AI could provide solutions for climate change
It takes approximately 40 minutes for 82,944 processors on the world’s fastest computer to compute what one percent of our brain calculates in a second.
Despite this lag, AI has more promising solutions than humans when it comes to addressing the issue of climate change. Climate change is a serious issue that needs immediate attention across the globe. Compelling pieces of evidence that point towards a bleak future include a rise in global temperatures, extreme events, shrinking of the ice sheet, warming oceans, sea level rise, and ocean acidification.
There are several good reasons for us to believe that the fourth industrial revolution powered by AI is a perfect opportunity for researchers to embrace AI as a transformative tool to address this grave issue.
In recent years, computers have helped store massive amounts of structured and unstructured data. With trillions of sensors around us and improved algorithms (thanks to advanced deep learning), the power of AI is only getting unleashed.
Headways in cloud technologies, quantum computing, distributed computing, and deep-learning chips and graphic processing units (GPUs) ensure that businesses can handle and integrate big data with agility and speed. Further social media platforms and open-source communities are helping create ‘democratization’ and connecting by building something called a ‘collective intelligence’ e.g. open-source communities developing AI tools and sharing applications.
A recent study by World Economic Forum (in collaboration with PWC and Stanford Woods Institute for the Environment) shows that extraordinary challenges can be dealt through the combination of the Internet of Things (IoT), robots, autonomous vehicles, biotechnology, nanotechnology, and quantum computing, among others.
Recently, the range of AI capabilities have expanded and now includes
Automated Intelligence — includes ‘labor-intensive tasks’ that systems can complete automatically e.g. recycling of household materials.
Assisted intelligence — is based historical input to detect patterns e.g. deep learning, natural language processing and anomaly detection for extreme events like hurricanes etc.
Augmented Intelligence — allows people to understand and predict an uncertain future. E.g. creating interactive climate simulators to help people establish connections on climate policy etc.
Autonomous Intelligence — automates decision making with no human interference.
Forest Observation after Hurricanes
Microsoft recently identified artificial intelligence as a “game changer” for the planet. With a $50 million commitment in five years, the Seattle-based firm launched AI for Earth and in a span of six months, awarded 35 grants in more than 10 countries. These projects are using AI to focus primarily on climate conservation, biodiversity, water, and agriculture. Through the grant, the projects are provided access to the company’s cloud platform, Azure and AI tools.
One such research funded by Microsoft is being conducted by Columbia University’s Maria Uriarte, a forest ecologist and a professor of Ecology, Evolution and Environmental Biology along with Tian Zheng, a statistics profession at Data Science Institute. Their study focuses on the effect of Hurricane Maria on El Yunque National Forest, the only tropical rain forest in the US Forest system. The researchers apply Microsoft’s machine learning tools to deeply analyze data on how tree species in the forest fared during the hurricane — which species withstood the strong winds and which didn’t.
Advanced technologies enable the researchers to survey the 28,000-acre forest through advanced imaging and remote sensing technologies. Such tools make the study more effective by making the ground observations easier. Data mining remains crucial to the task since the plethora of images needs to be clustered, sorted and analyzed. By partnering with NASA, the researchers have access to millions of images captured by NASA’s satellites and fly-over planes. These high-resolution pictures are mapped, matched and identified with data that has been collected by Uriate which consists of every single tree in given plots.
The team relies on AI to identify to distinguish and classify the species of trees. According to State of the Planet, Uriarte describes
Then we can use that information to extrapolate to a larger area. We use the plot data both to learn [i.e. to train the algorithm] and to validate [how well the algorithm is performing].
Interestingly, AI is not only being used as a tool for analyzing species in forest areas but also looking at innovative ways to gather and examine ocean data.
Oceans Data Platform
Due to their massive formation underneath the ocean, coral reefs play a crucial role in absorbing elements coming from the ocean, resulting in the reduction of coastal erosion. Corals help in reducing the damage caused by tropical storms, hurricanes, and energy of tsunamis. But today, they are among the most threatened ecosystems on Earth, largely owed to unprecedented global warming and climate changes.
A collapse of coral reefs could be catastrophic for human civilization since it provides food security to half a billion people, contributing around US$375 billion per year to the global economy.
With the earth’s atmosphere is heating, changes in the frequency and intensity of tropical storms, rising sea levels, altered ocean circulation patterns, the world’s coral reef ecosystems could fully collapse as soon as 2050. This could mean that the coral reefs in all 29 reef-containing World Heritage sites would cease to exist by the end of this century if we continue to emit greenhouse gases under a business-as-usual scenario (UNESCO).
Interestingly, AI’s usage is helping 50 Reefs initiative identify climate change-resistant corals to create a portfolio of key reef ecosystems that are
Researchers apply AI tools to understand the types of coral reefs that have a better survival rate all across the world. Bloomberg Philanthropies, the Tiffany Co. Foundation, and Paul G Allen Philanthropies jointly fund the idea of applying technology and science to prioritize such protection efforts.
Research scientist Dr. Emma Kennedy, is a part of the 50 Reefs project. A combination of 360-degree imaging tech and AI allows scientists to gather and analyze images of shallow water reefs within seconds. The time efficiency can be significant especially if the images are more than 56,000. Latest deep learning techniques further permit AI to recognize different types of corals through pattern detection from complex contours and textures of reef imagery.
Since 1970, the number of natural disasters has quadrupled, causing more than 3.3 million deaths and resulting in trillions of catastrophic damages. By 2030, 60% of the world’s population will live in cities with 1.4 billion facing the highest risk from natural disasters. Given the startling stats, climate science is progressing towards using statistics, machine learning, and data mining to forecast weather changes. This is known as Climate Informatics.
Deep learning and AI are working in collaboration to ensure that weather predictability is accurate, especially in cases of extreme weather conditions. Since the capacity for variables (ocean and atmospheric dynamics and ocean and atmospheric chemistry) is close to being unlimited, many simulations can be created to prove useful. Rising baseline sea levels (due to climate change) mean larger storm surges from hurricanes.
By comparing the past predictions to actual outcomes, AI can improve its simulation capabilities, that has greater accuracy than weather forecasting by humans.
IBM’s Deep Thunder is a great example of how big data relies on machine learning tools to initiate a global forecasting model. The Deep Thunder group is also aligning work with research centers in Brazil and India to accurately predict flooding and potential mudslides due to the severe storms. Deep Thunder is powered by IBM’s own platform, Watson which aims to forecast the weather with improved accuracy. The data through the acquisition of The Weather Company only creates additional data including extensive weather records and weather statistics from all over the world. Using the WIoT weather model through the API, repair crews and utility companies can get proactive and move in quickly to repair damage or restore services during severe storms. Artificial Intelligence is also helping in selecting more reliable weather prediction model. Traditionally, averaged and equal weights were given to predictions but AI helps in assigning more weights to predictions that are more reliable and accurate.
Soon, hybrid AI techniques could be the answer to addressing serious concerns regarding climate change. It may not be wrong to say that this could mean a human-machine collaborative learning.
For artificial intelligence, this would mean that developing advanced, human-like capabilities like emotions and the ability to learn more quickly.
Leveraging AI to address climate change issues calls for collaborative efforts and needs to be harnessed efficiently for the benefit humankind. While the AI applications to the environment remain beneficial, once unleashed, AI researchers, tech pioneers, AI adopters, industry leaders, and governments need to ensure its ‘safety, explainability, transparency, and validity’. If used with caution and the right sensibility, AI could encapsulate a future that will create a safe environment for humanity and the planet.