In our paper entitled “Unintended consequences of conservation actions: managing disease in complex ecosystems”, we used an investigation of a possible impact of a vaccination campaign in the Serengeti to argue for a holistic approach towards conservation management. In their comment, Packer and colleagues agree with our overall conclusion about the merit of a holistic approach to conservation; however they take issue with the example we use. Our example centres on a vaccination program on domestic dogs against Canine Distemper Virus (CDV) and its impact on the Serengeti lion-cheetah relationship and cheetah population viability. This vaccination campaign, originally named “Project Life Lion” [1-3], was launched in 1996. Its stated aim was “to vaccinate tens of thousands of domestic dogs in villages around the Serengeti National Park against canine distemper and rabies in an attempt to prevent further epidemics of these diseases in wild carnivores” [4].

Packer et al. suggest that we used “incorrect correlative inferences between dog vaccinations and changes in lion numbers”. The authors then question the approach we used to link lion and cheetah population dynamics, stating that we “fit a questionable relationship between lions and cheetahs to infer negative impacts on cheetah populations”. Here, we address the issues raised by Packer and colleagues about our work and conclusions.

"Correlative inferences between dog vaccination and changes in lion numbers"
We did not make any specific inference about any correlation between dog vaccination and changes in lion numbers. Our only assumption was that the vaccination campaign should, at least in part, address its overall aim to prevent further epidemics of CDV in wild carnivores in the Serengeti National Park [4]. As we state in our paper, we agree that it is a reasonable assumption that establishing a vaccination cordon sanitaire might prevent such epidemics in lions, and we know from available published information that CDV outbreaks can cause large mortality in lions, as proved by the lethal outbreak of 1994 that killed a third of the Serengeti lions [5]. We therefore used a scenario-based approach to explore the impact of a potential change in the likelihood of a lethal CDV outbreak on lion and cheetah dynamics, varying the number of lethal outbreaks that could occur over the 60-year time frame from 0 to 6. In our scenarios, the only inference was that a CDV vaccination campaign could reduce, to some extent, the likelihood of a potential lethal CDV outbreak. This inference is in line with the stated aim of the vaccination campaign. We made no assumptions or correlative inferences about any direct link between vaccination and lion numbers.

Whether or not CDV was detected in the lion population after the start of the vaccination program does not necessarily mean that the campaign has had no impact on lion abundance, and most importantly, on the likelihood of a lethal CDV outbreak. As acknowledged by Munson and colleagues [6], the 1999 and 2007 outbreaks were “silent” as no associated mortality was reported. Available peer-reviewed literature and the personal communications we received from Packer and colleagues provided no additional information on the impact of the CDV vaccination campaign on lion numbers and on the likelihood of a lethal CDV outbreak. Therefore, although the arguments presented by Packer and colleagues support the importance of tackling the issue using a scenario-based approach, they do not invalidate our original assumption, i.e., that a CDV campaign might have affected the likelihood of a lethal CDV outbreak.

We accept that if the vaccination campaign has had no impact on the likelihood of a lethal CDV outbreak on lions, then there should be no correlation between vaccination and cheetah dynamics in our analysis; but this then also raises a wider issue about whether the vaccination campaign has been able to address its stated aims.

"Relationship between lions and cheetahs"
As extensively discussed in our study, there is substantial evidence that lions have a negative impact on the survival of cheetah cubs [7,8]. There is also evidence of a negative correlation between cheetah and lion population size [8]. This was also supported by our analysis, where including lion abundance to our cheetah population dynamics model improved the model fit from explaining only ~10% of variability in cheetah numbers to explaining nearly half of this variability (Table 2). We agree that precise manner in which we modelled the impacts of the lion population on cheetah cub survival had a relatively small effect. Here we selected the most plausible relationship by selecting a threshold value that yielded the highest r-squared between modelled and observed cheetah abundance (Figure S1). The statement that we “made no attempt to assess the goodness of fit under the assumption that cheetah numbers vary independently of lion numbers” is however incorrect: goodness of fit of the cheetah population model was assessed under three different situations, namely (1) without incorporating lion dynamics, i.e., no lion impact; (2) coupled to the lion population model; and (3) coupled to published lion abundance. We emphasise again that model (3) provided a substantially improved fit (45%) compared to model (1; 10%), demonstrating that lions are likely to have a key impact on cheetah population dynamics.

In conclusion, we welcome the support of Packer and colleagues for holistic approaches to disease management and their efforts to clarify our work. We hope we have addressed their concerns in this response.

References
1. Cleaveland S (1997) Dog vaccination around the Serengeti. Oryx 31: 13-14.
2. Cleaveland S (2011) Project Life Lion - A Synopsis By Dr. Sarah Cleaveland. http://www.pwpark.com/abo.... Accessed 2011 Feb 16.
3. African Lion Conservation (2011) http://www.african-lion.o.... Accessed 2011 Feb 16.
4. Oryx (1996), “Dogs to be vaccinated around Serengeti” 30(3) :169
5. Roelke-Parker ME, Munson L, Packer C, Kock R, Cleaveland S, et al. (1997) A canine distemper virus epidemic in Serengeti lions (Panthera leo). Nature 379: 441-445.
6. Munson L, Terio KA, Kock R, Mlengeya T, Roelke ME, Dubovi E, Summers B, Sinclair ARE & Packer C (2008) Climate extremes and co-infections determine mortality during epidemics in African lions. PLoS One 3: e2545.
7. Laurenson MK (1994) High juvenile mortality in cheetahs and its consequences for maternal care. Journal of Zoology 234: 387-408.
8. Laurenson MK (1995) Implications of high offspring mortality for cheetah population dynamics. In: Sinclair ARE & Arcese P (eds) Serengeti II Dynamics, management, and conservation of an ecosystem. The University of Chicago Press.


Alienor Chauvenet
Sarah Durant
Nathalie Pettorelli
Institute of Zoology, Zoological Society of London,
Regent’s Park, London, NW1 4RY

Ray Hilborn
School of Aquatic and Fishery Sciences,
University of Washington, Seattle, Washington 98195 USA