Sébastien Bélanger ', Éric Bauce ', Richard Berthiaume ', Bernard Long2, Jacques Labrie2, Louis-Frédéric Daigle2 and Christian Hébert3
'Département des sciences du bois et de la forêt, Faculté de foresterie, de géographie et de géomatique, Pavillon Abitibi-Price, Université Laval, Québec, Canada
Institut National de la Recherche Scientifique, Centre Eau, Terre et Environnement, Québec, Canada
Natural Resources Canada, Canadian Forest Service, Laurentian Forestry Centre, Québec, Canada
Résumé
En Amérique du Nord, le longicorne noir, Monochamus scutellatus scutellatus (Say), est l'un des insectes xylophages causant le plus de dommage dans la forêt boréale suite au passage d'un feu. Au Canada, la récupération du bois dans les brûlis peut contribuer à maintenir un volume de bois adéquat requis par l'industrie forestière, mais les larves de cet insecte occasionnent d'importantes réductions de la valeur économique des produits de sciage. Cette étude visait à estimer la progression des dommages en fonction de la température dans des sections troncales d'épinette noire {Picea mariana Mill.) et de pin gris {Pinus banksiana Lamb.). Grâce à la technologie de la tomodensitométrie axiale (CT- Scan), nous avons modélisé les taux de développement sous-corticaux et de progression de la profondeur des galeries en fonction de la température pour les deux essences. En général, ces taux étaient légèrement plus élevés chez l'épinette noire que chez le pin gris. Les œufs pondus sur les bûches placées dans la chambre de croissance à 12°C n'ont pas éclos et les larves ne se sont jamais développées sous l'écorce. La progression des galeries dans les bûches soumises à 20°C ont montré différents patrons de progression indiquant que les larves peuvent entrer en diapause sous l'écorce ou dans l'aubier. À 28°C dans l'épinette noire, 21 jours ont été nécessaires aux larves afin d'entrer dans le bois d'aubier et les galeries ont progressé à un taux de 1,441 mm/jour. Lorsque les essences d'arbres ont été combinées, les seuils de température de 13,6°C et 15,1°C ont été obtenus respectivement pour le développement sous-cortical et la progression des galeries. Les larves ont pénétré rapidement le bois d'aubier des arbres brûlés lorsque la température atteignait 24 ou 28°C. À la lumière de ces résultats, des dommages importants peuvent survenir l'année même suivant le passage des feux soulignant ainsi l'importance de débuter la récupération du bois aussitôt que possible afin de limiter les pertes économiques.
Abstract
The whitespotted sawyer, Monochamus scutellatus scutellatus (Say), is one of the most damaging wood-boring insect in recently burned boreal forest of North America. In Canada, salvage logging can contribute to maintain timber volumes required by the industry but larvae of this insect cause important damage reducing the economic value of lumber products. This study aimed to estimate damage progression as a function of temperature in recently burned black spruce {Picea mariana Mill.) and jack pine {Pinus banksiana Lamb.) trees. Using the axial tomograph technology (CT-Scan), we modeled subcortical development and gallery depth progression rates as function of temperature for both tree species. Generally, these rates were slightly higher in black spruce than in jack pine logs. Eggs laid on logs placed at 12°C did not hatched or larvae were unable to establish under bark as no larval development was observed. Gallery depth progression at 20°C showed various patterns indicating that larvae may enter diapause under bark or into sapwood. At 28°C in black spruce, 21 days were necessary for larvae to enter into sapwood and then, the gallery depth progressed at a rate of 1.441 mm/day. When tree species were combined, temperature thresholds of 13.6°C and 15.1°C were obtained respectively for subcortical development and gallery depth progression. Whitespotted sawyer larvae enter rapidly into sapwood of burned trees when temperatures reached 24°C or 28°C. Thus, important damage may occur within few months after fire, highlighting the importance of beginning salvage logging as soon as possible after wildfires to reduce economic losses.
Introduction
Salvage logging is considered as an efficient way to minimize the economic impact of large-scale wildfires in all forested regions of the world (Lowell et al., 1992; Saint-Germain and Greene, 2009). In USA, this is an important issue because of the reductions in land base available for timber harvesting (Lowell et al., 1992). In Canada, it is viewed as a measure that can contribute to maintain timber volumes required by the industry, particularly within the context of recent reductions in the annual allowable cuts (Lowell et al. 1992; Saint-Germain and Greene, 2009).
Insects are major agents of deterioration of fire-damaged and fire-killed trees (Lowell et al., 1992). They rapidly attack trees, loosen the bark by feeding on phloem and some species, such as the whitespotted sawyer, Monochamus scutellatus scutellatus (Say) (Coleoptera: Cerambycidae), excavate deep galleries into the wood (Raske, 1972). This insect is considered as the most damaging xylophagous species in recently burned boreal forest of Canada (Gardiner 1957). Cerezke and Volney (1995) reported up to 30% losses in wood value from downgrading in several burned forests of the prairie provinces in Canada.
Monochamus s. scutellatus is widely distributed throughout North America, except in the central part of USA, where it attacks pines {Pinus spp.), spruces {Picea spp.) and balsam fir {Abies balsamea (L.) Mill) (Raske, 1972; Cerezke, 1977; Yanega, 1996). Its life cycle is completed in one year in the southern part of its range but it may takes 2 years in the northern part (Peddle, 2000). In Alberta, Cerezke (1977) reported that 25% of the individuals went through a 1-yr life cycle; in the boreal forest of Québec, only 10% completed their life cycle in 1-yr (unpublished data). Other Monochamus species also have a 1 and/or 2-yr life cycle. For example, Monochamus carolinensis (Olivier) has a 1-yr life cycle in eastern USA (Pershing and Linit, 1986b) and Monochamus saltuarius Gebler, which is distributed in East Asia, Russia and central/eastern Europe, has a 1-yr life cycle but some individuals require two years (Togashi et al., 1994). Most species of Monochamus enter an obligate diapause during their last larval instar but species or individuals going
through a 2-yr cycle also need to enter diapause at an earlier larval instar (Rose, 1957; Togashi, 1991).
Larvae of 1st and 2nd instars feed on the subcortical tissues while the 3rd instar begins to excavate into the sapwood near the end of the first summer (Raske, 1972). After overwintering, larva resumes activity and continues to dig towards the heartwood before turning back during the second summer and digging back up to 5 mm from outer bark where it prepares a pupal cell and overwinters as a 4th instar larva (Peddle, 2000). This forms a U-shaped gallery which is characteristics of the Monochamus genus (Wilson, 1962; Alya and Hain, 1985; Akbulut et al., 2004). Larvae pupate in early spring (May) of the second year and adults emerge from the bole at the beginning of the third summer (June- July), thus completing the 2-yr life cycle.
The preparation of salvage logging plans requires time and often the construction of new roads to reach inaccessible burned forests, particularly in the boreal region. Meanwhile, woodborer deposit eggs and larval development start and progresses rapidly because temperature in burned forest is much warmer than in unburned forest (Hossack et al., 2009). Therefore, it is a race against time to salvage burned trees before woodborer's galleries reduce wood value to such an extent that it would become worthless to salvage. The expected climate warming will reduce the temporal window where salvage logging will be profitable, thus increasing pressure on forest managers for taking rapid and efficient decisions. If several studies have estimated economic losses caused by woodborer damage on fire-killed trees (Richmond and Lejeune, 1945; Prebble and Gardiner, 1958), none has addressed the issue of predicting damage progression over time. This crucial information for improving management of burned forests is the aim of our study. It requires studying damage progression into the wood as a function of temperature in order to acquire basic information that might be incorporated later into a predictive model.
Considering the small size of tunnels and the speed at which damage progresses within the bole, one should rely on a non-destructive and highly precise technology to characterize the process. Axial tomography (CT-Scan) is a non-destructive technology that has been
originally designed for medical applications but it is now used for other purposes. For instance, sedimentologists use this technology to identify and quantify the space occupied by benthic organisms in marine sediments (Gagnoud et al., 2009). More recently, Jennings and Austin (2011) also used a micro-CT scanner to characterize the shape and position of tunnels formed by larval xiphydriid woodwasps into the wood. In our study, we used 3-D high resolution images processed by CT-scan in order to follow whitespotted sawyer larval galleries inside logs over time. Specifically, the objectives of our study were to determine 1) the time of larval entrance into the sapwood of burned black spruce {Picea mariana Mill.) and jack pine {Pinus banksiana Lamb.) logs to determine subcortical development rates and 2) the progression of gallery depth to estimate ongoing damage, both a as a function of temperature and tree species.
Materials and methods
Origin of logs used in the experiment. To realize this experiment, burned black spruce and jack pine trees were cut on July 16 2008 in a 10 ha bum (50° 19 'N; 71° 10' W) located near Peribonka Lake at 200 km north of Lac St-Jean, Québec, Canada. Fire was ignited on July 12 by lightning and fire severity was low, with trees still having a green crown in the higher canopy but reddish needles in the lower canopy and bark being slightly carbonized. The fire bum mostly mature stands of >120 years old where black spruce was the most abundant tree species (70%) but where jack pine was also a significant component (30%). For each species, seven trees with a diameter at breast height (hereafter DBH) of 15-20 cm were randomly selected and cut down. Five log sections of 40 cm long were cut between 0.5 and 2.5 m from each bole. A total of 35 black spruce logs and 35 jack pine logs were thus collected and then brought back to the Laurentian Forestry Centre in Quebec City and stored in a cold room at -4°C. The ends of each log were waxed with paraffin to reduce moisture loss. After that, logs were stored at -40°C during three days to kill any potential woodborers.
Whitespotted sawyer breeding. Whitespotted sawyer adults were collected on recently cut black spruce and balsam fir log piles between the 6 and 18 July 2008 at the Montmorency Forest, the experimental forest of the Université Laval (47°19'N; 71°08'W). Each specimen collected was kept individually into a 150 ml plastic bottle. In the laboratory, isolated specimens were fed with a balsam fir twig placed into each bottle. Bottles were stored in a growth chamber at 12°C to keep adults alive as long as possible. Breeding cages consisted of a transparent plastic container (15 x 10 x 50 cm) with one screened opening on each side. One 40 cm log of black spruce or jack pine was randomly selected and placed within one of these cages. Three males and two females of the whitespotted sawyer were placed into each cage during three days for mating (if not already done on the field) and oviposition on logs. Two balsam fir twigs were inserted inside each container as food and water was sprayed in the cages every day. During this 3- day period, rearing cages were stored at 22°C with a photoperiod of 14L:10D. Meanwhile, any died specimen was replaced by another of the same sex.
Rearing and CT scanning. After the aforementioned 3-day period, seven logs of each tree species were stored in one growth chamber at a specific temperature (12, 16, 20, 24 and 28°C), a constant relative humidity (60%) and a 14L:10D photoperiod. In order to avoid any bias, a limited randomization was used to attribute the 5 logs of each tree to different growth chambers, thus ensuring that logs of each tree would be exposed to each temperature. Four weeks after the beginning of the experiment, logs were scanned for the first time using a Siemens Somatom Volume Access scanner. Afterwards, logs were scanned regularly but at different frequencies depending on temperature at which logs were placed. A permanent mark (a pin) was put on each log to ensure consistent positioning and alignment of the log at each scanning session. During such session, a log is placed on the scanner bed which slides into a gantry made of a round tube equipped with an X-ray source in one part and by 32 arrays of 600 detectors each on the opposite part. In a rotating movement, the detectors receive the x-ray radiations in different angles (each 5 degrees) around the log and produce a series of 2D transversal images (Ketcham and Carlson, 2001; Michaud et al., 2003). Each picture is the result of the variable permeability of different materials to x-rays, mostly due to their density (Pierret et al., 2002). This
measured X-ray intensity is attenuated through material following the Beer-Lambert law [1]:
I = Ioe-V [1] where I is the attenuation intensity captured by receivers, Io represents the emitted
radiation intensity, x is the length of the X-ray path trough the material, and \AS is the mass absorption coefficient of the sample. When this coefficient is compared with the coefficient IT wat_r in the equation [2], we get the tomographic intensity (TI) expressed in Housefield units (HU) for each scanned point (or voxel).
T I = [ ( us/ U w a t e r ) - l ] 1 0 0 0 [2]
By comparing TI with the TIwater (TIwater= 0 and TIajr = -1000 by convention), the resulting TI corresponds to a tomographic density value which varies as a function of the wood material and empty spaces into galleries (Capowiez et al., 1998; de Montety et al., 2003; Michaud et al., 2003). Tomographic intensity of each corresponding voxel (volumetric pixel) was transmitted to the CT-scan computer of the scanner and was represented on a gray scale of 4096 values with 1 value per pixel (Michaud et al., 2003). About 1000 axial slices of 0.6 mm thickness were obtained per 40 cm long log with a pixel resolution of 0.4 x 0.4 mm on a matrix of 512 x 512 pixels. A radiation energy of 140 kV at an intensity of 40 mA was used (Dufour et al., 2005).
Data analysis. Pictures of each log were reconstructed in 3D with a program developed using Matlab, a software developed for treating medical images. Using successive pictures of the same logs made it possible to identify each gallery and measure its depth (from inner bark toward center in mm), which was considered as the best estimate of woodborer damage. It also allowed estimating the time of entrance of each larva into the wood as well as their progression toward tree center at different rearing temperatures and tree species. Logs in which whitespotted sawyer females have not laid eggs or in which it was not possible to monitor gallery progression at each scanning session were discarded from the analysis (Table 2-1).
Statistical analysis. Temperature-dependent development rates of insects are usually nonlinear (Hagstrum et Milliken, 1988; Subramanyam et Hagstrum, 1993) and the progression of Monochamus s. scuttelatus galleries at different temperatures also followed a sigmoid curve (see Figure 2-1). Thus, the nonlinear Chapman-Richards model, often applied to describe forest growth (Pienaar and Turnbull, 1973; Zhao-gang and Feng-ri, 2003), was used and was expressed as [1]:
Y = a ( l - e x p ( - n x ) f [1] where Y is the depth of galleries, x is the number of days since the beginning of the
experiment, _* is a parameter controlling the maximum depth of galleries (curve asymptote), 77 is a parameter expressing the progression rate of the gallery and 6 is a parameter shaping the curve. As log was our experimental unit, we first determined the average depth of galleries in each log after each scan session. Then, we used these log averages to calculate tree species averages at each temperature. The nonlinear function was then applied to these averages to model gallery depth progression as a function of temperature for each tree species. Model performance was based on the number of logs in which the model converged, as well as on the average coefficient of determination (r2) and the mean square error (MSE) (Colbert et al., 2002). We used PROC NLIN (SAS Institute version 9.2), with the Gaussian-Newton iterative (maximum number of iteration=100) method (Fekedulegn et al., 1999), to find the best fit of each function. As SAS did not provided the coefficient of determination, it was calculated using the DataFit 9.0 computer program (Nowatzki et al., 2002). A multivariate analysis of variance (MANOVA) was conducted to compare model parameters in order to determine if gallery depth progression into sapwood varied similarly or not between tree species and temperature.
The Chapman-Richards function was used to estimate 1) larval subcortical development rate and to describe 2) the average depth progression rate of galleries at each temperature into the sapwood and xylem. The reciprocal of the time required to reach 5% of the maximum depth (1/time to reach 5% of maximum depth), determined from each Chapman- Richards function, was used as the larval subcortical development rate. The depth progression rate of galleries was estimated as the reciprocal of the time elapsed for tunneling between 5 and 95% of the maximum depth on curves for each temperature.
Linear regressions were used to model larval subcortical development rate and gallery depth progression rate as function of temperature and tree species. These models allowed estimating temperature thresholds for subcortical development and gallery depth progression for each tree species.
Results
Model fitting. No gallery was dug into sapwood at 12°C and debarking did not revealed any subcortical larval development. The multivariate analysis of variance applied to parameters of the Chapman-Richards function showed that the average gallery progression depth was significantly influenced by temperature (F6>36 = 12.77, p = 0.001). The first root was responsible for 91% of the variability caused by temperature, in which the parameter 17, which expresses the depth progression rate of the gallery, had the strongest influence (Eigenvector = 4.5), followed by a (Eigenvector =0.025) and 3 (Eigenvector of - 0.00000449) which had weak or almost no influence. No significant effect was detected for tree species (F3J8 =2.11, p = 0.1349), but a trend favouring faster progression into black spruce was noted at 16°C and 20°C (Figure 2.1 C-D). Therefore, we decided to compute equations for each tree species. The depth progression rate at 20°C was clearly slower and much more variable than at any other temperature for both tree species (Figure 2-1). It will be presented in more details in the next section.
Chapman-Richards model. The high coefficients of determination obtained by Chapman- Richards model revealed good fits of this function to our data (r =0.96 to 0.99). At 16, 24 and 28°C, the curves were similar for both tree species but gallery depth progression was always higher in black spruce (Figure 2-1). Furthermore, the maximum depth reached by whitespotted sawyer larvae increased with temperature. At 16°C, the maximum gallery depth averaged 25.8 and 21.7 mm for black spruce and jack pine respectively (Figure 2-
1D). At 24°C, it averaged 42.3 mm for black spruce and 38.2 mm for jack pine (Figure 2- 1B), while at 28°C the maximum gallery depth was 51.8 mm in black spruce and 46.6 mm in jack pine (Figure 2-1A).
At 20°C, the larval gallery depth progression was abnormally lower and was not similar between tree species; the adjustment of Chapman-Richard's curves to our data was difficult to obtain and consequently, this model function was inadequate to model gallery depth progression as a function of time (Figure 2-1C). By analyzing graphs presenting depth progression for each gallery, we observed that they should be classified into two groups (Figure 2-2).
1) Early entrance into the sapwood: slightly more than half of the larvae entered rapidly into the sapwood; between days 30 and 70 in black spruce and slightly later in jack pine, between days 48 and 133. In black spruce, galleries stopped their depth progression for about 80 days (Figure 2-2A) but resumed after day 133. In jack pine, after early larval entrance, galleries stopped depth progression or continued very slowly (Figure 2-2C).
2) Late entrance into the sapwood: slightly less than half of the larvae entered late