Materials and Methods
Standardized procedures for conducting the annual SPB prediction survey involve placing from one to three multiple-funnel traps (Lindgren 1983), available from Phero-tech, Inc., Delta, British Columbia, and other sources, (e.g., Synergy Semiochemicals Corp., Burnaby, British Columbia, Canada) in pine forests within a county or national forest ranger district in early spring. The traps consist of 12 funnels superimposed over a collection cup (Figure 1 - A) that is partially filled with soapy water. The number of counties or ranger districts surveyed varies from State to State, depending on availability of pest management personnel, supplies, travel distances, and other factors.
The survey is initiated when flowering dogwood (Cornus florida) begins to bloom or loblolly pine pollen appears. These physiological events coincide with the long-range dispersal of SPB (Billings, unpublished data). Each spring, when ambient temperatures consistently exceed 59 °F, (the flight threshold for SPB and its predators) (Moser and Dell 1979), populations of SPB disperse from overwintering sites in search of weakened pine trees in which they initiate new infestations (Billings and Kibbe 1978). Due to regional variations in climate, trapping usually occurs from early March to mid-April in the Gulf Coastal States and in April or May in States in the northern portions or mountain regions of the beetle’s range. It is during this spring dispersal period that airborne populations of adult SPB and associated insects can be most effectively monitored with pheromone-baited traps for predictive purposes (Billings 1988).
Each trap is baited with a single pheromone packet containing two 400 microliter “bullet” capsules of racemic frontalin (Phero-tech, Inc., Delta, British Columbia) (Figure 1 - B) , the SPB aggregation pheromone (Kinzer and others 1969, Payne and others 1978), and a rapid-release dispenser of steam-distilled southern pine turpentine (Billings 1985) (Figure 1 - C). From 1987 to 2005, the turpentine dispenser consisted of a 250 ml amber Boston round bottle (Fisher Scientific Company, Pittsburg, PA) with a cotton wick. Various brands of commercial-grade southern pine turpentine (Klean-stripTM (1987-1997), Star-texTM (1998), and HerculesTM (1999-2006)) were used, depending on availability. These insect- and host-produced volatiles attract in-flight adult SPB (Figure 1 - D, 1 - E) as well as a major predator, the clerid beetle Thanasimus dubius (Coleoptera: Cleridae) (Figure 1 - F) (Billings and Cameron 1984, Moser and Dell 1980, Payne and others 1978, Vite and Williamson 1970). More than any other associated insect, T. dubius is believed to play a major role in the population dynamics of SPB (Moore 1972; Reeve and Turchin 2002; Thatcher and Pickard 1966; Turchin and others 1991, 1999).
Each trap is installed in a pine-forested area, preferably in stands having sawtimber trees (> 30 cm in d.b.h.) with a sparse hardwood understory. The traps are purposely placed outside of SPB infestations to attract long-range dispersing SPB and its predators. Traps are suspended from a 3-m metal pole or from a cord strung between two hardwood trees, at least 10 m from any live pine (Figure 1 - A). If multiple traps are deployed in a county or ranger district, they should be at least 1 mile apart. The turpentine bottle is placed within the top funnel of the trap, and the frontalin packet is hung from a funnel stanchion near the middle of the trap.
Insects are collected from traps weekly for 4 consecutive weeks, and the numbers of adult SPB and clerids are counted and recorded. Upon completion of the survey, the data from each State and Federal cooperator are sent to the authors for compiling. With these data, the authors or cooperators or both make predictions of SPB infestation trend and relative population level for the current year, based on mean numbers of SPB per trap per day and percent SPB. The latter is defined as the number of SPB X 100 percent divided by the combined number of SPB plus clerids caught per trap (Billings 1988). The data are plotted onto the SPB prediction chart (Figure 2). This chart was developed and refined over the years by comparing the relationship between the two independent variables (SPB per trap per day and percent SPB) and actual numbers of SPB spots detected in a given county or ranger district for the current year to ascertain SPB population level. Actual SPB infestation trend is obtained by comparing the number of spots reported in a given locality or State in the current year with the number reported for the same locality or State in the previous year.
Where trapping data are available for the previous year from the same county or ranger district (preferably from the same specific trap location), a comparison of mean SPB per trap per day and percent SPB for the current year with that for the previous year provide further insight into the direction of infestation trends (increasing, static, or declining). In turn, the number of SPB spots detected in a given county, ranger district, or State in the previous year is useful for predicting population levels (high, moderate, or low) in the current year. For purposes of this paper, the severe outbreak level shown in Figure 2 was considered the same as the increasing or high level. If no spots were detected in a given county the previous year, and few SPB are caught this year suggesting a declining trend, than the prediction would be static or low (rather than declining or low, since infestation levels cannot decline below 0). Trap catch data for all individual counties or ranger districts monitored within a State are averaged to make SPB predictions at the State level. Once all the data are received and processed, the Southwide predictions at the local and State levels are sent to each cooperator and also are made available on the Texas Forest Service Web page.
The accuracy of State-level predictions made since 1987 are summarized in this paper. The Southwide SPB Prediction System was initiated in 1987 in 11 Southern States (Arkansas, Texas, Louisiana, Mississippi, Alabama, Georgia, Tennessee, Virginia, Florida, South Carolina, and North Carolina). Federal and State pest managers in several other States joined at a later date (Maryland in 1988, Oklahoma in 1996, Kentucky in 2000, Delaware in 2001, and New Jersey in 2002) and have participated ever since. Florida did not conduct SPB pheromone surveys from 1989 to 1994, but returned as a cooperator from 1995 to the present. At the end of each calendar year, SPB cooperators provide the authors with total numbers of SPB spots detected from aerial surveys in each county or ranger district monitored with pheromone traps, as well as for the entire State. These data are used to evaluate the accuracy of each year’s predictions and provide valuable information for making predictions for the next year.
For purposes of the SPB Prediction System and to evaluate its accuracy, the following variables are defined.
Trend:
- Increasing = > 25-percent increase in total spots in a county, ranger district, or State from previous year;
- Static = ≤ 25-percent change in total spots from previous year;
- Declining = > 25-percent decrease in spots from the previous year.
Level:
- Low = < 0.40 spots per 1000 acres of host type in a county, ranger district, or State during a given year;
- Moderate = 0.40-1.50 spots per 1000 acres of host type;
- High = > 1.50 spots per 1000 acres of host type;
- Severe outbreak = combined with increasing or high level for purposes of this paper.
Error:
- Minor = discrepancy between prediction and actual trend or population level was a single category, (e.g., prediction was for static trend whereas actual trend was increasing; prediction was for low level whereas actual level was moderate, etc.);
- Major discrepancy between prediction and actual trend or level was two categories, (e.g., prediction was for declining trend whereas actual trend was increasing or outbreak, prediction was for low population level whereas actual level was increasing or outbreak, etc.).
Host Type:
- Total acres of loblolly, shortleaf, and slash pine (Pinus elliottii) from 1990 to 2000 Forest Inventory and Analysis data for a specific State, county, parish, or national forest ranger district.
Encyclopedia ID: p3291
