Young Climate Voices

Climate diplomacy: Harnessing AI for damage mitigation in Africa

By Sharon Tshipa

The African Union’s Climate Change and Resilient Development Strategy and Action Plan (2022-2032) highlights the use of computers and the internet as critical in building climate resilience across Africa. SHARON TSHIPA reviews research on the potential of Artificial Intelligence (AI) in mitigating or reducing loss and damage from extreme weather and also natural events (such as earthquakes and volcanic eruptions) across the continent.

INTRODUCTION

Young African climate activists such as Vanessa Nakate of Uganda, Lesein Mutunkei of Kenya, Yero Sarr of Senegal and Remi Zehiga of the Democratic Republic of Congo are among many youth voices calling on global leaders to urgently redress the climate crisis and limit global warming to 1.5°C in order to reach net zero CO2 emissions by 2050 (Masson-Delmotte et al., 2019). This bottom-up approach is inspired by the fact that although Africa’s global contribution to harmful greenhouse gas emissions is only 3.8%, the continent is enduring climate change’s most destructive effects (CDP, 2020).

While Africa’s bottom-up approach is still in its infancy, Africa’s top-down approach spans decades. The presence of Africa in the climate change debate two decades prior to the adoption of the United Nations Framework Convention on Climate Change (UNFCCC) in 1992 shows that African heads of state and government have long appreciated the gravity of the climate change challenge (Lisinge-Fotabong et al., 2016). African political leaders (through institutions such as the African Union) have made a number of groundbreaking decisions to help member states deal effectively, efficiently and equitably with the risks posed (Lisinge-Fotabong et al., 2016). Though plausible, their efforts are continuously threatened by, for example, lack of scientific data.

The lack of data has resulted in Germany, Japan and India being placed at the top of the list of countries suffering from extended periods of heat although sub-Saharan Africa is dubbed a “literal hotspot of heatwave activity” – a weather-related loss event listed as a major cause of damage (Eckstein et al., 2019; Harrington & Otto, 2020).

Lisinge-Fotabong et al. (2016) posit that African political leadership has recognised the importance and timeliness of Africa’s active engagement in coordinated climate diplomacy and design of robust policies for a collective effort to confront complex climate change challenges. This article argues that for climate diplomats to succeed, they need to deliberately and actively take advantage of various modern technologies such as Artificial Intelligence (AI) to achieve sustainable goals (Galaz et al., 2021; Javaid et al., 2022; Yigitcanlar et al., 2020).

In May 2022, an experts’ consultative meeting on developing a continental strategy for AI in Africa was held in Dakar, Senegal, enabled by the African Union High-Level Panel on Emerging Technologies (APET). APET views AI as worth harnessing for Africa’s socio-economic development (AUDA-NEPAD, 2022). The APET AI for Africa report provides guidelines for African countries to exploit AI-based technologies for the continent’s advancement (AUDA-NEPAD, 2022). Its AI strategy for Africa is due to be released in the first quarter of 2023 (AUDA-NEPAD, 2022). This article contributes to the ongoing AI for Africa dialogue, and seeks to inform the development of strategic policies that will consider the uptake of AI in mitigating or reducing loss and damage. Taking into consideration the constant rise of extreme climate-related events and the world’s limited ability to prepare for them, this article argues in favour of AI, as it can play an important role in addressing the climate emergency (see Nordgren, 2022; Taddeo et al., 2021). To date, AI has been found to enable the estimation of extreme yield loss when crops are harmed by unusual weather conditions (see Zhong et al., 2022) and evacuation planning, while assisting with damage assessment (see Rolnick et al., 2022). This article aims to answer the following question: How can AI be used to mitigate or reduce climate-related loss and damage in Africa?

A review of the literature (using open-access journals) shows that AI can effectively be used to mitigate or reduce loss and damage caused (directly and indirectly) by earthquakes, floods, meteorological droughts, locusts, soil erosion, bark beetles, lightning strikes, volcanoes, cyclones, tsunamis and wildfires. AI can also be used to manage forest destruction, landslides, electricity outages, food shortages and communication network damage. It can also boost the climate resilience of critical infrastructure, improve urban resilience, strengthen the resilience of vulnerable people in disasters and enhance smart farming to enable people to cope with the impacts of climate change extreme weather events on supply chains.

Floods

Among the articles reviewed, there is a strong consensus that floods are the most common natural hazards that frequently damage natural systems, threatening human rights to food, health and development (see Darabi et al., 2021; Fang et al., 2021; Lei et al., 2021; Shahabi et al., 2021 among others). Between 1996 and 2015, for example, 150,016 floods were recorded worldwide and they continue to increase in frequency and intensity, largely due to climate change (Auliagisni et al., 2022; Fang et al., 2021; Ghosh et al., 2022; Molina et al., 2022). To save properties and lives, accurate identification of remedial measures for a flood are necessary (Waseem & Manshadi, 2020).

In assessing the potential of AI in mitigating or reducing loss and damage, the research for this article showed that machine learning and deep learning models¹ can be used to produce reliable flood susceptibility maps (see Darabi et al., 2021; Fang et al., 2021; Ghosh et al., 2022; Lei et al., 2021; Lin and Billa 2021; Shahabi et al., 2021). Studies analysed further concede that identifying floods and producing flood susceptibility maps are crucial steps for decision-makers to prevent and manage disasters — a shift from the flood defence approach. Since current flood maps are not easy to acquire and are hard to understand, climate diplomats should motivate for the development of more comprehensive and user-friendly flood maps using community-based information, which could help ensure that communities that don’t have the necessary knowledge don’t develop a false sense of security (Auliagisni et al., 2022). These can also assist local administrators to adopt proper sustainable management plans to reduce future flood-related damage as this is one of the costliest natural disasters (Ghosh and et al., 2022; Wanghe et al., 2022).

Tropical storms

Cyclones, typhoons and hurricanes are also among the most intense and costly natural disasters (Sun et al., 2022). To prevent or reduce loss and damage caused by these storms, research findings show that machine learning models have been used to predict hurricane paths and assess damage using reanalysis data, while deep learning has successfully detected many severe storms and has also been used with other data sources like passive microwave satellite data for monitoring tropical cyclones (Sun et al., 2022). McCarthy et al. (2020) suggest that machine learning methods have demonstrated high-accuracy mapping capabilities on small spatial scales but require a large amount of robust training data. Efforts to map larger areas at finer resolutions may require hundreds of thousands of images and therefore encounter Big Data processing challenges.² Ruckelshaus et al. (2020) note that combining satellite- derived imagery with machine learning classification algorithms improves the opportunity to understand, map and respond to natural, climatic and human- caused impacts.

Since responsiveness and resilience to coastal disasters are critical for all development stakeholders, climate diplomats should consider the uptake of AI and Big Data technology by African countries, for example Google HydroNets, a specialised deep neural network architecture built specifically for water levels forecasting, and Google MetNet, a Neural Weather Model for precipitation forecasting. Hydrometeorological risk forecasting and early warning is viewed as key in typhoon-related disaster risk reduction and preparedness (Yu et al., 2021). If people know “what the weather will do” as opposed to just knowing “what it will be” they will be better prepared to evade severe loss and damage and sustainable climate resilient development will be possible (Yu et al., 2022).

Wildfires

Among natural disasters with great loss and damage implications are wildfires caused by lightning, volcanic eruptions and spontaneous ignition, some of which are the results of global warming and all of which result in forest ecosystems destruction (Mohajane et al., 2021; Park et al., 2022; Waseem & Manshadi, 2020). Early fire detection and intervention are vital for damage minimisation. To this end, scholars such as Sun et al. (2022) note that machine learning methods have been used to:

• improve fire detection and prediction,
• classify and map wildfire severity,
• detect wildfires automatically using unmanned aerial vehicles (UAVs) or satellite images,
• improve smoke plume forecasting,
• assess human health issues connected to poor air quality, and
• trace human-caused wildfires.

Among other methods that can tackle various aspects of wildfires, Waseem and Manshadi (2020) highlight the use of the artificial neural network (ANN), which is an AI method that teaches computers to process data in a way that is inspired by the human brain. However, the scholars note application of ANN in large areas takes a lot of time, hence they point to the multi-threshold algorithm as a better option in such a scenario (Zhang et al., 2021). Since forests are a primary source of livelihood, to prevent forest fire disasters Mohajane et al. (2021) also point to the application of remote sensing and machine learning algorithms for forest fire mapping.

Earthquakes

The seismicity of Africa, according to Algerian Professor Djillali Benouar (n.d.), is mainly concentrated in South- East Africa and North Africa.

Given their frequency and devastating effects, extensive applications of AI techniques are considered crucial for earthquake forecasting and probability, and hazard, risk mapping and mitigation purposes. Freddi et al. (2021) also suggest that machine learning techniques are advantageous as they may be able to find relationships between event parameters and corresponding event losses that would otherwise elude ad-hoc and “traditional” statistical approaches. For example, the World Bank’s Geospatial Operations Support Team (GOST) in Guatemala City fed drone and street-level imagery to machine learning algorithms to automatically detect “soft-story” buildings or those most likely to collapse in an earthquake.

On the other hand, deep learning methodologies have accelerated the development of more reliable and efficient algorithms for earthquake monitoring (Sun et al. 2022). AI-based earthquake monitoring methods can result in advancing seismic hazard safety by empowering Earthquake Early Warning (EEW) systems with faster and more reliable estimations of earthquake parameters, and providing more complete and precise earthquake catalogues used for improving long- term seismic hazard assessments.

CONCLUSION

The research reviewed in this article determined the potential of AI in mitigating or reducing climate change- induced loss and damage and found that AI is already positively influencing the management of natural disasters. This worldwide use of AI points to the degree of ease associated with the use of the systems by varied sustainable development stakeholders. The article urges African climate diplomats to consider advocating for and adopting the best AI technology for use in their countries (see Venkatesh et al., 2003). Research revealed that machine learning and deep learning approaches can be used to produce reliable flood susceptibility maps. Deep learning approaches can also successfully detect severe storms as they develop and improve earthquake risk assessment. Machine learning approaches can be used to detect and trace human-caused wildfires.

These study results are timely and relevant given that the Sendai Framework for Disaster Risk Reduction (SFDRR) 2015-2030 highlights the importance of scientific research, supporting the “availability and application of science and technology to decision making” in disaster risk reduction (Freddi et al., 2021).

This review of the literature shows that science and technology can play a crucial role in the world’s ability to reduce casualties, physical damage and interruption to critical infrastructure caused by natural hazards and their complex interactions, thereby ensuring the realisation of sustainable development goals. According to Freddi et al. (2021), the Sendai Framework encourages better access to technological innovations combined with increased disaster risk reduction investments in developing cost- effective approaches and tackling global challenges induced by climate change.

For this to be achieved on the African continent, African climate diplomats should support the progressive development of AI with relevant, adequate and correct data sets to evade flawed and damaging results (see Galaz et al., 2021). AI uptake by various African governments, the private sector and individuals, can ensure loss and damage mitigation and reduction across all industries on the continent whose economies are already facing challenges from the climate crisis.

REFERENCES

AUDA-NEPAD. 2022. The African Union Artificial Intelligence Continental Strategy For Africa AUDA- NEPAD. Available at https://www.nepad.org/ news/african-union-artificial-intelligence- continental-strategy-africa

Auliagisni, W., Wilkinson, S. & Elkharboutly, M. 2022. Using community-based flood maps to explain flood hazards in Northland, New Zealand, Progress in Disaster Science, 14. Available at https://doi. org/10.1016/j.pdisas.2022.100229

Benouar, D. n.d. Earthquake Risk in Africa: A school’s guide. https://www.humanitarianlibrary.org/ resource/earthquake-risk-africaa-schools-guide-0

CDP. 2020. CDP Africa Report Benchmarking Progress Towards Climate Safe Cities, States, And Regions. https://www.cdp.net/en/research/global-reports/ africa-report

Darabi, H., Torabi Haghighi, A., Rahmati, O., Jalali Shahrood, A., Rouzbeh, S., Pradhan, B. & Tien Bui, D. 2021. A hybridized model based on neural network and swarm intelligence – Grey wolf algorithm for spatial prediction of urban flood- inundation, Journal of Hydrology, 603. Available at https://doi.org/10.1016/j.jhydrol.2021.126854

Eckstein, D., Künzel, V. & Schäfer, L. 2021. Global climate risk index 2021, Germanwatch e.V., 28. Available at https://germanwatch.org/sites/ default/files/Global Climate Risk Index 2021_2.pdf

Eckstein, D., Künzel, V., Schäfer, L. & Winges, M. 2019. Global Climate Risk Index 2020: Who Suffers Most from Extreme Weather Events? Germanwatch. Available at www.germanwatch.org

Fang, Z., Wang, Y., Peng, L. & Hong, H. 2021. Predicting flood susceptibility using LSTM neural networks, Journal of Hydrology, 594. Available at https://doi. org/10.1016/j.jhydrol.2020.125734

Figueiredo, M. M. 2019. Artificial Intelligence Acceptance: Morphological elements of the acceptance of Artificial Intelligence. https://repositorio.ucp.pt/

Freddi, F., Galasso, C., Cremen, G., Dall’Asta, A., di Sarno, L., Giaralis, A., Gutiérrez-Urzúa, F., Málaga-Chuquitaype, C., Mitoulis, S. A., Petrone,, C., Sextos, A., Sousa, L., Tarbali, K., Tubaldi, E., Wardman, J. & Woo, G. 2021. Innovations in earthquake risk reduction for resilience: Recent advances and challenges, International Journal of Disaster Risk Reduction, 60. Available at https://doi. org/10.1016/j.ijdrr.2021.102267

Galaz, V., Centeno, M. A., Callahan, P. W., Causevic, A., Patterson, T., Brass, I., Baum, S., Farber, D., Fischer, J., Garcia, D., McPhearson, T., Jimenez, D., King, B., Larcey, P. & Levy, K. 2021. Artificial intelligence, systemic risks, and sustainability, Technology in Society, 67. Available at https://doi. org/10.1016/j.techsoc.2021.101741

Ghosh, S., Kumar, D. & Kumari, R. 2022. Cloud- based large-scale data retrieval, mapping, and analysis for land monitoring applications with Google Earth Engine (GEE), Environmental Challenges, 9. Available at https://doi.org/10.1016/j.envc.2022.100605

Ghosh, S., Saha, S. & Bera, B. 2022. Flood susceptibility zonation using advanced ensemble machine learning models within Himalayan foreland basin, Natural Hazards Research. Available at https://doi.org/10.1016/j.nhres.2022.06.003

Harrington, L. J. & Otto, F. E. L. 2020. Reconciling theory with the reality of African heatwaves, Nature Climate Change, 10, 9:796–798). Available at https://doi.org/10.1038/s41558-020-0851-8

Javaid, M., Haleem, A., Pratap Singh, R., Khan, S. & Suman, R. 2022. Sustainability 4.0 and its applications in the field of manufacturing, Internet of Things and Cyber-Physical Systems, 2, 82–90. Available at https://doi.org/10.1016/j. iotcps.2022.06.001

Javaid, M., Haleem, A., Singh, R. P. & Suman, R. 2022. Enhancing smart farming through the applications of Agriculture 4.0 technologies, International Journal of Intelligent Networks, 3:150–164. Available at https://doi.org/10.1016/j. ijin.2022.09.004

Jena, R., Pradhan, B., Naik, S. P. & Alamri, A. M. 2021. Earthquake risk assessment in NE India using deep learning and geospatial analysis, Geoscience Frontiers, 12 (3). Available at https://doi. org/10.1016/j.gsf.2020.11.007

Lei, X., Chen, W., Panahi, M., Falah, F., Rahmati, O., Uuemaa, E., Kalantari, Z., Ferreira, C. S. S., Rezaie, F., Tiefenbacher, J. P., Lee, S. & Bian, H. 2021. Urban flood modeling using deep-learning approaches in Seoul, South Korea, Journal of Hydrology, 601. Available at https://doi.org/10.1016/j. jhydrol.2021.126684

Lin, J. M. & Billa, L. 2021. Spatial Prediction of Flood- Prone Areas Using Geographically Weighted Regression, Environmental Advances, 6. Available at https://doi.org/10.1016/j.envadv.2021.100118

Lisinge-Fotabong, E. & Diakhité, M. 2016. Climate Diplomacy in Africa Historical Presence of Africa in the climate change debate prior to the UNFCCC process. https://climate-diplomacy.org/sites/default/ files/2020-10/2016-NEPAD-Climate-Diplomacy-in- Africa-Policy-Brief_updated_v3.pdf

Masson-Delmotte, V., Zhai, P., Pörtner, H-O., Roberts, D., Skea, J., Shukla, P. R., Pirani, A., Moufouma-Okia, W., Péan, C., Pidcock, R., Connors, S., Matthews, J. B. R., Chen, Y., Zhou, X., Gomis, M. I., Lonnoy, E., Maycock, T., Tignor, M. & Waterfield, T. 2019. Global warming of 1.5°C An IPCC Special Report on the impacts of global warming of 1.5°C above pre-industrial levels and related global greenhouse gas emission pathways, in the context of strengthening the global response to the threat of climate change, sustainable development, and efforts to eradicate poverty. Available at www.environmentalgraphiti.org

McCarthy, M. J., Jessen, B., Barry, M. J., Figueroa, M., McIntosh, J., Murray, T., Schmid, J. & Muller- Karger, F. E. 2020. Mapping hurricane damage: A comparative analysis of satellite monitoring methods, International Journal of Applied Earth Observation and Geoinformation, 91. Available at https://doi.org/10.1016/j.jag.2020.102134

Mohajane, M., Costache, R., Karimi, F., Bao Pham, Q., Essahlaoui, A., Nguyen, H., Laneve, G. & Oudija, F. 2021. Application of remote sensing and machine learning algorithms for forest fire mapping in a Mediterranean area, Ecological Indicators, 129. Available at https://doi.org/10.1016/j.ecolind.2021.107869

Molina, J. L., Espejo, F., Zazo, S., Molina, M. C., Hamitouche, M. & García-Aróstegui, J. L. 2022. HydroPredicT_Extreme: A probabilistic method for the prediction of extremal high- flow hydrological events, Journal of Hydrology, 610. Available at https://doi.org/10.1016/j. jhydrol.2022.127929

Nordgren, A. 2022. Artificial intelligence and climate change: ethical issues, Journal of Information, Communication and Ethics in Society. Available at https://doi.org/10.1108/JICES-11-2021- 0106

Park, M., Tran, D. Q., Bak, J. & Park, S. 2022. Advanced wildfire detection using generative adversarial network-based augmented datasets and weakly supervised object localization, International Journal of Applied Earth Observation and Geoinformation,114. Available at https://doi.org/10.1016/j. jag.2022.103052

Rolnick, D., Donti, P. L., Kaack, L. H., Kochanski, K., Lacoste, A., Sankaran, K., Ross, A. S., Milojevic- Dupont, N., Jaques, N., Waldman-Brown, A., Luccioni, A. S., Maharaj, T., Sherwin, E. D., Mukkavilli, S. K., Kording, K. P., Gomes, C. P., Ng, A. Y., Hassabis, D., Platt, J. C. & Bengio, Y. 2022. Tackling Climate Change with Machine Learning, ACM Computing Surveys, 55, 2, Association for Computing Machinery. Available at https://doi. org/10.1145/3485128

Ruckelshaus, M., Reguero, B. G., Arkema, K., Compeán, R. G., Weekes, K., Bailey, A. & Silver, J. 2020. Harnessing new data technologies for nature-based solutions in assessing and managing risk in coastal zones, International Journal of Disaster Risk Reduction, 51. Available at https://doi.org/10.1016/j.ijdrr.2020.101795

Shahabi, H., Shirzadi, A., Ronoud, S., Asadi, S., Pham, B. T., Mansouripour, F., Geertsema, M., Clague, J. J. & Bui, D. T. 2021. Flash flood susceptibility mapping using a novel deep learning model based on deep belief network, back propagation and genetic algorithm, Geoscience Frontiers, 12(3). Available at https://doi.org/10.1016/j.gsf.2020.10.007

Sun, Z., Sandoval, L., Crystal-Ornelas, R., Mousavi, S. M., Wang, J., Lin, C., Cristea, N., Tong, D., Carande, W. H., Ma, X., Rao, Y., Bednar, J. A., Tan, A., Wang, J., Purushotham, S., Gill, T. E., Chastang, J., Howard, D., Holt, B. & John, A. 2022. A review of Earth Artificial Intelligence, Computers and Geosciences, 159. Available at https://doi. org/10.1016/j.cageo.2022.105034

Taddeo, M., Tsamados, A., Cowls, J. & Floridi, L. 2021. Artificial intelligence and the climate emergency: Opportunities, challenges, and recommendations, One Earth, 4(6), 776–779. Available at https://doi. org/10.1016/j.oneear.2021.05.018

Tanaka, K., Zhu, M., Miyaji, K., Kurokawa, T. & Akamatsu, T. 2022. Spatial distribution maps of real-time ocean observation platforms and sensors in Japanese waters, Marine Policy, 141. Available at https://doi.org/10.1016/j.marpol.2022.105102

Venkatesh, V., Morris, M. G., Davis, G. B. & Davis, F. D. 2003. User acceptance of information technology: Toward a unified view, MIS Quarterly: Management Information Systems, 27(3), 425–478. Available at https://doi.org/10.2307/30036540

Wanghe, K., Guo, X., Ahmad, S., Tian, F., Nabi, G., Igorevich Strelnikov, I., Li, K. & Zhao, K. 2022. FRESF model: An ArcGIS toolbox for rapid assessment of the supply, demand, and flow of flood regulation ecosystem services, Ecological Indicators, 143. Available at https://doi. org/10.1016/j.ecolind.2022.109264

Waseem, M. & Manshadi, S. D. 2020. Electricity grid resilience amid various natural disasters: Challenges and solutions, Electricity Journal, 33(10). Available at https://doi.org/10.1016/j. tej.2020.106864

Yigitcanlar, T., Desouza, K. C., Butler, L. & Roozkhosh, F. 2020. Contributions and risks of artificial intelligence (AI) in building smarter cities: Insights from a systematic review of the literature, Energies, 13(6). MDPI AG. Available at https://doi.org/10.3390/en13061473

Yu, J., Liu, J., Baek, J-W., Fong, C. & Fu, M. 2022. Impact-based forecasting for improving the capacity of typhoon-related disaster risk reduction in typhoon committee region, Tropical Cyclone Research and Review, 11(3):163–173. Available at https://doi.org/10.1016/j.tcrr.2022.09.003

Yu, J., Liu, J. & Choi, Y. 2021. Review and prospects of strategies and measures for typhoon-related disaster risk reduction under public emergencies in TC region, Tropical Cyclone Research and Review, 10(2): 116–123. Available at https://doi.org/10.1016/j.tcrr.2021.05.002

Zhang, G., Wang, M. & Liu, K. 2021. Deep neural networks for global wildfire susceptibility modelling, Ecological Indicators, 127. Available at https://doi.org/10.1016/j.ecolind.2021.107735

Zhong, R., Zhu, Y., Wang, X., Li, H., Wang, B., You, F., Rodríguez, L. F., Huang, J., Ting, K. C., Ying, Y. & Lin, T. 2022. Detect and attribute the extreme maize yield losses based on spatio-temporal deep learning. Fundamental Research. Available at https://doi.org/10.1016/j.fmre.2022.05.006

ENDNOTES

1. In layman’s terms, machine learning and deep learning are both types of AI. Machine learning is AI that can automatically adapt with minimal human interference. Deep learning is a subset of machine learning that uses artificial neural networks to mimic the learning process of the human brain. Where machine learning algorithms generally need human correction when they get something wrong, deep learning algorithms can improve their outcomes through repetition, without human intervention. A machine learning algorithm can learn from relatively small sets of data, but a deep learning algorithm requires Big Data sets that might include diverse and unstructured data (see https://levity.ai/blog/difference-machine- learning-deep-learning).
2. Big Data refers to massive, complex data sets that are rapidly generated and transmitted from a wide variety of sources and that are too large or complex to be dealt with by traditional data- processing application software.

Citation: New Agenda: South African Journal of Social and Economic Policy No 87, First Quarter 2023, March: p33. https://ifaaza.org/new-agenda/new-agenda-issue-88/

Sharon Tshipa is a development practitioner, researcher, award-winning multimedia freelance journalist and fiction writer based in Gaborone, Botswana. She holds a Master’s degree in Development Practice and has a BA in Media Studies. She is the co-founder and Chairperson of the Botswana Society for Human Development.