Insights and Success Stories

A new evidence-based system to strengthen drought resilience in the Horn of Africa

Written by Earthpulse Team | 27 May 2026

Drought is considered one of the world’s most economically damaging natural hazard. It is also one of the hardest to see coming and the least understood. In fact, unlike floods or storms, it builds slowly, making it hard to detect, and most importantly to understand the full social impact and economic cost.

In the three-year period from 2020 to 2023, the Horn of Africa faced the most severe drought of the last 70 years. Brought on by five consecutive failed rainy seasons, this extreme period left catastrophic consequences. With over 31.9 million people in urgent need of assistance, this prolonged period of drought caused the death of an estimated 43,000 people in Somalia alone and wiped-out entire livestock herds worth more than $1.5 billion in Kenya.

The communities hit the hardest during this time, such as those of pastoralists and smallholder farmers, were also the least equipped to absorb the economic losses. So far, the damage has been even more structural, as recurring years of drought created an even more noticeable gap on the financial tools needed to mitigate and also to deal with its devastating impact.

Insurance products, contingency funds and payout mechanisms are the most important recovery tools that governments use, and households need, to recover from drought. However, effective drought risk financing depends on transparent, consistent, long-term data.

The World Bank’s bet on better data: the Next Generation Drought Index

In response to this recurring issue, The World Bank’s Disaster Risk Financing and Insurance Program (DRFIP) developed the Next Generation Drought Index (NGDI) Dashboard to address this gap directly.

Developed with ESA, the platform bridges satellite technology, risk modelling and decision-making giving governments, insurers, and development practitioners a common, transparent framework of evidence with state-of-the-art drought risk information for targeted financing.

Critically, it is important to note that drought risk indicators can be computed months before the actual harvest, allowing for better planning and anticipatory action before the worst impacts materialise. For the most vulnerable communities, that early warning can be the difference between a managed, and mitigated response, and a humanitarian crisis.

The World Bank’s $360.5 million DRIVE project, which protects communities across Kenya, Ethiopia, Somalia, and Djibouti, already uses the NGDI to design and calibrate index insurance products.

The platform’s effectiveness rests on the quality of the data that feeds it. As satellite missions age and new sensors come online, that quality cannot be taken for granted.

From global mandate to local action: how ESA’s GDA facility is supporting the World Bank

To keep the platform operational and scientifically robust as satellite missions evolve, the World Bank engaged ESA’s Global Development Assistance Fast EO Co-Financing Facility (GDA FFF) to upgrade the platform.

Earthpulse led the data science pillar of this upgrade, working alongside DHI, the platform’s original developer. Our mandate: replace the legacy MODIS vegetation dataset, which is no longer available, with the higher-resolution Sentinel-3 NDVI300 product and do so in a way that preserves the integrity of two decades of historical drought records.

Earthpulse delivered three things:

  • A production-ready data pipeline that downloads, cleans, gap-fills, and harmonises Sentinel-3 NDVI300 data, covering both historical baseline generation and near-real-time operational updates. We tested four gap-filling algorithms and selected the hybrid spatio-temporal method, which reconstructs missing data accurately in under two minutes, far outperforming the alternatives in both speed and precision.
  • A cross-sensor harmonisation calibration that aligns Sentinel-3 readings with the MODIS baseline, achieving a correlation of 0.985 and near-zero systematic bias across four and a half years of data. This ensures that drought severity scores calculated today are directly comparable to those from 2001.
  • An exposure analysis module that intersects drought severity maps with LandScan 2023 population data and ESA WorldCover land cover data – quantifying, for any drought event, how many people and how much agricultural land are affected. The outputs feed directly into the NGDI Dashboard.

All code is open-source under the GNU GPL licence, publicly available and fully documented, so the World Bank and its partners can maintain and extend these capabilities independently.

A drought risk platform built to last and to scale: outcomes that matter for drought risk financing

This upgrade directly strengthens the evidence-base that financial institutions need to make informed decisions. The World Bank now has a vegetation data pipeline fit for the next generation of satellite infrastructure – one that produces cleaner data with global coverage, catches more of the errors that distort financing decisions, and connects drought severity to real-world exposure in a way the platform could not do before.

The NGDI Dashboard, already active across the Horn of Africa, can now operate with greater confidence. For index insurance designers, that means payout thresholds calibrated against cleaner, more reliable data reducing the risk of triggering payments when there is no drought or missing them when there is. For governments, it means better-calibrated early warning. For the millions of people whose livelihoods depend on these financial tools working when drought hits, it means a system built on better science.

Built with, not just for: co-creating solutions with the World Bank and partners

The NGDI upgrade was not developed in isolation. Throughout the project, Earthpulse worked in close coordination with the World Bank’s Disaster Risk Financing and Insurance Program (DRFIP) team and DHI to ensure that every technical decision, from the choice of gap-filling algorithm to the design of the exposure analysis module, responded to real operational needs. Outputs were reviewed and validated against the World Bank’s existing workflows, ensuring that the new pipeline integrates seamlessly into the NGDI Dashboard and meets the standards required for use in sovereign risk financing decisions.

What comes next: scaling evidence-based drought resilience globally

The same pipeline that runs for the Horn of Africa today can be deployed for any drought-prone region in the world – extending the reach of evidence-based drought risk financing to new geographies as the World Bank and its partners expand the NGDI’s coverage.

 

 

Article originally published on https://gda.esa.int/story/a-new-evidence-based-system-to-strengthen-drought-resilience-in-the-horn-of-africa/
Featured image credits: European Space Agency