A multi-sensor array for detecting and analyzing nocturnal avian migration

Abstract

Avian migration has fascinated humans for centuries. Insights into the lives of migrant birds are often elusive; however, recent, standalone technological innovations have revolutionized our understanding of this complex biological phenomenon. A future challenge for following these highly mobile animals is the necessity of bringing multiple technologies together to capture a more complete understanding of their movements. Here, we designed a proof-of-concept multi-sensor array consisting of two weather surveillance radars (WSRs), one local and one regional, an autonomous moon-watching sensor capable of detecting birds flying in front of the moon, and an autonomous recording unit (ARU) capable of recording avian nocturnal flight calls. We deployed this array at a field site in central Oklahoma on select nights in March, April, and May of 2021 and integrated data from this array with wind data corresponding to this site to examine the influence of wind on the movements of spring migrants aloft across these spring nights. We found that regional avian migration intensity is statistically significantly negatively correlated with wind velocity, in line with previous research. Furthermore, we found evidence suggesting that when faced with strong, southerly winds, migrants take advantage of these conditions by adjusting their flight direction by drifting. Importantly, we found that most of the migration intensities detected by the sensors were intercorrelated, except when this correlation could not be ascertained because we lacked the sample size to do so. This study demonstrates the potential for multi-sensor arrays to reveal the detailed ways in which avian migrants move in response to changing atmospheric conditions while in flight.

Keywords: Avian migration; Data integration; Remote sensing; Weather surveillance radar.

Figure 1. Relationship between the number of birds detected by LunAero and the regional avian migration intensity from KTLX in the atmosphere above KAEFS on select nights in March, April, and May of 2021.

Each data point corresponds to a 1-h period on the night of March 26, March 27, March 28, March 29, April 24, or May 25, 2021. Each date is associated with a different color.

Figure 2. Relationship between the local avian migration intensity detected by PX-1000 and the regional avian migration intensity detected by KTLX in the atmosphere above KAEFS in the spring of 2021.

(A) Relationship on the night of May 25, 2021. (B) Relationship on the nights of March 26, March 27, March 28, March 29, and April 24, 2021. Each data point corresponds to a ~10-min period, which is the temporal resolution of the KTLX weather surveillance radar (WSR). The regional migration intensity on the night of May 25, 2021 significantly predicts and explains 59.1% of the variance in local migration intensity (β = 5.00 ± 0.321, P < 0.001).

Figure 3. Direction and velocity of the wind in the atmosphere above the Kessler Atmospheric and Ecological Field Station (KAEFS) on select nights in March, April, and May of 2021.

This polar plot shows wind direction as the angle in degrees clockwise from north and velocity as the distance between the data point and the origin. Each data point corresponds to a 3-h period for a given date, represented by a color. This data stems from the National Centers for Environmental Prediction (NCEP)’s North American Regional Reanalysis (NARR).

Figure 4. Relationship between regional avian migration intensity detected by KTLX and wind velocity in the atmosphere above KAEFS on select nights in March, April, and May of 2021.

Each data point corresponds to a ~10-min period, which is the temporal resolution of the KTLX weather surveillance radar (WSR), for the nights of March 26, March 27, March 28, March 29, April 24, and May 25 of 2021. The wind data stems from the NCEP’s NARR.

Figure 5. Difference between the flight direction of individual birds detected by LunAero and the direction of the wind as a function of wind velocity in the atmosphere above KAEFS in the spring of 2021.

(A) Southerly and (B) northerly winds. Southerly winds are winds with direction angles between 90 and 270 degrees clockwise from north, while northerly winds are winds with direction angles between 270 and 0 degrees and between 0 and 90 degrees clockwise from north. The black dots represent mean differences for discrete wind velocity values. The data correspond to the nights of March 26, March 27, March 28, March 29, April 24, and May 25, 2021. The wind data stems from the NCEP’s NARR.