Enenkel M, Bollettino V, Vinck P. Preparing for consecutive climate events. HPHR. 2021;31.
The next generation of disaster financial instruments seeks to provide social protection mechanisms for the world’s most vulnerable populations and help avoid public health emergencies. Anticipatory action is designed to mobilize resources in risk-prone areas when there is a forecast for extreme weather events. In the case of drought risk, for instance, weather-index insurance can trigger payouts based on satellite-based monitoring of environmental variables and pre-agreed thresholds to speed up the availability of funds to prepare for and mitigate disasters. Such insurance models are usually designed to cover the ‘statistical tails’ of weather and climate – the most extreme events. However, there are cases where consecutive moderate to severe drought events result in socioeconomic impacts that are comparable or more severe than the impact of an isolated extreme event. Covering these consecutive events reveals unique challenges and data gaps, particularly because there is currently no socio-economic baseline that matches the granularity of available climate hazard datasets. We argue that longitudinal data collection exercises that are complementary to state-of-the-art vulnerability or food security surveys would enable researchers and index designers to explore new opportunities. A better understanding of pre-existing vulnerabilities would help to better translate non-extreme hazard information into impact estimates and predict critical socioeconomic developments more reliably. It would also support the evaluation and improvement of anticipatory financing strategies via feedback loops at household level.
While the recognition of the risks posed by extreme weather events is not new, climate change has raised concerns about the ability to forecast and prepare for an increase in the severity, frequency, and duration of such extreme climate anomalies. Extreme weather events have direct and at times complex, indirect impacts on societies and economies (Fankhauser and S.J. Tol, 2005; Beier, Brzoska and Khan, 2015; Nkhonjera and Dinka, 2017). From a public health perspective, anomalies in the earth’s atmosphere can, for example, lead to droughts, potentially resulting in crop failure or water shortages, leading to food insecurity and potential adverse consequences on nutritional status of a population. Extreme weather or climate events can also be a risk multiplier for conflicts or other forms of violence (Kelley et al., 2015) and create compound risks for communities already affected by other disasters, including infectious diseases epidemics like SARS-Cov-2. Climate change has significantly increased the economic, social and public health costs of extreme weather-related events by affecting the severity, frequency, and duration of climate anomalies, such as droughts (Spinoni et al., 2014). These changes threaten to double the number of people in need of humanitarian assistance by 2050, at a cost of around 20 billion US dollars per year (IFRC, 2019).
Efforts to protect populations from climate change-related risks have been insufficient to meet the burgeoning cost of disaster. To date, climate-related disaster risks have relied on two important measures:
Disaster risk financing or anticipatory financing is an early response mechanism, anticipating possible impacts in risk-prone areas and mobilizing (as well as distributing) resources when critical threshold values are detected early via monitoring systems (index insurance) or when there is a forecast for extreme weather events (forecast / impact-based financing). We argue that knowledge gaps related to climate hazards and socioeconomic vulnerabilities that are partly caused by antecedent shocks may undermine the ability of disaster risk financing to achieve its potential as a social safety net to avert public health crises. Specifically, the current focus on extreme events ignores the risks associated with smaller scale (non-extreme) consecutive climate events. The repetition of such events can result in disaster impacts that are comparable to or even collectively worse than isolated extreme events.
Extreme weather events are rare events that often result in the loss of lives, property, or equipment. The annual cost of such events ranges from a few billion U.S. dollars to more than $200 billion (Guan and Zhu, 2017) and the relative financial impact of these events has been increasing over the past several decades. To stem these losses, experts have sought to improve their ability to monitor and forecast severe weather-related events. A common feature of weather and climate monitoring systems is their heavy reliance on information from satellites to detect extreme conditions as early as possible as a trigger to facilitate early action (Revilla-Romero et al., 2015; Enenkel et al., 2019). Lead times of forecasting systems may range from several hours to days in the case of typhoons and hurricanes, all the way up to 12 months and beyond in the case of drought. Methods to enable and improve such forecasts include ensemble forecasting (Pepler et al., 2015), the ‘combination’ of up to 52 models, and through models using advanced machine learning (Jones, 2017).
Detecting extreme moisture deficits to trigger insurance payouts or triggering cash-transfers based on forecasts can prevent emergencies from turning into a severe crisis. However, it can also create a significant blind spot for non-extreme, sequential climate events that may be as damaging or even more damaging than extreme events (Darge Delbiso et al., 2017). For example, repeated, non-extreme weather shocks have been identified as a cause of acute food insecurity that can result in chronic malnutrition, leading to physical and economic consequences for communities at risk (Brown et al., 2020). Consecutive moderate to severe events can result in socioeconomic impacts that are comparable or worse than one isolated extreme event – not only because they do not make it into the headlines of global newspapers (Moke and Rüther, 2015), which can have an effect on available humanitarian funding, but also because most response instruments, including disaster financing, focus on ‘statistical tail’ or major, rare events only.
This phenomenon is visible in regions in sub-Saharan Africa that are largely rainfed and dominated by subsistence farming. Chronic weather events in this context increases the susceptibility to weather shocks, affecting an estimated 33 million farms, whose owners often live below the official poverty line, (Gassner et al., 2019). More generally, the climate impact pathways related to food insecurity can be extremely complex so that a focus on extreme events is insufficient to avert major public health crises. Current approaches focus on sub-seasonal and seasonal forecasts (Vitart et al., 2019), from detecting the onset of the season, often via moisture-related indicators (Enenkel et al., 2018), to the harvest. Yet, the socioeconomic impact of one agricultural season can have severe consequences in the next, as food from the previous year’s harvest runs out, but the next harvest is not yet ready.
In addition to the tendency to focus on extreme weather events, a second flaw of current forecasting approaches is that they generally focus exclusively on environmental factors, ignoring the important social dimensions of risk (representing vulnerabilities). While satellite-derived information to characterize weather hazards is already available at field-scale with several updates per week (depending on the satellite’s sensor, orbit and constellation), socioeconomic information needed to characterize vulnerabilities and understand risks is not (Enenkel et al., 2020). Socioeconomic data is often collected too infrequently or is too aggregated, making it difficult to take targeted action. In short, while we have routine updates on data about threats from weather and climate, we know little about what this means for the impact weather-related events will have on the families and communities in their path.
There are efforts to harmonize data collection at a global scale, partly via machine learning (Carletto, Zezza and Banerjee, 2013; Swamy et al., 2019). However, socioeconomic data are still collected without interagency data collection and data governance standards, at low temporal frequency (several years between surveys), insufficient granularity (national scale) or during emergencies. This lack of sub-national longitudinal studies results in the lack of a socioeconomic baseline, with far-reaching consequences for early warning and early action as well as financial instruments. It is possible to estimate how many people in a certain satellite pixel and at a certain point in time are affected by a drought event that is for instance classified as ‘severe’. However, it is virtually impossible to robustly relate the current climate hazard to socioeconomic vulnerabilities that might or might not have their roots in climate and would in any case exacerbate the impact of a drought or other climate hazards at household level. This limitation is likely already having an impact on the performance of all kinds of disaster risk reduction instruments, but it would have an even bigger impact once inter-seasonal conditions are considered.
While arguably simple, the idea that financial protection strategies can be leveraged to reduce disaster risk is relatively recent in humanitarian action. In a speech in Ireland in 2018 (Casement Lecture: A Collective Call towards Innovation in Humanitarian Financing), then United Nations Under-Secretary-General for Humanitarian Affairs and Emergency Relief Coordinator, Mark Lowcock discussed how to make international humanitarian response “better, faster, and cheaper.” He emphasized the need for humanitarian agencies to shift from a posture of responding to disasters to one of anticipating and preparing for disasters ahead of time. An important focus of this effort is in anticipatory financing, the provision of resources before an event occurs based on forecasts.
A six-part action plan was articulated to enhance the ability of the international humanitarian system to anticipate and prepare for disasters based on anticipatory financing. The first is to utilize disaster risk insurance, an insurance scheme to cover costs incurred from extreme weather and natural disasters. While often available in the wealthiest countries, disaster risk insurance is virtually unheard of in poorer countries, making recovery efforts from disasters difficult. This is particularly true in areas that are impacted by smaller but frequent crises that fall off the radar of international support.
The second action rests on pre-agreed contingency and forecast-based financing windows by donors. This allows funding to arrive when needed quickly, rather than applying for funds after a disaster and waiting months for the support to arrive. These mechanisms can be further strengthened in time through analysis of quality empirical data and analysis. Third, is greater connectivity with market-based solutions like social impact bonds or other similar initiatives. This allows some risk to be transferred to international financial markets. Fourth is a move to purposefully support resilience-building measures through ensuring that humanitarian response delivers long term benefits and contribute to enhance resilience to disasters. Fifth, is a focus on making the existing humanitarian finance system more efficient and effective. Aid could be made to be more flexible and effective through the use of local organizations that are better poised to respond quickly and who know the communities they work in far better than international organizations (Inter-Agency Standing Committee, 2020). Finally, is the need for a more organized, streamlined effort to share empirical findings of what finance mechanisms work best in which circumstances.
While the evidence for the effectiveness of anticipatory financing is mounting, its application is still often limited by soft triggers, such as the declaration of an emergency based on expert assessments or hard triggers, such as pre-agreed thresholds for parametric insurance (Clarke and Mahul, 2016). These triggers, however, are largely conditioned by exceptional conditions such as extreme weather events and are not currently taking into account the compounded effects of back-to-back moderate events.
The limited ability to monitor consecutive, non-extreme, climate events and to accurately capture the socio-economic terrain, including vulnerabilities, is creating severe limitations for what is rapidly emerging as a critical tool for disaster risk reduction. Financial instruments, like parametric insurance, however, are generally established to pay out if the hazard parameters hit a certain high threshold or pre-agreed triggers corresponding to extreme events (Collier, Skees and Barnett, 2009). In this they are missing the need to also support those affected by chronic events that result in far-reaching socioeconomic impacts at household level due to pre-existing high levels of vulnerability. There is value in layering approaches that aim to complement extreme events risk transfer mechanisms via financial instruments that can cover more frequent, less severe events. However, independent from the financial instrument, it is vital to understand the links between climate shocks and vulnerabilities, whose origin might lie years in the past, as well as to develop the technical capacities to apply and adapt risk layering in an operational setting.
Understanding vulnerabilities and the continuous monitoring of dangerous chronic conditions is therefore critical. Alternatives to costly and time-consuming loss assessments could lower premiums and speed up the payout process. It is in fact unrealistic to send assessment teams to vulnerable countries or regions every one to three months to update the socioeconomic baseline. Yet, there are promising tools for self-reporting at household level, which 1) work in the local language 2) do not require an internet connection, because data can be cached on the device 3) can be coupled with an incentive to report truthfully. SMS-based surveys or interactive voice response (IVR) calls can be used alternatively or in parallel as long as the same questions and answers are used. Ideally, such longitudinal data would also be complementary to existing surveys, for instance by including simple measures of child malnutrition via mid upper arm circumference (MUAC) measurement band, or by including novel questions related to climate risk perception. Most importantly, data collection follows a ‘no regret approach’, meaning that there is an added value in understanding the characteristics of the human terrain of subsistence farmers independently from the existence or impact of an extreme event.
Assuming such socioeconomic baseline dataset exists, we can take advanced methods, even predictive analytics, to a new level. Pattern recognition algorithms could help to identify which climate hazard levels are critically conditioned on the current vulnerability level of the community at risk. Predictive analytics could estimate a likelihood that climate impacts on agricultural production would increase the hunger gap, even early on in the season. A more localized understanding or climate risk perceptions could contribute to bridging the gap between risk financing and social protection mechanisms, resulting in advantages for risk taking (e.g. agricultural investments) or the uptake of community-based weather index insurance products
The next generation of drought risk financing tools will focus on the co-design of indices and a hybrid trigger approach that complements the assessments of local experts with a quantitative perspective (Bavandi, Aubrecht and Enenkel, 2021). The need to continuously monitor the livelihood conditions of a representative sample among vulnerable communities will likely increase. The effort to understand changes in livelihoods conditions, either linked to or independent from disaster impacts, should not be neglected as it can result in multiple advantages. First, the likelihood that a crisis is ‘overlooked’ is reduced, because alerts can be linked to critical socioeconomic thresholds, including primary and secondary effects of the climate hazard conditions in parallel. Second, not focusing the monitoring only on isolated agricultural seasons increases the chances of detecting upcoming hunger gaps earlier. Third, working with local institutions to coordinate the data collection at household level instead if international assessment teams is in line with the requests of the humanitarian community (Inter-Agency Standing Committee, 2020), helping communities at risk to strengthen their risk ownership. There is hope that climate risk financing instruments can support climate change adaptation successfully and cost effectively. Hence, starting to collect, harmonize, analyze and visualize socioeconomic data with the help of the people who know their own livelihoods best at higher frequencies is an endeavor with many potential benefits. Since drought impacts and their root causes cannot be regarded as isolated environmental or socioeconomic issues, data collection serves multiple analytical purposes, resulting in the potential cross-fertilization with other disciplines and extreme events (e.g., floods or wildfires).
ltimately, the development of a socioeconomic baseline, might lead to certain risks for financial instruments, such as strategic reporting on household level. Sophisticated approaches, such as incentivized reporting (Sheriff and Osgood, 2010) will need to be tested to ensure both data quantity and quality. In the long run, however, the designers of financial instruments could decrease the uncertainty in their simulations and even evaluate if payouts had the desired effect with regard to providing a social safety net, such as increasing agricultural investments or a decrease in child wasting and stunting.
Bavandi, A., Aubrecht, C. and Enenkel, M. (2021) ‘Faster and Better Risk Indicators: Introducing the Next Generation Drought Index (NGDI) Project, World Bank Financial Protection Forum, accessible at: https://www.financialprotectionforum.org/blog/faster-and-better-risk-indicators-introducing-the-next-generation-drought-index-ngdi-project’.
Beier, D., Brzoska, P. and Khan, M. H. (2015) ‘Indirect consequences of extreme weather and climate events and their associations with physical health in coastal Bangladesh: a cross-sectional study’, Global Health Action, 8. doi: 10.3402/gha.v8.29016.
Brown, M. E., Backer, D., Billing, T., White, P., Grace, K., Doocy, S., & Huth, P. (2020). Empirical studies of factors associated with child malnutrition: highlighting the evidence about climate and conflict shocks. Food Security, 1-12.
Carletto, C., Zezza, A. and Banerjee, R. (2013) ‘Towards better measurement of household food security: Harmonizing indicators and the role of household surveys’, Global Food Security, 2(1), pp. 30–40. doi: 10.1016/j.gfs.2012.11.006.
Clarke, D. and Mahul (2016) ‘Clarke, Daniel and Mahul, Olivier, Disaster Risk Financing and Contingent Credit: A Dynamic Analysis (June 1, 2011). World Bank Policy Research Working Paper No. 5693, Available at SSRN: https://ssrn.com/abstract=1871589’.
Collier, B., Skees, J. and Barnett, B. (2009) ‘Weather Index Insurance and Climate Change: Opportunities and Challenges in Lower Income Countries’, The Geneva Papers on Risk and Insurance – Issues and Practice, 34(3), pp. 401–424. doi: 10.1057/gpp.2009.11.
Delbiso, T. D., Rodriguez-Llanes, J. M., Donneau, A. F., Speybroeck, N., & Guha-Sapir, D. (2017). Drought, conflict and children’s undernutrition in Ethiopia 2000–2013: a meta-analysis. Bulletin of the World Health Organization, 95(2), 94..
Enenkel, M., Farah, C., Hain, C., White, A., Anderson, M., You, L., … & Osgood, D. (2018). What rainfall does not tell us—enhancing financial instruments with satellite-derived soil moisture and evaporative stress. Remote Sensing, 10(11), 1819. doi: 10.3390/rs10111819.
Enenkel, M., Osgood, D., Anderson, M., Powell, B., McCarty, J., Neigh, C., … & Brown, M. (2019). Exploiting the convergence of evidence in satellite data for advanced weather index insurance design. Weather, Climate, and Society, 11(1), 65-93. doi: 10.1175/WCAS-D-17-0111.1.
Enenkel, M., Brown, M. E., Vogt, J. V., McCarty, J. L., Bell, A. R., Guha-Sapir, D., … & Vinck, P. (2020). Why predict climate hazards if we need to understand impacts? Putting humans back into the drought equation. Climatic Change, 162(3), 1161-1176. doi: 10.1007/s10584-020-02878-0.
Fankhauser, S., & Tol, R. S. (2005). On climate change and economic growth. Resource and Energy Economics, 27(1), 1-17. doi: 10.1016/j.reseneeco.2004.03.003.
Gassner, A., Harris, D., Mausch, K., Terheggen, A., Lopes, C., Finlayson, R. F., & Dobie, P. (2019). Poverty eradication and food security through agriculture in Africa: Rethinking objectives and entry points. Outlook on Agriculture, 48(4), 309-315. doi: 10.1177/0030727019888513.
Guan, H. and Zhu, Y. (2017) Development of Verification Methodology for Extreme Weather Forecasts in: Weather and Forecasting Volume 32 Issue 2 (2017). Available at: https://journals.ametsoc.org/view/journals/wefo/32/2/waf-d-16-0123_1.xml (Accessed: 5 March 2021).
IFRC (2019) ‘The costs of doing nothing – the humanitarian price of climate change and how it can be avoided, available at: https://media.ifrc.org/ifrc/wp-content/uploads/sites/5/2019/09/2019-IFRC-CODN-EN.pdf’.
Inter-Agency Standing Committee (2020) ‘Grand Bargain Annual Independent Report 2020, accessible at: https://interagencystandingcommittee.org/grand-bargain-official-website/grand-bargain-annual-independent-report-2020’.
Jones, N. (2017) ‘How machine learning could help to improve climate forecasts’, Nature News, 548(7668), p. 379. doi: 10.1038/548379a.
Kelley, C. P., Mohtadi, S., Cane, M. A., Seager, R., & Kushnir, Y. (2015). Climate change in the Fertile Crescent and implications of the recent Syrian drought. Proceedings of the National Academy of Sciences, 112(11), 3241-3246. doi: 10.1073/pnas.1421533112.
Moke, M. and Rüther, M. (2015) ‘Media and Humanitarian Action’, in Gibbons, P. and Heintze, H.-J. (eds) The Humanitarian Challenge. Cham: Springer International Publishing, pp. 253–263. doi: 10.1007/978-3-319-13470-3_13.
Nkhonjera, G. K. and Dinka, M. O. (2017) ‘Significance of direct and indirect impacts of climate change on groundwater resources in the Olifants River basin: A review’, Global and Planetary Change, 158, pp. 72–82. doi: 10.1016/j.gloplacha.2017.09.011.
Pepler, A. S., Díaz, L. B., Prodhomme, C., Doblas-Reyes, F. J., & Kumar, A. (2015). The ability of a multi-model seasonal forecasting ensemble to forecast the frequency of warm, cold and wet extremes. Weather and Climate Extremes, 9, 68-77. doi: 10.1016/j.wace.2015.06.005.
Revilla-Romero, B., Hirpa, F. A., Pozo, J. T. D., Salamon, P., Brakenridge, R., Pappenberger, F., & De Groeve, T. (2015). On the use of global flood forecasts and satellite-derived inundation maps for flood monitoring in data-sparse regions. Remote Sensing, 7(11), 15702-15728. doi: 10.3390/rs71115702.
Sheriff, G., & Osgood, D. (2010). Disease forecasts and livestock health disclosure: A shepherd’s dilemma. American Journal of Agricultural Economics, 92(3), 776-788. doi: 10.1093/ajae/aap042.
Spinoni, J., Naumann, G., Carrao, H., Barbosa, P., & Vogt, J. (2014). World drought frequency, duration, and severity for 1951–2010. International Journal of Climatology, 34(8), 2792-2804.doi: https://doi.org/10.1002/joc.3875.
Swamy, V., Chen, E.,Vankayalapati, A., Aggarwal, A., Liu, C., Mandava, V., Johnson, S. (2019) ‘Machine Learning for Humanitarian Data: Tag Prediction using the HXL Standard, conference paper, KDD ’19, August 04–08, 2019, Anchorage, AK; available at: https://www.kdd.org/kdd2019/docs/Humanitarian_Data_tagging_KDD2019_SocialImpactTrack_HXLTagPrediction.pdf (last accessed May 10th, 2020)’, p. 3.
Vitart, F., Cunningham, C., DeFlorio, M., Dutra, E., Ferranti, L., Golding, B., … & Tippett, M. K. (2019). Sub-seasonal to seasonal prediction of weather extremes. In Sub-seasonal to seasonal prediction (pp. 365-386). Elsevier. doi: 10.1016/B978-0-12-811714-9.00017-6.
Markus Enenkel, PhD has a background in remote sensing, humanitarian decision-support and climate risk financing. He has worked with different humanitarian aid organizations, such as the IFRC, Doctors without Borders or the UN World Food Programme. Dr. Enekel’s research interests concentrate on the development of novel approaches for anticipatory action related to climate and conflict.
Dr. Vincenzo Bollettino is the Director of Resilient Communities Program at the Harvard Humanitarian Initiative. His research and professional experience include disaster resilience, humanitarian action, civil-military engagement in emergencies, and humanitarian leadership.
Patrick Vinck, PhD is the Director of Research of the Harvard Humanitarian Initiative and an Assistant Professor in the Department of Global Health and Population at the Harvard T.H. Chan School of Public Health and in the Department of Emergency Medicine, Harvard Medical School. He leads a team conducting research on resilience, peacebuilding, and social cohesion in contexts of mass violence, conflicts and natural disasters. After starting his career working on food security in Central Africa, he has specialized in socio-behavioral research and evaluations on the effects of violence and recovery efforts in fragile and conflict-affected states.