The U.S. network of 160 weather radars known as NEXRAD (NEXt generation RADar) is one of the largest and most comprehensive terrestrial sensor networks in the world. To date, the National Climatic Data Center (NCDC) has archived about 2 petabytes data from this system. Although designed for meteorological applications, these radars readily detect the movements of birds, bats, and insects. Many of these movements are continental in scope, spanning the entire range of the network. It is unclear whether biological or meteorological data comprise the bulk of the archive. Regardless, the biological portion is sufficiently large that it likely represents one of the largest biological data archives in the world, perhaps second only to satellite-based remote sensing data.
Principal Investigator : Robert H Diehl
Cooperator/Partner : Todd M Preston
Ongoing research continues to demonstrate that this radar network shows enormous potential to inform on a wide range of applications including wildlife biology, conservation, and management; weather-related modeling and forecasting; energy development; human safety; and public education (see http://www.nrmsc.usgs.gov/staff/diehl/research/remote_sensing/weather_radar - content no longer available). Harnessing this potential of NEXRAD data for biological applications depends on empowering prospective users in gaining access to that subset of the NCDC archive in usable form. It is a goal of the biological radar research community to advance software and database capability to make these data more accessible to the wider scientific community. To that end, several technological hurdles must still be overcome before the full potential of the NEXRAD data archive is fully realized. Specifically, NEXRAD’s potential as biological instrumentation is far from realized owing to the size of the data archive, the challenges faced by prospective users in distinguishing biological from non-biological data, and the absence of a mechanism for making that data available in a more accessible form. This project pursues a proof-of-concept demonstrating how biological weather radar data can be post-processed into value-added biological metrics that are accessible to the public and that can address large spatial and temporal scale biological questions.
To achieve this goal, we acquired data from NCDC and hand screened these data to retain only those attributable biological organisms. (Apart from this project, colleagues of ours have initiated research to develop algorithms to mine biological targets from the NEXRAD data stream. When successful, the resulting algorithms will obviate the need to hand screen biological targets.) Additional processing removed lingering sources of systematic bias (Diehl and Larkin 2005, Buler and Diehl 2009, Sheldon et al. 2013), separates vertebrates (mostly birds) from invertebrates (mostly insects), and calculate biologically meaningful metrics including animal density, speed, direction of travel, and height distribution.
The goal of the project was to show a series of data processing steps toward making derived radar-based biological metrics available to the public for scientific purposes. To demonstrate the utility of such an archive, we chose for this proof-of-concept to retrieve data that addressed two large scale biological questions, one based on temporal data and one on spatial data. The temporal dataset focused on documenting the phenomenon of reverse migration - where birds migrate in a direction that is not seasonally appropriate - over several years. We retrieved six years (>350 days) of fall data from the Mobile, AL radar station. Fortuitously, data on bird migration from this radar was well positioned geographically to coincide with six corresponding years of existing radio telemetry data on migrating birds that effectively served as ground-truth for radar observations of reverse migration. The technologies proved highly complementary. Although radio tracking data show that occasional individual birds engage in reverse migration (Smolinsky et al. 2013), weather radar data show that the phenomenon is much more widespread, occurring among thousands of birds and across an area larger in extent than could be practically covered by other monitoring methods.
Our large spatial dataset focused on how migrating birds respond to tropical storms in the Gulf of Mexico. We retrieved regional/continental-scale snapshots of data from >20 radars when widespread bird migration was present during the passage of tropical storms of varying intensity. Conventional wisdom considers such storm activity as one of many challenges birds must face during migratory passage. We show, however, that tropical storm activity also promotes the movements of birds passing to the west of storm centers where winds are strong and favorable for fall migration. We also explore how this benefit varies with the intensity of these storms and in relation to typical migratory conditions.
As a result of this project multiple metrics on the biology of aerofauna from hundreds of so-called radar volume scans will be uploaded to ScienceBase where they will be publicly available for download, increasing data discovery and integration of NEXRAD data to broader scientific communities both within and outside of the USGS. The project advanced methods for the processing, analysis, preservation, and publishing/sharing of this valuable, but underused, data source, and did so in the context of addressing biological questions that will themselves be published in the primary literature.
Note: this description is from the FY14 report
