The promise and challenges of small unmanned aerial systems (sUAS) for agricultural research in the Long-Term Agroecosystem Research (LTAR) network.

Authors: Coffin, Alisa; Clark, Pat

Submitted to: Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: 10/2/2018
Publication Date: N/A
Citation: N/A

Interpretive Summary:

Technical Abstract: Small unmanned aerial systems (sUAS) are now being promoted extensively as a tool to monitor crop conditions with fine resolution and low latency. Rangelands are also benefiting from the increased resolution and frequency of sUAS vegetation monitoring systems. The sUAS industry has mushroomed in recent years with the introduction of relatively low-cost package solutions that promise high-performance decision tools for farmers. While these ensembles may provide adequate data quality to quickly spot problems, such as pest and disease outbreaks, their use as research instruments requires additional steps to calibrate sensors, validate data, and manage the entire sUAS data collection process, from mission planning and preparation to data handling, processing and storage. Further steps are then required to curate data that is incorporated into research projects. Scientists in the USDA Long-Term Agroecosystem Research (LTAR) are integrating sUAS datasets as inputs to models that characterize and forecast agroecosystem dynamics. Across LTAR, sUAS data provide a means of “bridging the gap” between low repeat, high resolution, and high repeat, low resolution satellite imagery for models that predict soil moisture and biomass, for example. At the Gulf Atlantic Coastal Plain LTAR site in Tifton, GA, calibrated multi-spectral data from a sUAS-borne sensor were collected over cropped areas on multiple dates in 2017 and 2018. In some areas, ground measurements coincided with sUAS collections. Initial reviews of the data indicate mixed results, providing frequent fine resolution wall-to-wall information across the entire farm, but with problems of uneven illumination from rapidly changing skies during the mid-summer noontime flights. At the Great Basin LTAR site in Boise, ID, sUAS imagery were collected at 1-mm ground sample distance (GSD) in several dominant shrub-steppe plant community types to assess disturbance-induced changes in species richness and abundance, fuel continuity and loading, and ground cover. Surface-from-motion (SfM) has also been applied to snow-off and snow-on imagery to assess snow depth in high-elevation catchments. The challenge of working with sUAS datasets across the LTAR network is being addressed with a network sUAS data management plan, currently under development. Main features of this plan, when it is fully developed, will include: best practices for mission planning, data collection, storage and naming protocols; metadata collection, and a mechanism by which final datasets can be transferred to and stored in an archive for network-wide research use.