How can we tackle interruptions to human-wildlife feeding management? Adding media campaigns to the wildlife manager’s toolbox

In recent years, wildlife managers have been seeking ways to reduce the occurrence of independent, recreational human-wildlife feeding interactions, which continue to gain global popularity and may have negative effects on the humans and wildlife involved. Current popular methods, such as signage and posters, have yielded mixed results and their application is often interrupted, though the effects of these interruptions on feeding levels are currently unknown. This has led to calls to both identify a management option that can be applied successfully from a distance and to determine whether this action may assist in recovering long-term programmes from the potential effects of interruptions. Marketing and media tools have been shown to successfully change human behaviours in conservation campaigns, flagging them as a potential tool that could be applied to human-wildlife feeding management. 

Here, we performed a 4-year study using a wild fallow deer population in a popular urban green space as our model system. We tracked changes in human feeding behaviours across four different management stages. These included pre-management (stage 1), during traditional management (i.e. “don’t feed the deer” signage, stage 2), mid-interruption (i.e. COVID-19 pandemic, stage 3), and during the application of a structured media campaign (stage 4). 

We found that feeding by visitors decreased during traditional management (stage 2), but rapidly returned to pre-management levels during the interruption (stage 3) despite traditional controls still being in place. However, we discovered that feeding dropped significantly after the release of a media campaign (stage 4), despite the audience and conditions being unchanged. We also identified which imagery and educational messages resonated with viewers; information that can be applied to future campaigns in other locations.

We, therefore, recommend that wildlife managers both investigate and be prepared for the negative effects that interruptions of any type (e.g. the recent COVID-19 pandemic, other interruptions to funding) may have on ongoing management campaigns of this ilk. We recommend that media campaigns be explored as a potential tool to reduce the occurrence of the unregulated feeding of wildlife by humans in these sites, thereby promoting better human-wildlife coexistence. 

Methods
A key thing to note initially is that this data was collected over 4 key stages based on the management activities currently in action, aiming to reduce human-deer feeding within the site. These stages occurred across 4 years (2018-2021), with each stage representing a summer collection (i.e. June-July) for a consecutive year. These stages included pre-management (Stage 1), introduction of internal controls such as signage, posters, and ranger patrols (Stage 2), interruption of management by Covid-19 (Stage 3), and release of a targeted media campaign across both traditional and social media (Stage 4).

All areas of the Park that are open to the public were identified and divided into sectors. As fallow deer are naturally sexually segregated during the summer months, this was performed by splitting the Park into the area utilised by the males (eastern side of the Park) and the area utilised by the females (western side of the Park) as outlined by the Park rangers. Both the male and female areas were then subdivided into four sectors each, of roughly the same size and habitat type, resulting in a final total of eight sectors.  A stratified a priori schedule was then made to ensure that all sectors were sampled across all times of the day and days of the week, alternating between starting in a male area and female area sector each morning. This prevented pseudoreplication from occurring due to oversampling the same herds and areas repeatedly. All observation periods were scheduled during peak visiting hours within the Park (i.e. 8:30am to 5pm), when interactions were most likely to occur. If no deer were present in the assigned starting sector, then the sectors were walked systematically until a herd was found. A herd was defined as a group of 2+ individuals within 50m of each other and broadly within view of each other (i.e. they had to be a cooperative unit rather than separate groups that were unaware of each other due to barriers such as hills or thickets).

When a herd was located, data collection was initiated. The herd size, the start time of the observation, and the location of the herd (recorded using a Garmin etrex 30 unit) were all immediately documented. Observations of the herd continued until the scheduled end time (which varied daily based on a stratified schedule to ensure that observations encompassed all times of the day on all days of the week for both the male and female sides of the Park) or until the herd diverged or merged with another herd or ran from the area, at which point we terminated the observation and began a new one. The new observation began on either the remaining herd present in that sector, a herd in that sector selected at random (if multiple herds were present), or else, if the sector had been vacated rapidly (i.e. the deer had run out of the sector and out of sight, often due to some disturbance), then the sectors were walked again until a new herd was located. The end time was noted anytime an observation was terminated.

Every 15 minutes, the number of people within a 250-metre radius of the edge of the herd was documented. The number of those people that were exercising or walking a dog was also counted and later removed from the total number of people observed in each scan, providing us with a proxy for the number of visitors available to interact with the herd over the course of the observation. Whenever people approached to feed the deer, the number of people who attempted to feed was recorded. The herd location was also taken every 15 minutes. These GPS points were used to calculate a final mean location for each herd; i.e. a single GPS location point for that herd over the total observation period.

Observations were non-invasive and conducted under research permit UCD AREC-E-18-28. Observers remained passive, maintained their distance, and did not interact with either the deer or the public. Full collection protocols can be found in Griffin et al. 2022.

近年来，野生动物管理者一直在寻求方法，以减少独立发生的休闲式人与野生动物投喂互动——这类互动在全球范围内愈发流行，却可能对参与其中的人类与野生动物造成负面影响。当前主流的管控手段（如警示标识与宣传海报）收效参差不齐，且实施过程常遭中断，而此类中断对投喂行为强度的影响目前仍不明确。因此学界呼吁，既要研发可远程有效实施的管理方案，也要探明此类措施能否帮助长期管控项目抵消中断带来的潜在影响。已有研究证实，营销与媒体工具可有效改变保护运动中的人类行为，这使其成为人-野生动物投喂管控的潜在备选方案。

本研究以某热门城市绿地中的野生黇鹿（fallow deer）种群为研究对象，开展了为期4年的调查，追踪了四个不同管理阶段下的人类投喂行为变化：分别为管理前阶段（阶段1）、传统管控阶段（即设置“请勿投喂鹿群”的警示标识，阶段2）、管控中断阶段（即新冠疫情期间，阶段3），以及系统性媒体宣传阶段（阶段4）。

研究结果显示，传统管控阶段（阶段2）游客投喂行为有所减少，但管控中断阶段（阶段3）尽管传统管控措施仍在实施，投喂行为却迅速回升至管理前水平。但我们发现，媒体宣传启动后（阶段4），投喂行为出现显著下降，即便受众与现场环境均未发生变化。此外，我们还明确了哪些视觉素材与宣教内容更易获得受众认同，相关结论可推广至其他地区的后续宣传活动。

因此，我们建议野生动物管理者提前调研并做好应对准备，以应对各类中断（如近期新冠疫情、资助中断等）对这类持续性管控项目造成的负面影响。同时，我们建议将媒体宣传作为潜在管控工具，以减少此类场地内人类无节制投喂野生动物的行为，进而推动人与野生动物的良好共存。

方法
首先需要明确的是，本研究的数据采集基于四类核心管理阶段，旨在管控场地内的人鹿投喂行为，整个数据采集周期为2018-2021年，共4年，每个阶段均对应连续一年的夏季（即6-7月）采样。四个阶段分别为：管理前阶段（阶段1）、引入内部管控措施阶段（如设置警示标识、宣传海报及护林员巡逻，阶段2）、新冠疫情导致的管控中断阶段（阶段3），以及针对传统媒体与社交媒体开展定向宣传的阶段（阶段4）。

研究首先将公园所有对外开放区域划分为若干采样单元。由于黇鹿在夏季会自然出现性别分离现象，我们按照公园护林员的划分方式，将公园分为雄鹿活动区（公园东侧）与雌鹿活动区（公园西侧）。随后，雄鹿活动区与雌鹿活动区各自再被细分为4个规模与生境类型基本一致的采样单元，最终共计8个采样单元。

我们制定了分层预先设定的采样计划，确保所有采样单元在每日的各个时段以及每周的各天都能被覆盖，且每日上午的首个采样单元交替选自雄鹿活动区与雌鹿活动区，以此避免因重复采样同一鹿群与区域而产生伪重复问题。所有观测时段均安排在公园的游客高峰时段（即上午8:30至下午5:00），此时段内人与鹿的互动发生概率最高。若指定的首个采样单元内没有鹿群，则系统遍历周边采样单元直至发现鹿群。本研究中鹿群的定义为：至少2只个体间距不超过50米、且彼此可视范围内的群体，即需为协同活动的单元，而非因山丘、灌丛等障碍物彼此未察觉存在的分散群体。

发现鹿群后，即刻启动数据采集工作，立即记录鹿群规模、观测开始时间以及鹿群位置（使用Garmin etrex 30手持GPS设备记录）。对该鹿群的观测将持续至预设的观测结束时间，该时间每日根据分层采样计划调整，以确保公园东西两侧的观测覆盖一周内每日的各个时段；或直至鹿群分散、与其他鹿群合并或逃离该区域，此时终止当前观测并开启新的观测。新观测将优先选择该采样单元内剩余的鹿群；若该单元内存在多群鹿，则随机选取一群开展观测；若该采样单元内鹿群迅速撤离，即因干扰导致鹿群跑出单元并消失在视野中，则再次遍历周边采样单元直至找到新的鹿群。每次观测终止时，均需记录结束时间。

每间隔15分钟，记录鹿群边缘250米半径范围内的总人数，同时统计其中正在锻炼或遛狗的人数，并在后续计算中从总人数中扣除该部分，以此得到观测期间可与鹿群产生互动的游客数量代理指标。每当有人靠近鹿群尝试投喂时，记录尝试投喂的人数。每15分钟同步记录一次鹿群位置，所有GPS点位将用于计算该鹿群在整个观测周期内的平均最终位置，即该鹿群在整个观测期间的单一GPS定位点。

本研究的观测均为非侵入式，且已获得研究许可UCD AREC-E-18-28。观测人员全程保持被动状态，与鹿群和游客均保持安全距离，未发生任何互动。完整的采样方案可参考Griffin等人2022年的研究。