AI Accelerating Climate Data for High-Quality Index Insurance
Climate Data is a critical component of managing and overcoming the climate crisis. Having high-quality climate data allows our team to develop tools to help farmers impacted by climate change to navigate the changing future. Sprout leverages AI models to improve its climate data and insurance.
Sprout uses climate data from multiple sources to build its insurance models and assist farmers throughout the season with climate-triggered agronomy advice. We use third-party climate data, like NASA Landsat 8 satellite data, and our in-house climate data. Our climate data is created from a combination unique data, satellite data and processed by AI algorithms.
Types of Climate Data
Let’s look at climate data in three buckets:
- Historical Data any weather observation in the past
- Current Data the weather right now
- Forecasted Data predicted weather in the future
Methods of Getting Climate Data
Translating the weather into actual data (numbers we can use) requires measurement tools and observing the weather when it happens. When there’s enough real weather observations, we can use other tools to fill in gaps where we have data or predict the future (without observations — i.e. no time travel).
- Weather Stations (hardware on the ground)
- Satellites (hardware in the sky)
- AI (software-generated algorithms that provide estimates)
Sprout’s Unique Crop Climate Data
Working directly with coffee cooperatives and coffee buyers in Kenya, our team is able to understand where coffee cooperatives are located and where there are individual coffee farms. Most farms are not addressed in Kenya, so we went farm-to-farm documenting various types of coffee farms. We couldn’t map every farm (that would take years!), but we sampled a good number across our focus region, Central Kenya.
In order for us to be able to make unique climate data across the whole coffee-growing region, we co-developed an AI-generated coffee map using our unique data and the help of our (very smart) partners at NASA Harvest. They do crop mapping all over the world, a work on important projects, such as monitoring the wheat production in Ukraine.
Our AI-generated coffee map allows us to understand where there is likely to be a coffee farm (in pink) and where there is a low likelihood of a coffee farm. The map is an extrapolation from the thousands of data points we collected across the region. This map is used in the process of making our high-quality insurance product so that we only pick up pixels from coffee farms.
Index Insurance and Climate Data
The impact of climate change on coffee, and other crops grown around the equator in lower income countries, are predicted to be double digit losses in production value. By 2050, significant coffee production is at risk. Farmers need to transition to a more resilient coffee production while facing the risk of more extreme weather conditions.
Sprout designs insurance that covers farmer from the risk of extreme weather. Our team works with local agronomists, farmers and buyers to understand the impact that weather has on crop production. Using that knowledge, we create unique climate data and build insurance models around the most important climate variables — rainfall, temperature and many combinations.
Climate Smart Coffee is Sprout’s flagship program, where we have designed a rainfall insurance product to protect farmers from the impact of drought on coffee production. We look across the four major phases of coffee growing, from developing buds in pre-flowering to final expansion of the cherry before harvest. Each phase has a different sensitivity to insufficient rain, and our actuary models take this into account.
The insurance product we’ve built is useful across the Central Kenya Region where an estimated 250,000 farmers grow coffee and hundreds of coffee cooperatives, like Mutira Coffee Cooperative in Kirinyaga County.
Sprout at the United Nations HQ
On May 2 ,2023, Sprout was invited to attend a session at the United Nations Headquarters in New York City. Sprout was among finalists in the AI for the Planet competition organized by the Alliance for the Planet.
The meeting was discussed the role of Artificial Intelligence to address the Climate Crisis. Highlights included discussions around how we generate climate data. Climate data comes from an accumulation of weather observations, whether through weather stations on the ground or satellites from the sky. Observing the weather is the first step, translating that observation into useful, accurate data is another complex process. Some countries have more been able to invest more in their data collection and technical trainings on collecting and analyzing data. Lower income countries have less access data and less resources working on to backfill their historical data and predict future data using algorithms. It will be a collaborative effort to create a fair and equal transition to a climate smart future.
The Permanent Mission of the Kingdom of Morocco and the Permanent Mission of France, co-chairs of the Group of Friends for Climate Action, host this side event to the Science Technology and Innovation (STI) Forum on the role of Artificial Intelligence to address the Climate Crisis, in collaboration with the AI for the Planet Alliance.
Sprout featured in BCG and Google Report: Accelerating Climate Action with AI
In November 2023, BCG and Google published a new report on How AI Can Speed Climate Action. The report dives into the critical role AI can play in climate action and how leaders can make the most of it, and also puts associated risks into perspective. The full version is here: https://lnkd.in/eH_dee_G.
