Effects of mobility on daily attraction to light traps: comparison between lepidopteran and coleopteran communities

Authors


Toshihide Hirao, Tomakomai Research Station, Hokkaido University Forests, Takaoka, Tomakomai, Hokkaido 053-0035, Japan. E-mail: hirao@fsc.hokudai.ac.jp

Abstract

Abstract. 

  • 1Light traps have been widely used for surveying insect community structure, but some biases are known to occur under certain meteorological conditions.
  • 2In addition to weather factors, we raise the novel hypothesis that if the daily movement distance of focal insects is shorter than the effective range of the light trap, then the species richness and abundance of the daily captures will increase during the course of the night.
  • 3This study examined the daily attraction patterns of Lepidoptera and Coleoptera to light traps and the factors affecting the attraction of these communities. Light traps were run for three consecutive nights in each month from April to October 2001 in a cool-temperate forest in Japan.
  • 4The species richness and abundance of Coleoptera increased during trap nights, whereas lepidopteran captures remained constant. Meteorological factors influenced the capture of both communities throughout the sampling season, but the daily increment in Coleoptera was not explained by the daily trends in weather conditions.
  • 5We argue that the daily augmentation of Coleoptera capture rate results from the daily movement distance of Coleoptera being generally shorter than the effective and perceivable range of the light traps. These results suggest that consideration of the typical daily movement of a focal taxon is required when conducting biological monitoring using light trap sampling.

Introduction

Pervasive species loss and anthropogenic alteration of natural habitats in recent years have emphasised the need to monitor changes in species diversity as one of the features of biological communities. The forest entomofauna, in particular, is one of the dominant components of species diversity in terrestrial ecosystems (Dajoz, 2000), and there have been frequent calls for the rigorous quantification of insect biodiversity (Samways, 2005). In temperate and tropical forests, Lepidoptera and Coleoptera are two major insect taxa of high species diversity (Schowalter et al., 1986; Novotny & Basset, 2005). Thus, illuminating the factors influencing their distribution and abundance is considered particularly important for the conservation of biodiversity.

Light traps have been widely used for monitoring Lepidoptera and Coleoptera population dynamics (e.g. Wolda, 1992; Watt & Woiwod, 1999; Kato et al., 2000) and surveying their community structure (e.g. Kato et al., 1995; Leps et al., 1998; Willott, 1999). Light trap sampling is relatively unbiased and more efficient than some other sampling methods (Raimondo et al., 2004), but sampling by light traps must be carefully designed, because sampling errors are caused by attractive efficiency under certain conditions, and, as with any sampling method, inherent biases must be accounted for (e.g. Holyoak et al., 1997).

The effects of meteorological conditions on the abundance or species richness of Lepidoptera caught in light traps have been particularly well examined (e.g. Holyoak et al., 1997; Yela & Holyoak, 1997). Trap efficiency for Lepidoptera is positively correlated with temperature and the thickness of cloud cover, and negatively correlated with wind speed, precipitation, and the fullness of the moon on the trap night (Bowden, 1982; Dent & Pawar, 1988; Yela & Holyoak, 1997; Butler et al., 1999). The daily variation in insects collected is therefore considered to mainly reflect changes in flight activity caused by weather conditions (Muirhead-Thomson, 1991), which has been demonstrated by comparisons between direct census counts and light trap captures (Riley et al., 1992; Quiring, 1994). On the other hand, few studies have directly addressed the effects of weather factors on the abundance or species richness of Coleoptera captured by light traps (Rodriguez-Del-Bosque & Magallanesestala, 1994; Rodriguez-Del-Bosque, 1998). Instead, bait traps and pitfall traps have been used to examine the effects of daily weather variation on the activity of Coleoptera (Davis, 1995; Honek, 1997), and have shown that the activity of coleopterans was influenced by temperature and wind speed as in lepidopterans, although it was positively correlated with rainfall or moisture (Rodriguez-Del-Bosque & Magallanesestala, 1994; Davis, 1995), which is in remarkable contrast to Lepidoptera.

However, we propose that a temporal factor related to movement and dispersal ability of organisms other than ones associated with meteorological conditions may also regulate the attractive efficiency of light traps. Specifically, it is hypothesised that shorter daily movement distance relative to the effective range of light traps may lead to increasing daily captures. For the purpose of clearly and logically building up this hypothesis, a schematic diagram is presented to explain the attraction process of the organisms with relatively shorter dispersal ability (Fig. 1). Under favourable conditions, the light emitted by traps should have far-reaching effects on insect attraction, and be concentrically radiated from traps. In the diagram, a light trap is placed at the centre and surrounded by three hypothetical concentric circles at equally spaced increasing distances, assuming that the dispersal distance (d m) of the organisms in one night is one-third of the effective range of the light trap (the outermost circle represents the effective range in Fig. 1). Black dots on the diagram denote individuals (J) which are hypothetically distributed at random. The triangular shaded area funnelling into the trap represents the area covering a typical flight path of individuals in a subregion of the circle. This triangular representation assumes that individuals roughly move along an average linearised trajectory towards the light trap, if it is assumed that they stop during the day and continue in their directed journey towards the trap during the night. The shaded area is divided to three subareas by the mobile distance in a night. Under this assumption of limited dispersal ability, individuals captured in the first night should reside in the innermost subarea (A in Fig. 1), and individuals trapped in the second and third night are in the middle subarea (B in Fig. 1) and outermost one (C in Fig. 1), respectively. The expected numbers of individuals in the subarea B (3aJ/9) and C (5aJ/9) are three and five times that expected in A (aJ/9), respectively, under the random distribution of individuals, because the subarea B = 3ad2π and C = 5ad2π, whereas A = ad2π, where a is the proportion of the triangular shaded area (A + B + C in Fig. 1) to the total area of the trap range. Thus, from this diagram, if the daily movement distance of focal insects is much shorter than the effective distance of light traps, the species richness and abundance of the entomofauna captured daily by light traps should increase in the course of a 3-day trapping period.

Figure 1.

Schematic diagram explaining the hypothesised attraction process of organisms with shorter dispersal ability relative to the effective range of a light trap. Under the assumption that the dispersal distance (d m) of the organisms in one night is one third of the effective range of the light trap (the outermost circle represents the effective range), a light trap is placed at the centre and surrounded by three hypothetical concentric circles at equally spaced increasing distances. Black dots on the diagram denote individuals (J) which are hypothetically distributed at random. The triangular shaded area funnelling into the trap represents the area covering a typical flight path of individuals in a subregion of the circle. The innermost subarea (A) should contain individuals captured in the first night, and then the middle subarea (B) and outermost one (C) should harbour individuals trapped in the second and third night, respectively. See the text for the details.

The daily movement distance would be governed by the life-history strategies and physiological traits of the taxon. For example, the daily movement of adult lepidopterans is mainly for mating and oviposition (Scoble, 1995), and then the flight capability and duration of Lepidoptera should be physiologically superior to Coleoptera (Wootton, 1992; Grodnitsky, 1999). However, the relationships between the sampling range of light traps and the daily movement of insects have never been investigated, although the effective range of pheromone-baited traps has been well studied (e.g. Schlyter, 1992; Dodds & Ross, 2002; Bacca et al., 2006). Thus, as a first step, the existence of the hypothetical patterns in daily catches should be examined by comparing the results between taxa that are clearly distinguishable in their flight capabilities.

In this study, the daily attraction patterns towards light traps in Lepidoptera and Coleoptera, and the factors affecting attraction were examined. First, the effects of seasonal meteorological factors (temperature, humidity and wind) and a date factor (order of consecutive trap nights) on species richness and abundance of Lepidoptera and Coleoptera captured daily by light traps were analysed. Then, to evaluate daily trends in meteorological factors, daily changes in these factors were analysed. The results are discussed by comparing the patterns of Lepidoptera and Coleoptera.

Materials and methods

Study site

The study was performed in the cool-temperate deciduous Tomakomai Experimental Forest (TOEF; 42°43′N, 141°36′E, c. 90 m above sea level) on Hokkaido, Japan, as part of the international project DIVERSITAS in the Western Pacific and Asia–International Biodiversity Observation Year (DIWPA-IBOY) to document biodiversity throughout the Western Pacific and Asian Region. IBOY is one of the core projects of DIWPA, an activity of DIVERSITAS, which is an international biodiversity science programme. TOEF, a primary research site for DIWPA-IBOY (cf. http://diwpa.ecology.kyoto-u.ac.jp/iboy.htm), has an average annual precipitation of 1161 mm and an average annual temperature of 5.6 °C. Oak (Quercus crispula), maple (Acer mono) and linden (Tilia japonica) dominate the forest. The canopy ranges from 15 to 25 m in height, and saplings of the dominant tree species grow on the forest floor. Deciduous trees break bud in early to mid-May and shed their leaves in late October.

Light trap sampling

The lepidopteran and coleopteran data collection for this study was performed in 2001. The sampling procedure followed the IBOY protocol for light traps (Nakashizuka & Stork, 2002). Three sets of two traps, one at ground level (1 m from the ground) and the other in the canopy (15.8, 16.3 or 17.1 m from the ground) were operated simultaneously at three randomly selected locations within a 1-ha plot (100 m × 100 m) in a forest stand. Trapping was repeated at monthly intervals from April to October 2001, and traps were run from about 17.00 hours throughout the night for 3 nights during each monthly sampling event. Sampling was avoided 1 week before and after the full moon and on rainy days. The traps were emptied daily during the trapping period. Thus, in total, 126 samples (6 traps × 3 days × 7 months) were collected. Adult insects were first sorted by order, and then lepidopteran and coleopteran specimens were identified to species and counted by taxonomic experts. H. Kogi identified and counted the Lepidoptera, and A. Kashizaki identified and counted the Coleoptera.

Meteorological observations

In this study, air temperature (°C; 1.5 m above ground – a.g.), relative humidity (percentage; 1.5 m a.g.), wind speed (m s−1; 33 m a.g.) and precipitation (millimetre) were recorded hourly as the meteorological factors potentially affecting the daily capture of insects. These weather data were collected at a meteorological station in TOEF about 4 km from where the light traps were set. Air temperature was measured by a platinum resistance thermometer (E-734, Yokogawa Weathac Inc., Tokyo, Japan), relative humidity by a capacitate hygrometer (P-HMP-45D, Ogasawara Inc., Tokyo, Japan), wind speed by a windmill anemometer with a wind vane (C-W154, Ogasawara Inc.) and precipitation by a tipping-bucket rain gauge (B-011, Yokogawa Weathac Inc.).

Statistical analysis

Patterns of daily attraction towards light traps in the lepidopteran and coleopteran communities, and factors affecting the attraction were analysed using a negative binomial log-linear model equivalent to the mixed Poisson model (Lawless, 1987). Species richness and abundance of Lepidoptera and Coleoptera in daily collected samples were regarded as response variables in the model of each community. Minimum air temperature (°C), the time over 100% relative humidity as potentially saturated time (h), the mean wind speed (m s−1) during the run time of light traps (continuous variables for meteorological factors), and the order of three consecutive trap nights in each month (ordinal variable for the date factor) were treated as explanatory variables of fixed effects. Although the negative correlation between an air temperature and a relative humidity at a time of day is generally well known, the seasonal relation between the minimum air temperature and the time over 100% relative humidity of day is obscure. Thus, it would be reasonable to assume that explanatory variables in these models are independent from each other. As the responses of Lepidoptera and Coleoptera to relative humidity were unknown, time over 100% relative humidity, rather than mean relative humidity, was adopted to represent the humidity factor. Precipitation was not incorporated in the models, because none of the rainfall was measurable during the run time of the traps. To manage overdispersion stemming from spatial pseudoreplication of the sampling design, the six traps were treated as a random effect. The significance of each explanatory variable in fitted models was evaluated by a likelihood-ratio test based on the χ2-test. Furthermore, daily variation in minimum air temperature, time over 100% relative humidity and mean wind speed were analysed by a linear mixed-effects model and a likelihood-ratio test based on the F test, respectively, to examine whether these meteorological factors showed a trend over three consecutive nights in each month. All analyses and graphics were performed under the R environment for statistical computing (R Development Core Team, 2006) and its extended packages aod (Lesnoff & Lancelot, 2006) and nlme (Pinheiro et al., 2006).

Results

In total, 51 742 specimens of 892 lepidopteran species and 11 633 specimens of 355 coleopteran species were collected in 126 samplings [original dataset not shown; downloadable from DIWPA-IBOY (Forest Ecosystem) database: http://diwpa.ecology.kyoto-u.ac.jp/IBOY/LoginForm.aspx]. For both Lepidoptera and Coleoptera, July had the greatest numbers of species and individuals.

The models of daily attraction for the lepidopteran community revealed that species richness and abundance of Lepidoptera attracted to light traps were determined seasonally only by the minimum air temperature during the run time of the traps, and there were significant positive responses of both species richness and abundance to minimum air temperature (Table 1). On the other hand, the daily attraction pattern of the coleopteran community was distinct from that of the lepidopteran community. In Coleoptera, species richness and abundance in the traps were significantly affected by minimum air temperature and time over 100% relative humidity as meteorological factors, and by the order of the three consecutive nights as the date factor (Table 1). Species richness and abundance in the coleopteran community responded positively to minimum temperature and time over 100% relative humidity (Table 1). Furthermore, both species richness and abundance of Coleoptera significantly increased with the order of consecutive sampling nights, although Lepidoptera capture levels were relatively constant for the 3 days (Fig. 2). The estimated daily species richness and abundance on the first, second and third night were 65.4 ± 8.4 (SE), 73.6 ± 9.7 and 73.5 ± 9.6 species (Fig. 2a) and 401.0 ± 59.5, 461.4 ± 69.7 and 445.9 ± 68.5 individuals (Fig. 2c) for Lepidoptera, and 8.4 ± 1.6, 16.1 ± 3.2 and 18.3 ± 3.4 species (Fig. 2b) and 46.6 ± 11.7, 91.0 ± 23.0 and 136.0 ± 34.6 individuals (Fig. 2d) for Coleoptera, respectively.

Table 1.  Results from negative-binomial models for factors influencing species richness and abundance of Lepidoptera and Coleoptera through light trap sampling in Tomakomai Experimental Forest (northern Japan). Estimates of meteorological variables for the fitted models are shown. The P-values were derived from a likelihood-ratio test based on a χ2-test. ‘Temperature’ is the explanatory term of minimum air temperature (°C), ‘humidity’ is the time over 100% relative humidity (h), ‘wind’ is the term of mean wind speed (m s−1) and ‘ordinal day’ is the order of the three consecutive sampling nights each month.
 Estimated.f.DevianceResidual d.f.Residual devianceP(χ2)
Species richness      
 Lepidoptera      
  Null   119592.4 
  Temperature0.053161.3118531.1< 0.001
  Humidity0.06813.2117527.90.071
  Wind 0.20811.0116526.90.325
  Ordinal day –22.2114524.70.331
 Coleoptera      
  Null   119476.7 
  Temperature0.1791175.1118301.6< 0.001
  Humidity0.074111.3117290.3< 0.001
  Wind0.09210.2116290.10.664
  Ordinal day –237.1114253.0< 0.001
Abundance
 Lepidoptera
  Null   119844.1 
  Temperature0.035126.1118818.0< 0.001
  Humidity0.04910.7117817.30.388
  Wind0.2381−2.1116819.40.146
  Ordinal day –2−3.0114822.40.223
 Coleoptera
  Null   119699.1 
  Temperature0.2551176.9118522.2< 0.001
  Humidity0.05613.9117518.30.046
  Wind−0.0611−1.1116519.40.286
  Ordinal day –217.1114502.3< 0.001
Figure 2.

Daily attraction patterns towards light traps in lepidopteran (a, c) and coleopteran (b, d) communities in Tomakomai Experimental Forest (northern Japan), based on species richness (a, b) and abundance (c, d) in 126 samples collected over 3 nights each month from April to October 2001. A median (bar), quartiles (box), a minimum and a maximum (whisker) and outliers of species richness or abundance for each night are represented as box-whisker plots. Furthermore, estimates and the standard errors derived from models are presented beside the box plots.

For daily variation in meteorological factors, significant decreases in the time over 100% relative humidity (P = 0.046; d.f.1 = 2, d.f.2 = 118, F = 3.161) and in mean wind speed (P = 0.002; d.f.1 = 2, d.f.2 = 118, F = 6.335) with the order of consecutive sampling nights were found. The estimated wind speeds on three consecutive nights were 1.71, 1.68 and 1.20 m s−1, respectively. There was no significant daily trend in the minimum air temperature (P = 0.318; d.f.1 = 2, d.f.2 = 118, F = 1.157). According to hourly measured meteorological conditions during the sampling season, the air temperature and relative humidity showed larger seasonal changes than the daily variations, although a seasonal change in wind speed was not observed (Fig. 3).

Figure 3.

Daily variation in air temperature (°C; a), relative humidity (percentage; b) and wind speed (m s−1; c) in Tomakomai Experimental Forest through the sampling season from April to October 2001.

Discussion

A daily increment in both species richness and abundance of Coleoptera captured by the light traps was demonstrated during three consecutive nights (Fig. 2b, d), whereas a constant number of daily captures of Lepidoptera was observed (Fig. 2a, c). Minimum air temperature sharply restricted species richness and abundance of Lepidoptera attracted to light traps throughout the sampling season (Table 1). On the other hand, the captured Coleoptera increased seasonally with minimum temperatures and relative humidity (Table 1). Although the relative humidity and wind speed exhibited daily tendencies of decreases within the course of any given sampling night, no daily trends in minimum temperature during the three consecutive trap nights in the monthly sampling were observed. These results showed that the observed daily increase in captured Coleoptera was not caused by weather effects on flight activity during the light trap sampling.

Among the meteorological factors, the result that air temperature limited the trap captures of Lepidoptera is consistent with previous studies (Bowden, 1982; Dent & Pawar, 1988; Holyoak et al., 1997; Yela & Holyoak, 1997; Butler et al., 1999). The effect of wind speed on captured Lepidoptera was not observed, in contrast to previous studies (Bowden, 1982; Dent & Pawar, 1988; Yela & Holyoak, 1997), because the wind was light during the sampling period. Relative humidity has been reported to have no effect on captures of Lepidoptera by light traps (Guedes et al., 2000); this study also suggested that relative humidity was not related to the activity of Lepidoptera. In Coleoptera, the temperature markedly influenced the number of captures, as shown in several preceding studies (Rodriguez-Del-Bosque & Magallanesestala, 1994; Davis, 1995; Honek, 1997); this reflected the seasonal trend in richness and abundance of the coleopteran community, which is phenologically more abundant in summer than in spring and fall (see Hirao et al., 2006). Wind speed was not a significant factor for Coleoptera as well as Lepidoptera. In this study, species richness and abundance of attracted Coleoptera, in contrast to Lepidoptera, were distinctively augmented by humidity, which suggests that the seasonal emergence of Coleoptera increases with humidity (Davis, 1995).

Daily collection in light traps represented the distinctive patterns potentially relevant to differences in daily movements between Lepidoptera and Coleoptera. Because they have the same levels for the fixed effect as the order of the consecutive trap nights in the models, their behaviours other than the daily movement to the light traps may be omitted from the interpretation of the results, and then the effect is likely to reflect only the daily movement distance to the light traps. Daily increments of coleopteran captures by light traps were not caused by an increase in the abundance of a few dominant species or the addition of rare species, because both species richness and abundance increased significantly during three consecutive trap nights (Fig. 2), which suggests that the daily movement distance of Coleoptera may be generally shorter than the effective and perceivable range of the light traps. The difference in daily movement distances between Lepidoptera and Coleoptera can be derived from differences in flight activity and capability. The flight activity of Lepidoptera should be much higher than that of Coleoptera. As lepidopterans typically have larger wings and correspondingly larger flight muscles than coleopterans, flight capability and duration of Lepidoptera should be superior (Wootton, 1992; Grodnitsky, 1999). This implies that a moth that has perceived the light may reach the trap in a single night, whereas a beetle may take several nights to arrive at the trap.

The daily increment in Coleoptera was not affected by the daily trends in meteorological factors. Relative humidity and wind speed showed tendencies to decrease within the course of a trap night, but relative humidity was positively correlated with captures of Coleoptera, as evident in the results of seasonal weather conditions, and wind speed was independent of the attractive efficiency (Table 1). This apparent discrepancy between the seasonal change and daily variation in relative humidity can be reconciled by the inference that the daily decrease in humidity would be insignificant to the activity of Coleoptera because the seasonal change is far larger than the daily variation (Fig. 3). Thus, the daily augmentation of captured coleopterans is likely to reflect the behaviour of daily movement in this case, although appropriate mark–recapture experiments would be required in order to test the hypothesis unequivocally. This would represent the most important goal of any potential future study testing the mechanisms underlying the observed pattern.

Alternatively, the contrasts in daily capture patterns by light traps between Lepidoptera and Coleoptera may be explained by a lower sampling efficiency among Coleoptera than Lepidoptera, which is caused by the architecture of light traps for capturing flying insects. Some beetles might be prone to escape from the traps, leading to a decline in species richness and abundance on the first night. Moreover, some beetles might move randomly during the day, which would make them move further from the traps. This aspect might also cause a gradual increase in species richness and abundance from night to night.

Sampling biases are incidental when any kind of trapping is used for biological monitoring (e.g. Raimondo et al., 2004). In this study, the results suggest that the typical life-history strategies relevant to flight activity and capability in a focal insect community may cause bias in the monitoring of entomofauna using light traps. Even over 3 nights, a smaller portion of the community was captured among Coleoptera (84.7% of the potential species pool; see Hirao et al., 2006) than Lepidoptera (92.7%). To clearly validate the hypothesis that the daily increment of captured individuals by light traps results from the shorter distance of daily movement than the effective range of the light, the range of daily movement and the distance for perceiving the trap light by focal taxa should first be measured and compared. More consecutive trap nights (> 3 nights) would then be required. Furthermore, if a difference in daily movement arises between communities, inflation of alpha diversity within habitat patches and reduction in beta diversity among patches would occur in the more mobile community by mass effects (cf. Shmida & Wilson, 1985) through migration among habitat patches. The mass effect through dispersal is a fundamental determinant of species diversity (e.g. Leibold et al., 2004), and therefore adequate control of sampling error by considering the movement range of focal taxa is necessary when monitoring is carried out using trapping methods.

Acknowledgements

We are grateful to H. Kogi, technician of Institute of Low Temperature Science, Hokkaido University, for identifying lepidopteran specimens, and the staff and graduate students of Tomakomai Research Station, Hokkaido University, for their support during the study, especially H. Asano and K. Ono for sorting invertebrate specimens. We also thank M. J. Toda, R. K. Didham and two anonymous reviewers for helpful comments. Financial support was provided by the Japanese Ministry of Education, Culture, Sports, Science and Technology (nos. 09NP1501, 11440224, and 15207008).

Ancillary