Equitable Earth’s Post

IBM just release the Environmental Intelligence (EI) Public Preview yesterday. My team works on the Geospatial API component of this platform. IBM EI Preview blog post here : https://lnkd.in/d-BhFphJ This release is important as it is providing a set of combined APIs and capabilities for developers and data scientists alike to generate insights for business climate related resiliency. How is this useful to us all? Well, here is the more interesting question: How come that many (most?) business processes or workflows still do not include environmental data (climate, weather, biomass, water changes, etc) into account today and therefore take the risk to be disconnected from environmental reality and its impacts on its workings : It is as if businesses are noodling along inside the Plato cavern, in an ideal world, with no harmful nor beneficial changes happening every minutes/hours/days to their customers, employees, machines, buildings, assets... Doesn't feel right does it?  Climate change is on us: I believe strongly that we need to use environmental data to help us all (business, people and the earth itself) understand what is happening, be more resilient and responsive, and overall reduce our overall impact. We do have the data ready for this in our platform and I believe every business on earth should leverage some of our data in one way or another: We have to shift our business mindset quickly as major disruptions from the outside are rushing into the cavern! Timely, there is also an interesting blog post from my UK colleague @Andy Stanford-Clark here (https://lnkd.in/dgAky4aP) about the work we have been doing in IBM EI Geospatial platform related to Voluntary carbon markets and Above Ground Biomass (aka AGB) calculation and predictions. Directly related to my question above obviously! Some real Above Ground Biomass code samples about deforestation released here: https://lnkd.in/dYV8xZrd We will show more use cases on how to use our data in the future: This is just the beginning ! Have Fun and be useful my developers friends! Get an access to IBM EI Geospatial platform here: https://lnkd.in/dfNeCqmG

Artificial intelligence has a significant, negative environmental impact. Experts propose energy-efficient algorithms, renewable energy, and sustainable hardware practices to mitigate this. However, it remains an ongoing issue.

🌍 Modeling Above Ground Biomass with Machine Learning and Remote Sensing I'm excited to share a recent workflow I developed for my MBA in Data Science capstone project ! This end-to-end geospatial data science process integrates optical and radar satellite data to predict and map above ground biomass, crucial for understanding forest carbon stocks and supporting sustainable land management. For more details, check out my post on Medium where I dive deeper into the methodology and results! You can follow the workflow and replicate the analysis using the Python scripts available on my GitHub repository. Here’s a snapshot of how I approached it: 📡 Data Collection & Feature Engineering: collected remote sensing data (optical and radar) to get key forest characteristics such as vegetation indexes, biophysical parameters and texture measures. 🛠 Data Pre-processing: pre-processed and stacked datasets by setting a default projection and using bi-linear resampling to smooth spatial resolution reduction. 🤖 Cross-Validation & Feature Scaling: applied techniques like K-fold cross-validation, feature scaling and feature selection with Mean Decrease in Impurity (MDI) and Mean Decrease in Accuracy (MDA). 💻 Model Development: built three predictive models, Random Forest, XGBoost, and Deep Neural Network and tuned their hyper-parameters through Bayesian Optimization and Random Search. 📊 Model Evaluation: evaluated these models by assessing accuracy with metrics such as RMSE, MAE, R², adjusted R². Additionally, performed a Mann-Whitney U test to check if the metrics differences are statistically significant. 🔍 Feature Interpretation: used SHAP (SHapley Additive exPlanations) to interpret features behavior for each model. 🌳 Field Data Validation: validated the models performance comparing predictions against field data from Brazil’s National Forest Inventory (NFI) to ensure reliability. 🌐 Mapping & Prediction: finally, I mapped the predicted biomass using Google Earth Engine, delivering valuable insights for forest monitoring. Feel free to reach out if you have questions or want to discuss the project further! #MachineLearning #RemoteSensing #GeospatialData #ForestMonitoring #Sustainability #DataScience #Biomass #Carbon #RandomForest #XGBoost #DeepNeuralNetwork

Digital Earth Solutions (DES) is a spin-off of the Spanish National Research Council (CSIC), founded on over 20 years of research into ocean dynamics and modelling. DES’s flagship product, SPOT, makes ocean modelling easy, providing extremely accurate simulations in minutes. The intuitive user interface offers real-time data integration and the ability to run both forecasts and hindcasts. Being cloud-based, SPOT is accessible 24/7 from anywhere in the world. With specific modules developed for; Oil Spill, Search and Rescue, Plastic Pollution and Algae, SPOT is able to serve a diverse range of customer needs. Automated reports deliver key information clearly, allowing decisions to be made at speed. Read the full article in Spill Alert Issue 27: https://buff.ly/3D2pU2e #ocean #marine #ecosystem

Satellite data and machine learning (ML) are opening up new methods to gather comprehensive, independent emissions data. In a study with Climate TRACE partners, TransitionZero, Global Energy Monitor, and collaborators from Pixel Scientia Labs and Georgetown University, we published an open-access paper in MDPI on remote sensing of emissions from power plants. This model estimates emissions from combustion power plants by looking at their water vapor plumes from satellites and training ML models to use those data to estimate power generation and carbon dioxide emissions. The plants used to validate this methodology account for 32% of global power plant CO2 emissions on the Climate TRACE dataset from 2015-2022. Open access to independent emissions data is a critical step for enabling emissions reductions from global power plants. We hope that by sharing this methodology scientists, policymakers, and corporations will be better able to assess and manage those carbon emissions. Read the paper to learn more about this exciting development in emissions monitoring. https://lnkd.in/e-PAgBY6 #Science #MDPI #PowerPlantEmissions #EmissionsData #ML 

AI and the Path to Sustainable Futures: Insights from the Global Conference - <p>Discover AI's role in sustainability at the Global Conference, exploring energy, agriculture, and urban planning innovations.</p> Read More: https://lnkd.in/gPsxa4C8

Revolutionising Sustainable Energy with Boncalor’s STG Technology AI Meets Sustainability: A Game-Changing Approach to Energy The rapid rise of AI and computational needs is revolutionizing industries—but it’s also driving a massive surge in energy demand. According to Earth.org: "By 2040, emissions from the Information and Communications Technology (ICT) industry are expected to reach 14% of global emissions, with the majority stemming from ICT infrastructure, particularly data centers and communication networks. These figures highlight the urgent need to address AI’s carbon footprint and its role in environmental deterioration." Read more here: Earth.org: The Real Environmental Impact of AI The environmental costs of advancing AI are undeniable unless we innovate—and that’s exactly where Boncalor AB comes in. What is Boncalor’s Solution? Solid-State Thermal Generators (STGs) are reshaping how energy is produced and consumed, merging AI-powered compute capacity with sustainable heating and cooling. Here’s how: Zero-Carbon Compute & Thermal Capacity: Boncalor’s patented STG technology offers energy solutions with zero carbon footprints. Energy Efficiency: By optimizing power usage and storing low-cost energy, STGs cut energy costs by up to 30%—benefitting industries, data centers, and residential users. Grid Stabilization: Their innovative mesh network not only stabilizes regional grids but also minimizes environmental losses. Why Now? With studies warning of AI’s environmental impact, sustainability must be at the core of innovation. Boncalor’s STG technology enables industries and governments to reduce emissions while meeting the surging demand for computational power and green energy solutions. Boncalor aims to deploy 160 MW of STGs globally by 2029, building a smarter, greener energy infrastructure. Their advanced AI optimization ensures energy is never wasted, keeping carbon footprints at zero. Together, we can tackle the environmental challenges posed by AI and energy use. Let’s build a future that’s smarter, greener, and more sustainable.   For insights or investment opportunities, reach out here on LinkedIn or at Ruben@capilon.se

🔍 Using Remote Sensing to Mitigate Methane Emissions from Africa’s Waste Sector 🔍 🚀 Clean Air Task Force (CATF) is leveraging the power of remote sensing technology to tackle methane emissions from Africa’s waste sector. 🌱 📊 Data Acquisition & Analysis: Utilizing GHGSat’s cutting-edge satellite constellation, the project captured high-resolution spectral imagery to detect and quantify methane hotspots. The satellites are equipped with sensors that operate in the short-wave infrared range, capable of identifying methane’s unique spectral signature even at low concentrations. 📡 Satellite Technology: The GHGSat satellites employ a technique known as Differential Absorption Spectroscopy. This method involves measuring the absorption of sunlight by methane molecules at specific wavelengths as the light passes through the Earth’s atmosphere. The data collected provides a spectral radiance measurement, which is then processed using advanced algorithms to isolate the methane signal and calculate emission rates with an accuracy of up to 10kg/hr. 🌍 Geospatial Mapping: The project generated georeferenced methane concentration maps, overlaying them onto high-resolution base maps for precise localization of emission sources. This geospatial analysis is crucial for identifying specific waste management sites contributing disproportionately to the overall emissions. 🔎 Algorithmic Processing: The data underwent rigorous processing using Machine Learning algorithms to enhance detection sensitivity and reduce false positives. These algorithms were trained on a vast dataset of known methane emissions, allowing them to distinguish between natural and anthropogenic sources effectively. 📈 Emission Estimation: The emission estimates were derived using the Integrated Mass Enhancement method, which calculates the total enhanced methane column by integrating the concentration over the plume area. This approach provides a more accurate representation of the total emissions compared to point measurements. 🤖 Automation & Scalability: One of the key technical achievements of this initiative is the development of an automated processing pipeline that can handle large volumes of satellite data, enabling near-real-time monitoring of methane emissions across the continent. 🔗 Integration with Ground-Based Observations: To complement the satellite data, ground-based sensors and mobile monitoring units were deployed to validate the remote sensing findings. This integrated approach ensures the robustness of the data and facilitates targeted mitigation strategies. #RemoteSensing #SpectralImaging #MethaneDetection #EnvironmentalEngineering #GeospatialAnalysis #MachineLearning #ClimateTech #Sustainability #Innovation #GHGSat #CATF https://lnkd.in/gVmz5tFY

Sustainability-in-Tech : World’s First Bio-Circular Data Centre French data centre company, Data4, says its new project will create a world-first way of reusing data centre heat and captured CO2 to grow algae which can then be used to power other data centres and create bioproducts. Why? The R&D project, involving Data4 working with the University of Paris-Saclay, is an attempt to tackle […] The post Sustainability-in-Tech : World’s First Bio-Circular Data Centre appeared first on Enhance Systems.

https://lnkd.in/ehGBiiBz Kudos to Meta and Google for both looking at liquid sorbent based DAC.... We at Deep Green are too... some exciting plans for an industrial heat R&D campus coming soon.... In the absence of any stable or coherent government support, the data centre industry (with free and subsidized heat) has the opportunity to accelerate the development of important new industries like DAC (as well as other sectors like urban agriculture - synthetic protein and fat production, aquaculture etc). In so doing (as well as heating pools of course...) the data centre industry can position itself as a key partner in the electrification of heat and the wider carbon economy...