Deciding which species “deserve” the most resources and attention in terms of conservation can be a contentious issue at the best of times.
Members of the general public tend to favour protecting the cute and cuddly however, are these characteristics really the most effective way to decide how to direct our efforts? Using a publication written by Chapple et al. (2011) and estimating white shark populations in California, USA, as a case-study, this two-blog series aims to highlight:
- The difficulties of estimating population sizes of marine species
- The importance of the production of the most accurate information possible from the scientific community, and
- The impact of these factors on the success of conserving endangered marine species.
The Challenges of Modelling the Population Size of Migratory Marine Species
In 2011, a study completed and published by Chapple et al. estimated that the central California population of white sharks consisted of only 219 mature and sub-adult individuals. Their study was conducted over a three year period in two focal areas (both of which are well-established seal rookeries) around the Farallon Islands and Tomales Point in which white sharks are known to aggregate. As the dorsal fin is the equivalent to a ‘fingerprint’ for individual white sharks, photographs of dorsal fins were used by Chapple et al. to generate abundance data of the number of individual white sharks seen across these two sites throughout the study period. Population estimates were then subsequently modelled for sharks in this area.
In order to model these population estimates, Chapple et al. made five key assumptions:
- Their study white shark population was closed
- There was homogenous sampling of animals and all individuals had an equal probability of capture
- The “tagging” process did not influence the chance of recapture
- There was zero “tag” loss
- Photo-identifications of white sharks at these two aggregation sites represented a random sample of the central California population
If correct, the low estimates of white sharks presented by Chapple et al. justifiably raised concerns for the status of the white shark population in California and the ENP. Just as with any scientific publication however,
“it is important to consider the potential for methodological flaws and assumption biases that may have resulted in an under-estimation of the actual population size” (Burgess et al., 2014)
In June of 2014, George Burgess and his colleagues published the results of an important re-evaluation of the size of the white shark population off of California, USA, based upon a data set collected at the same sampling locations as Chapple et al.’s 2011 study (Jorgensn et al., 2010). The differences in the estimates of the white shark population size off of California between the two studies was staggering, with Burgess et al.’s estimate (2014) indicating a minimum all-life stages population size of over 2000 individuals.
Burgess and his colleagues found that all five key assumptions made by Chapple et al. (2011) had been violated, resulting in an underestimation of the white shark population in this area.
Assumption #1: Closed population
By assuming that the study population was closed, Chapple et al. effectively assumed that all mature and sub-adult white sharks would return annually to the seal rookeries of their study. However, considering that:
- Multiple examples in the literature have shown that most sharks do not return annually to particular aggregation sites
- Some sharks within the ENP may be located within the region but not in either of the two study areas and therefore will not be accounted for in the study
- The number of unique individuals identified by Chapple et al. increased each year (suggesting an immigration of sharks during the study)
It was concluded by Burgess et al. that both the existing literature and Chapple et al.’s own data demonstrated that the study population was in fact open over the three years of data collection; thus violating a key model assumption.
Assumption #2: Homogenous Sampling of Individuals
In more layman’s terms, Chapple et al. assumed that every shark had an equal chance of being sighted, and that sharks at the two aggregation sites which they collected data mix in a homogenous or “equal” manner with each other.
Although it is well known that white sharks do show site fidelity to seal aggregation sites or rookeries, research has also shown that individual sharks show preference to specific locations and may restrict their movements as a result of this; limiting the mixing of individuals over small spatial distances. Based on these findings alone, it is clear that Chapple et al. should not have assumed homogenous sampling as they in fact had an increased probability of re-sampling previously observed sharks at each of their sample sites. This assumption violation results in a low bias towards their population estimates.
Assumption #3: Tagging method does not affect subsequent chance of sampling
One of the main issues with using photo-identification to model white shark populations is that it requires the sharks to be lured and baited to the surface where they can be photographed. It is therefore assumed that all sharks have an equal probability of being attracted, and that all sharks in the area will have an equal chance of being attracted again. However:
- In comparison to acoustic detections done by the Oceans Research team in Mossel Bay, photo-identification has been found to show the lowest probability of detection and tended to underestimate residency times and local abundance for white harks
- Intra-specific dominance patterns between large and small sharks may exclude some individuals from the area or inhibit their approach to the surface, leading to greater probability of low quality dorsal fin photographs and under-reporting of shark numbers
- Some individuals may learn that the bait does not lead to a food reward through negative attraction conditioning. These “trap-shy” sharks will therefore have a smaller probability of attraction during subsequent sampling events and be less likely to be re-sighted.
Although being unable to re-sight previously ‘tagged’ sharks may lead to population over-estimations, the latter two factors combined would lead to an underestimation of white shark population size, which is what Burgess et al. believes to be the case in Chapple et al.’s study.
Assumption #4: Zero tag loss
Unlike a fingerprint, the primary characteristic markings on the dorsal fin of white sharks – the pattern of notches on the trailing edge of the fin – have the potential to change with time. Although this “tag-loss” as a result of marking changes does not necessarily result in over-estimation of population size, it introduces the probability that individuals can be present but not correctly identified; violating the fourth assumption made by Chapple et al. (2011).
Assumption #5: Random sampling of the central California sub-population using photo identification
Burgess et al. commented in their review of Chapple et al.’s study that this was the most serious of all assumptions violated by the research group. As white sharks in the Californian region have been observed to reside predominantly at a single aggregation site and less often at others, the sharks sampled by Chapple et al. cannot and do not represent random samples of the central California white shark population. Instead, sampled sharks represent only sharks that visit these two aggregation sites either a single time or consistently over the study period. This violation is further compounded by the lack of sampling elsewhere in this oceanic region, such as the known white shark aggregation site of Año Nuevo Island which was excluded from population modelling for ‘logistical reasons’.
This post highlights how difficult it can be to not only collect abundance data for marine species such as white sharks but also how difficult accurately modelling and monitoring their population sizes can be. As Burgess suggested models cannot be applied blindly, but careful consideration of the assumptions behind is urged.
But what does this all mean in terms of conserving the white shark and other marine species? Does it really matter if the numbers aren’t quite right?
Keep posted for my next blog outlining the issues of protecting, conserving and monitoring marine species and why it is important that the community is supplied with the most accurate information possible in regards to population sizes and dynamics.
Burgess GH, Bruce BD, Cailliet GM, Goldman KJ, Grubbs RD, Lowe CG, Macneil MA, Mollet HF, Weng KC, O’Sullivan JB (2014) A Re-evaluation of the Size of the White Shark (Carcharodon carcharias) Population off California, USA. PLoS ONE 9(6): e98078.
Chapple TK, Jorgensen SJ, Anderson SD, Kanive PE, Klimley AP, et al. (2011) A first estimate of white shark, Carcharodon carcharias, abundance off central California. Biol Lett 7: 584-583.
Jorgensen SJ, Reeb CA, Chapple TK, Anderson S, Perle C, et al.. (2010) Philopatry and migration of Pacific white sharks. Proc Roy Soc B 277: 679-688.
Delaney DG, Johnson R, Bester MN and Gennari E (2012) Accuracy of using visual identification of white sharks to estimate residency patterns. PLoS ONE 7(4): e34753. doi:10.1371/journal.pone.0034753