Spatial Informatics Group, LLC

Double the meetings, double the focus on #resilience! Today we hosted discussions on both wildfire and drought preparedness. As California continues to be impacted by #climatechange, we're working together to develop proactive solutions. From sharing resources to setting intentions, we're committed to building a more resilient future for all Californians. #WildfirePreparedness #DroughtResilience California Department of Water Resources CA Wildfire & Forest Resilience Task Force

🎉 The 2024 Geospatial Landscape is here! Over 1,600 companies, 972K working in the field, 50 distinct categories, and $29 billion in funding. (Get your copy using the link in the comments) ✅ One of my two big goals for the year was to create a comprehensive picture of the geospatial field. Our industry is so unique with many different focus areas and verticals, and I believed that it warranted a clear picture of how big this space is. 📄Over the years I have seen lots of different industry reports that are very high level and, in my opinion, only scratch the surface of the depth and data of the industry, trying to focus too much on capturing the 10-year growth rate of the industry, not the real current landscape or trends. On top of that they are inaccessible to most, costing thousands of dollars for a single, static, PDF. 😊 My goal was to make this data accessible as an easy-to-view dashboard for both industry professionals and job seekers with regular updates, news, and insights. I also wanted it to be a community with the ability to share ideas together. You can download the landscape using the link below and if you want, there are upgraded tiers for industry professionals and job seekers. Let me know your thoughts and I hope this is valuable to you! #gis #moderngis #geospatial #earthobservation #spatialanalytics #remotesensing #UAV #engineering #gisindustry #agtech #locationdata #realestate #cre #esg #climatechange #climaterisk #gps #forestry #hydrology

🌿 Advancing Wetland Conservation Through Innovative Technology: Insights from Our Latest Research I’m excited to announce that my co-authors Dr. Narcisa Pricope, Dr. Joanne Halls, MSc, Asami Minei, Dr. Cuixian Chen, and Dr. Yishi Wang and I recently published an article entitled “Precision Mapping of Coastal Wetlands: An Integrated RemoteSensing Approach Using UAS LiDAR and Multispectral Data”. The article focuses on how coastal wetlands play a pivotal role in global biodiversity and climate adaptation, offering crucial ecosystem services such as carbon storage and flood protection. Yet, these vital habitats face escalating threats from natural changes and human activities. This study, published in Journal of Remote Sensing, leveraged unmanned aerial systems (UAS) equipped with LiDAR and multispectral sensors to revolutionize wetland monitoring across diverse landscapes in North Carolina. This research was funded by the North Carolina Department of Transportation, contract number RP 2020–04 awarded to Dr. Narcisa Pricope (Lead PI) and Dr. Joanne Halls in the Department of Earth and Ocean Sciences at the University of North Carolina Wilmington. 🔍 Key Findings: Through high-resolution elevation data and detailed vegetation analysis, augmented by sophisticated machine learning algorithms, we achieved precise classifications of wetland types. Estuarine Intertidal Emergent wetlands demonstrated the highest accuracy, benefiting from clearer spectral signatures and consistent tidal influence adjustments. In contrast, palustrine forested and scrub-shrub wetlands presented challenges due to complex vegetation structures and variable inundation levels. 🌐 Impact and Implications: An integrated UAS-derived approach not only enhances wetland mapping accuracy but also offers scalability, efficiency, and cost-effectiveness. This methodology is pivotal for informing policy-making and conservation strategies aimed at bolstering coastal resilience and mitigating climate change impacts. By showcasing the transformative potential of advanced aerial technologies in ecological monitoring, our research underscores their critical role in sustaining natural habitats worldwide. 🚀 Looking Ahead: As we continue to refine and expand these methodologies, we aim to further empower conservation efforts and promote sustainable environmental practices. By bridging the gap between technology and ecology, we pave the way for more informed decision-making and proactive measures to safeguard our planet's precious ecosystems. Read the full article and explore the findings in depth: https://lnkd.in/evFs4kVk Please also check out this storymap initially designed by James Giddens and later finished by Salman Bashit here: https://lnkd.in/dg5jbcZ8 #WetlandConservation #EcologicalMonitoring #ClimateAdaptation #UAS #LiDAR #MachineLearning #EnvironmentalScience #ConservationTechnology

New preprint! PiRPLO: a single-run method for calibrating wildland fire simulations like statistical models https://lnkd.in/eX3fwg5p Since I started working for Spatial Informatics Group, LLC and Pyregence Consortium, I have been involved in various projects that do simulation-based prediction of wildfire risk (a.k.a. "Wildfire Monte-Carlo", a.k.a. "Burn-probability modeling"). Each time, the same questions arose: 1. How well are our predictions matching the historical record? 2. How do we achieve better goodness-of-fit? 3. How do we avoid giving too much importance to the vast majority of historical fires that are relatively small / inconsequential? 4. Which among the historical fires are we having a hard time reproducing? 5. How do we avoid wasting too much time and resources in an endless cycle of calibration simulations? 6. Is our calibration process overfitting the training data? This pre-print, which I co-authored with Chris Lautenberger, PhD, PE and David Saah, offers principled answers to all of these questions, by adapting methods from mathematical statistics. The key ingredients are useful for fire modeling, far beyond the scope of this paper: 1. using sample-reweighting techniques to avoid running calibration simulations over and over again; 2. embracing Poisson point process models and random measures as the mathematical foundation. I'm enthusiastic about the organization of research that this enables: simulation experts can publish datasets of simulated fires, after which modeling experts can fit models of fire occurrence, without having to run simulations themselves. This is a framework for model validation and calibration; the results are only as good as the models, data, and importance weights we feed to it (garbage in, garbage out). Since statistical learning is essentially the most inefficient form of learning and requires extreme parsimony in modeling assumptions, this framework should not replace modeling insights informed by physics or ecology; instead, it should complement them on aspects that physics and ecology have difficulty answering, such as the distribution of how wildfires start and stop. We haven't heard back from reviewers yet; in the meantime, feedback is welcome (do not spare us).

New preprint! PiRPLO: a single-run method for calibrating wildland fire simulations like statistical models https://lnkd.in/eX3fwg5p Since I started working for Spatial Informatics Group, LLC and Pyregence Consortium, I have been involved in various projects that do simulation-based prediction of wildfire risk (a.k.a. "Wildfire Monte-Carlo", a.k.a. "Burn-probability modeling"). Each time, the same questions arose: 1. How well are our predictions matching the historical record? 2. How do we achieve better goodness-of-fit? 3. How do we avoid giving too much importance to the vast majority of historical fires that are relatively small / inconsequential? 4. Which among the historical fires are we having a hard time reproducing? 5. How do we avoid wasting too much time and resources in an endless cycle of calibration simulations? 6. Is our calibration process overfitting the training data? This pre-print, which I co-authored with Chris Lautenberger, PhD, PE and David Saah, offers principled answers to all of these questions, by adapting methods from mathematical statistics. The key ingredients are useful for fire modeling, far beyond the scope of this paper: 1. using sample-reweighting techniques to avoid running calibration simulations over and over again; 2. embracing Poisson point process models and random measures as the mathematical foundation. I'm enthusiastic about the organization of research that this enables: simulation experts can publish datasets of simulated fires, after which modeling experts can fit models of fire occurrence, without having to run simulations themselves. This is a framework for model validation and calibration; the results are only as good as the models, data, and importance weights we feed to it (garbage in, garbage out). Since statistical learning is essentially the most inefficient form of learning and requires extreme parsimony in modeling assumptions, this framework should not replace modeling insights informed by physics or ecology; instead, it should complement them on aspects that physics and ecology have difficulty answering, such as the distribution of how wildfires start and stop. We haven't heard back from reviewers yet; in the meantime, feedback is welcome (do not spare us).

🌿 Advancing Wetland Conservation Through Innovative Technology: Insights from Our Latest Research I’m excited to announce that my co-authors Dr. Narcisa Pricope, Dr. Joanne Halls, MSc, Asami Minei, Dr. Cuixian Chen, and Dr. Yishi Wang and I recently published an article entitled “Precision Mapping of Coastal Wetlands: An Integrated RemoteSensing Approach Using UAS LiDAR and Multispectral Data”. The article focuses on how coastal wetlands play a pivotal role in global biodiversity and climate adaptation, offering crucial ecosystem services such as carbon storage and flood protection. Yet, these vital habitats face escalating threats from natural changes and human activities. This study, published in Journal of Remote Sensing, leveraged unmanned aerial systems (UAS) equipped with LiDAR and multispectral sensors to revolutionize wetland monitoring across diverse landscapes in North Carolina. This research was funded by the North Carolina Department of Transportation, contract number RP 2020–04 awarded to Dr. Narcisa Pricope (Lead PI) and Dr. Joanne Halls in the Department of Earth and Ocean Sciences at the University of North Carolina Wilmington. 🔍 Key Findings: Through high-resolution elevation data and detailed vegetation analysis, augmented by sophisticated machine learning algorithms, we achieved precise classifications of wetland types. Estuarine Intertidal Emergent wetlands demonstrated the highest accuracy, benefiting from clearer spectral signatures and consistent tidal influence adjustments. In contrast, palustrine forested and scrub-shrub wetlands presented challenges due to complex vegetation structures and variable inundation levels. 🌐 Impact and Implications: An integrated UAS-derived approach not only enhances wetland mapping accuracy but also offers scalability, efficiency, and cost-effectiveness. This methodology is pivotal for informing policy-making and conservation strategies aimed at bolstering coastal resilience and mitigating climate change impacts. By showcasing the transformative potential of advanced aerial technologies in ecological monitoring, our research underscores their critical role in sustaining natural habitats worldwide. 🚀 Looking Ahead: As we continue to refine and expand these methodologies, we aim to further empower conservation efforts and promote sustainable environmental practices. By bridging the gap between technology and ecology, we pave the way for more informed decision-making and proactive measures to safeguard our planet's precious ecosystems. Read the full article and explore the findings in depth: https://lnkd.in/evFs4kVk Please also check out this storymap initially designed by James Giddens and later finished by Salman Bashit here: https://lnkd.in/dg5jbcZ8 #WetlandConservation #EcologicalMonitoring #ClimateAdaptation #UAS #LiDAR #MachineLearning #EnvironmentalScience #ConservationTechnology

🎉 The 2024 Geospatial Landscape is here! Over 1,600 companies, 972K working in the field, 50 distinct categories, and $29 billion in funding. (Get your copy using the link in the comments) ✅ One of my two big goals for the year was to create a comprehensive picture of the geospatial field. Our industry is so unique with many different focus areas and verticals, and I believed that it warranted a clear picture of how big this space is. 📄Over the years I have seen lots of different industry reports that are very high level and, in my opinion, only scratch the surface of the depth and data of the industry, trying to focus too much on capturing the 10-year growth rate of the industry, not the real current landscape or trends. On top of that they are inaccessible to most, costing thousands of dollars for a single, static, PDF. 😊 My goal was to make this data accessible as an easy-to-view dashboard for both industry professionals and job seekers with regular updates, news, and insights. I also wanted it to be a community with the ability to share ideas together. You can download the landscape using the link below and if you want, there are upgraded tiers for industry professionals and job seekers. Let me know your thoughts and I hope this is valuable to you! #gis #moderngis #geospatial #earthobservation #spatialanalytics #remotesensing #UAV #engineering #gisindustry #agtech #locationdata #realestate #cre #esg #climatechange #climaterisk #gps #forestry #hydrology