The Microsoft Planetary Computer
Susan Leighton spotlights the Planetary Computer, combining a multi-petabyte catalog of global environmental data with intuitive APIs, a flexible scientific environment that allows users to answer global questions about that data, and applications that put those answers in the hands of conservation stakeholders.
In Microsoft’s 2021 Environmental Sustainability Report, in addition to discussing the early progress on commitments around being a carbon negative, water positive, and a zero-waste company by 2030, we stated that we would build a Planetary Computer. So, what is it? This blog contains a consolidated set of details pulled from our Sustainability Report and docs that describe the Planetary Computer and some of the early uses of it.
Background
The Planetary Computer enables data-driven decision-making by scientists, researchers, policy makers, and developers by providing global-scale environmental monitoring capabilities. We believe that we will be putting forth our biggest contribution to the protection of ecosystems by identifying, investing in, and orchestrating a coherent cloud infrastructure, bringing together stakeholders and offering best-in-class science-driven tools, AI (based upon the earlier AI for Earth), machine learning, sustainability apps, distributed computing frameworks, and petabytes of environmental data and satellite images.
With these resources, people will be able to measure, monitor, and model the data, exposing the impact of our historic actions or inactions locally and globally, on the world’s ecosystems. We will then have information and insights that will facilitate our ability to manage the world’s ecosystems, as the modeling results are analyzed and interpreted.
“Only when we have a massive amount of planetary data and compute at a similar scale can we begin to answer one of the most complex questions ever posed – how do we manage natural resources equitably and sustainably for a prosperous and climate-stable future?” Dr. Lucas Joppa, Chief Environmental Officer.
Data
Private Preview of the Planetary Computer began as planned in April 2021. Over 500 users were signed up and able to access 24 petabytes of environmental data across more than 30 key environmental and Earth observation datasets with APIs for access and scalable compute capacity. The available datasets are accessible via Azure Blob Storage after ingestion and transformation into a cloud-optimized format, and the datasets can be used within our Planetary Computer Hub or outside of our hub – on the user’s own compute resources.
The data that we have aggregated is from organizations around the world, and we will continue to expand the available datasets as more data is collected. Some examples of datasets available to the Planetary Computer include the following:
- Global Flood Risk dataset (used to assess the risk of flooding in coastal areas),
- Global Water Reservoirs dataset (used to determine water availability for drinking and for hydro power planning for reservoirs around the world),
- the Global Biodiversity Information Facility (GBIF) global biodiversity occurrence dataset documenting over 1.6 billion species occurrences, and
- the European Centre for Medium-Range Weather Forecasts (ECMWF) catalog of real-time meteorological and oceanographic datasets.
The entire set of data currently available is here in the Planetary Computer data catalog.
Applications
The development of applications has already begun by our partner network. In order for the application development by the partner network to be most efficiently enabled, the Planetary Computer has been built on open-source: open-source tools and open standards. Some of these tools have been developed by the open-source community, like the work to streamline and standardize the cataloging and discovery of geospatial data by the STAC (Scientific and Technological Activities Commission) community or the set of data science tools developed by the Pangeo community for earth science.
Some examples of the initial applications that have been running on our Planetary Computer are focused on conservation planning, forest risk management, and land cover classification. Here are more details on some examples:
- The Nature Conservancy (TNC) is using the Planetary computer to protect and manage land, oceans, and freshwater biodiversity. Microsoft and TNC are enabling more equitable access to the Marxan Planning Platform (spatial planning software used by conservationists that helps with decisions about the designation of protected areas) by implementing it on the Planetary Computer. Originally launched in the early 2000’s, Marxan is now able to scale using the Planetary Computer. This is the first open-source spatial conservation planning cloud computing platform with free access for all users globally. For more information, see this recent announcement: MARXAN Platform Changes the Nature of Conservation Planning.
“Over time, the demand for Marxan outgrew our ability to service the global community, we needed to scale Marxan, capitalizing on new innovations in computing, data science, and technology- that’s why we partnered to develop this new, first of its kind platform,” Dr. Jennifer McGowan from TNC.
- Conservation Science Partners (CSP) has also been using the Planetary Computer to analyze changes across forests, habitats, and the impacts of human land use by leveraging earth observation data, artificial intelligence, and their own deep subject matter expertise. The CSP Analytics Labs is affiliated with the Planetary Computer and has been making global-scale environmental capabilities accessible to scientists, developers, and policy makers in order to facilitate data-driven decision making. The CSP’s Analytics Lab runs applications that provide answers to complex questions in real time, using earth observation data, AI, data science, and their own subject-matter expertise. For more information, see CSP launches its new Analytics Lab! – Conservation Science Partners (csp-inc.org)
- Another application example is CarbonPlan using the Planetary Computer to enable more transparent and scientifically rigorous evaluation of forest-based climate solutions. (Microsoft even used this app to visualize the climate risks of forest carbon offsets to help us more effectively manage our carbon removal purchasing program.)
- The 2020 Global Land Cover map is another application that has benefited from the Planetary Computer. Esri has published this new high-resolution 2020 Global Land Cover map for global users. The creation of this map was dependent upon the collaborative effort of the Planetary Computer’s compute and data resources, along with a deep learning AI and classification model developed by Esri’s partner, Impact Observatory. Our Planetary Computer hosted approximately 500 terabytes of imagery to facilitate the development. For more information, see Most current and comprehensive global imagery map shared by Esri
What’s Next and Key Trends
The following are our next steps along with the trends we are seeing:
- Make publishing and using spatial data in the cloud simple – make it easier to ingest and access data in a way that lowers the complexity. This is especially important given the exponential growth trend of the volume and variety of spatial data.
- Solve data access challenges for science and operational applications – help ensure the focus remains on building the apps. We are seeing enthusiastic data providers, making their data more accessible for environmental sustainability use cases.
- Expand the catalog of datasets hosted in the Planetary Computer – continue identifying valuable, openly licensed datasets and accelerate their availability. Along with the increase in available data, there needs to be an increase in accessibility in sharing the data via cloud-optimized formats and APIs that will be provided.
- Support the growing ecosystem of applications powered by the Planetary Computer – grow our network of partners developing meaningful apps using the services of the Planetary Computer
- Continue to invest in and grow the open-source ecosystem of tools working with Earth Science Data – we will continue growing and supporting the open-source tools used for processing and analyzing the data on the Planetary Computer. The best mechanism for the needed collaboration between governments, academia, and industry to solve the world’s most complex problems is to ensure everything is based upon open standards and open-source software, facilitating more productive, frictionless, integrated relationships.
For more information about the Microsoft Planetary Computer, check this out – Home | Planetary Computer (microsoft.com)
For a more historical perspective on the Microsoft Planetary Computer, check this out – A Planetary Computer to Avert Environmental Disaster – Scientific American.
Light
Dark
0 comments