Biological Control of the Winter Moth

0 downloads 0 Views 579KB Size Report
Alsophila pometaria, and the Bruce spanworm, Operophtera bruceata, were present in virtually each situation, and interestingly, sizable outbreaks of each.
ANNUAL REVIEWS Annu. Rev. Entomol. Copyright

© 1995

Further

Quick links to online content

1995. 40:475-92

by Annual Reviews Inc. All rights reserved

BIOLOGICAL CONTROL OF THE WINTER MOTH

Annu. Rev. Entomol. 1995.40:475-492. Downloaded from www.annualreviews.org by University of Alberta on 11/11/14. For personal use only.

lens Roland Department of Biological Sciences, University of Alberta, Edmonton, Alberta, Canada T6G 2E9

Douglas G. Embree Canadian Forest Service, Maritimes Region, PO Box 4000, Fredericton, New Brunswick, Canada E3B 5P7 KEY WORDS:

population dynamics, parasitism, predation, host-plant interactions

ABSTRACT

The biological control of the winter moth in North America, by the intro­ duction of parasitic insects, is reviewed with the aim of identifying the common factors leading to successful control. These patterns are assessed in light of the large literature on the biology of the host insect and its natural enemies, both in the introduced populations and in endemic populations in Britain. Successful control has arisen from the combined effect of added mortality from the introduced agents and a suppression of the pest to the level where generalist predators already in the system regulate the winter moth at a new, low density.

INTRODUCTION Of all biological control attempts implemented against insects, control of the winter moth

(Operophtera brumata)

in Canada is one of the most frequently

cited (1,23, 27,33) and among the most successful. Detailed population studies accompanied release of biological control agents in Nova Scotia (8, 10, 11, 16, 34, 35) and British Columbia (26, 28, 36-45), and native populations in Britain had also been the subject of long-term study of the role of natural enemies in population dynamics (6, 17,18,22,24,30,31,48-53).

0066-4170/95/0101-0475$05.00

475

476

ROLAND & EMBREE

Interestingly, despite the wealth of infonnation on the biology of the winter moth and its natural enemies, the mechanism responsible for the success of this biological control system remains unclear (14, 38, 41). This review evalu­ ates the evidence from previous (9-1 1, 16) and recent (26, 32, 34-45) studies of winter moth biological control and attempts to provide a comprehensive explanation of why this control was so spectacularly successful.

Annu. Rev. Entomol. 1995.40:475-492. Downloaded from www.annualreviews.org by University of Alberta on 11/11/14. For personal use only.

HISTORY OF RELEASES IN NORTH AMERICA Winter moth invasion in North America had very different impacts in the three affected regions. The pest was a major threat to the hardwood forests in Nova Scotia, to urban shade trees in British Columbia, and to orchards, notably filberts, in Oregon.

Nova Scotia The initial work on parasitoid releases began in Europe in 1952 with investi­ gations into likely candidates for release conducted by the Canadian Depart­ ment of Agriculture and the Commonwealth Institute of Biological Control. A total of 63 species of parasitoids, either discovered by means of field collection or previously recorded in the literature, were considered (57). Col­

Cyzenis Agrypon jlaveolatum.

lections in Gennany and France consisted mostly of the tachinid flies

albicans

and

Lypha dubia

and the ichneumonid wasp

Because of the numbers obtainable, C.

albicans

appeared to be the most

promising candidate, and its ability to overwinter in Nova Scotia was con­ finned by studies in 1955 of caged insects in an immature stand of red oak (plot 1-2) at Oak Hill, Nova Scotia. The initial release of 3 1 C.

albicans adults

in 1954 may or may not have been successful, but the second release of 1008 definitely established the parasitoid in Nova Scotia. A. jlaveolatum was estab­ lished after the first release of 250 individuals in 1956 (21). Releases of both species continued until 1965 ( 11). Until 1963, the released parasitoids were from Europe, but after that year, local stock was released along the perimeter of the winter moth-outbreak area, which had by then extended into New Brunswick. The strategy was to release massive numbers in single locations, and the total area in which parasitoids were released encompassed 10,000 km2 (see Table 1 for summary of releases). Of the other species of parasitoids, the tachinids L.

dubia and Phorocera obscura and the ichneumonids Phobocampe crassiuscula and Pimpla turionellae were released in large numbers but none were recovered; Pimpla contemplator failed to overwinter in field cages ( 12). British Columbia Major releases of C.

albicans and A. jlaveolatum in British Columbia occurred

from 1979 through 1980 ( 15). The majority of the parasitoids came from Nova

WINTER MOTH BIOLOGICAL CONTROL

Table 1

477

Summary of release of C. albicans and A. flaveolatum in North America Averilge no. Total release

Species

Time span

area

(lan2)

No. of

of adults

Year of first

sites

per site

recovery

Nova Scotia

Annu. Rev. Entomol. 1995.40:475-492. Downloaded from www.annualreviews.org by University of Alberta on 11/11/14. For personal use only.

C. albicans A. flaveolatum

1954-1965

10000

16

1860

1958

1956-1962

6000

9

330

1957

C. albicans

1979-1981

500

25

240

1981

A. flaveolatum

1�79-1981

500

33

145

1981

C. albicans

1982

100

II

345

1983

A. flaveolatum

1982

100

11

145

British Columbia

Oregon

Scotia, with a small number from Germany. Over 200,000 fifth-instar larvae were collected. Pupae were shipped to quarantine services of Canada Agricul­ ture in Ottawa and then to British Columbia for release. The Nova Scotia experience suggested that a high percentage might become established, so the strategy in British Columbia was to make small releases in numerous sites, thereby saturating the area of infestation. In all, releases were made at 33 sites ' (Table 1) scattered across an area of 600 km2. Both species were recovered within two years after' the initial release.

Oregon Releases were made in 1982 only (29). Parasitoids were collected in Nova Scotia and shipped to the US Department of Agriculture Agricultural Research Service Beneficial Insects Research Laboratory in Delaware. Releases were made in commercial filbert orchards at II locations (Table I) across a lOO-km2

C. albicans was recovered the year following the release, but A. jlaveo­ tatum was not recovered. Population studies were unfortunately discontinued

area.

after 1983.

DYNAMICS OF NATIVE POPULATIONS Studies of native populations of winter moth in Britain were initiated with the intent of identifying the role played by natural enemies (including parasitoids later released for biological control in North America) in population dynamics. Biological control practitioners rarely have the chance to compare the popu­ lation dynamics of an insect in a native location with its dynamics in a

Annu. Rev. Entomol. 1995.40:475-492. Downloaded from www.annualreviews.org by University of Alberta on 11/11/14. For personal use only.

478

ROLAND & EMBREE

concurrent biological control program. Such an opportunity should provide good insight into factors promoting successful biological control. In Britain, winter moth populations peaked at 9- to lO-year intervals over the 19-year span of the study by Varley & Gradwell (51, 53). The dynamics involved three principal factors: the degree of synchrony between winter moth egg hatch and host-plant budburst, the density-dependent pattern of predation on pupae in the soil, and the minimal impact of parasitoids and disease. First, in years when egg hatch occurred prior to oak budburst, first-instar larvae starved (50, 51, 53) or dispersed by ballooning and became established on other host plants (56); the density of larvae on oak was therefore low. If hatch occurred after buds opened, then larvae survived well and tended not to disperse, which resulted in high population densities on oaks. Winter moth also exhibits this pattern on other plant species such as apple (25), and even on atypical host plants such as Sitka spruce (5, 54). This variation in synchrony added stochastic variation to the annual densities. The timing of budburst among plant species also affects the density of winter moth; species that leaf out early, such as apple, have a higher caterpillar density than do those, such as oak, that leaf out late (25, 56). Second, pupae of endemic populations in Britain suffered relatively high levels of predation from generalists in the soil (6, 18, 30,48, 49, 51-53), and pupal mortality was higher in years of high pupal density. Although this density-dependent pattern of predation was not sufficient, by itself, to regulate density, it did buffer the annual variation in abundance caused by the effect of synchrony betweep budburst and egg hatch. Pupal mortality was also strongly spatially density dependent (6, 30, 31). Finally, parasitoids, including those species introduced for biological control in Canada, caused little mortality of winter moth, typically less than lO% parasitism (48, 53). This magnitude was not correlated with winter moth abundance (53), although parasitism was 20% in the first year of the study, the year with the greatest densities. Microsporidian disease occurred at rates of about 75% among adults (2), but its direct effect on host mortality was low (48, 49, 53). The indirect effects of microsporidian disease on factors such as fecundity are not known (2). Although virus was present in high density populations, it typically caused less than 3% mortality (55). Research on endemic populations in Britain was concurrent with the studies of introduced populations in Nova Scotia. Predictions on the outcome of biological control in Nova Scotia populations based on the endemic studies in Britain were perhaps premature but, as we indicate later, may have contained the essence of future patterns seen in the Canadian populations. Varley & Gradwell predicted in 1968 that "There will be a strong parasite-host oscillation giving periodic outbreaks causing defoliation at nine or ten year intervals. Only

WINTER MOTH BIOLOGICAL CONTROL

479

if some potent density dep�ndent relationship becomes established cari the population be stabilized perinanently at a low level" (52, p. 141).

IMPACT OF INTRODUCED PARASITOIDS IN NORTH AMERICA

Annu. Rev. Entomol. 1995.40:475-492. Downloaded from www.annualreviews.org by University of Alberta on 11/11/14. For personal use only.

Nova Scotia Between 1950 and 1965 the winter moth infestation spread from the south shore throughout most of the province and into Prince Edward Island and New Brunswick (3, 10); this pest's range has not increased since. Large populations ' denuded large tracts of hardwood forests, principally red oak, urban shade trees, and roadside bushes. Wild apple trees were particularly susceptible, and the burnt appearance of these trees earned the winter moth the nickname, the fire bug. The outbreaks generally collapsed five years after C.

albicans became

detectable in host populations. Collapse occurred over the entire range of winter moth sites, and by 1966, the insect was present in numbers too small to be recorded accurately throughout the region. Winter moth has persisted in commercial apple orchards but at levels below the limits of its food supply. It remains a problem because it feeds on flower buds, which results in deformed apples (32). Populations have remained low or virtually nonexistent in forested areas ever since, but small flare-ups have occurred on other sites. Three have occurred since 1966, peaking at intervals of 8, 9, and 9 years, respectively. Defoliation was localized-limited to orchards, urban shade trees, and road­ side bushes, and particularly to early-leafing species with which the hatching winter moth larvae are well synchronized. In addition, the fall cankerworm,

Alsophila pometaria,

and the Bruce spanworm,

Operophtera bruceata,

were

present in virtually each situation, and interestingly, sizable outbreaks of each of these latter species have occurred at the same time intervals as those of the winter moth (14) (Figure 1). While by

C. albicans, A. pometaria (10)

O. brumata is susceptible to parasitism O. bruceata are not. Hence, the

and

concurrent eruptions of these species are probably independent of the intro­ duced parasitoid. In the initial outbreak, parasitism by the two parasitoids averaged 78% at its highest levels. Of the nine study plots at Oak Hill, the highest level of parasitism was 100% at one site. At four other sites, the maximum recorded level ranged from 80 to 86%. The lowest parasitism rates recorded prior to population collapse were 50 and 52% in two of the plots. On one plot (plot

1-2) where at least 50% of the prepupal larvae were parasitized in th� final year of the collapse, mortality of the surviving prepupae, pupae, and adults was 100% (8). At other sites such as Fox Point, popUlations began declining

480

ROLAND & EMBREE 12 10

FALL CANKERWORM

8 8 4

Annu. Rev. Entomol. 1995.40:475-492. Downloaded from www.annualreviews.org by University of Alberta on 11/11/14. For personal use only.

2 1975



L5

ex: OJ l::::> 0 u..

0

� en Z W IZ

12

1993

1984

1_

BRUCE SPAN WORM

10 8 e 4 2

1975 2

WINTER MOT N.S. H(

J



1988

Figure 1

1984

)

��

' NnR MOT H(

.

) .

1975

.

.

,,� 1984

YEAR

�I

1993

Schematic patterns of eruption of geometrid defoliators over time in Canada (vertical

scales are arbitrary, based on Forest Insect and Disease Surveys, Canadian Forest Service).

WINTER MOTH BIOLOGICAL CONTROL

481

after parasitism had only reached 26% (8), although parasitism ultimately rose to 80%. The population dynamics of the winter moth have not been studied in oak stands since the population collapse, although studies of associated oak defo­ liators continued until 1970 (11, 13). However, studies in apple orchards continued until 1980 (32) and were also conducted in 1991-1992 (34, 35). Parasitism by C.

albicans

during decline ranged from 26 to 72%, and later

Annu. Rev. Entomol. 1995.40:475-492. Downloaded from www.annualreviews.org by University of Alberta on 11/11/14. For personal use only.

parasitism by A. jlaveolatum reached a maximum of 19% in a single orchard. Parasitism at the time of the general population collapse was on average lower on apples than on red oak. After host decline, populations of the two parasitoids persisted at low levels, and the mortality they caused was not temporally density dependent (32). Parasitoids were abundant enough, however, to allow the collection of more than 30,000 individuals for release in British Columbia and Oregon in 1979-1982. Since the winter moth population collapsed in Nova Scotia, parasitism has been less than 5% in orchards and has not increased temporally in response to the intermittent flare-ups occurring in orchards (32). The key factors in orchard populations are mortality of pupae in the soil, followed by mortality of dispersing first-instar larvae, and in small populations, mortality of large larvae presumably caused by bird predation (32). Studies in four orchards during the 1992 flare-up failed to detect

C. albicans in two orchards with small

populations and found only minimal parasitism (28% was the highest recorded on a single tree) by this species in two more-heavily infested orchards. Soil predation caused the greatest mortality, up to 98.4% where host populations were low and 74.5% where popUlations were more abundant (35). Hyperparasi­ tism of C.

albicans

in orchards in Nova Scotia has not been found.

British Columbia Winter moth became an urban problem, particularly on Garry oak

garryana),

(Quercus

in Victoria on Vancouver Island about 1970 (19) and has since

spread about 100 km northward. Five years after the release of C.

albicans

and A. jlaveolatum, a general population collapse occurred. As in Nova Scotia, this collapse was not as pronounced on apple as on oak. Parasitism by

albicans

C.

reached levels of 70--80% on oak and 50--80% on apple (39), but

unlike in Nova Scotia,

A. jlaveolatum

was ineffective, causing only 1-2%

parasitism. At the same time as the collapse of the initial outbreak, the pro­ portion of winter moth numbers taken by soil predators increased dramatically, reaching levels of 90-98%. Life-table data indicated no difference in density of unparasitized pupae between orchards and oak stands (Le. survivors from parasitism) but a large difference in adult density (high in orchards and low in oaks) (39). Since the collapse in 1984, the winter moth has remained at low densities except for a single flare-up starting in 1990, six years later (41, 47).

Annu. Rev. Entomol. 1995.40:475-492. Downloaded from www.annualreviews.org by University of Alberta on 11/11/14. For personal use only.

482

ROLAND & EMBREE

Hyperparasitism of C. albieans by the ichneumonid Phoboeampe sp. (28) is estimated at 18% at one site (Mt. Tolmie) (J Roland, unpublished data). Avian predators caused heavy local mortality in orchards (44). The winter moth and C. albieans spread to the mainland of British Columbia in 1984, but the pest did not cause serious damage until 1989 when white birch stands, fruit trees, and deciduous ornamentals were heavily defoliated. The outbreak persisted until 1991 and coincided closely with the most recent flare-up on Vancouver Island. Studies of populations on apple and blueberry showed that parasitism by C. albicans reached levels of 55% on birch (Betula sp.) and 35% on blueberry (Vaccinium corymbosum) (26). Pupal predation at these sites was in the range of 90% and rose sharply during the first two years of decline (26). Use of virus to control winter moth in British Columbia (4) resulted in a 20-40% reduction in caterpillar density in the year of application, but this reduction was not carried over to the next generation. Oregon The winter moth has been present in Oregon and Washington since 1950, but its rate of spread has been slow. In Oregon where the attempts at biological control were made, the insect feeds on at least a dozen species of deciduous trees, as well a rhododendrons, roses, and blackberry. The commercial crop most affected is filbert, Corylus avellana, along with rosaceous ornamentals, particularly flowering plum (Prunus) and crab apple (Malus sp.). C. albicans was recovered the year following its release on one of the seven release sites, and in the year of the release at another. Parasitism was typically less than 10% (29). A. flaveolatum was not recovered. No further studies have been conducted to determine the degree to which the introduced parasitoids have been established or their impact (29). SIMILARITIES AMONG INTRODUCTIONS Although the two major releases of insect parasitoids against winter moth occurred in widely separate locations three decades apart, they share several similarities. Outbreaks had lasted at least 15 years prior to parasitoid introduc­ tion in both cases. The two species released in British Columbia (c. albicans and A. jlaveolatum) are the same as those that had measurable impact on winter moth populations in Nova Scotia, and in both cases four to five years passed between the release of the parasitoids and the collapse of the host population. This lag is strongly correlated with the time taken for parasitism rates to exceed 30% in both populations (8, 41). It also matches the time interval needed for predation in the soil in both provinces to rise above 90% at most sites (8, 41) (Figure 2). In more recent introductions of winter moth to the mainland of

483

WINTER MOTH BIOLOGICAL CONTROL

d " O D

Annu. Rev. Entomol. 1995.40:475-492. Downloaded from www.annualreviews.org by University of Alberta on 11/11/14. For personal use only.

site 1-2

site 4

ktotal

V

V



ar

58

59

60

61

58

59

60

61

,ee,

Sit

� eo

61

62

59

eo

61 yea'

61

62

62

-------­ /

oL-

/-f

�-===���

__

69

61

60

y..,

Sit

60

,ea,

site 5

59



59

62

59

year

60

61

62



__

£