Asia Green Corridors: A Zero-Backend Browser Observatory for Port and Shipping Decarbonisation Decision Support in Asia

Green shipping corridors are increasingly central to maritime decarbonisation in Asia, yet quantitative evidence to inform corridor choices remains limited and often requires paid data services. This paper introduces Asia Green Corridors, a zero-backend browser observatory that integrates previously separate public datasets: satellite-AIS port activity for 787 ports in PortWatch's Asia-Pacific classification (including 654 with Asian country codes) and its documented disruption log; satellite-estimated emissions from 2,885 Asian international-shipping sources (Climate TRACE); 534 western North Pacific cyclone tracks since 2005 (NOAA IBTrACS); and a live sea-state feed (Open-Meteo). The 2025 emissions layer totals 266.7 Mt of CO₂-equivalent on a 100-year global-warming-potential basis. The observatory provides one-click port screening, compares six Middle East–Asia corridor hubs by vessel calls and satellite-estimated emissions, and monitors three strategic chokepoints daily. A transparent scenario sandbox and checkpoint solver estimate the green-fuel shares implied by the IMO Revised Strategy's emissions checkpoints under an illustrative 2025-normalised pathway. At the model's default settings, meeting the indicative 2030 checkpoints on time requires deployed green-fuel shares of approximately 18%–29% of energy. These exceed the IMO's separate 5%–10% fuel-uptake ambition, but differing indicators, accounting assumptions and the model's restricted mitigation levers preclude interpreting the comparison as a direct policy shortfall. Anomaly detectors are evaluated against 83 eligible documented disruption events. At an operating point chosen by a transparent, predefined rule, they detected 59.0% of events in sample, generating 1.41 alerts per monitored port per year (2.04 per actively graded port); this event-level sensitivity is not a measure of comprehensive detection accuracy. The full sensitivity grid, event-type results and limitations are disclosed. Implemented as one HTML file and five static data bundles, the application requires no backend, account or API key and explicitly identifies unavailable feeds. It is openly released as an empirical companion to the author's corridor working paper.