Predicting Disasters and Saving Lives
With SEEDS and Microsoft
SEEDS wanted to make disaster warnings and other risk-related information up to date and obvious down to a neighborhood scale
SEEDS partnered with Microsoft and Gramener to create an advanced machine learning-based model that could help them plan and respond more effectively to disasters
PARTNERSHIP - SEEDS & GRAMENER JOIN HANDS
SEEDS is a not-for-profit organization that helps make communities resilient through comprehensive interventions in the areas of disaster readiness, response, and rehabilitation.
Gramener has partnered with SEEDS and Microsoft to develop the Sunny Lives AI model under Microsoft’s global program ‘AI for Humanitarian Action.
This predictive model aids in planning better risk reduction strategies against worsening climate emergencies and disasters.
Saving One House At A Time Before Disaster Strikes
When disasters like floods and heatwaves occur, warnings and other risk-related information are often vague and not up to date. Much of the current risk information work at a macro-level, covering hundreds of square meters of area and being too hard to understand by at-risk populations.
There was a need to localize the risk information down to a neighborhood scale. This would help to mobilize targeted immediate response and build long-term resilience in the vulnerable communities. SEEDS wanted to automate, scale, and code their vast experience accumulated over the decades responding to various disasters at the ground level.
The SEEDS Sunny Lives model is a proactive attempt to address this challenge. The solution can deliver granular risk scores down to house level. This means that authorities can warn the high-risk houses earlier, allowing them to safeguard their property and reduce the damage caused. For instance, residents can take pre-emptive measures such as moving to the top floor of the house or evacuating to a safe site. Such actions reduce the damage, lowering the overall time and cost of recovery.
Technology Used For This Solution
Creating the Sunny Lives model needed work from scratch due to some challenges:
- The population in those areas was dense enough to make accurate differentiation between houses difficult. As geographies change, so do the roof types, and identifying these different roof types was of prime importance for the model’s concept.
- There was no readily available training data to tag roof types for forecasting damage intensity.
Gramener, in partnership with Microsoft, built a Machine Learning model to counter these issues. To create the training data, Gramener accessed high-resolution satellite imagery and manually tagged over 50,000 houses. The roofs were classified under 7 categories based on the material used for their construction.
The final model could identify roofs with up to 90% accuracy. On top of this, Gramener added other geographic layers or parameters to the model.
The following parameters were considered – Waterbodies – River, Lake, Pond, and Sea, Distances from Road Network, Topographic Wetness Index (TWI), Elevation and Slope Vegetation (NDVI) – Sparse, Moderate, and Dense, Impervious surface, Landslide Risk, Building Footprints.
Developing the model took about 4 months. It was piloted in 2020 during cyclone Nivar and Burevi that hit the southern Indian states of Tamil Nadu and Kerala, respectively. The solution was deployed at scale in Puri during Cyclone Yaas in May 2021.
The solution required using a technology that could provide information that can be easily comprehended by the disaster response teams and the affected communities.
Considering that many disaster-prone areas are off the technology grid with low mobile penetration and network connectivity, it was necessary to break down the output from complex technologies like machine learning and artificial intelligence into simple and widely understood information. For example, SEEDS created specific advisories for dwelling types, suggesting the course of action for the individuals to reduce the risks.
A mock-up of the advisory shared by SEEDS with at-risk families. The solution utilizes high-resolution geospatial imagery and open datasets. It also uses geospatial analytics tools such as Kepler and Microsoft Azure to host and represent the results.
SUNNY LIVES BECAME A BOON TO LOCALS
LIVING IN DISASTER PRONE AREAS
The solution produces hyper-localized risk information that can be used by different stakeholders responding to the disasters. These stakeholders include national-level disaster management professionals, climate change adaptation experts, government agencies, and at-risk communities. Another highlight of the solution is its scalability. It can respond to different kinds of disasters such as floods, heatwaves, and earthquakes.
During cyclone Yaas, once the path of the cyclone was predicted, Gramener procured high-resolution satellite imageries of low-income settlements that would be impacted and ran the AI model for these areas. With the model, it was possible to get a risk profile of the area at a household in a span of a few hours, a process which otherwise would have required manual surveys for a few days.
With this solution, SEEDS could alert individuals with detailed directions on how to safeguard their belongings, house, pets, and livelihood. For instance, when Cyclone Yaas hit the shores of Puri in May 2021, the Sunny Lives model was able to help the impacted communities with success. One of them was Satyavati.
A resident of Odisha state in India, Satyavati spent over six decades of her life surviving cyclones. In November 2020, she lost all her belongings, including a food stall, to Cyclone Nivar.
This year, however, things were different. Satyavati, along with many others in her community, received a personalized advisory from SEEDS printed in the local languages a day before the cyclone made landfall. She followed the advisories as her family secured away from their belongings and left for a nearby government shelter.
She moved the contents of her food stall to a neighbor’s concrete house. After the cyclone, even though her hut was impacted as predicted by the model, she was able to resume her food stall immediately.
In this way, the model offers actionable insights and directly helps the communities most at risk.
Quantifiable Impacts That Saved Millions of Lives
SEEDS could improve their dwelling detection rate from 52% to an impressive 88%
High impact assessment accuracy with >90% of damaged houses detected as high-risk
Model scaled for heatwaves sharing pre-emptive advisories to 50K individuals from at-risk communities
1.1K families were evacuated on time using the advisories generated by the model