Saturday, June 15, 2013

Census of the GWS in Gaansbai - Paper!

Really???  Source

Looks like the Megalobomb is human after all - did I just detect a soft spot for the researchers at DICT?
Read her latest post and check out the paper!

Where is the critical analysis?
That is, apart from whinging about the dismal number of GWS! Talk about the lady being totally jaded  - where I come from, 900-odd big Sharks from one single location is actually totally awesome!
Especially considering that previous white shark population estimates suggest that white shark numbers are small relative to other apex predators which is a statement I totally support see below!

Mind you, the paper is quite nice.
I was actually impressed - especially by the meticulous data collection, and by the various programs used for the analysis that are really quite epic. Boy things sure have changed since I did my few semesters of Biology courses back then in the 70ies!

BUT - of course, there has to be a but! :)
I'm slightly underwhelmed by the conclusions and also by the infographic, and this is why.

1. Mark-recapture

And I cite
A common bias in many mark-recapture studies is capture heterogeneity.
In our study we attracted sharks by bait, thus some individuals may have become ‘trap happy’ or ‘trap shy’ over time. This may lead to bias on estimates, but the effects of baiting on individual sharks remains undetermined. To address this, future work should focus on the effects of shyness or boldness in individual white sharks and assess whether they are more or less likely to appear close to a baiting vessel over time as well as incorporating such heterogeneity in behaviour in population size estimation.
This is also one of the principal flaws of the controversial (= flawed) Chapple paper, ie that over time, individual Sharks may develop gradual repulsion or special affinity for the vessels - but much more than that: some Sharks may never approach the bait, decoy or vessel to start with, let alone come to the surface for a picture of their fin, and this despite of very much being in the vicinity!
I don't know about those GWS - but when it comes to our Bulls, I could tell you stories about several known individuals that have been visiting for years and not once approached the feeder. And Juerg's paper also teaches us that whilst we're feeding, there are several individuals in the vicinity that we never get to see, quite possibly because they don't like approaching humans or also, because other individuals may out-rank them and thus monopolize the feeding area.

Long story short? 
The postulated bias is highly likely, is likely to result in numbers that are too low and the effect, whilst unknown in its magnitude, could potentially be highly significant = there may be significantly more GWS than postulated!

2. Population Size and Conservation Status.

And then there is the whole rare vs endangered controversy.
Yes, globally, there are likely not a lot of GWS  - but isn't that what is to be expected simply from their position atop the trophic web? Think about the famous pyramid - the volume at the top is tiny!
Also keep in mind that albeit being temperate water Sharks with a potentially enormous range, they have not at all colonized all available habitats, likely because philopatry is largely preventing them from straying too far from their established hot spots and migration corridors. Thus there are no reliable reports about established populations around South America or in the Eastern Atlantic (and David, don't you start...), with the only exception, i.e. the the Med having likely been colonized by a possibly single Aussie female that had lost its way.

And what about the number of GWS in their known ranges?
We don't dispose of reliable population estimates about most of them, namely the Med (likely very small); the NW Atlantic (more and more sightings but no census); the NE Pacific (currently in revision when it comes to California, but with no published census for the Mexican population), the NW Pacific (where there is little literature apart from sporadic reports of sighted and killed GWS from Japan, Taiwan and Vietnam - tho in view of the status of Asian fisheries, the situation is likely to be grim); Australia that boasts two populations but no published census (tho after so many years of protection, they may well be on the upswing); and New Zealand where research is still very much in its infancy.
And finally, South Africa where the present paper smells like the start of a country-wide revision of previous assumptions.

So yes, indeed, we don't know whether globally, they are 3,000, 5,000 or 10,000 - with my money being on the latter!

And if so, what does that mean?
Does rarity automatically imply that the animals are endangered?
To a certain degree, the answer is yes, and this for the rather trivial reason that smaller populations are generally more prone to be effected by risk - which is quite possibly why the IUCN has classified them as vulnerable despite of not really disposing of much supporting data.

Further conclusions, at least at this point in time, are however problematic.
Thus, comparisons to terrestrial species whose massive population declines and partial extinctions are amply documented, or that are trapped in a population bottleneck like the Cheetah are highly questionable. 
We simply don't know, and I've also not seen any plausible guesstimates about the global rate of depletion of GWS or about the carrying capacity of their global habitats - back then and especially now that the latter are likely equally depleted, see the comments about bottom-up effects! 
There are simply not enough data allowing us to make any such assertions, let alone proffer that this already threatened species may be closer to extinction than we previously thought!

I say, be careful with such statements!
Only because one local census results in numbers that are 50% lower than previously assumed, this cannot just simply be extrapolated for global populations! Just think of the California numbers that are likely to be trebled, or think about the dramatic increase of GWS sightings on the East Coast of the US!
I don't believe for a picosecond that Michelle Wcisel has stated that it was possible that the great white could be one of the most endangered species in the world - but the ingress to the infographic is certainly highly misleading, to the point that the sharktivists are already sharing it as the latest fact, see the image at the top!
It is not!

Anyway, just my 2 cents as usual.
But read the paper and make up your own mind!

PS: Michael Domeier's take here.
In all fairness and at the risk of committing sacrilege by posting a dissenting opinion, the paper does not claim to be anything else than a census for Gansbaai - hence if there are principally subadults, than that's what has been counted.
The question about the numbers of YOY, juveniles and adults becomes only relevant once somebody will publish a paper about the entire GWS population of SA.. Incidentally, that's another major flaw of the Chapple paper where the only animals that were actually recorded, and this in only two of the known aggregation sites were adult GWS.

PS2: more lousy journalism here!


Unknown said...

Great analysis Mike, as always..; for this paper I completely agree with you and your causion. Everybody use this info to share the idea that GWS are in low numbers in SA, but this estimate is only restricted to Gansbaii and 800-1000 GWS at on aggregation site is quite important for a top predator... also I thing the methods revealed a lot of incertaintly in the last papers... But it's a good way to assess the trends in populations there and follow them over time.

DaShark said...

I actually like the paper!

The authors do not state that it is anything more than a census for Gaansbai - and you are right, the method is excellent for detecting trends over time.

The problem IMO is the infographic that infers that the global status of GWS may be somehow dramatic, an assertion the paper does neither suggest nor support.

In reality, most published and also anecdotal observations point to an increase of many populations - which if you consider how long they have been protected both locally and internationally is not at all surprising!

Michelle W. said...

Thanks for the analysis DaShark, and since I know he reads the blog too, thanks for the comments Dr.D!

It's important to remember that mark-recapture analysis assumes that every animal you documented survived your study period. While our count of 532 which became the super pop. estimate of 908 may seem quite large, it assumes all sharks survived the study period. We know at least two of the GWSs from our database are dead - both killed in the past 12 months – and that 30-40 GWSs are killed every year in the KZN shark nets/hooks. The odds that all 532 sharks that we used for the estimate survived the 5 year study period are low. But I take your point - it's not a number to sneeze at.

This is why we reached out to the only other regional estimate conducted in South Africa (Cliff’s) that was conducted 13 years prior. Since nearly every white shark in Gansbaai goes next door to Struisbaai (where Cliff et. al. tagged most their animals), its somewhat fair to compare the estimates. Since they are quite similar, GWSs may not have rebounded since their protection in 1991. Now, if I read you correctly, you may be making the argument that since the estimates are similar, there may be enough GWS in S.Africa and that their protection wasn’t necessary back in 1991? A very interesting debate which hopefully will be resolved by the national estimate that’s currently underway. Of course, no baseline population information exists, and even if it did, the ocean is a completely different animal now than what is was in 1991 anyway.

As for mark-recapture analysis, it’s the bane of empirical workto find man-made equations that match nature. If you look at the assumptions your system must fit to run certain analysis - there are probably only a handful of systems that fit them (i.e. a petri dish of bacteria or a barn full of cats). Hence, we went with an open population that nearly doubled our actual count because we cannot assume that we saw every white shark that came to Gansbaai (unlike Chapple's closed pop. assumptions).

Or did we? From tracking sharks in Gansbaai, we know that not every tagged shark would appear at the surface for cage diving operators (similar to Juerg), however, every white shark that has ever been tagged and tracked (whether acoustic, SPOT, or PAT) was baited to the surface by a research boat to get tagged. They are surface hunting animals around pinniped colonies, so perhaps they are much more likely to fin at a seal colony than a bull is likely to go to a baiting area? For GWSs, it’s simply a bias no one knows enough about to account for. However, if you consider the extraordinary effort that went into the data collection (5 years worth of daily commercial trips which operate in all but the very worst conditions, year-round, in coastal and pinniped colony areas) my gut tells me we saw most.

As for MCSI’s comments about no mention of life history, it’s hard for us to mention information that simply doesn’t exist in S.Africa. California is spoiled with data, and I can speak for all S.African researchers when I say we are totally jealous of them and their money, and please send us some USD (currently trading at R10 = $1!). We feel that the year round effort accounts for the various motivations (seasonal seals, seasonal inshore) to come to Gansbaai and the 5-year study period accounts for the possible Guadalupe-esque 3-year rotation. Sure, we could have held our breath until, 1) more SPOT tags that don’t totally foul are deployed on GWSs, and 2) after -/+ 10 years of data collection someone actually analyses that data - but we'd be dry crusty corpses by then.

And finally - the media – nope, I never said that BS and have asked for my ‘quote’ to be removed and for various ‘facts’ to be corrected. You can quote me here: The GWS is not as threatened with extinction as good journalism is. :)

DaShark said...

No thank YOU Michelle!

Fine re the accuracy of your mark-recapture analysis.
As I said I dunno about GWS (thank god for prophylactic caveats huh! :)) and also, the long uninterrupted observation period and the fact that contrary to Chapple, you did not only deploy decoys but also chum and bait are quite convincing!

I'm looking forward to the nationwide census - not only in terms of absolute numbers but also in terms of population structure, i.e. the ratio of YOY vs juveniles vs subadults vs adults!
Does California teach us anything about that - and more importantly, are there any assumptions about how the ratios should be so that one can assess the effect of past and present anthropogenic effects?

With that in mind, NO, I do not at all suggest that GWS protection should not have been enacted in SA, or anywhere else!
If MD is correct and your GWS are principally subadults, one could argue that you only get to see them owing to SA's GWS protection measures after the slaughter in the 70ies and 80ies mentioned in the paper!

I also do not at all contest your conclusion that the Gansbaai numbers may have dropped, and that may be an indication that GWS protection of the SA stock may need improvement within their range - specifically in Mozambique where the situation appears to be deteriorating instead of improving!

But maybe, you guys are battling with issues beyond strict GWS conservation, ie overfishing (of Fishes) that may have reduced the carrying capacity of their habitat, meaning that the possibly reduced SA population is already the best possible outcome - but of course I'm speculating.
And yes totally agree - the ocean is NOTHING like it used to be 20 years ago!

As I said, i like the paper!
The infographic, not so much as it postulates global effects that are neither suggested by this local census nor by global observations.
In fact, it's more of a disinfographic, the bad news being that those memes are already being picked up and propagated by the media and will end up becoming the "truth".

Unknown said...

To DaShark and Michelle: I wasn't saying the paper is not good... It's a good paper (with a better methodology than the one of Chapple et al.). And it's a good baseline for GWS census in SA (a country where there is a lot of data, contrary to what you say Michelle..) and if other aggregation sites such as False Bay or Mossel Bay conduct the same analysis, compring these will add precious informations for the understanding of the GWS population dynamics.
However, I think that the infographic was pretty bad (although I understand that it was a first try) as it let non- specialists doing shortcuts about the conservation status of GWS and population dynamics in SA, and it's easy to see the consequences by reading the commentaries in response to it...
But anyway, good job for the paper and hope the next infographic will be done with better causion...
And waiting next year at Shark International Conference to go back to SA and their GWS!

DaShark said...

@ Anonymous
I'm not gonna post your comment "as is".

But of course I'm intrigued.
Mind elaborating on the scientific aspect whilst forgetting about the BTWs?

Anonymous said...

Think carefully about the discovery curve - figure 2.

and what has been omitted from the paper.

DaShark said...

Aahh Anonymous - riddles riddles! :)

But OK lemme try:
One would expect a lot of detections of new sharks at the start and over time, once more and more sharks are being identified, diminishing returns.

Hence, the curve should be logarithmic, i.e steep at the beginning and flat at the end.

Because the curve is not plateauing, one could argue that not all sharks have yet been identified = it is premature to publish a final number = the published number is too low?

Is that what you mean???
And, what has been omitted?

Anonymous said...

yes indeed, curve should go flat ... theoretically

patience for the rest :-)

DaShark said...

Phewww... so it looks like I'm not totally fossilized quite yet... :)

Very interesting!
Which of course begs the question, how long does one have to count - or does the fact that most sharks are only seen once/a few times (Fig. 1) maybe tell us that the population is completely open with close to zero residency, meaning that one should rather focus on the question of "how many sharks visit per year"?

The other corollary of this high population turnover and low residency is that it appears to indicate that cage diving is not unduly conditioning those sharks, at least not when it comes to large temporal and spatial scales - which confirms most other findings on this subject & is a good thing!

Michelle W. said...

Yep, the riddler and DaShark are on to something here!

There are actually three possibilities why our discovery curves doesn't fall perfectly flat, they could be:

1) Constant recruitment in the population - unlikely
2) The study has not identified every individual in the system - which would be fairly incredible considering the dataset - but will be addressed once the national estimate comes out (this is why you run an open population estimate on the data - which accounts for this partially).

Or the third option which our study eludes to...

3) The fins change, so as much as we try and avoid it, we are re-identifying the same individuals over and over again as different sharks - thus, we are always 'discovering' new animals.

We are not the only white shark fin-ID residential estimate struggling with a never-quite-flattening discovery curve, therefore I think #3 might be the case as it's equally effecting us all. The leopards keep changing their damn spots!

Anonymous said...

According to the curve : in the beginning every second shark is a new one and in the end every third shark or so is a new one ?

So maybe there are at least 2 more reasons why the curve is not quite flattening ?

DaShark said...

Maybe just state what you want to state and be done with it?

Anonymous said...

Later , elsewhere.

Good blog.