REU: Tad Nakatani
 
Indirect effects of climate change on the Symmorphus wasp at Big Pine Creek

ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
CONCLUSIONS
REFERENCES
ACKNOWLEDGEMENTS

POWERPOINT

ABSTRACT

Upward elevational range shifts are a well documented consequence of warming global climate. The purpose of this study was to examine how upward shifts by the willow leaf beetle Chrysomela aeneicollis in eastern Sierra Nevada Mountains are impacting one of its specialist predators, the parasitoid wasp Symmorphus cristatus. S. cristatus hunting times were measured at varying elevations and compared with local beetle larvae abundances as well as temperature. Trap-nests were placed at varying elevations as well, but were not used by the wasp. Finally, beetle larvae on willows were manipulated to test for any S. cristatus preferences in beetle host plant. Hunting time was increased at the low elevation site with low beetle abundance. At this site, hunting time also increased with temperature. Wasps also showed no preference in beetle host plant. It is apparent that climate induced changes in the distribution of one species affect interactions between species and, thus, have the potential to significantly alter community structure.

INTRODUCTION

For decades, the topic of global warming has attracted much attention from climatologists, ecologists, and physiologists alike. Past studies have shown that the earth is indeed experiencing a warming trend (Beniston 1997, Mann et al. 1999, Hughes 2000). In light of this knowledge, the challenge for the ecologist is to determine how a warming climate will affect both individual species and whole communities. Numerous studies have already shown that climate and atmospheric changes can significantly alter a species’ physiology, distribution, and life cycle patterns (Parmesan 1996, Myneni et al. 1997, Visser et al. 1998, Parmesan et al. 1999, Pounds et al. 1999). Long-term studies of communities are ideal for mapping changes in species and community structure. However, short-term studies examining how factors that are changing or are likely to change with global warming affect species are also powerful tools for predicting the fates of communities.

One expected consequence of warming is a shift in the ranges of many species. Various studies have shown that species have shifted up in elevation and toward the poles in latitude (Parmesan 1996, Parmesan et al. 1999, Pounds et al. 1999). Parmesan (1996) demonstrated that this range shift may be driven by significantly higher extinction rates at lower elevations and latitudes. Changes in the range and local abundance of one species also affect its interactions with other species (Visser et al. 1998). This can lead to further shifts in range and even local extinctions.

The Big Pine Creek drainage of the Sierra Nevada Mountains provides an excellent arena for examining such species and community effects along an elevational gradient. The willow leaf beetle Chrysomela aeneicollis, and its host plants Salix spp have been studied in detail at this site since the early 1980’s (Smiley et al. 1985, Smiley and Rank 1986, Rank 1992a, Rank 1992b, 1994, Rank and Smiley 1994, Rank and Dahlhoff 2002). However, comparatively little work has been done on the parasitoid wasp Symmorphus cristatus that specializes on C. aeneicollis (Sears et al. 2001). S. cristatus adults are nectar feeders preferring flowers of the family Apiacae, such as cow parsnip (Smiley and Rank 1986). Females make nests in holes bored by beetles in fallen lodgepole pines (Pinus murrayana). A female lays an egg on the roof of the hole and provisions it with third instar C. aeneicollis larvae. Once provisioned the female closes off the chamber by building a wall with mud and plant fibers, and then she may begin a new chamber. Larvae of the C. aeneicollis feed on willows and convert salicin from the leaves into a defensive secretion consisting mainly of salicylaldehyde (Pasteels et al. 1983). Larvae feeding on willows with high salicin content produce more salicylaldehyde (Smiley et al. 1985, Rank et al. 1998). Generalists predators generally prefer Chrysomelid larvae feeding on low-salicin willow, while specialist predators often prefer larvae on high-salicin willows (Rank 1994, Kopf et al. 1997, Zvereva and Rank 2003).

In recent years, C. aeneicollis populations have been disappearing at lower elevations and increasing at higher elevations (Smiley, personal communication). Since S. cristatus specializes on C. aeneicollis, it is possible that the range shift of the beetles may result in a range shift of S. cristatus.

The purpose of this study was to determine how S. cristatus is affected by factors that are, in all likelihood, changing with the global climate. Hunting time, the amount of time needed by a female S. cristatus to capture a C. aeneicollis larva and return it to her nest, was measured at varying elevations. Local prey abundance and temperature were also recorded, and analyses were performed to test whether they were good predictors of the wasps’ hunting efficiency. Flight speed in insects has been shown to increase with temperature (Vogt et al. 2000), and S. cristatus hunting rate has previously been observed to increase with temperature (Smiley, personal communication). Therefore, it is predicted that S. cristatus should have shorter hunting times at elevations with the best combination of high prey abundance and warm temperature.

Additionally, a host plant preference study was conducted to test whether S. cristatus preferentially preys upon C. aeneicollis larvae feeding on either high or low-salicin willows. More willows at high elevations tend to have high salicin content than at low elevations (Smiley et al. 1985, Smiley and Rank 1991). Therefore, a preference for larvae on certain willows may affect the predator-prey interaction if S. cristatus were to shift its elevational range. Because specialist predators have been shown to prey at higher rates on Chrysomelid larvae feeding on high-salicin willows, it is predicted that S. cristatus should also prefer larvae feeding on high salicin willows.

MATERIALS AND METHODS

NEST PROVISIONING ALONG AN ELEVATIONAL GRADIENT

Hunting time -- Nest provisioning behavior was observed at three sites along an elevational gradient. At the Falls site, the lowest site at 2950 m, behavior was recorded at two fallen lodgepole pines. At the Pond Bog site (3128 m) and the Upper site (3229 m) provisioning was observed at one lodgepole pine per site. Preliminary observations were conducted at the Second Falls site (2540 m) but were discontinued because no nest provisioning behavior was recorded. All recorded behaviors occurred from 22 July to 6 August 2004 between 0921 and 1518 h. Many nest holes had been marked in previous years by John Smiley. Holes that had not previously been marked but were being provisioned were given a marking. Hunting times were obtained by recording the time between a female’s exit from her nest hole and her re-entrance into the nest with a Chrysomela aeneicollis larva. During observation, a digital thermometer was placed in the shade of the log, and temperature was recorded from this at the time of each re-entrance with a C. aeneicollis larva.

Prey abundance -- On 8 August 2004 prey abundance was measured at 7 different elevations ranging from 2500 m to 3360 m. At each site, C. aeneicollis larvae on willows were counted for 15 minutes. For purposes of comparison, counts of adult abundance conducted by N. Rank and E. Dahlhoff on 17 June 2004 were also obtained. Each of the 2 observers conducted a 5-minute count of adult C. aeneicollis at each site. At the Sam Mack site there was only one observer, so a 10-minute count was performed.

Trap-nests – Trap-nests were constructed from clear Douglas fir blocks approximately 4.5 cm x 4.5 cm x 15.2 cm with a 4.8 mm hole (Sears et al. 2001). The blocks were split in half through the hole and taped back together so they could be reopened and examined later. 16 blocks were placed at each of 7 sites ranging in elevation from 2500 m to 3360 m. Four fallen lodgepole pines were chosen at each site and 4 trap-nests were placed on or next to each log. The original design called for the trap-nests to be collected so that the number of cells provisioned, the number of prey per cell, and the total prey mass per cell could be examined at varying elevations. However, no traps were in use on the final day of observation, so they were not collected.

Analysis of hunting time – To test whether hunting times varied between sites, an ANOVA was performed. However, variances did not appear equal, so the Levene test for unequal variance and a Welch ANOVA were used. ANCOVAs were used to test the combined effects of elevation and temperature as well as C. aeneicollis larvae abundance and temperature on hunting time. Because only 3 elevations and 3 prey abundance counts were used in this model, these factors were treated as an ordinal variables. Additionally, data from the Falls site was combined with data collected by John Smiley in August 2000. All hunting times occurring above 17 ºC (the lowest observed temperature for a hunting time at the Falls site in this study) were tested in a linear regression against temperature.

HOST PLANT PREFERENCE

The preference study was conducted at the Falls site on two species of willow: the SG-rich S. orestera and the SG-poor S. geyeriana (Smiley and Rank 1991). Five pairs of adjacent willows were selected. Crawling predators such as ants were excluded from the experimental area of each plant by a sticky resin (Tree Tanglefoot, Tanglefoot Company). A mixture of first, second, and third instar larvae was gathered from the Upper site where they were highly abundant. 50 of these larvae were transferred to the experimental area of each plant except for Block 5 which only received 34 larvae per willow. Blocks 1-4 were set up on 27 July, and Block 5 was set up on 22 July 2004. The number of surviving larvae was counted 3 times (days 3, 6, and 10). A blocked repeated measures design was used to test for any difference in survival between willow types. All statistical tests were performed using JMP IN version 4.0.4 statistical software.

RESULTS

HUNTING TIME

Abundance counts revealed that prey was most abundant at the Upper site, intermediate at the Pond Bog site, and quite low at the Falls site, with C. aeneicollis larvae counts of 898, 457, and 13, respectively (Fig. 1). Mean hunting times at the Upper, Pond Bog, and Falls sites were, respectively, 11:45 min, 10:02 min, and 28:05 min. Although the prey abundance at the Upper site was almost twice as high as the Pond Bog site, hunting times at these sites did not differ in variance or mean. S. cristatus at the Falls site, however, took longer to successfully hunt a C. aeneicollis larva than wasps at the two higher sites (Fig. 1: df = 2, 45; p = 0.0001). The Falls site also had greater variance in hunting times than the other two sites (Fig. 2: df = 2, p = 0.0003). Treated as ordinal, elevation and prey abundance were statistically equivalent, and there was a marginally significant trend revealing that temperature had an effect on hunting time that varied with either elevation or C. aeneicollis larvae abundance (Table 1: df = 2, 41; F = 2.9; p = 0.067). At the Upper and Pond Bog sites hunting time was not influenced by temperature, but at the low elevation, low prey abundance Falls site, hunting took longer with increasing temperature (Fig. 3). When combined with Smiley’s data from 2000, the hunting times occurring above 17 ºC exhibited a strong increase with increasing temperature (Figure 4: df = 1,24; F = 8.7; p = 0.007).

HOST PLANT PREFERENCE

C. aeneicollis larvae were preyed upon at equal rates on both SG-rich S. orestera and SG-poor S. geyeriana (Fig. 5). By day 3, only 2 of the 10 plants had greater than 50 percent survival, but they were both S. geyeriana. However, by day 10, there were no larvae remaining on 3 of the 5 S. geyeriana.

DISCUSSION

HUNTING TIME

As predicted, hunting time did vary with the abundance of C. aeneicollis larvae. S. cristatus takes much longer to capture prey when abundance is very low, but it does not vary in hunting time at the two higher sites, even though the Upper site has nearly twice the abundance of the Pond Bog site. It appears that S. cristatus may decrease its hunting time as prey abundance increases, but there is a threshold above which the addition of prey does not help females to hunt any faster. Hunting time consists of the time it takes the female to search for the prey as well as the time it takes her to gather the prey and return to the nest. Gathering and returning time is not likely to vary much between the hunting times collected in the study. Therefore, the major differences in hunting times are likely to be caused by differences in time spent searching for prey. At the higher sites, where prey abundance is above the threshold, S. cristatus should have no problem locating prey, and search times should be consistently near the minimum. This would explain why hunting times were less variable at the Upper and Pond Bog sites. The greater variability of hunting times at the Falls site is also likely to be determined by abundance. At the low abundance site, search time was clearly not consistently near the minimum. When prey is scarce, some females may spend a long time searching. However, parasitoid wasps usually can remember a successful foraging site and return to it (van Nouhuys and Ehrnsten 2004). Therefore, if a female locates a willow with multiple third instar C. aeneicollis larvae on it, her return trips will have near minimal search times. Thus, when abundance is low there is a greater potential for variation in hunting times.

The fact that temperature only had an effect on hunting time at the Falls site may also be due to differences in abundance. If temperature is affecting the rate at which S. cristatus can fly, then this effect will be more apparent when the wasp is flying for longer periods, as is the case at the Falls site where hunting times were longer. The trend at the Falls site, however, was the opposite of what was expected. Smiley’s (personal communication) work with S. cristatus at Big Pine Creek showed that females’ hunting rate increased with temperature. This study, on the other hand, suggests that hunting rate decreases with increasing temperature. Smiley’s study, however, measured hunting rate over a much wider range of temperatures, whereas most of the hunting times in this study were gathered near mid-day when temperatures were generally high. If the effect shown in this study is real, it would not be incompatible with Smiley’s results. It is possible that, like abundance, increasing temperature may increase hunting rate but only up to a certain threshold point, and most of the temperatures at the Falls site may have been above the threshold. This theory is supported by the fact that all of Smiley’s observations above 17 ºC combined with observations from this study show decreased speed with temperature increase. This model can be explained by a balance between two factors that influence hunting time: flight speed and prey location. At colder temperatures, the flight muscles of S. cristatus may not be warm enough to function maximally. This explains Smiley’s results showing a positive relationship between temperature and time. On the other hand, if S. cristatus uses odor of C. aeneicollis to detect prey, the odor may break down faster at higher temperatures, making prey location more difficult. Thus, at high temperatures, prey location is the limiting factor in hunting speed, but at low temperatures flight speed may be more important. It is also likely that prey location is much more difficult at low abundance sites, so at high abundance sites prey location time would be less variable. Thus, it is not surprising that the high temperature effect was not as apparent at the Upper or Pond Bog site or in Smiley’s work at the Falls site when C. aeneicollis was more abundant there.

TRAP-NESTS

In the study by Sears et al. (2001), trap-nests in Big Pine Creek were used by S. cristatus. Traps were filled at sites where wasp abundance was high thus creating a higher demand for suitable nest holes. It may be that currently there are no sites in the drainage where S. cristatus population density is high enough to create a high demand for suitable nests, and females simply do not have to “settle” for inhabitable, but less desirable trap-nests. If this is indeed the case, it may be that S. cristatus populations have not yet adjusted in response to changes in local C. aeneicollis abundance, or that, under current conditions at sites with high C. aeneicollis abundance, S. cristatus populations can not reach abundances high enough to create a high demand for nest holes. An alternative explanation is that wood-boring beetles have actually created more holes in logs, so nest supply is greater than in past years.

HOST PLANT PREFERENCE

Equal C. aeneicollis survival on high and low-salicin willows suggests that S. cristatus, unlike other specialist predators, does not prefer larvae with more secretion. However this finding is consistent with a previous study which demonstrated that a different species of Symmorphus wasp also did not show a preference for high or low-secreting Chrysomelid larvae (Rank et al. 1998). However, due to the design of this study, S. cristatus may actually have a preference that was not revealed. The test was performed at the Falls site where prey abundance was very low. Because it is harder for females to find C. aeneicollis larvae there, they would be more likely to take whatever, they could find rather than being choosy. Additionally, the blocks were not checked until day 3, and any signs of preference may have disappeared by then. The sticky resin also only kept out crawling predators. The C. aeneicollis larvae that were eaten were therefore not all necessarily taken by S. cristatus. Their survival counts could have been significantly affected by other flying predators such as a species of Hemiptera that has been observed feeding on C. aeneicollis.

CONCLUSIONS

From this study, it is apparent that S. cristatus is indeed being affected by shifts in range and local abundance of C. aeneicollis. At lower elevations where C. aeneicollis abundances have become very low, interactions between S. cristatus and C. aeneicollis appear to have been altered from times when C. aeneicollis was more abundant there. Due to increased hunting times at low prey abundance sites, the wasps exert more energy and leave their nests open to parasitoid invasion for longer periods of time. Although a succussful trap-nest study would be needed to determine whether these factors will decrease the amount of prey a female can capture or the number of eggs she will lay, it is likely that S. cristatus populations will decrease when local abundances of C. aeneicollis drop enough. It is also likely that if local C. aeneicollis populations continue to disappear at low elevations, S. cristatus will become locally extinct at these elevations as well. Thus, the case of C. aeneicollis and S. cristatus demonstrates that as warming climate trends cause changes in a single species’ range, there will also be much broader effects on community structure due to integral changes in interspecies interactions.

ACKNOWLEDGEMENTS

I’d like to thank the following people for making the WMRS REU program an incredible experience:

John Smiley
Nathan Rank
Elizabeth Dahlhoff
Eric Berlow
All the REU interns
WMRS staff
ESICE

About Us Science Fridays Staff Corie Cann: Bio ESI Home Tad Nakatani: References Tad Nakatani: About Me Tad Nakatani: Home