Four Continent of Africa Problem Statements
Problem Statement 1: Mitigating Human Wildlife Conflict
Human-Wildlife conflict is a well-known phenomenon which occurs when growing human populations overlap with established wildlife territory, creating reduction of resources or life to some people and/or wild animals. It is broadly defined as “any human-wildlife interaction which results in negative effects on human social, economic or cultural life, wildlife social, ecological or cultural life or the conservation of wildlife and their environment” (adapted from the IUCN/SSC African Elephant Specialist Group) and may comprise conflicts over resources, especially food and water that humans perceive to be owned by them, and wildlife attacks on humans. Including ‘any perceived or actual threats to human safety or property by wildlife and ensuing retaliatory responses by affected humans.
Examples of human-wildlife conflicts in Africa include destruction of fishermen’s nets by sea turtles or crocodiles; crop raiding or destruction in farming communities by elephants, chimpanzees, baboons, bush pig, manatees; transmission of disease between livestock and wild animals; killing of livestock by lions or buffalo; Chimpanzee, snake or elephant attacks on women going to farm or children going to school; crocodile or hippopotamus attacks on humans on river banks or in water; competition for water between village communities and wildlife; or habitat destruction by fire, logging, urbanization that are human driven.
Ultimately, these interactions result in loss of life and injury to humans; loss of life and injury to livestock; loss of farm production and revenue; injury or death of wild animals; negative attitudes and mutual distrust between humans and wild animals. There are well documented cases such as the reduction of lion populations across Africa’s drylands, increasing incidence of poisoning of carcasses and depletion of vulture populations, intentional hunting of sharks, manatees, crocodiles and sea turtles by fishermen.
The Challenge: Human populations will continue to grow, thus requiring even greater surfaces for farming and urban development and increasing loss of forest and other habitats for already endangered species. As such the risk of conflicts will become increasingly higher. The challenge is to develop tools and models for managing multiple land uses including real time monitoring and mapping of potential/actual conflict zones and species, in order to develop innovative policy, technical and technological solutions to target and mitigate the various drivers, forms and consequences of human-wildlife conflicts. These may include but should not be limited to:
· Low cost but effective methodologies for mapping critical habitats and threats to priority species such as Pangolins, Rosewoods, sea Turtles, etc.
· Tools and technologies for tracking hotspots and new incidences of human-wildlife conflicts that make rapid and response and monitoring of outcomes easier for conservation managers
· Algorithms to scan websites and social media pages and collate existing data into usable formats for management and decision making
· Models for strengthening stakeholder awareness and attitudinal change with respect to the problem and promoting community involvement and benefits derived through its mitigation.
Problem Statement 2 -Managing the spread of Zoonotic diseases by the bushmeat trade
The bushmeat industry has been a topic of increasing importance among both conservationists and public health officials for its influence on zoonotic disease transmission and animal conservation Bushmeat, or wild animal meat provides an essential source of protein and income for human livelihoods throughout much of Africa. However, its consumption is linked to the transmission of zoonotic diseases, such as Ebola, and its over-harvest is a major threat to many wildlife species.
In order to reduce the spread of zoonotic diseases from the bushmeat trade, it is necessary to achieve one or several of three objectives:
· Prevention of disease spread from wild animals to a human host,
· Prevention of disease spread within local populations where a human host now exists
· Prevention of disease spread from a local to the global population
These require tracking and analysis of various types of variables in order to develop models, applications, systems and tools for more targeted analysis, decision making, screening and vector control. Many African states unfortunately do not have the resources for large scale deployment of health workers or facilities to address the growing threat of zoonotic disease spread that is facilitated by the harvest and consumption of bushmeat.
Nevertheless, little information exists on patterns of contact with wildlife in communities that practice bushmeat hunting, and capacity to predict emergence of new zoonoses has been limited so far especially in terms of understanding the environmental changes that drive them. Consequently, outbreaks of zoonotic diseases when they occur are often identified after the spread has begun to other localities. Response options are largely reactive and poorly planned, rather than predictive and well planned. Strengthening the ability to collect the requisite data, undertake appropriate analysis, coordinate cross sectorial exchange of data and expertise beyond country borders, create predictive models, deliver health and safety education toolkits to grassroots hunting communities, provide online screening of suspicious cases or hotlines for reporting and tracking suspicious cases would improve efficiency of prevention, response options and early containment of outbreaks.
The Challenge: Lack of extensive public awareness about the risks posed by zoonotic diseases, and links to bushmeat harvest and consumption or to processing and handling techniques is a limiting factor to prevention of zoonotic disease transfer to humans via the bushmeat trade.
Limited ability to collate data on factors such as rates and location of deforestation, distribution of species that are known to be potential vectors of zoonotic diseases, human demographic patterns, location of bushmeat markets, certification and monitoring of hygiene standards in bushmeat markets, access to awareness raising programs, or access to health screening facilities constrains the establishment of predictive models and efficient deployment of limited human resources and material supplies for rapid response towards zoonotic disease outbreaks.
Poor communications coverage in secluded rural areas and lack of permanent access to internet-based communications options, weakens response coordination and containment strategies.
The purpose of this challenge is to promote development of databases, maps, applications, communication systems or other tools to aid awareness raising, prediction, detection, and coordination of responses towards transmission of zoonotic diseases from bushmeat harvest, trade and consumption.
Improving access of local NGO’s to conservation funding
The Earth Summit in 1992, was a turning point in the role of NGOs in addressing development and Environmental issues. Agenda 21 which emerged as the end document, highlighted the important role of NGOs in sustainable development and preservation of the environment. Consequently, the late 90’s and early 2000’s saw increased levels of involvement of NGOs in shaping development agenda and implementing solutions at the grassroots level. Indeed, NGOs are now recognized to be an important stakeholder who need to be empowered technically, materially and financially in order to achieve sustainable development goal No. 15 which addresses sustainable management of forests, combating desertification, halting and reversing land degradation, and halting biodiversity loss.
Despite the increased recognition and involvement of NGO’s in sustainable development generally, and in conservation of threatened species specifically, a major handicap remains their access to funds for attending trainings and conferences, funds for conducting research, advocacy and behavior change activities within local communities and with other stakeholders involved in activities that threaten biodiversity. Indeed, every survey made on challenges for implementation of strategies or action plans, invariably comes up with access to financial resources as a major problem for NGOs and other stakeholders. This comes against a backdrop of increased funding levels committed by bilateral and multilateral agencies for the conservation of nature.
This situation is particularly critical for local/grassroots NGOs and volunteer organizations who do not often have the money to fund refreshments during sea turtle beach patrols, to pay for an office, a computer or phones for informants to use in providing alerts on criminal activities. Indeed, there seems to be a large disconnect between available funds, and the amounts that actually get to support local / grassroots Conservation NGOs in Africa. Arguably, these NGOs suffer from lack of visibility and credibility, or even simply do not know what sources of funding exist that are tailored for their circumstances. Likewise, many funding agencies do not have a clear identification of active even if under-resourced NGOs that could be boosted through targeted grants and mentoring programs.
With the significant increase over the last decade, in trafficking of wildlife, spread of zoonotic diseases through unregulated trade in bushmeat, and overall global impacts of unsustainable human activities around forests and wetland environments, it is more expedient now, than ever before, to bridge the donor community and grassroots NGO’s that are on the ground and willing to make a difference for conservation. With their low overhead rates, active engagement and exceptional mastery of local context, such NGO’s could achieve significant returns on investment if they were to receive targeted support and mentoring from funding agencies and capacity building institutions.
The Challenge: Despite major announcements over the years on availability of funding to address the challenges outlined at Rio+20, and subsequent development and / or conservation summits, it is often not clear how NGOs will have access to such funds and improve their role in reducing the impact of identified challenges. Grassroot NGOs play a critical role in supporting the poor and disadvantaged groups, and in conducting grassroots awareness raising, monitoring and advocacy or capacity building for conservation of threatened species, but funding constraints have prevented them from fulfilling their commitments. These constraints can be linked to lack of knowledge about potential sources and modalities for accessing funds, availability and cost of internet connections, absence of mentoring agencies specialized in facilitating institutional development and fundraising capabilities for grassroots NGOs.
· Tools and / or technologies that build, update and share databases of funding sources/opportunities
· Tools and / or technologies that build, update and share databases of local NGOs that would benefit from conservation funding and mentoring
· Databases of conservation mentors and volunteers who could provide online or physical services to local NGOs
· Online systems for that permit monitoring and direct reporting from the field, on use of funds acquired by local NGOs
· Tools and or technologies that permit the use of the abovementioned options in areas of limited internet connectivity
Problem Statement 3 - Preventing livestock depredation at pasture by large predators
The Problem: Livestock-carnivore conflict is a leading cause of retaliatory killings of large predators including lion, leopard, cheetah, and hyena. Pastoral communities often rely on the health of their cattle for their livelihoods and when cattle are killed by predators, the communities may retaliate by spearing, poisoning, or shooting the predators. Our data show that of the depredation events that occur at pasture, over 50% occur when livestock do not return to the corral after grazing and remain lost at pasture overnight. There are several reasons why livestock may not make it back to the corral at night: - Herders find themselves in an emergency and must return to the corral quickly - Livestock stray from the herd and are not reported lost until nightfall - Herders are unable to locate lost livestock before returning to the corral We believe these problems could be solved by better herding techniques, including livestock tracking, as well as rapid response of teams for finding lost livestock and responding to herder emergencies. Improved and immediate communication between herders and rapid response teams would also be critical for warning herders of approaching predators, often tracked via satellite collar. However, mobile network is often limited in areas where predators roam and livestock graze and most herders will not have access to a smartphone.
The Challenge: The goal of this challenge is to provide a technical solution to one or multiple of the situations described above. For example, you could consider the development of a device or app that would serve as a hotline to the nearest ranger team. You could also consider a livestock tracking device that would allow herders and support teams to locate lost livestock quickly. Additional considerations for this challenge include: - Most herders are young (about ages 7 – 14) and have limited experience with technology - Most herders will not have access to a smartphone and will not be literate, so solutions using app-based surveys may not be feasible. - Cellphone network is limited in many areas where livestock and predators roam, so solutions using radio, satellite, or other non-mobile network based technology are preferred. Bonus consideration will go to those who develop a solution both for herders to communicate with rapid response teams and for livestock tracking. Criteria: Develop a solution to either improve communications between herders and rapid response teams or track livestock (or both!). Competitors should provide: 1. Concept note for the proposed solution 2. Description of required technology 3. Any code with documentation used to facilitate the communication or tracking solution
Organization: Tanzania People & Wildlife Problem Statement POC: Elizabeth Naro, enaro@tanzaniapeoplewildlife.org Title: Preventing livestock depredation at pasture by large predators
Problem Statement 4: Identifying camera trap captures and remotely sending data
The Problem: Motion-triggered cameras (also called camera traps) can provide invaluable evidence of wildlife movements and behavior without the influence of human or vehicular presence. Camera traps can also be used to monitor human movement on less frequently trafficked animal paths or tracks. While this human movement is often benign, motion-triggered cameras located on animal tracks and paths can capture photographic evidence of individuals carrying weapons, materials for traps, or other paraphernalia related to wildlife and wild lands crime. However, in a heterogeneous landscape with limited vehicular access to remote animal tracks and paths, placing camera traps strategically to monitor wildlife that do not travel on roads (carnivores may be seen on dirt roads but herbivores will often be located in thicker vegetation), can be challenging. Many tracks and animal paths used by herbivores and people with nefarious agendas are far from roads and not easily accessible by a car. Therefore, many camera traps are located on roads for ease of changing batteries and downloading SD cards with photos. Due to the work involved and the battery life, downloading data from the camera traps occurs about once per month. If poachers or others committing wildlife or wild lands crime are observed in photos, it is often much too late to deploy game scouts to apprehend the individual. We believe we might be missing very important data by locating camera traps only along roads and gathering data too late to take action if and when individuals with weapons are observed. We believe these problems could be solved by better data collection techniques including locating camera traps in important inaccessible areas, increasing the frequency of data download, and speeding up the tedious work of identifying animals and entering data from camera traps photos (identifying thousands of photos per camera manually can take weeks of work, slowing down the process of identifying potential poachers). Automating the process of identifying animals and sending data remotely in 4 – 6 hour intervals would be ideal for deploying more camera traps to other areas, capturing animals in their natural movements apart from roads, and observing people that might have weapons for poaching or other illegal activities. If data are received in the office at shorter intervals and specific alerts sent to key individuals when an algorithm identifies potential illegal activity, organizations could deploy antipoaching teams rapidly, and potentially before any incident occurs.
The Challenge: The goal of this challenge is to provide a technical solution to one or multiple of the situations described above. For example, you could consider the development of a solution to remotely send photos from the camera traps every 4 – 6 hours (possibly via satellite or mobile network), instead of manual downloading of the data once per month. You could also consider creating an algorithm to identify animals, people, and vehicles from the photos based on an existing list of species in the area and an existing database of identified photos. Particularly, identifying individuals carrying weapons would improve anti-poaching activities. The identified photos would then have to be sorted by species and the data (including date, time, camera ID, species, etc.) written out into a readable and analyzable format (i.e. excel or google sheets).
An example of a camera trap photo is below:
Date: October 6, 2020
Time: 7:52:36
Camera ID: R011
Species: African Buffalo
Count: 1
(Note: the software does not allow a photo attachment here, but a full range of photos can be access on dropbox below.)
https://www.dropbox.com/sh/qv05v7p6y6rty3f/AADHGVANS3HygNZvndROCuI3a?dl=0
Criteria: Develop a solution to either improve data collection frequency by sending data to the office remotely and/or develop an algorithm for identifying wildlife, people, and vehicles from the photos. All competitors should provide:
1. Concept note for the proposed solution
2. Description of required technology If proposing a solution to identify camera trap photos, competitors should also provide:
3. Any code with documentation used to identify wildlife, people, and vehicles
4. Any code with documentation used to enter these data automatically into an excel sheet, google sheet, or other output format.
5. Estimate of the accuracy of the algorithm so we can understand the extent to which manual ground-truthing is required.
Four Global Problem Statements
Global Problem Statement 1
Profiling and Combating Zoonotic Disease Risk from Wildlife Trafficking, Wildlife Markets, and Human Encroachment: A Decision-Making Tool
Organization: The Bureau of Oceans and International Environmental and Scientific Affairs (OES) at the U.S. Department of State and USAID Regional Development Mission for Asia
Overview of the Problem:
A majority of emerging infectious diseases with pandemic potential originate from wildlife, such as coronaviruses, Influenza, Ebola, and HIV/AIDs. Amid the ongoing COVID-19 pandemic, experts around the world have emphasized importance of characterizing spillover risk in different locales and in different species.
Wildlife trafficking, wildlife consumption for food and medicine, and encroachment into wildlife habitat are forms of contact that drive the emergence of zoonotic disease. High-risk wildlife wet markets, in which wildlife is slaughtered and sold alongside many different species, are hotbeds for wildlife trafficking and create key zoonotic disease risks. These markets keep many different species, which would never otherwise be found in nature, together in cramped conditions. Often these markets are not well-regulated or inspected for legality or public health and hygiene standards.
Destruction of protected wildlife habitat and encroachment into wildlife habitat areas also increase the risk of zoonotic disease. Poorly regulated forms of encroachment, such as from illegal mining and illegal logging operations into wildlife habitat puts humans closer to wildlife — which leads to greater transmission risk of pathogens into human communities. Encroachment also makes it easier for wildlife traffickers to poach protected species. Construction of illegal roads not only paves the way for illegal mining and logging, wildlife trafficking, and illicit drug production, but also increases zoonotic disease risk.
Understanding risk of spillover from wildlife – through wildlife trafficking, consumption, high-risk markets, and encroachment- is fundamental to developing effective prevention and risk mitigation. A better understanding of risk can inform policy and regulations aimed at reducing emerging disease risk associated with wildlife trade. The assessment of specific markets as “high risk” or “low risk” with regard to spillover can also inform targeted interventions by public health or law enforcement professions, which is important in the face of scarce financial resources.
The Challenge: Identify, manage, and reduce the risk of zoonotic diseases, from wildlife trafficking and consumption, wildlife markets, or human encroachment into wildlife habitat. For example:
· The risk profiles of wildlife markets with regard to pathogen spillover vary among markets. Some markets will be “high risk” and others will be “low risk”. The challenge is to:
a) evaluate the most relevant criteria to be part of a risk profile (e.g. certain species or practices, such as slaughter on site), and evaluate a process for obtaining this information (preferably by “citizen scientists”), and propose a pipeline or platform for this data collection,
b) map networks of trade and other activities (e.g. patterns of consumption or encroachment) that contribute to spillover and amplification risk, and
c) to incorporate parts 1 and 2 into a decision making tool for interventions in wildlife markets, based on the risk of zoonotic spillover and transmission in specific wildlife markets. The ranking guidelines and methodology should be validated in different countries with varying wildlife trade and marketing contexts.
Resources:
Woods, M., Crabbe, H., Close, R. et al. Decision support for risk prioritisation of environmental health hazards in a UK city. Environ Health 15, S29 (2016). https://doi.org/10.1186/s12940-016-0099-y
Huong NQ et al. Coronavirus testing indicates transmission risk increases along wildlife supply chains for human consumption in Viet Nam, 2013-2014. PLoS One. 2020. https://doi.org/10.1371/journal.pone.0237129
UN Food and Agriculture Organization. Characterizing Livestock Markets for Real Time Decision Making: The Market Profiling Application. Sept 2019. http://www.fao.org/3/ca6132en/ca6132en.pdf
Global Problem Statement 2
Organization: Vulcan Inc.
Problem Statement 2: Organizations trafficking wildlife products can be highly complex, involving many people and functional units. Disabling these organizations effectively is a challenging task, often tackled by identification of key individuals or relationships for intervention via known connections (following the money). Identifying these leverage points often involves synthesizing a huge amount of data, such as car registrations, gun registrations, financial transactions, informant information, known personal relationships, and when people are seen together. Graph representations are often helpful in making sense of it all, but collecting and processing all relevant data to generate these graphs can be hugely time-consuming. A tool to automate preprocessing, ingestion, and display of this data would save valuable investigator time.
Possible bonus features:
a. Low-impact way for informants to submit additional information
b. Graph Machine Learning to automatically classify individuals or connections of interest, or to hypothesize missing connections.
Semantica AI is a possible source of inspiration, though their pricing model is prohibitive for many organizations.
Graph theory, conceptual overview: https://medium.com/basecs/a-gentle-introduction-to-graph-theory-77969829ead8
Survey on machine learning graph applications: https://arxiv.org/pdf/2005.03675.pdf
Possible datasets: Synthesize similar to datasets at https://icon.colorado.edu/, and the output goal of the ingestion process should be data in forms similar to those there.
Possible raw input files could include spreadsheets of vehicle identification numbers and license plates, separate spreadsheets of ownership information, lists of known associate groups, incident reports with names, dates, vehicles, and weapons, financial transactions between people or companies, and so on.
Global Problem Statement 3
Organization: Vulcan Inc.
Problem Statement 3: Many animal populations vulnerable to poaching are tracked to some degree with collared or tagged individuals. Identification of when and where these groups are threatened would enable better-targeted interventions on their behalf. The ability to ingest and at a high level characterize the behavior of animal groups such as those available at movebank.org would support the protection of wide-ranging or remote animal populations. Create a herd/group behavior identifier algorithm that can ingest this track information along with a list of poaching incidents with a wide time and location window (self-generated for testing, as this information is sensitive) and outputs an estimation of where in the track these incidents occurred.
Behavior/anomaly detection primer: https://towardsdatascience.com/a-note-about-finding-anomalies-f9cedee38f0b
An example of track-based behavior detection: https://www.researchgate.net/publication/265964999_Automatic_detection_of_suspicious_behavior_of_pickpockets_with_track-based_features_in_a_shopping_mall
Possible datasets:
Those available at movebank.org
Relevant sets available at https://www.ncei.noaa.gov/ or data.gov, such as https://catalog.data.gov/dataset/sea-turtle-satellite-telemetry-data
Global Problem Statement 4
Organization: TRAFFIC
Problem Statement POC: Giavanna Grein (Giavanna.grein@traffic.org), Senior Program Officer
Title of Problem Statement: Protecting Wildlife and People from the Risks of Online Trafficking in Wildlife
The Problem:
The current pandemic has revealed the fragile link between human health and wildlife exploitation, and how poorly regulated and illegal trade in wildlife can catalyze disease transmission and shatter global economies. The World Health Organization determined that COVID-19 is a zoonotic disease, meaning it originated from an animal. Other zoonotic diseases to date have included SARS, Ebola, Bird Flu, and MERS. COVID-19 is suspected to have originated in bats and may have jumped to humans via an intermediary wild species like the pangolin. With physical wildlife markets under scrutiny or suspended and people under lockdown in many countries to stop the spread of COVID-19, sellers are turning to online marketplaces and social media platforms to offload stockpiles of live wildlife, wildlife products and meat originally destined for physical markets. The sale of these items online further increases the risk of disease transmission to human populations through the use of delivery and express courier services, or direct selling to interested buyers in person. Online market places have been increasingly exploited by wildlife traffickers over the last decade, with social media platforms now serving as the main mechanism to connect buyers and sellers. TRAFFIC and NGO partners WWF and IFAW convene the Coalition to End Wildlife Trafficking Online which unites the tech sector to reduce wildlife trafficking online. This Coalition, which is comprised of 36 member companies including Google, Facebook, eBay, Alibaba and Tencent, works to standardize prohibited wildlife policies, train company staff to better detect illicit wildlife products such as elephant ivory and live tiger cubs, enhance automated detection filters, and educate and empower users to report suspicious listings. The Coalition has achieved great success to date, including over 3.3 million listings blocked or removed by company members in two years, though the widespread trade in live animals online is still a challenge that requires technology solutions.
The Challenge:
Convenors of the Coalition, as well as tech company enforcement teams, are limited in capacity for manually searching online platforms for prohibited live animals for sale. This process is very resource-intensive and inefficient when searching on a global scale. To reduce the trade in high-risk live animals that may transmit zoonotic diseases to humans across global supply chains, the challenge is to develop a tool that will identify and alert these sales taking place on one social media platform in one language as a starting point, with the potential to scale to additional platforms and languages in future. This will allow TRAFFIC, NGO partners and law enforcement agencies to flag new demand trends and emerging markets — and therefore target where action is needed to mitigate or eliminate risk.
Criteria:
· The tool developed should be able to scan one social media platform to identify high-risk live animals offered for sale. If feasible in the time provided, it would be great to capture information on where the seller is located, contact details, where they will ship the animal, any buyer information available, any reference to the health of the animal, the number of animals available, and which species are offered.
· The tool should be able to search in one language to start, with the ability to scale in future to include multiple languages. Zoohackathon participants may choose which language to include based on location. Sample languages include: Arabic, English, French, Portuguese and Vietnamese.
· The tool should include a warning system that will alert TRAFFIC of these risks to coordinate a response.
Things to consider that make online monitoring challenging:
Sellers are able to create, delete and recreate accounts and profiles as needed to avoid detection.
Not all sellers list an animal as for sale, or even include the name of the species. Some may simply use images of the animal as a means of advertising and let interested buyers comment on their posts. From there the conversations are taken into private communications like WhatsApp chats. Language in the comments will include things like ‘PM/DM’ for price and how much?
Many of the listings for live animals are found in private groups which will require admission by a group administrator.
It is important to note that the sale of endangered wildlife is an illegal activity and that many buyers and sellers involved are criminals. No member of your team should engage with these sellers directly, or like/comment any of their content. Nor should any member attempt to purchase any animals or products.
Data Sets and Other Resources:
To learn more about illegal wildlife trade online visit the Coalition to End Wildlife Trafficking Online website.
To review online monitoring reports from TRAFFIC visit the publication page.
To learn more about the link between COVID-19 and wildlife trade see TRAFFIC’s Wildlife trade, COVID-19 and zoonotic disease risks: shaping the response report.