allows users to visualize and calculate trends for historical and future projections of annual streamflow, precipitation, and temperature. Similar constructs and solutions could be repurposed for localities where trained AI systems provide climate assessments and recommendations based on livability and workability parameters that are within the acceptable thresholds for a community. Communities could access environmental information and predictions, ranging from seasonal, extreme heat and precipitation forecasts to drought conditions. Consolidating this data with information from other domains, such as demographics, infrastructure costs, and cultural context, would provide comprehensive, actionable insights; improve understanding; and, ultimately, help communities ask the right questions and provide feedback on issues that matter to them. A Catalyst for Climate Resilience The data-driven science and AI scenarios envisioned here are only the beginning. By fully operationalizing the science, we will achieve a level of scale and agility where agencies, the private sector, academic and community partners, citizen scientists, and the public at large will be able to play a key role in building climate resilience. The pathways described here have the potential to solve the challenges already at our doorstep and those that Carl Sagan alluded to in his 1985 remarks where he urged for international amity. Almost 40 years later, the need for operationalization of science couldn’t be more urgent. Using novel AI approaches to address this need can be the catalyst that makes climate resilience a reality.
the affected area, such as downed powerlines, wind speeds, and hail. A digital twin of the locality visualizes the likely course of action, continuously updates the model as new information from the field is streamed in and allows planning according to different scenarios and outcomes. Post-event, the feedback “bots” or intelligent automations work in the background to refine the recommendations, based on the data and learnings from this event, to build resilience toward potential future extreme weather events. Extreme weather events have been causing billions of dollars in damages to infrastructure. Consider the conventional approaches to managing a large-scale infrastructure rebuilding project. From building to security and IT systems to operations, the process is often siloed and manual, forcing decisions based on incomplete or obsolete data. However, for critical infrastructure—particularly in regions that are prone to climate adversities—new AI-based capabilities are allowing for insight-driven planning and sustainment that dramatically improves operational efficiency, safety, and overall resilience. For example, AI and digital twins are being used to provide a digital thread that links authoritative data sources, such as enterprise datata systems, geographic information system layers, utility assets, and building information model (BIM) updates to enable integrated analysis; real-time feeds; and human-centered, contextualized visualizations for diverse stakeholders to make critical design and engineering decisions for improved climate resilience. Agencies are already implementing climate resilience solutions that assess risk across key climate threats. For example, DoD Climate Assessment Tool (DCAT) allows facility managers of DoD installations to assess risks from environmental threats under various climate scenarios. The U.S. Army Corps of Engineers utilizes relevant climate information to enhance infrastructure climate resilience. The Army’s Climate Hydrology Assessment Tool (CHAT)
Reducing or eliminating barriers to
operationalizing climate science could lead to astonishing advances in various sectors. Significant opportunities include infusing climate AI into our daily lives and multi- domain insights for decision-making, rapid preparedness, scenario planning, and recovery.
Novel Data-Driven Science and AI Scenarios Where the Science Is Operationalized While many of us may not be conscious of the technological changes around us, AI has already penetrated many moments in our day-to-day lives—from avoiding highway collisions to communicating with our home devices through speech recognition. Leveraging these already pervasive technologies and infusing important environmental and climatological information would be game-changing. Envision global positioning system (GPS) navigation with an AI application for a long road trip that recommends an optimal route to the driver of an electric vehicle (EV) based on weather conditions, overall traffic-flow management, locations of charging stations, and personal preferences. We are seeing the beginning of these types of innovations already. Imagine an AI application that dynamically aggregates and controls thousands of devices that are plugged into the electricity grid (including EVs) so they can supply
power when demand is greatest while minimizing inconvenience to vehicle owners. This would decrease costs by minimizing the need to build new power plants and increase the reliability of electricity supply by adding grid resources at scale. This aggregation of distributed energy resources—called Virtual Power Plants (VPPs)— could reduce peak demand in the U.S. by 60 gigawatts, almost 5% of current total electric grid capacity, with EV batteries potentially playing a significant role. Embedding AI algorithms with even small impact, when multiplied across hundreds of millions of devices, can compound the effect, facilitating optimization of electricity generation and load, creating a more intelligent electricity grid, and revolutionizing the energy sector. Delivering nuanced information in near-real time is possible with the combination of AI and climate science. Here’s an example: A storm is approaching a region and the region’s grid operators are trying to predict where damages and outages are most likely to occur. The operators pull up a dashboard showing historical data about the area, such as average and maximum rainfall. AI also displays predictions within the dashboard for
Prachi Sukhatankar leads climate and infrastructure engagements for clients, with a focus on using technology solutions such as climate intelligence, smart transport infrastructure, and advanced energy technologies to meet mission objectives.
“Not only is reliable, accessible, compelling data necessary to build long-term resilience and sustainability planning, it also provides context and timely warnings for those living in areas vulnerable to high-impact events in the here and now. Existential matters related to water and energy are certainly in this mix as are immediate threats including flooding, fire, and excessive heat. None of these issues respect state boundaries, thus demanding high levels of collaboration and a thoughtful, coordinated exchange of information and analytics.”
The future era of climate resilience requires a commitment to translating advanced climate science and purposefully embedding it into operational use; this is where novel AI approaches can help in powerful ways.
Barriers exist to operationalizing climate science, including limited access to scientific data and discoverability hindering collaboration and the limited infusion of scientific products into daily lives at scale.
Democratization of AI can help overcome these barriers and bridge the gap between science and operations. Examples include application of edge AI, NLP, predictive analytics, and generative AI toward advances that range from Earth-observation data acquisition, scientific collaboration, climate risk assessments, and community resilience recommendations to grid optimization and large-scale electrification. These are the catalysts for making climate resilience a reality—they are here and available now.
— Wellington “Duke” Reiter , executive director of Ten Across, a regional resilience initiative that focuses on proactive decision making around social, economic, and climate change.
VELOCITY | © 2023 BOOZ ALLEN HAMILTON
Powered by FlippingBook