Species around the world are in trouble, but we can help

Apr 18, 2018   |   Bonnie Lei, Project Manager, Microsoft AI for Earth & Jennifer Marsman, Principal Engineer, Microsoft AI for Earth

Cell phone photographing flowers

One of the greatest global challenges of our time is avoiding the impending extinction of a growing number of species around the world.

Human life is sustained by the diverse range of other species on this planet. We rely on photosynthetic organisms like plants and phytoplankton to produce the oxygen we breathe, and insects like bees and butterflies to help pollinate our food crops. Abundant lifeforms in watersheds help filter and protect freshwater. Even species that do not have an obvious benefit to humans are important to protect – they play a role in maintaining delicate ecosystems, other species, and the resiliency of areas to extreme weather and climate change, to name just a few.

Unfortunately, biodiversity worldwide is rapidly declining. In March, the U.N.’s Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services (IPBES) released a new set of reports finding that animal and plant species are under severe threat in nearly every region of the world.

That’s just what we know – the value of many known species has yet to be determined. Consider the Pacific Yew tree’s bark. It is now the source of a cancer drug, yet its medical properties were not discovered until 1977. More than that, we know precious little about the full picture of life on Earth. Scientists have observed and described only 1.5 million species of the estimated 10 million on earth. We should be trying to maintain the fullest, most diverse array of species on this planet, to maintain the most resilient overall ecosystem.

To accomplish this, we will need more people using technology – it’s how humanity has solved every major problem we’ve ever faced. Human ingenuity, paired with advanced technology like artificial intelligence, can help us close the knowledge gap, make new discoveries, and more effectively conserve and protect species.

That is our goal in supporting the work of more than 65 AI for Earth grantees across more than 20 countries, many of whom are working to address key threats to global biodiversity. Here are three of these heroes for Earth:

FeiFang-headshotFei Fang, Assistant Professor at Carnegie Mellon University (Malaysia & China)
Poaching has increased in the last decade, often leading to the extinction of species and destruction of ecosystems. In protected areas, rangers are faced with the difficult task to try to catch poachers before they harm wildlife. With a limited number of rangers, and a finite amount of time they can spend patrolling, how can they travel the most effective patrolling routes?

 While she was a graduate student at the University of Southern California’s Center for AI in Society (USC CAIS), Fei Fang started working on the Protection Assistant for Wildlife Security (PAWS), using machine learning to process data from past ranger patrols and predict where poaching is likely to occur in the future. Now as an assistant professor at Carnegie Mellon University, Fang has continued her work to improve the speed and accuracy of this tool. She is using her AI for Earth grant to help train the algorithm to incorporate real-time information collected on rangers’ patrol routes, like footprints, to help them make changes to optimize their patrols and more effectively prevent poaching.

MichaelSchmidt-headshotMichael Schmidt, Deputy Director at Long Live the Kings (LLTK) (Puget Sound, WA)
Salmon are iconic to life in the Pacific Northwest, having sustained the local economy throughout history. However, in the Salish Sea, the network of coastal waterways in southwestern British Columbia and northwestern Washington State, salmon and steelhead populations have declined sharply in the past 30 years. LLTK is a non-profit organization that works to restore wild salmon and steelhead and support sustainable fishing in the Pacific Northwest.

 They are using their AI for Earth grant to improve the speed and accuracy of an ecosystem model to help protect salmon populations. This is a big task – LLTK helps aggregate thousands of data points across more than 60 participating organizations. Cloud and AI helps pull the data into information, and information into insights to assess young salmon and steelhead growth, track fish and marine mammal movements through acoustic telemetry, monitor marine conditions, and develop an intensive ecosystem model of the Salish Sea. Results from the Salish Sea ecosystem model will be used to improve hatchery, harvest, and ecosystem management, and inform habitat protection and restoration efforts.

UriarteZheng-headshots

Tian Zheng & Maria Uriarte, Professors at Columbia University (Puerto Rico)
Hurricane Maria, which hit Puerto Rico and the U.S. Virgin Islands in September and October of 2017, has inflicted an estimated $90 billion in damages. The Category 5 winds also caused widespread damage to forests across the islands. Forests serve as one of our planet’s carbon sinks, as trees utilize carbon dioxide from the atmosphere for photosynthesis. After a forest is damaged, the first tree species that start to grow back are smaller, younger, and tend to store less carbon. As climate change increases the occurrence of extreme weather events such as hurricanes, disturbances to forests will also increase and forests may not have enough time to ever fully recover.

Previously, forest monitoring often meant researchers had to travel by foot to tree plots to make manual measurements. But the type of large-scale, long-term monitoring needed in the face of climate change can be expedited by using remote sensing data, such as satellite images. Professors Tian Zheng and Maria Uriarte, members of the Data Science Institute at Columbia University, are using their AI for Earth grant to apply machine learning tools to assess the damage of Hurricane Maria by more efficiently processing satellite images from before and after the storm. Specifically, they are using ground observations of forest plots to create a deep learning model that’s capable of correctly classifying the species of individual trees in these satellite images in order to better understand how damaged forests recover over time, and the impact to their ability to store carbon and aid in climate change mitigation.

We’re just beginning to see the payoff of applying AI to these issues. We believe AI can have many other useful applications for addressing biodiversity. For example, motion-capture cameras take photographs of animals as they move throughout sanctuaries, and computer vision technology can be applied to these images to recognize each individual animal and track its movements, without requiring tagging. This is both less intrusive for animals and helps conservation groups conserve their limited human and financial resources. If a city buys a shipment of new trees, machine learning models can tell us the best places to plant them to optimize tree cover and reduce the heat from reflective roads.  Additionally, reinforcement learning can be used to tweak variables in a simulated ecosystem to better understand and predict future outcomes, enabling precision conservation and better resource management.

For AI to make a difference, though, it has to be readily accessible to these researchers and they need the skills and training to deploy it effectively in the field. That’s why we’re committed to lending Microsoft’s AI technology and other tools to those on the front lines—who have the experience and expertise to effectively leverage our technology for conservation. If you’re interested in partnering with Microsoft in your conservation work, we invite you to apply for one of our AI for Earth grants here.

The rest of us can help as well. Some of our AI for Earth partners and grantees are also leveraging technology to engage everyday people as citizen scientists. They have digital tools that you can use with your friends or family to record and upload your own species observations, contributing to vast repositories of species’ data that supports the work of scientists around the world. Here are two that you can use today —iNaturalist’s app (Android, Apple) or the Cornell Lab of Ornithology and National Audubon Society’s eBird tool.

So this Earth Day, as you reach for that shovel to go plant a tree, or strap on binoculars to help count birds in your area, don’t forget to grab your phone or tablet, so you can join us in harnessing technology to help understand and preserve biodiversity.

To learn more about Microsoft’s broader efforts to protect biodiversity this Earth Week, please check out tomorrow’s Microsoft on the Issues blog post. 

Tags: , , , , , ,