The Center for Data Innovation spoke with Max Evans, co-founder and chief technology officer of ClimateAi, a company that uses machine learning and artificial intelligence (AI) to identify and mitigate climate risks in the supply chain. Evans spoke about the data used in AI forecasting and how risk intelligence can mitigate and adapt industries to a changing climate.
Becca Trate: What is risk intelligence, and how does it relate to climate change?
Max Evans: Risk intelligence is a long-term, systematic approach that identifies, monitors, and quantifies risk incidents. First used in military situations, it’s since become commonly used by businesses to gather information and create a map of any potential everyday or high-risk hazards, from cyber threats to consumer dissatisfaction to extreme weather events to supply chain failures. It is actionable intelligence that businesses can use to optimize decisions in operational environments—ones that mitigate risks, take advantage of opportunities, and create new value for stakeholders.
That’s where climate change comes into play. Climate change is an unprecedented challenge facing businesses and their supply chains because climate volatility is a risk multiplier. Climate impacts, from frequent and severe weather events to subtle shifts in temperature or precipitation, can interrupt production and contribute to rising costs. It disrupts everything from commodity prices to supply chains.
Luckily, advances in machine learning (ML) techniques, AI, data availability, sensors, and climate science create a new paradigm for climate intelligence. Climate risk intelligence is an emerging field that helps businesses monitor their exposure to climate risk more accurately.
Trate: What types of data sources does Climate AI use?
Evans:Using proprietary ML and AI modeling, ClimateAi delivers unmatched weather and climate predictions. ClimateAi’s technology uses a hybrid ML technique called physics-driven AI. Physics-driven AI models are trained by a patented combination of public and proprietary data to recognize patterns and make faster, more accurate regional or hyper-local forecasts that can predicate changes for today, tomorrow, or 40 years from now. ClimateAi uses ML to analyze climate models, historical weather patterns, weather forecasts, climate information such as current temperatures and pressures, and other data.
Our historical data sources include the European Centre for Medium-Range Weather Forecasts’ research institute, the Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS) for historical precipitation data, and public observational weather data from the Global Historical Climatology Network daily (GHCNd), including the California Irrigation Management Information System. We also connect with most of the main private weather station providers.
For historical climate forecasts, we use several historical scenarios from the sixth Intergovernmental Panel on Climate Change report’s “Shared Socioeconomic Pathways,” including the low-emissions, middle-of-the-road, and high-emissions scenarios. It also connects to World Climate Research Programme’s CMIP6 climate projections.
Finally, for daily, sub-seasonal, and seasonal forecasts and hindcasts, we use an ensemble method of statistical and dynamical models from the Global Ensemble Forecast System, created by the National Centers for Environmental Prediction (NCEP).
Trate: Climate change impacts all sectors of the economy, but in what sector is climate data most impactful?
Evans: Economic sectors that depend on weather and climate across multiple time scales are especially vulnerable to climate change. Agriculture is a common example, as players across the value chain, from growers to retailers to financiers, make several key decisions based on weather and climate.
These decisions include what crops to plant, planting times, the potential for drought, over-precipitation, heat, and potential weather extremes. These decisions have traditionally been made using historical data, but climate change has drastically accelerated shifts in weather patterns. That means that better climate data can help optimize these decisions.
But other, seemingly unrelated industries also make decisions based on climate. For example, in the retail industry, understanding forecasts for the season ahead can help inform production decisions. An apparel company will be able to produce more raincoats and fewer summer clothes for the fall season if there is a colder, rainier forecast for the fall. This can cut overproduction, saving emissions and money. Thinking about the climate in multiple lead times can help inform product life cycle decisions.
Trate: What role does risk intelligence play in optimizing climate adaptation?
Evans: As climate change accelerates, we must not only cut emissions and slow the pace of global warming, but we must also adapt to the physical consequences. Adapting, at its core, means changing to better survive in a new environment. Climate change’s impacts vary by location, but risk intelligence can support decision-making and enable more effective climate adaptation responses. Climate risk intelligence shows regionalized short-term and long-term climate hazards and risks. It can show which locations are vulnerable to gradual sea-level rise, enabling communities to build infrastructure like sea walls, or project the severity and incidence of drought, enabling decision-makers to install new water-efficient technology to combat scarcity.
It also identifies new opportunities for adaptation. Agricultural decision-makers can, for example, make new crop choices suited to the changed climate. Within each crop, there are often many varieties with different tolerances and abilities. One of the least invasive ways to thrive in spite of increasing climate risk is choosing the right variety for the season. So, decision-makers can use climate intelligence to make an educated decision on which variety or varieties to plant, given the weather and the probability of different outcomes.
Trate: What role can risk intelligence play in mitigating the impact of climate change?
Evans: Risk intelligence can play a key role in mitigating the impact of climate change. Mitigation refers to efforts to reduce or prevent the emission of greenhouse gases. Climate risk intelligence has knock-on effects that mitigate emissions. For example, making the right decisions around agricultural inputs reduces water, fertilizer, and pesticide use, which reduces greenhouse gas emissions per ton produced. In the retail example, reliable weather forecasting translates easily into demand forecasts. Demand forecasting enables companies to produce and manufacture the correct amount of goods, which reduces overproduction.
Risk intelligence can also identify the long-term disastrous impacts of climate change, which spurs mitigation efforts to avoid these impacts. The technology behind climate risk intelligence like AI and ML can also be used in energy efficiency measures throughout transportation, agriculture, and industry.