The flight activities of insect migrating over the Bohai Sea during night were observed with an entomological radar (KC-2) at Beihuang from May to October of 2005 to 2007. The meteorological data, i.e. wind and temperature up to 2000 m above the ground level (AGL) were measured by balloon ascent over the period of 21 h to 23 h every night. The relationship between layering occurrence and meteorological causes was analyzed, thus the mechanism of layering was clear and a model for simulating the height of insect layers was made. The main points are as follows.1. The occurrence of layering phenomenon was coincident with the activities of insect migration through seasons, namely, the frequency of layering was higher during peak migrations of insect than during other time. In addition, the pattern of variation of layering frequency through time was similar to that of light-trap catches. These indicated that layering was a common phenomenon for the migratory insects during their long-distance migration.2. The insects migrating over the sea during night were mainly lepidopteran, odonata, beetles and neuropteran. Most of them were agricultural insect pests or natural enemies in northern China. Cotton bollworm, armyworm, beet armyworm, dragonfly, green lacewing predominated the migratory insects. Asparagus setaceus, Asian corn borer, mugwort cutworm, clover cutworm, bean semilooper, silver ingot semilooper, cabbage moth were less predominant insect pests, which were thought also migratory. The light-trap catches natural enemies, such as dragonfly, ladybug, hoverfly, green lacewing, and carabid beetles increased and decreased suddenly, which indicated the obvious accompanying migration.3. The altitudes of layers were generally below 500 m AGL, with the highest frequency of layering occurred at altitude of 200 - 300 m. The frequency of layering at altitude of 500 - 1500 m was also high. However, there was rare layering at altitude of above 1600 m. The frequency of layering at altitude of above 500 m increased as the month proceeding. This is because that the occurrence of maximum wind speed at these altitudes increased and insects tended to form layers at the height with wind maximum.4. The frequency of insect layering was significantly correlated with the wind direction in May, June, September, and October. The frequency of the insect layering formed in southern, southwestern winds was much higher than in other winds in May and June. The frequency of the insect layering formed in northern, northeastern, and eastern winds was much higher than in other winds in September and October. These indicated that insects tended to select suitable wind directions for flight during northward migration in spring and southward migration in autumn, so as to achieve the longest distance during migration. However, the frequency of the insect layering was not related to the wind direction in July and August. During of this period, insects tended to concentrate at the altitude with wind shear.5. The height where the maximum insect aerial density occurred was significantly related to the height where the maximum wind speed occurred,but not related to the height where the temperature inversion occurred. Insects tended to fly at the altitudes where the winds speed was maximum, or the altitudes with wind shear rather than the altitudes with temperature inversion in Jul and Aug. The insects tended to form layers at the top of temperature inversion only when the temperature inversion occurred at the same altitude as the wind maximum or wind shear occurred in suitable wind direction in July and August.6. Based on the relationship between insect layering and the wind and temperature, a model for estimating altitude of layering was established. At time T and height H, the probability of layeringP(T, h)=Max(Pd(T, WDh), Pd(T, Wh))×Pv(WSh)×Pt(th), where Pd(T, WDh), Pd(T, Wh), Pv(WSh) and Pt(th) was the layering probability function of the wind direction, the wind shear, the wind maximum, and the temperature, respectively. This model fitted our dataset well and could predict the altitudes of layering. |