Back to index of invertebrate studies

Back to index of 2002 research reports

BUTTERFLIES AND FOREST DISTURBANCE

Aim:

To determine the effect of disturbance on the population sizes of the butterfly species present on Buton Island.  

Objectives:

-         To determine the population densities of butterfly populations in three study grids, each with a different level of disturbance.

-         To conduct a comparison of the data collected in the morning and the afternoon to establish what time is most effective for the collection of butterfly data for use in a long term monitoring scheme.    

Introduction:         

The islands of Sulawesi are found within the Wallacea region, home to many butterfly species a few of which are endemic to these islands. The Wallace line forms a divide between the Australasian and Asian mainland, with the Wallacea region containing much unique invertebrates, uncharacteristic of anywhere else

Butterflies are useful early indicators for disturbance, with many species being highly susceptible to alterations within habitat. Tree felling produces forest glades, which can effect butterfly populations which are sensitive to the light intensity and temperature changes which will be changed upon anthropogenic influence.

Three study grids are used in the survey, each receiving a different level of disturbance due to logging. From a similar study conducted in Trinidad it has been shown that the amount of disturbance has increased butterfly populations by enhancing the glade effect, which is similar to the edge effect produced on the periphery of protected areas.

As no similar studies have previously been conducted on the Island of Buton it is yet to be shown that the same effect will exist here. If so, this does not mean that disturbance should be considered as beneficial to butterflies as it may be that whilst some of the more common glade species are increasing, those which prefer shady conditions are in decline.

By combining the data collected with habitat survey data conducted by other projects in the study grids, it should be possible to determine the effect of disturbance upon the butterfly population, and from this continue a monitoring program to ensure the protection of all present species. 

Methods:

Three different study groups were used to collect the data, each with a different level of disturbance.

•       The Kakenawe forest reserve, which is selectively logged, and can therefore be classed as semi- disturbed.

•       The Lambusanga forest reserve, in which no logging occurs, and so is classed as an undisturbed grid.

•       The forest bordering the village of Kaweli, in which there are no limitations to the amount of felling that can be conducted, providing an example of a very disturbed plot of forest.

Each of the above forest plots contains a 1km2 grid comprised of transects at every 100m throughout the grid. To enable collection of butterflies that would give a representation of the population size and species composition that would reflect the grid as a whole, transects were walked every 200 meters across the length of the grid.In cases where the transect terrain was too uneven to be able to conduct an accurate butterfly count, the walk would be replaced by the nearest possible alternative so five transects per grid could be obtained. Each of the transects spanned the whole grid giving each transect a length of 1km. To standardize the amount of effort per butterfly sighting each of the walks took 1 hour to complete, with the time transect being split into 100m sections to aid in the monitoring of the amount of time taken per transect. Each of these walks was conducted in the morning during the hours of 08:00 and 11:00 and then repeated in the afternoon between the hours of 12:00 and 16:00.

If there had been heavy rain before or during the time transects needed to be walked the count was abandoned as early trials showed that butterfly sightings dropped considerably during and just after a period of heavy rain. In these circumstances the transect was completed on a different day when weather conditions were more suitable for data collection.

During the standard butterfly walk a note was taken of all the butterflies which were sighted within an approximate 5m box of that point on the transect, as well as observing the flight path of the butterfly to help lower the chance of repeat counts. If the butterfly was high up in the canopy a note of this was made though identification could not be made due to the distance away.

Where possible the butterflies that could not be identified on sight were caught in a swoop net and taken back to the laboratory in an envelope for identification. If the specimen could not be caught, a note was taken of the distinguishing features of the butterfly to attempt identification, which if still not possible was given a rough description and put at the bottom of the count list so all sightings could be counted.  If a butterfly was too quick or there was too much foliage in the way to be able to make an accurate identification the butterfly was marked as unknown.  A note was also made about the weather conditions at the time of identification i.e. sunny, overcast and whether there had been any rain at any point during the day.  Baited traps were not used during data collection as there had been problems in previous years with macaques raiding and pulling down the traps to get to the banana bait.

Initial results: 

There has been no statistical analysis of the data collected at present, though it was evident during the walks that there were high butterfly population sizes in the Kakenawe study grid. At Kawele the population size showed a considerable decline possibly due to the lack of forest cover and therefore increased predation by birds. The Lambusanga study grid also showed a lower count of butterflies that the Kakenawe grid, which is probably high due to the increase in glades, which are produced by the felling of trees and are very appealing to many butterfly species.

At present there does not seem to be marked differences in the data collected during the morning or early afternoon, the count is more dependant on the weather conditions at the time.

However if transect was conducted after 16:00 there did appear to be a slight drop in the amount of sightings even in good weather as the angle of the sun was too low to penetrate the canopy. 

Final Report Title:

A dissertation entitled The effect of disturbance on the size and composition of SE Sulawesi butterfly populations will be produced by Kay Jones, University of Leeds by June 2003.