A better future for the planet means we need better maps today – the case for combining geospatial data and AI

Image credit: NatureServe

Imagine setting a budget without knowing your current income. It’s certainly possible to complete the exercise – one could simply list all the planned and dream expenditures for the year and then try to reconcile this as the charges pile up, or a more frugal option would be to make an educated guess as to what your income may be and budget accordingly. Neither is very precise, and there’s substantial room for error that could far overshoot your budget.

That is essentially how we are managing our natural resources today, and that lack of precision, granularity and accuracy means that we as a society are making ‘withdrawals’ out of the global carbon budget and our planet’s future far faster than is responsible. Top climate scientists from around the world have given humanity an 11-year deadline to prevent catastrophic climate impacts that will result in floods, drought, habitat and species loss, and water and food scarcity.

This reality is sobering. But we’re also in the middle of another age, that sees technology and AI accelerating at an equally staggering pace, scale and magnitude. The combination of AI and GIS is required to accurately observe, classify, and predict changes to environmental systems, from watersheds to agricultural land to forests. Without the ability to see and understand these changes, it is exceptionally difficult for researchers, conservationists and policymakers to take effective actions to protect and conserve our lands and waterways.

Now, with breakthroughs in geospatial and AI technology, we can get a better, more accurate and detailed view of our natural world. It was this point of view that led us to launch AI for Earth two years ago at the 2017 Esri User Conference. It’s now grown into a $50M, five-year commitment by Microsoft to empower anyone to use AI to solve global environmental challenges. Esri is a partner in this work, offering grantees access to ArcGIS Pro and working with us on the Microsoft Data Science Virtual Machine, which combines ArcGIS Pro with Microsoft AI on Azure to provide advanced geo AI capabilities.

Today, I’m at the 2019 Esri User Conference with some good news. With more than 380 AI for Earth grantees, many of which are working with Esri as well, it’s clear that this investment is yielding results that are providing new, in-depth insights about the world around us. Here are a few leaders in this space:

 A new map of the most ‘at risk’ places in the United States

Better conservation strategies depend on the answer to a simple question - where are the most important places for the protection of threatened species? Yet the answer to that deceptively simple question is challenging to produce because we have a basic but considerable information challenge in identifying where at-risk species are found, and where they are not. While we have a great deal of data, making sense of it is a seemingly impossible task given the sheer number of species to be identified and tracked and disparate data collection methods.

NatureServe has been working to address this challenge since 1974, and their network now spans all 50 U.S. states, the provinces and territories of Canada, and much of Latin America. Using a standardized scientific methodology for tracking at-risk species over many decades, NatureServe has built, in the words of Chief Scientist Healy Hamilton, “an internationally standardized dataset on at-risk species that is both taxonomically rigorous and geographically precise, which we are now leveraging in new ways to inform conservation actions.” Without these standard data, decision-makers would be without the best available data to identify and effectively protect threatened species and their habitats.

At the User Conference, they shared an update on their Map of Biodiversity Importance (MoBI) and it’s incredibly impressive! You can check it out here:

For this project, NatureServe selected more than 2,600 species for distribution modeling, ranging beyond the “cute and fuzzy” mammals to include endangered plants, pollinators like bumblebees and butterflies, and even freshwater invertebrates like mussels and crayfish, all of which play important ecological roles in maintaining healthy functions in the communities where they occur. With support from Esri, The Nature Conservancy, and Microsoft, NatureServe modeled habitat (including 30-meter resolution on soils, topography, climate, land cover and hydrologic systems, as well as predictor layers for aquatic and terrestrial species) and synthesized the models into a map that identifies the most important places for conserving at-risk species. Armed with this information, decision-makers can make more effective and precise conservation action plans.

Real-time mapping of the impact of climate change on species

Climate change is impacting marine species, as sea levels rise and temperatures and acidity change. Even small changes can impact species in big ways, as smaller species form a vital part of the food chain. Consider whales - eight species of whales live in Antarctic waters due in large part to the abundance of krill, very small crustaceans that are near the bottom of the food chain. But the peninsula’s pristine ecosystem, including the krill, is being disrupted by the effects of climate change. Duke University’s Marine Robotics and Remote Sensing Lab is developing machine learning models, pairing the Azure Data Science Virtual Machines (DSVM) with Esri’s ArcGIS Pro to develop models that can manage the massive amounts of image and video data on the whales that are collected via drones and satellite imagery.

In other cases, whole habitats are being altered or lost entirely as climate change is playing a role in more frequent and intense storms. The National Audubon Society is using Microsoft’s cloud and AI tools, paired with Esri’s ArcGIS tools, to improve bird monitoring after weather-related disasters. Audubon’s Tim Meehan created a tool that geotags certain areas and then uses AI to perform a before-and-after comparison to assess storm effects. This enables researchers to quickly and efficiently evaluate damage and act to preserve endangered birds.

Join us

It’s never been more urgent to develop efficient, effective, and scalable solutions to our environmental challenges. By harnessing the powerful combination of data and technology, many organizations are stepping up to the plate to solve different aspects of our climate crisis. But we need more people to join in and share tools, solutions and knowledge. Here are a few ways you can join in this important work:

·      Apply for an AI for Earth grant

·      Explore AI for Earth APIs

·      Check out the Data Science Virtual Machine

·      Learn more about AI for Earth