Over the past two decades, there has been a sharp increase in climate hazards, and forest fires in particular have occurred alarmingly often. Forest fires now result in an annual tree cover loss of three million more hectares than they did in 2001; in 2021 alone, 9.3 million hectares of tree cover were lost to forest fires.[1] The World Economic Forum estimates that the annual global cost of wildfires is about $50 billion.[2]
Forest fires have severe impacts on local ecosystems, including but not limited to biodiversity loss, soil erosion, loss in soil nutrients and porosity, human loss, and property damage. Globally, governments are working on implementing technology-based systems to manage and monitor forest fires. However, in low-resource contexts of developing countries it is challenging to implement such systems.
Frontier technologies are helping countries respond to fire disasters
Applications based on Artificial Intelligence (AI), Machine Learning (ML), the Internet of Things (IoT), Augmented Reality and Virtual Reality (AR/VR) – collectively referred to as frontier technologies – can be used to combat climate change, making countries more resilient to forest fires, with the potential to mitigate them in the future.
The United Nations Framework Convention on Climate Change (UNFCCC) defines frontier technologies as those that will ‘reshape industry and communications and provide urgently needed solutions to global challenges and have the potential to displace existing processes’.[3] What makes frontier technologies especially promising are their shared characteristics of portability, replicability, and affordability.
Frontier technologies are being applied to combat forest fires by creating detailed geo-spatial maps indicating the fire susceptibility of specific areas, mapping soil use and forest health to identify deforestation and other misuse of forests, detecting forest fires and triggering early warning systems (EWS) to alert local communities, and predicting where fires are likely to spread when they have occurred. These applications can also be used to monitor areas after the fire has been suppressed to check for re-ignition and collect information to inform future fire-fighting procedures.
Pakistan is highly vulnerable to forest fires
Pakistan, with its diverse topography, ecosystems, and climate zones ranks high on the ND-GAIN Index which measures a country’s vulnerability to climate change. Pakistan suffers from consistent high temperatures and low rainfall, resulting in extreme climate events. Human activities such as agriculture and mining pose a further threat of losing forests. Northern and Eastern areas have the greatest forest cover, and a majority of fires occur in the Khyber Pakhtunkhwa region, which has a forest cover of 15%, and the capital region of Islamabad.
Detected incidences of annual fire events in Pakistan (2012-2020)
Source: VIIRS data from Fire Information for Resource Management System (FIRMS), NASA.
Pakistan currently relies on manual processes to tackle forest fires
Forests in Pakistan are governed by several laws and policies at the national and provincial government level, which lay out the framework for the management and conservation of forest areas. Forest fire prevention and management is undertaken by representatives of several government entities along with local communities at the divisional level, who form Forest Fire Management Teams (FFMTs).
In the event of a fire, FFMTs are responsible for raising the alarm and ensuring that relevant processes of fire management, communication, and evacuation are followed. Most of these processes are manual, such as the deployment of personnel in high-risk zones to monitor any occurrences of fire, or manual dispatches of alerts and dissemination of information within government agencies. Relying on manual processes leads to delayed responses and loss of life owing to human error, among other challenges. Further, the processes are focused more on fire detection and response rather than fire prevention.
An example of an AI-based solution for Pakistan
Efforts are already underway to develop AI-based solutions for forest fires in Pakistan. An example is an ongoing project to apply AI and IoT solutions for forest protection, funded by UK Aid via the Frontier Tech Hub initiative and implemented by World Wildlife Fund (WWF) Pakistan and the Lahore University of Management Sciences (LUMS).
The output of the pilot of this project was an EWS, which was later adapted into a suite of solutions for fire mitigation, preparedness, and response. The EWS is based on fire risk maps developed using real-time IoT weather stations and open-source satellite images. The project is focusing on ensuring that the solutions that emerge can be adapted to low cost and in low resource settings, with the participation and buy in of local communities.
There are several barriers in the development of frontier technologies
For greater adoption of these solutions across Pakistan, there is a need to develop a more enabling environment. Key elements of an enabling environment are the availability of good quality datasets, mobile and internet connectivity, access to digital infrastructure (such as computing servers), and relevant human capital. A comprehensive framework, which lays out the roles of government agencies and local communities is also important.
Creating an enabling environment requires investing in research and development facilities, providing partnerships and funding for solution providers (e.g., start-ups, NGOs), building digital infrastructure, and introducing dedicated policies and procedures on the use of technology in forest fire management. Further, the deployment of AI-based information systems would require specialised expertise within the government, and investments in building the capabilities of local government agencies as well as communities.
Efforts to adapt AI applications in the context of Pakistan
The Central Insights unit at GSMA Mobile for Development is currently working with the UK FCDO country office in Pakistan on a research project to identify use cases of AI solutions for forest fire management in Pakistan, with a focus on the Khyber Pakhtunkhwa (KP) province and capital region of Islamabad. The team also plans to build a Pakistan Wildfire Framework as a forest fire management tool for government agencies and local communities in these regions.
[1] World Resources Institute (2023). The Latest Data Confirms: Forest Fires are Getting Worse.
[2] Wildfires refer to fires impacting all land areas including forests, grasslands, and tundra. Forest fires are a type of wildfire.
[3] UNFCCC website (2023). Frontier Technology.
THE CENTRAL INSIGHTS UNIT IS CURRENTLY FUNDED BY THE UK FOREIGN, COMMONWEALTH & DEVELOPMENT OFFICE (FCDO), AND SUPPORTED BY THE GSMA AND ITS MEMBERS.
If you are interested in or working on a similar area, please reach out to us at [email protected] to find out more about the project or to share your insights with us.