关键词: 杜仲, 杜仲梦尼夜蛾, 危害, 风险分级, 预警技术

Abstract: [Objective] This study was to establish an early warning technology for forecasting outbreaks of defoliator insects Orthosia songi Chen et Zhang based on a recommended standard "Grading Methods of Natural Disaster Risk" issued by Civil Affairs of China. The main aim was to provide relevant information for risk grading and establish a warning technology of outbreak for forest and agriculture defoliator insects.[Method] This study used O. songi-host tree Eucommia ulmoides system as a model located at Linbao City, Henan Province, China. Leaf area loss rate (%) and fruit dropped rate (%) were measured under different larval densities (larvae per hundred leaves). Probability index (P) of O. songi population occurrence was determined using the expected larval densities. Quantitative indexes of loss (C) caused by O. songi occurrence were used in basis of the expected scope of leaf area loss rate, fruit dropped rate, outbreak area rate and expected increasing management cost rate. A risk classification matrix of O. songi occurrence risk was set up based on the normalized risk probability index (P) and loss index (C). Meanwhile, the expected leaf area loss scope, fruit dropped loss scope and occurrence area was predicted using the observed egg and larval densities in permanent plots in Eucommia plantations. These data were used to forecast the risk grade. Feasibility and maneuverability of risk grading system were then evaluated through actual leaf area loss rate and fruit dropped rate during late development stage of the same generation.[Result] Leaf area loss rate and fruit dropped rate differed statistically significant among different larval densities. A larval density at 16 larvae per hundred leaves led to more than 60% of leaf area loss. Every single larva caused 2.8% of average leaf loss. The 17% or more of fruits premature drop when larval density was 24 larvae per hundred leaves or above. Every larva caused average 0.8% of fruit premature drop. Based on survey of egg density and larval density in May 2015, we predicted that larval density ranged 11 to 35 larvae per hundred leaves with a probability index P of 2, and the expected occurrence area rate of 75% at a loss index C of 2. The expected scope of leaf area loss rate at C=2 ranged from 16% to 49%, and the expected fruit drop rate at C=3 varied from 6% to 10%. The expected increase in management cost was 0 with a loss index C=4. Therefore the smallest value should be selected as loss index C when C had a different value in the risk classification matrix. Risk score R should be calculated as P×C, or 2×2=4 in the corresponding risk classification matrix. A high risk of orange alert was predicted and did not rule out extremely high risk in some plantations. The actual average leaf area loss rate and fruit dropped rate caused by O. songi surveyed in July 2015 were 17.2% and 9.4%, respectively, among which plantations accounted for 17% of the total area suffered extremely high damage. The results fitted well the prediction, indicating that the risk grading method is feasible.[Conclusion] Risk score R of O. songi disaster was divided into four grades and was given four different colors for different R values, or four different levels of risk of O. songi occurrence. Red color represent extremely-high risk with R value ranging 1-2; orange color, high risk with R from 3-4; yellow means medium risk with R from 6 to 9; green represents low risk, R ranging 12-16. The risk grading method in this study was well suited for short-term predictions, i.e. prediction of risk caused by one generation. Refining the indices for long-term prediction will be a focus of future studies.

Key words: Eucommia ulmoides, Orthosia songi, damage, risk grading, warning technology