In this first tech blog post we describe our growing confidence in our key sustainability signal, water stress, and why we've decided to narrow our focus.
As former academics, Alex and I know that, in the long term, our integrity is our reputation. And while climate risk is going mainstream, a veracity crisis in ESG offerings threatens to divert progress into greenwashed business intelligence and reporting products. As such, Alex and I had been conscious not to overpromise the sustainability signals that we would be able to deliver to investment decision makers. To do so would undermine the very reason we founded oxeo: to drive sustainability in primary capital market decision-making.
In the last few months of prototyping and development, we have demonstrated to ourselves that for water stress, our first key sustainability metric, we can measure and predict it with accuracy and quantifiable uncertainty, anywhere on our planet Earth.
Drought prediction is hard. Not only must you handle the uncertainty and stochasticisity of weather, and the long-term drivers of the climate, you must also consider human behaviour, extraction by companies and their assets, and interaction with land cover changes and the built environment. Basin hydrology is unique across the globe, and infiltration to the water table is not directly observable. Water basins are connected via upstream and downstream relationships, so must be analysed in the context of their network arrangement. The scarcity of in-situ groundwater data and static conditioning data like geology, soil type, and land cover creates a generaliseability problem across the globe.
I recently led a team of OxEO and University of Oxford researchers in the Wave2Web hackthon, organized by the World Resources Institute and sponsored by BlackRock and Microsoft. We enhanced state-of-the-art computational hydrology methods predicting 90-day water availability in six reservoirs west of Bengaluru, India. Our innovation was to seamlessly pass hydrological state variables through time horizons, conditioning first on historic weather data, then on short-term forecast data, and finally climate data.
With OxEO we go one step further and solve the generaliseability problem. Using Earth Observation, we obtain surface water extents which we can use as targets for our prediction problem. The result is a continuously learning system, making short- and long-term predictions of water stress, tuned to weather forecasts and climate projections, back-cast against 40 years of observational data, and continuously deployed for every network of basins on the planet.
It is our view that as sustainability intelligence offerings mature, there may be many companies offering business intelligence-type solutions, but the primary sustainability signals, requiring deep sustainability and technical expertise, will be concentrated in the hands of only a few. With OxEO, we intend to own the space of near-real-time global water stress intelligence.
If climate change is a shark, water stress is its teeth. The coming decade will see water crises unfold as growth, climate change, and water scarcity collide. If we can change how investment decisions are made today, perhaps we can avoid tomorrow's crises. And so with narrowed focus we set out to change how water stress is measured and managed, something that affects almost every company, investor, and person alive today on our shared planet.
- Lucas Kruitwagen, Cofounder & CTO