Altitude choice by night migrants in a desert area predicted by meteorological factors

The height distribution of nocturnal migrants in southern Israel was determined by con-ically scanning the sky with the pencil-beam of an X-band radar at different elevation angles. Altitudinal profiles of meteorological parameters were derived from radio sondes launched at midnight and from pilot balloons launched every 4 h. A model to predict the height distribution of birds by means of meteorological variables was developed by assuming that the observed proportions of birds within a height zone, compared with the neighbouring height zones, reflect the degree of the birds' preference for that height zone. Only one among the variables included in the multiple regression analysis proved to have a significant influence on the height distribution of migrants: the difference of tailwind speed between height zones. Simulations with 1000 birds choosing altitudes by means of the night's altitudinal profile of tailwind speed closely traced the observed distributions. The fact that all the other meteorological factors which were previously suggested to have an influence on the flight range in trans-desert migration were not selected as relevant factors is discussed. The following basic information on nocturnal bird migration in the Negev is provided as a background for the statistical analysis: Directions of migration are within very narrow limits. During the first hour after take-off, 60% of autumn migrants and 75% of spring migrants are climbing, with vertical speeds of 0.1–2 m per s and 0.1–4 m per s, respectively. During the rest of the night, climbing and descending birds are in nearly equal proportions. Thus, there is a high potential of sampling atmospheric conditions at different altitudes. Height distributions in spring and autumn show the influence of the trade wind situation, autumn migrants making use of the northerly winds at low levels in spite of high temperatures, while spring migrants tend to reach the southwesterly winds at higher levels.