Post by Gijs van den Dool

🌍 𝗦𝗽𝗮𝘁𝗶𝗮𝗹 𝗗𝗮𝘁𝗮 𝗦𝗰𝗶𝗲𝗻𝗰𝗲 • 𝗘𝗰𝗼𝗹𝗼𝗴𝗶𝗰𝗮𝗹 𝗠𝗼𝗱𝗲𝗹𝗹𝗶𝗻𝗴 • 𝗥𝗲𝗺𝗼𝘁𝗲 𝗦𝗲𝗻𝘀𝗶𝗻𝗴 𝗦𝘁𝗿𝗮𝘁𝗲𝗴𝘆: 𝘜𝘯𝘭𝘰𝘤𝘬𝘪𝘯𝘨 𝘵𝘩𝘦 𝘱𝘰𝘸𝘦𝘳 𝘰𝘧 𝘤𝘭𝘰𝘶𝘥 𝘤𝘰𝘮𝘱𝘶𝘵𝘪𝘯𝘨 𝘧𝘰𝘳 𝘣𝘪𝘰𝘥𝘪𝘷𝘦𝘳𝘴𝘪𝘵𝘺 𝘢𝘯𝘥 𝘩𝘢𝘣𝘪𝘵𝘢𝘵 𝘤𝘰𝘯𝘴𝘦𝘳𝘷𝘢𝘵𝘪𝘰𝘯! 🛰️🌿 Today, I just wrapped up a 𝗡𝗔𝗦𝗔 𝗔𝗥𝗦𝗘𝗧 training on "Species Distribution Modelling with Google Earth Engine" (July 2026 - link in the comments). With growing biodiversity challenges, being able to access and analyse large collections of satellite imagery like Landsat, MODIS, and Sentinel) entirely in the cloud fundamentally changes how we can assess a species' likely habitat.  In the past, downloading large datasets and the need for powerful computers made advanced spatial analysis difficult. Now, using Google Earth Engine (GEE) removes many of these barriers and lets us create reproducible, machine-learning-based habitat maps much more quickly. The modules provided immediate cross-disciplinary inspiration, but a few points were especially useful for my own work: 🔹 𝗔𝗯𝘀𝗲𝗻𝗰𝗲 𝗠𝗼𝗱𝗲𝗹𝗹𝗶𝗻𝗴 𝗙𝗿𝗮𝗺𝗲𝘄𝗼𝗿𝗸𝘀: Learning how to handle missing data by comparing true absences, pseudo-absences, and background data for each algorithm. Getting this right is essential for training strong binary classifiers like Random Forests. 🔹 𝗙𝗶𝘁𝘁𝗶𝗻𝗴-𝘁𝗵𝗲-𝗡𝗶𝗰𝗵𝗲 𝗧𝗵𝗲𝗼𝗿𝘆: Mapping geographic distributions by relating presence points to environmental drivers. It gave me new ideas about how these concepts apply to Urban Ecology and how changed landscapes affect which species can live there. 🔹 𝗦𝗽𝗮𝘁𝗶𝗮𝗹 𝗕𝗹𝗼𝗰𝗸 𝗖𝗿𝗼𝘀𝘀-𝘃𝗮𝗹𝗶𝗱𝗮𝘁𝗶𝗼𝗻: This validation step was especially helpful for me; it confirmed that I am using the right method in my current canopy height project. Using spatial partitioning rather than standard random cross-validation is absolutely critical to avoid overestimating how well a model predicts across different areas. Whether examining macro-scale wildlife translocations, localised canopy structure, or the spread of invasive species, cross-disciplinary applications of these remote sensing techniques open up many new avenues for impactful ecological and landscape mapping.  Huge thanks to the NASA - National Aeronautics and Space Administration 𝗔𝗥𝗦𝗘𝗧 Team and the guest speakers for creating such a practical, code-focused workflow #EarthObservation #GoogleEarth Engine #SpeciesDistributionModeling #SDM #RemoteSensing #GIS #DataScience #UrbanEcology #ConservationTech #NASA #ARSET 🔸🔸🔸 I'm Gijs, an Independent Earth Observation & Geospatial Technical Advisor. I support climate, environmental, nature, and biodiversity projects that need the right technical expertise applied. 🔸🔸🔸