Combining offtake and participatory data to assess the sustainability of a 1 hunting system in northern Congo

28 Research suggests that bushmeat is hunted at unsustainable rates throughout much of the


45
Wildlife is thought to be hunted unsustainably across much of Central Africa, and, indeed,

54
Quantitative measures are needed to assess the ecological sustainability of hunting (Sirén 55 2015). Hunting sustainability assessments often rely on models based on comparing hunter offtake in 56 a given area and time-period with the maximum wildlife population production (e.g., population 57 growth models; Robinson & Redford 1991). However, this approach to understanding hunting 58 sustainability can be limited. These models contain inherent uncertainty because of poor biological 59 knowledge of wildlife species and the different sampling methods used to calculate wildlife offtake 60 (Ingram et al. 2021). Sustainability is also time and context specific, requiring tailored assessments 61 that account for spatial and temporal variation in hunter offtakes (Clayton & Radcliffe 1996).

62
However, some models are commonly used to assess sustainability during a short period of time,

67
Whilst such models can be invaluable to conservation scientists working in an area, they may be of 68 limited benefit to community-based hunting management, which may require a more adaptive and 69 less technically challenging approach.

70
An alternative approach to measuring sustainability, is to infer sustainability by using 71 indices/proxies (Robinson and Redford, 1994). For example, monitoring changes in harvesting rates 72 over time, using the overall number of animals and/or biomass harvested in a given area and time 73 period, has been used to provide insights into whether a hunting system is moving towards or away

86
Offtake data are, however, notoriously hard to interpret as they represent the outcome of 87 several processes (Crookes et al. 2005

96
Here, we test the degree to which participatory hunter surveys can be used to distinguish 97 between changes in prey populations and changes in hunting strategy in a low-resolution offtake 98 dataset, therefore improving the effectiveness of long-term offtake datasets to assess the sustainability 99 of hunting. To examine this, we ask whether sustainability inferences based on offtake surveys, and 100 hunter focused participatory surveys with explicit knowledge of hunter identity (here ethnicity) and 101 behaviour (hereafter combined as 'hunting strategy'), resulted in different conclusions. We use data 102 gathered from a site that underwent socio-economic transformations in northern Republic of the 103 Congo due to new roads and the establishment of commercial forestry, where long-term offtake data 104 were collected as part of a Protected Area management program. The overall hypothesis of this 105 research is that long-term low-resolution offtake data combined with participatory hunter surveys can 106 explain changes and sustainability of a hunting system undergoing socio-economic change.
107 Specifically, we pose the following hypotheses: 1) New roads and associated socio-economic 108 developments increase hunting levels (number of animals, biomass), but without accounting for 109 hunting strategy we can only provide limited inferences about sustainability; 2) New roads and 110 associated socio-economic developments affect prey profile as medium-large animals are hunted out 111 around the village (halo effect) and small ungulates, primates, and other prey increase in the hunt 112 profile, but without accounting for hunting strategy we can only provide limited inferences about 113 sustainability; 3) Hunters can clearly articulate how they respond to changing abundance of wildlife 114 and local socio-economic changes and these changes can be incorporated into the design of offtake 115 surveys; and 4) Comprehensive assessment of hunting strategy in relation to hunt offtake (number of 116 animals per hunt, biomass per hunt, prey profile per hunt, and CPUE per hunt), illustrates that hunting 117 strategy significantly influences offtake.

131
Traditionally the Aka were semi-nomadic, spending between four and eight months a year in forest 132 camps (Kitanishi 1995), although due to influences of commercial forestry and the conservation 133 project, the Aka now increasingly spend far more time in the villages and gun hunt for the Kaka, as 134 Aka rarely own guns or cartridges themselves.

136
Hunting regulations and monitoring of wildlife offtake 137 Given that NNNP is uninhabited, its integrity depends largely on effective management of its

206
The long-term records are incomplete due to administrative (e.g., occasional staff shortages) 207 and political factors; in total the dataset contains 88 of 120 possible months (6693 animals) between 208 1997 and 2006, although no hunting zone data were available for 1998 or 1999. Therefore, to fill gaps 209 in the data, we extrapolated yearly totals from these data by calculating monthly averages from 210 existing data within each year, and multiplying these averages for missing months within each year to 211 account for differences in the number of hunters per year. Extrapolating assumes that monthly offtake 212 is constant throughout the year, and that there are no seasonal peaks, which is rarely the case.

431
Overall, 59% of day hunts by Aka hunters contained medium-large ungulates, compared to 432 only 16% of day hunts by Kaka hunters (Figure 6). This supports claims by Kaka hunters that the 433 Aka's superior hunting skills make them more able to hunt medium-large ungulates. However, this 434 difference is not apparent when lamping at night, where hunting a medium-large ungulate requires 435 less skill than during the day.

436
Similarly, Aka return with a marginally higher proportion of small ungulates during the day 437 compared to the Kaka, although this difference was not apparent at night. Kaka instead often returned 438 with primates, present in 66% of Kaka hunts, compared to only 36% Aka hunts.  Hunters provided strong evidence for hypothesis 3, that hunters can clearly articulate how 501 they respond to changing abundance of wildlife and local socio-economic changes and these changes 502 can be incorporated into the design of offtake surveys. The results from the participatory techniques 503 supported the results from the long-term monitoring data: hunters reported a reduction in medium-504 large ungulates in Sombo hunting zone, and reported that they increasingly used Loundoungou zone.

505
Although hunters were not able to provide us with the quantitative data generated by the long-term 506 monitoring dataset, the general conclusions based on the findings of both methods would have been 507 the same. The additional advantages of the participatory techniques in this case were their ability to 508 describe changes in hunter strategies, and give explanations for these changes, and this understanding 509 is vital to informing management. PRA techniques should be seen as a useful tool to complement 510 long-term scientific data collection, involve hunters in sustainability assessments, and therefore 511 community management of wildlife.

512
We find strong evidence for hypothesis 4, that wildlife offtake is linked to hunting strategy, containing small ungulates and more than one animal than day hunts. In addition, based on low CPUE 519 and a low proportion of medium-large ungulates hunted when hunts occurred close to the village, the 520 high-resolution offtake dataset was able to provide some evidence for a 'halo' of depletion around 521 Makao-Linganga. Based on this knowledge, and the fact that hunters reported increasing their travel 522 distance over the years, it is highly likely that the long-term dataset is suffering from hyperstability.

523
However, the lack of monitoring of these behavioural factors in the long-term dataset means it is 524 impossible to quantify the effect of these factors. Nevertheless, by showing that hunting strategy 525 affects offtake, and providing qualitative data indicating that hunting strategy has changed over time, 526 the high-resolution offtake dataset has helped us understand that the long-term dataset is compounded 527 by changes in hunting strategy. Despite this, the comparisons of prey profile between the three 528 principal hunting zones using the different datasets all reached the same general conclusion.

529
The question remains as to which offtake metrics are the most accurate indicators of 530 sustainability. Kümpel et al. (2010) argue that changes in prey profile and CPUE are the most 531 accurate. In our case study, CPUE yielded a different result from prey profile data. This is possibly 532 due to the measure of effort we used, which assumed that hunters hunt for an average of eight hours 533 for every hunt, regardless of their distance from the village. However, Rist et al. (2008) demonstrate 534 that CPUE is sensitive to the measure of effort used: the authors illustrate that the proportion of total 535 time spent hunting actually decreases with increasing distance from a village in Equatorial Guinea. 536 Importantly, it is evident from our findings that the likelihood of catching each prey size/functional 537 group was influenced by different processes. For example, the ethnicity of the hunter during day hunts 538 influenced the likelihood of catching medium-large ungulates, while the likelihood of catching small 539 ungulates and primates was instead influenced by whether the hunt was conducted during the day or 540 night. However, the presence or absence of medium-large ungulates in hunt harvests yielded the same 541 results for all three methods of assessment. Medium-large ungulates are preferred prey species in this 542 region, and preferred species are killed when encountered by hunters, and therefore are more likely to 543 represent wildlife densities than non-preferred species. Hunters based the quality of different hunting 544 zones based on their perceptions of the abundance of medium-large ungulates, indicating this is a 545 locally significant indicator of sustainability. Ultimately, however, accurate indicators of 546 sustainability from offtake data may vary between sites, but must be one which represents local 547 processes that are understood, e.g., in this example the dynamics between Aka and Kaka hunters.

578
Interpretation of offtake data, when used to make inferences about sustainability, requires an 579 understanding of the degree to which offtake reflects the relative abundance of wildlife populations.

580
Put simply, changes in hunters' behaviour can mask changes in wildlife abundance as hunters adapt 581 their hunting strategies to evolving socio-economic contexts and wildlife depletion. We have shown 582 how hunters adapted their strategies in light of the arrival of a commercial forestry road, and that 583 including data on these hunting strategies in offtake monitoring protocols provided a qualitative 584 understanding of the effect of hunting strategy, therefore providing explanations for offtake trends in 585 the long-term low-resolution monitoring data. Incorporating hunting strategy into data collection 586 protocols, and using participatory techniques to understand changes in hunting strategy, is one way to 587 control for the effect of hunting strategy. We believe this method provides a valuable way to increase 588 the reliability of inferences about sustainability made from offtake data, and could easily be 589 incorporated into community-based hunting monitoring efforts and management.