- Top of page
- Materials and methods
The spruce bark beetle Ips typographus (L.) is one of the most important pests of mature Norway spruce Picea abies L. (Karst.) in Europe (Christiansen & Bakke, 1988). At endemic population levels, I. typographus reproduces in breeding material with no or weak defences such as wind-felled or otherwise damaged trees. During outbreak conditions, the species is able to kill living trees in large numbers. The male initiates the attack by boring into the bark and releasing an aggregation pheromone, which attracts both sexes. Each male is generally joined by two to three females. Each female excavates an egg gallery under the bark. After having established the egg galleries, the parent beetles may leave to establish a second brood later in the same summer. In Sweden, I. typographus is generally univoltine.
Storm disturbance is one of the most important factors triggering outbreaks of I. typographus (Christiansen & Bakke, 1988; Schroeder, 2001; Bouget & Duelli, 2004). Storms provide I. typographus with a sudden surplus of breeding material, with no or weak defences. Hence, the densities of I. typographus egg galleries in the colonized wind-felled trees will be low, which results in longer egg galleries (more eggs laid per gallery) and that the larvae are released from the strong intraspecific competition they usually experience (Anderbrant et al., 1985; Anderbrant, 1990). This means that the number of progeny per female increases. Thus, in the summers after a storm-felling, the beetle populations may increase to densities high enough for the beetles to overcome the defence of living trees (Christiansen, 1985; Mulock & Christiansen, 1986; Christiansen et al., 1987).
Salvage of wind-felled trees is the main forest management priority directly after storm disturbances. The two main reasons for salvage operations are: (i) to save the value of the fallen trees before their quality are degraded by bark and ambrosia beetles and (ii) to reduce the risk for subsequent tree mortality caused by I. typographus as a result of propagation in the wind-felled trees. However, after large-scale storm disturbances, it might not be possible to salvage all wind-felled trees before the flight period of I. typographus starts or even before the emergence of the new generation. In this situation, it would be of great value for the forest owners to be able to assess the risk for bark beetle colonization of wind-felled trees in different kinds of storm-damaged stands. Models assessing this risk, based on easily accessible environmental variables, could be used to optimize the use of limited logging resources (i.e. to decide which damaged stands to log first). Risk assessments could also be helpful for choosing stands in which wind-felled spruces could be retained to increase the amount of dead wood with a low risk of subsequent tree mortality. Dead wood is an important substrate for many threatened species (Gärdenfors, 2005). Also, management decisions for protected areas, such as national parks and nature reserves, hit by storm disturbances may be improved. Such areas may host rare species, which requires the presence of mature spruce forest. Thus, if the risk for colonization of wind-felled spruces by I. typographus is high, and thus also the risk for subsequent attacks on the remaining living spruces, one option might be to remove some of the wind-felled trees or to make them unsuitable for bark beetle reproduction.
On 8–9 January 2005, southern Sweden was hit by the storm Gudrun and an estimated volume of 50–75 million m3 of forest were felled (Anonymous, 2006, 2007a). This was the largest storm damage ever recorded in Sweden (Nilsson et al., 2004). The volume of downed trees corresponded to three yearly fellings for southern Sweden and six yearly fellings for the most severely affected regions. The majority of the downed trees (80%) consisted of Norway spruce (Anonymous, 2006). This large scale of the storm disturbance presented an opportunity to compare the colonization patterns of I. typographus in different kinds of storm gaps spatially distributed over a large area. The studied gaps were selected to include as large variation as possible in size (i.e. the number of felled spruces) because gap size can be expected to influence immigration rates and thus also colonization probability of I. typographus.
The resource concentration hypothesis (Root, 1973) predicts that specialist herbivores should reach higher densities in large patches as a result of larger emigration rates from smaller patches and larger immigration rates into larger patches. Later studies have explored how the rate of successful migration into habitat patches may vary between species as a result of differences in search behaviour, mode of patch detection and mobility (Bowman et al., 2002; Englund & Hambäck, 2007). Newly-created storm gaps, with felled spruces, can be viewed as habitat patches for I. typographus. Because of the high dispersal capacity of I. typographus (Forsse & Solbreck, 1985) immigration is an important mechanism for the colonization of storm gaps, although, in some cases, local populations may already be present in the affected forest stands before the storm disturbance. The flight behaviour of I. typographus (and other bark beetle species) is not known in detail, although Forsse and Solbreck (1985) estimated, by extrapolation, that less than 14% of the population flew more than 20 m above the ground. This means that a negative relationship between patch size and immigration per unit area patch could be expected as a result of the probability of intercepting a patch being proportional to the linear dimension of the patch rather than the area of the patch (Bowman et al., 2002; Englund & Hambäck, 2007). In addition, after large-scale storm disturbances (resulting in a sudden large surplus of breeding material at the landscape level), a limited supply of potential immigrants should also result in a negative relationship between patch size and immigration per unit area.
Only one similar well replicated study, including 65 storm gaps in Finland, has been published (Eriksson et al., 2005). In that study, the mean diameter of wind-felled spruces and the basal area of recently dead standing spruces were the most important variables influencing the percentage of wind-felled spruces colonized by I. typographus. These gap variables were also included in the present study. Because of the high dispersal capacity of I. typographus (Forsse & Solbreck, 1985), the composition of the surrounding landscape can also be expected to influence the colonization probability in a specific storm gap. Thus, in addition to gap variables, other variables reflecting the landscape around the gaps were included in the present study.
The present study aimed to determine which gap and landscape characteristics are correlated with the colonization of wind-felled trees of I. typographus in storm gaps. Both colonization percentage and egg gallery density were included as response variables in the analyses. In Scandinavia, I. typographus can reproduce in unattacked wind-felled trees in the two first summers after storm disturbance (Göthlin et al., 2000; Eriksson et al., 2005). The influence of gap and landscape variables on the colonization of I. typographus was thus analysed for the first summer (2005) and for the first and second summer pooled (2005 + 2006). There were two reasons for the analyses of the two pooled years instead of treating 2006 separately: (i) in an earlier study, the sum of the wind-felled trees colonized in the first and second summer after a storm was almost perfectly correlated with the subsequent tree mortality occurring in each stand (Schroeder & Lindelöw, 2002) and (ii) the sum of colonized trees represents the loss of timber quality caused by I. typographus in storm gaps not salvaged until after the second summer.
- Top of page
- Materials and methods
In 2005, 18 (50%) of the 36 gaps were colonized by I. typographus (Table 1). In 2006, this figure increased to 35 (97%). The percentage of bark samples colonized by I. typographus was low in the first summer and increased strongly in the second summer. The mean ± SD percentage of colonized bark areas per gap was 3.5 ± 5.4 in 2005 and 44.6 ± 21.6 in 2005 + 2006. The mean percentage of colonized trees (i.e. at least one of the bark areas colonized) was slightly higher than colonized bark areas; 4.8 ± 7.0 in 2005 and 50.1 ± 21.3 in 2005 + 2006. The mean ± SD attack density (egg galleries per m2 bark) of I. typographus, only including colonized bark areas, was 58.5 ± 34.9 in 2005 and 78.5 ± 25.8 in 2005 + 2006. A total number of 255 egg galleries from 2005 and 3910 from 2006 were recorded. The mean ± SD percentage of bark samples colonized, including also other bark-and wood-boring insects, was 80.4 ± 16.0 in 2005 + 2006. Of these other species, P. chalcographus dominated. The mean ± SD percentage of bark areas colonized per gap by this species was 54.0 ± 19.8. The mean ± SD percentage of colonized bark areas covered by galleries of P. chalcographus was 46.7 ± 17.0.
Significant models were obtained for all four dependent variables (Table 3). For the probability of bark area being colonized and the density of egg galleries in 2005, the only predictor variable included in the final model was the landscape variable STORM GAP_2000 m. An increase in area of storm gaps in the adjacent landscape decreased the probability of colonization and the egg gallery density (Fig. 1). The models accounted for a rather small part (13%) of the observed variation. For the ten gaps with ≥ 56 ha of STORM GAP_2000m, only one was colonized, whereas, for the 26 with < 56 ha of STORM GAP_2000m, 17 gaps were colonized (Fig. 1A).
Table 3. Variables included in the four final models and the results of the statistical analyses
|Models and variables included||Parameter estimate||Standard error||F-value||P-value||R2|
|Probability of bark area being colonized 2005||—||—||4.95||0.033||0.127|
|Probability of bark area being colonized 2005 and 2006||—||—||7.48||0.0002||0.491|
|BASAL AREA LIVING SPRUCE||−0.02734||0.01223||—||0.033||—|
|Density of egg galleries 2005||—||—||4.77||0.036||0.123|
|Density of egg galleries in 2005 and 2006||—||—||5.25||0.028||0.134|
Figure 1. Relationship between the area of storm gaps within 2000 m (STORM GAP_2000m) and (A) the percentage of bark areas colonized by Ips typographus and (B) the density of I. typographus egg galleries in the first summer (2005) after the storm disturbance. Egg gallery densities include both colonized and not colonized bark areas. Each point represents one gap.
Download figure to PowerPoint
For the probability of colonization in 2005 and 2006, the variables included in the model were DIAMETER (positive relationship), BASAL AREA LIVING SPRUCE (negative), DOM SPRUCE_1000m (negative) and VOL SPRUCE_2000m (positive) (Table 3). The model explained almost half (49%) of the variation between gaps. The first variable to be included in the model was VOL SPRUCE_2000m, which explained 16% of the variation (Fig. 2A). For the density of egg galleries in 2005 and 2006, the only predictor variable included in the final model was VOL SPRUCE_2000m (Table 3). An increase in volume of living spruce in the adjacent landscape increased the egg gallery density (Fig. 2B). The model accounted for a rather small part (13%) of the observed variation. There was a significant correlation between observed and predicted percentages of bark areas colonized (R2 = 0.48; P < 0.001) (Fig. 3).
Figure 2. Relationship between the volume of living spruce spruce forest older than 40 years within 2000 m (VOL SPRUCE_2000m) and (A) the percentage of bark areas colonized by Ips typographus and (B) the density of I. typographus egg galleries in the first and second summer (2005 + 2006) after the storm disturbance. Egg gallery densities include both colonized and not colonized bark areas. Each point represents one gap.
Download figure to PowerPoint
Figure 3. Correlation between predicted and observed percentages of colonized bark areas in 2005 and 2006. The predicted percentages were based on a model including the variables DIAMETER, BASAL AREA LIVING SPRUCE, DOM SPRUCE_1000m and VOL SPRUCE_2000m. Each point represents one gap.
Download figure to PowerPoint
The number of wind-felled spruces (Fig. 4) or the other variables reflecting gap size were not significantly correlated with I. typographus colonization.
Figure 4. The percentage of bark areas colonized by Ips typographus in relation to the number of wind-felled spruces per gap (NO. STORM-FELLED SPRUCES) in (A) the first summer (2005) and in (B) the first and second summer pooled (2005 + 2006) after the storm disturbance. The variable was not included in any of the final models. Each point represents one gap.
Download figure to PowerPoint
- Top of page
- Materials and methods
The results obtained in the present study demonstrate that both gap and landscape variables influence the colonization of wind-felled trees by I. typographus. In the first summer after the storm disturbance, the area of storm gaps within a radius of 2000 m (STORM GAP_2000m) was the only predictor variable significantly associated with colonization by I. typographus. When the data from the first and second summer are combined, both gap (DIAMETER and BASAL AREA LIVING SPRUCE) and landscape variables (DOM SPRUCE_1000m and VOL SPRUCE_2000m) were included in the final model for probability of colonization of bark areas. Thus, the inclusion of landscape variables in the analyses increased the explanatory power of the models. An earlier similar study conducted by Eriksson et al. (2005) in Finland reported that the mean diameter of wind-felled spruces and the basal area of recently dead standing spruces were the most important gap variables influencing the percentage of colonized wind-felled spruces in storm gaps by I. typographus (no landscape variables analysed). In the present study, basal area of recently dead standing spruces was not included in the final models.
The amount of breeding material in the surrounding landscape (estimated by the landscape variable STORM GAP_2000m) was negatively correlated with both the probability of colonization and the density of egg galleries of I. typographus in the first summer. Obviously, the immigration of beetles to gaps located in areas with large amounts of wind-felled trees (i.e. large areas of storm gaps) decreased as a result of the large quantities of alternative breeding material. The beetles responded to this variable at a scale of at least 2000 m (larger scales not tested), which indicates rather long dispersal distances, even in a situation with a large surplus of breeding material. This is the first time that a negative relationship between amounts of alternative breeding material and probability of colonization by I. typographus has been demonstrated at a landscape scale. The landscape variable STORM GAP included all storm gaps irrespective of whether they had been cleared or not before the summer (information about when individual storm gaps were cleared was not available). Because a considerable part of the wind-felled was salvaged before the flight period of I. typographus in 2005, the variance explained possibly would have been greater if these cleared areas could have been excluded from the analysis.
The first variable (and the only one for egg gallery density) included in the models for the colonization in the first and second summer pooled was the landscape variable VOL SPRUCE_2000m. The positive relationship with this variable could be a result of areas with high volumes of mature living spruce forest sustaining larger populations of I. typographus and thereby increasing the number of beetles immigrating into the gaps. In an earlier study, conducted after a large storm disturbance in France, Gilbert et al. (2005) demonstrated a positive relationship between colonization probability and the number of coniferous forest patches within 5000 m but not with total coniferous area. In the present study, there was also a positive relationship between the gap variable DIAMETER and the probability of colonization. This is in accordance with the study by Eriksson et al. (2005) and can be explained by the preference of I. typographus for large-diameter trees (e.g. Butovitsch, 1971; Schroeder et al., 1999; Göthlin et al., 2000). In the study by Eriksson et al. (2005), the minimum diameter requirement of sample trees was 10 cm compared with the 15 cm reported in the present study. This may explain the stronger influence of mean diameter in their previous study. There are no obvious explanations for why the two variables BASAL AREA LIVING SPRUCE and DOM SPRUCE_1000m, with negative signs in the model, should influence the probability of colonization.
The percentage of trees colonized by I. typographus in the first summer was much lower (5%) in the present study compared with that reported from earlier storm disturbances in Scandinavia. The result obtained in the present study is, however, in accordance with the findings from three other estimates conducted in southern Sweden in 2005 (the first summer after the storm) in which 2–4% of the wind-felled spruces were colonized by I. typographus (Anonymous, 2006; Schroeder et al., 2006). After the second largest storm-felling recorded for Sweden, amounting to 36 million m3 (Nilsson et al., 2004), 12% of the wind-felled trees were colonized by I. typographus in the first summer (Lekander, 1971), whereas, for smaller storm disturbances, when generally most of the trees have been salvaged before the summer, the percentage of colonized trees often exceeds 20% (Schroeder & Lindelöw, 2002; Eriksson et al., 2005). The low percentage of colonized trees in the present study could be explained by the large volume of wind-felled trees still remaining in the forest landscape during the flight period in 2005 and a low initial population level of I. typographus. Tens of millions m3 of wind-felled trees remained unsalvaged during the summer and considerable quantities of the salvaged trees were stored at forest roads as a result of a lack of transportation capacity (Anonymous, 2006). In addition, the catches of I. typographus in monitoring traps were relatively low in 2004 (the summer before the storm), indicating that the initial population levels were low (Anonymous, 2007b).
The percentage of bark areas and trees colonized by I. typographus increased strongly from the first (4% of bark areas, 5% of trees) to the second summer (45% of bark areas, 50% of trees). This demonstrates that, in Fennoscandia, wind-felled spruce trees can be utilized to a large extent also in the second summer if not attacked in the first summer. This is in accordance with earlier studies, even though the percentage of wind-felled trees colonized in the second summer were lower in those studies as a result of a higher colonization percentage already in the first summer (Göthlin et al., 2000; Eriksson et al., 2005).
There was no significant relationship between the number of storm-felled spruces and the probability of colonization by I. typographus in the first summer despite the large variation in gap size. Nor in the study by Eriksson et al. (2005) was the number of wind-felled spruces significantly related to the percentage of trees colonized in the first summer (gap sizes in the range 1–3284 wind-felled spruces). As a result of what is known about the flight behaviour of I. typographus, and the large surplus of wind-felled trees in the first summer, a lower probability of colonization for the largest gaps could have been expected (see Introduction). However, another factor, the timing of colonization, may have differed between small and large gaps and counteracted the effect of the flight behaviour. Large gaps should on average be colonized by immigrants somewhat earlier than small gaps as a result of their larger sizes and they should also have a higher probability of hosting local populations before the storm disturbance. Thus, the release of aggregation pheromones can be expected to start earlier, from large than from small gaps, resulting in an increased immigration rate (cf. Englund & Hambäck, 2007). Ips typographus is only weakly attracted by host volatiles alone (odours from wind-felled spruces in the gaps), whereas the aggregation pheromone, released after colonization, is highly attractive, and there is also a synergistic effect on attraction of the combination of host volatiles and aggregation pheromones (Austarå et al., 1986; Lindelöw et al., 1992; Schroeder, 2003; Erbilgin et al., 2007). In addition, the overall low percentage of colonizations, resulting in a low variability in the response variable, could have contributed to the lack of a significant relationship in the first summer. In the second summer, beetles produced in the large amounts of wind-felled spruces that remained in the forest landscape during the first summer may have immigrated into the gaps, thereby ensuring a high percentage of colonization also in the largest gaps.
After large-scale storm disturbances, large quantities of fallen trees may remain in the forest during the first summer and, in some cases, even during the second, as a result of limited logging resources. Thus, in such a situation, the foresters have to decide which storm gaps to salvage first. Important factors to consider when making this decision include: (i) the possibility to optimize the use of the limited logging resources with respect to amounts of salvaged trees per unit time; (ii) the increased risk for subsequent tree mortality caused by I. typographus in the subsequent years in areas not salvaged; and (iii) the risk for degradation of the quality of the fallen trees by bark and ambrosia beetles vectoring blue stain fungi. In Sweden, almost all cuttings of storm-felled trees are conducted with harvesters. Harvesters are most efficiently used in areas with many large storm gaps where less time is spent on transportations and preparations before logging can start. In earlier studies, the risk for subsequent tree mortality in stands bordering storm gaps with retained wind-felled spruces has been demonstrated to be linearly correlated with the number of colonized wind-felled spruces when the amount exceeds a threshold of some tens of colonized trees (Schroeder & Lindelöw, 2002; Hedgren et al., 2003; Eriksson et al., 2007). For very large areas with fallen trees, there may also be a second threshold above which the tree mortality will be even higher (Schroeder, 2001). Thus, this factor also implies that areas with many large storm gaps should be salvaged first. In Fennoscandia, the storm-felled trees generally represent a much higher value than the living trees subsequently killed by I. typographus. Thus, a high priority should be given to salvage as many as possible of these trees before they start to degrade. Ips typographus, acting as a vector for blue-stain fungi, is one of the major factors causing wood degradation. Both in the present and in an earlier study (Eriksson et al., 2005), there was no relationship between the percentage of colonized trees and gap area. This means that when considering all the three factors discussed above, areas with many large storm gaps should be salvaged first. Among such areas, the highest priority should be given to gaps with a high mean diameter of the fallen trees (Eriksson et al., 2005; present study), comprising areas known to have harboured large local populations the year before the storm disturbance (Eriksson et al., 2005) and areas with large volumes of mature spruce forest within 2000 m (present study). The present study demonstrated a negative relationship between percentage of colonized bark areas per gap in the first summer and the areas of storm gaps within 2000 m in the surrounding landscape, which, to some extent, appears to contradict the recommendation to salvage areas with many large storm gaps first. The relatively low explanatory power of this variable, however, implies that larger weight should be given to the factors discussed earlier.