IBM IBM is proud to unveil a new platform, IBM Environmental Intelligence, designed to empower data scientists and application developers with environmental data sets. This platform, currently under public preview, features curated global environmental data from public and commercial data sets, out-of-the box models, and connectors, that align with specific events and generate alerts, supporting data integration into mission critical software and daily operations.
Organizations need meaningful data to make informed and timely decisions that support effective responses to the current environmental challenges. Accessing geospatial, GHG emissions, weather, and climate impact data sets, for example, can be crucial for businesses to build scalable and climate resilient solutions. However, getting access to the specific data sets and insights is challenging.
IBM is proud to unveil a new platform, IBM Environmental Intelligence, designed to empower data scientists and application developers with environmental data sets. This platform, currently under public preview, features curated global environmental data from public and commercial data sets, out-of-the box models, and connectors, that align with specific events and generate alerts, supporting data integration into mission critical software and daily operations. Environmental Intelligence leverages cutting-edge technologies and expertise to help customers mitigate potential financial impact from operational losses and liabilities caused by climate disruption by proactively managing their environmental risks and uncertainties across their business activities.
Through this new technology, IBM seeks to empower data professionals to reach competitive advantage. Environmental Intelligence can help them accomplish that through the following main features.
Geospatial APIs
IBM Environmental Intelligence provides a wide range of Geospatial APIs to help application developers and data scientists gather what is happening on Earth's surface and then use the data to predict various outcomes and design a proactive course of action. There is a wealth of public satellite data readily accessible to obtain weather and climate related insights. But to use this information, data scientists need to invest considerable time and effort in processing and manipulating this data to make it readily usable for further applications. This is where IBM Environmental Intelligence comes in, simplifying this procedure by processing, normalizing, and organizing the public data sets into geospatial layers for scientific analysis and visualization. By leveraging APIs or Python Geospatial SDK, we can work with data sets like: high-resolution imagery (ESA Sentinel 2 and NASA Landsat 8), Global weather (ERA5), PRISM daily US weather, Wildfire risk potential, Soil Properties USA, and JAXA ALOS 3D Global. This means that users can more easily query the data and swiftly gain valuable insights. Explore the list of APIs to begin your journey today!
Exploring Geospatial AI Foundation Model
There is a vast amount of remote sensing data, measured in petabytes, that can be very complex to handle. To use this data effectively for specific environmental objectives data scientists need to preprocess data, often employing artificial intelligence (AI) techniques for specific tasks, train AI algorithms, and develop deep learning models.
To address this challenge, IBM has introduced the pretrained Geospatial AI Foundation Model. Powered by Harmonized Landsat Sentinel-2 satellite data, this model offers a more consistent, comprehensive view of Earth, in 2-3 intervals. This foundation model's uniqueness lies in its adaptability, accuracy, self-supervised learning abilities and its reliance on a blend of high-resolution satellite imagery and LiDAR for downstream predictive insights.
Leveraging this kind of Geospatial AI model can help in mainly two ways: by reducing the data labeling burden and by providing a consistent base from which to build targeted solutions. Instead of spending valuable time labeling large amounts of data and building bespoke models, data experts and developers can now focus on fine-tuning AI models for a suite of applications like detecting floods and estimating tree canopy height.
Geospatial Foundation Model-based inferences
The Above Ground Biomass model uses the pretrained AI foundation model to help track carbon in invested biomass and respond to organizations that use carbon credits to support sustainability goals. For example, a researcher studying the impact of deforestation on carbon sequestration might use the API to calculate above ground biomass in various forested areas, then compare these values to historical data to assess changes over time. IBM is providing APIs for Quellia, covering historic and future Above Ground Biomass baselines including species-specific data. Environmental Intelligence accelerates land plot analysis using geospatial imagery and AI assuring transparency with certified models. Quellia aims to establish a secure presence in the voluntary carbon market, leveraging digital tech for reliable certification and monitoring. Teaming up with IBM, Quellia is creating an all-encompassing solution to support stakeholders, promoting transparency and traceability.
Another example of how this technology can be leveraged is by helping organizations address regulatory requirements. Europe Deforestation Regulations (EUDR), expected to come into effect by December 2024, would require traders and operators of certain commodities (palm oil, soy, beef, and wood products) to ensure that these are sourced from deforestation-free land and have not caused forest degradation. The Above Ground Biomass model is designed to assist companies trading with EUDR-targeted commodities to overcome challenges associated with compliance demonstration such as those related to inconsistent geospatial data and the inability to get forestation data over time. Companies can also leverage the biomass estimation capability to determine if any deforestation or forest degradation has occurred to support their compliance requirements.
IBM is also working with Yara International to provide geospatial data that aspires to reach 100 million hectares of farmland around the world. This area is equivalent to twice the size of Spain, or close to 7% of all arable land worldwide. Having data of this farmland, which includes millions of smallholder farms, can help provide insights on crop yield which would help to increase food production on existing farmland and avoid deforestation.
Historical on demand APIs
The Historical on demand APIs available in Environmental Intelligence provide the weather historical context that are needed to extrapolate relationships and make correlations with past business and operational results which helps scientists predict future business needs and outcomes. It provides multiple data sources for historical weather and one data source for weather forecasting. The payloads are customizable and leveraging this platform, scientists can train the model for predictive analysis.
For example, Bayer is collaborating with IBM to use historical weather data to develop readymade capabilities, AgPowered Services, for agri-food companies. Industry use cases include portfolio risk assessment in crop insurance underwriting, more accurate forecasting of crop production changes year over year, and a critical data set to train agronomic models.
GHG Emissions API
Lastly, Environmental Intelligence can help organizations calculate internal carbon emissions and estimate supply chain emissions by using the GHG Emissions APIs provided within the Carbon Performance Engine. In accordance with the GHG protocol, the APIs are a collection of six endpoints that can be used against a structured data set to return scope 1, 2, and 3 emission details and calculations. For example, an engineer who works for a company specializing in natural gas distribution might use this API in support of his efforts to ensure safe and efficient delivery of natural gas to customers while minimizing methane leaks.
A step towards sustainability
Climate change is a great threat to our planet, with major implications for ecosystems, economies, and societies. According to the annual report from NOAA s Global Monitoring Lab' the global average atmospheric levels of carbon dioxide was 419.3parts per million, increasing 2.8 ppm since 2022. This is the 12th year in a row where there is an increase of over 2 ppm. IBM Environmental Intelligence gives data scientists and engineers the environmental data sets and advanced analytics they need to build solutions to track the effect of climate and weather change across industries and keep building more resilient business practices.
Attachments
- Original Link
- Permalink
Disclaimer
IBM - International Business Machines Corporation published this content on 21 May 2024 and is solely responsible for the information contained therein. Distributed by Public, unedited and unaltered, on 21 May 2024 04:02:04 UTC.
