi-modal data integration; focused on coupling local-scale datasets from managed ecosystems with DOEs regional-scale datasets on GHG fluxes, plant genotypes, and soil microbiome (Ameriflux, KBase and ESS-Dive) We will evaluate the impact of local and regional-terrestrial heterogeneity based on different sensing technology, such as remote sensing for plant phenology, and geophysical sensors that capture the spatial heterogeneity of soil properties LBNL has extensive expertise characterizing soil-plant interactions and other terrestrial ecosystem properties across scales Arva has developed machine learning capabilities in integrating multi-modal data to investigate the relationships between soil-plant interactions and crop yield, funded in part by a prior DOE SBIR In Phase I, we will develop prototype software and demonstrate its utility through three tangible applications: (1) water management based on in-situ soil sensors, geophysics and UAV images coupled with the ET estimates derived by Ameriflux; (2) the evaluation of farm practices (tilling/no-tilling, water manipulation) on GHG fluxes (carbon, methane) in rice fields based on high-resolution imagery and Ameriflux; (3) the identification of soil biogeochemical properties and the link to soil functional types through Kbase and ESS-Dive
