page 250

Topic 5. Possum Population Assessment

CUMMINS. One of the main topics to come up in this symposium is possum population assessment techniques which is quite a crucial aspect. After all we are looking at an animal supposed to be one of our major pest problems, but we seem to have very little idea of how many there actually are.

MEADOWS. I feel much of the scientific work has been based on assumptions for which there is very little evidence. It's like a large building being put on very muddy ground. What evidence is there for accurate estimation?

CUMMINS. I was interested in one particular point - the quite significant loss of a possum population in an average year - some 15% of the population. Dr Brockie, did you in fact have any idea what caused this loss? Was it loss through migration, through natural death or what?

BROCKIE. In the Orongorongo Valley where we've been working there is a loss of possums through emigration, but this is compensated for by immigration; immigrants are mostly males in their first year. The dead animals were mostly found in winter on the ground by Dr Ben Bell and A.J. White. They all I understand had very low body weights and they appeared not to have died of disease or predation. By a process of elimination these animals appear to have died by starvation, through poor food supply at that time of the year. It's probably a complex situation though, with many factors contributing to it. I suspect that the cold wet weather in winter curtails the animals' feeding. There must be a cut-off point where the possums have to decide whether to stay inside their den to conserve what energy resources they have, or to take a risk and go outside to look for something with which to support themselves.

B.D. BELL. Further to what Dr Brockie has said about the Orongorongo Valley population, I examined weights of adult animals known to die and there was certainly a very marked decline in their weight prior to death. We set traps in the main study area for four successive nights each month and one needs to pose the question regarding our effect on the study population through this trapping activity. I looked at mean weight trends in animals which were successively trapped on all four nights each month or even three of those four nights; there were successive declines in mean weight, as you might expect, since over much of the night the animals could be cooped-up in a cage trap.

page 251

However, the animals which showed a serious decline in condition and became comatose or dead in the traps were generally those animals which were already in relatively poor condition and underweight; healthy animals appeared able to regain the weight-loss from successive captures. Thus I conclude that our trapping activities were not a major factor inducing low weights and deaths in the study population.

COLEMAN. To respond to Mr Meadows, I think we have got to remember that work on assessment methods is on-going. We are dealing with a generally arboreal, nocturnal, forest dwelling animal so to a large extent we are limited to indirect methods of assessment. What you use is dependent on the effort you have available and what you want to get out of it. D.J. Bell has given us an estimate of the confidence you can place on the different techniques which are available to us. If you are prepared to put the manpower into it, if you are prepared to estimate the decay rate of pellets, then you have a pretty useful tool. It's no good considering assessment in a purely cursory fashion; after one night you cannot come up with something which can stand up to close scrutiny.

CHARLESTON. Do we have any idea of how reliable ground trapping is for an arboreal animal like the possum? Until you know this, how can you say how effectively pellets estimate the population?

COLEMAN. That's not quite true. We measure pellet estimates against trapping as you say; we also measure trapping against total kill, and Mr Pracy is one who has trapped a population to extinction, and we intend to do the same in the population we've worked on. We shall trap to extinction and relate our pellet numbers to our trap catch in fact.

CHARLESTON. But you still have problems of immigration and emigration unless you have an isolated area.

CLOUT. I have computed a trapping estimate with a removal-kill estimate and I found quite good agreement.

FITZGERALD. Isn't it in fact true that by any of these methods you will never know what extinction is - you have no way of knowing when a population is extinct with any degree of certainty.

CUMMINS. Except on an island.

page 252

GREEN. I think we are trying to split a hair four ways which is really not our purpose. If one traps in an area where pellets decay mostly within a few weeks, I would submit that if one runs a pellet line in that area after you have trapped to extinction and one finds no pellets, then there are no possums. So I submit that we can probably trap to extinction with a fair degree of confidence, and by extinction I mean the animals normally resident and those passing through at the time have been killed. If you do your trapping to extinction in an area where you already have done pellet counts and have established a certain relative density, then a further pellet count provides a fairly good estimate of the effectiveness of your killing.

FITZGERALD. Yes, I feel I'm not very good at finding pellets.

YOUNG. I don't see any difference in us trying to get an estimate for possum numbers than for any other estimate we make in the field. The problem last year for the project I was involved in was estimating the numbers of a fruit-fly, and it came back to exactly the same topics. We do as many types of estimates as we can so one can check out the other. The advantages of the possum situation over the fly is that you can get a higher percentage of your animals marked in the habitat and this should give a pretty good idea of the numbers in the trapping area. I don't see why we are subject to this difficulty. Any decent study that has been run for 2–3 years comes out with a very good estimate of the numbers present, unless you talk about hypothetical aerial animals which have aerial pellets.

CUMMINS. This leads us to one of the perennial questions which is approximately how many possums do we have in this country? How many do we have to deal with? Les Pracy I've already elicited a tentative estimate from you.

PRACY. Yes, I was once directed to undertake this rather painstaking estimate. Fortunately, I had done a national survey of possums and covered some of the areas in New Zealand two or three times over. Most of this work strangely enough came back to the original population studies in the Orongorongo Valley in 1946-47 where there was a complete trap out of an area with marginal lines to stop re-infestation. There was a correlation between two methods of estimating density: the Lincoln index gave 2.2 possums per acre (5.4/ha), the trap-out gave 2.6 possums per acre (6.4/ha). There were 2 study areas and a marked difference in their carrying capacity in relation to vegetation types. Now in assessing the population for the whole of New page 253 Zealand what this meant was that you would have to assess all the various vegetation types and the carrying capacities at the time. There were 2 or 3 areas in the North Island and about 3 in the South Island where we could reckon on good reference populations. Knowing the possums' distribution and with population densities plotted as 'scattered', 'light', 'moderate', 'heavy', and 'dense¹, then on forest types and carrying capacity you could work out the national population. At that time, it worked out at 46 million. I didn't like this as a figure and I remember saying to Ralph Kean, "To hell with this, we'll cut it in half" so we arrived at a figure of 22 million. Now there are over 187 areas of post-peak populations, with 80 in the South Island, according to the last national survey of 1971-72.

CUMMINS. Would anybody from the Forest Research Institute like to stick out their necks on this one?

GREEN. They're not that long.

YOUNG. In regard to population estimates, our problem has not been so much to find the number in an area, but to find a geographic area to relate that number to. In fact as soon as possible we require some sort of standard technique to let us apply our trapping grid data (which simply gives a figure of animals' home ranges) to an area. Until we can do this, we don't have any density measurements that are reasonable. The standard method is to add half of the grid-spacing, but some papers don't even mention how they convert to a density. Clout working in consultation with Seber got quite different densities than if done by other methods. So this seems to me a critical issue that has to be resolved if anybody is using grids or lines as trap systems. We can get a number, and then that has to be converted to a density, and there's no realisation of how this should be done and certainly no conformity.

CLOUT. In fact I added half the estimated home range radius to the edge of my trap-revealed range area; obviously even that is going to be a gross underestimate of the area we are actually dealing with. The problem is how far are the animals moving outside of your grid-trapped area. It would be interesting if anyone else has comments.

YOUNG. It's only important because everybody quotes densities.

page 254

JANE. We did some bait interference work at Kaingaroa and two bait interference lines were separated by a road and fire-break, probably about two chains (40 m) distance. The bait line in a tall stand on one side reached a 100% take within a week, the bait line across the road reached a 5% take in three days but didn't increase beyond that. This indicates that you can get a very sharp boundary of a range in such a situation.

YOUNG. One of the standard papers everybody is using is the 1973 one of Dr Crawley^*. We do not see clearly what the basis of his density figures were. Does he remember that?

CRAWLEY. Yes, at least as far as the trapping grid is concerned. I simply used half the grid spacing to increase the area around the grid. I think one can choose one of a whole variety of methods or several of these methods. It's by no means a peculiar problem to possum density estimation, or even to mammals as you've pointed out. There aren't any easy answers. I think all one can do is to apply as many methods as possible and check estimates against one another. If you can trap to extinction against a previously trapped population estimated in any other way, that's tremendous, for you can get another check there.

There has of course been a great deal done with rodents; areas have been completely cleared and then known numbers of animals introduced; systems have then been tried out to estimate what at least for a short period was a known population. Of course, knowing the propensity of rodents and similar animals to breed the population may rapidly increase, or may possibly decrease rapidly from mortality factors. But I think one can continue discussions like this to the point of sterility. You have to get on with the work.

CLOUT. I'm really concerned that most of the density estimates may be gross overestimates, for instance in the Orongorongo Valley and in my own case as well. We have heard from Dr Green about very extensive movements of possums which would not have been picked up by standard live-trapping. It seems possible that ranges may have been much larger than we had thought, so that densities are lower than have been estimated.