Expertise
Operational systems that keep observations, satellite data and model output in step, and turn them into decisions.
A useful environmental decision-support system stays current: it ingests observations as they arrive, fuses them with satellite products and model output, checks the result against thresholds that matter to someone, and lets users ask what happens under different scenarios. Most organisations already hold several of these parts. What they lack is the plumbing that keeps the parts in sync and the interface that makes them useful.
Suman has spent more than seven years building those parts for environmental agencies, regional centres and development programmes in South Asia: automated ingestion of ground observations, satellite products and model outputs into operational air-quality dashboards at ICIMOD; WRF-Chem and HYSPLIT forecast workflows on HPC; a temperature sensitivity alert system and agricultural weather decision support at RIMES for the World Bank CARE project; and, currently, a geospatial system for an SDC-funded parametric insurance initiative. His decision-support work starts from these proven building blocks and adds only what a specific decision needs.
Scheduled and event-driven pipelines that pull ground-station observations, satellite products and model runs into one spatial database, with QA/QC checks and gap reporting so operators know when an input is stale.
Bringing point observations, gridded satellite retrievals and model fields (for example WRF-Chem output) onto common grids and time steps, so they can be compared, bias-checked and shown side by side.
Translating agreed thresholds or index definitions into automated checks, then into alerts, maps and bulletins that reach the right users. Useful for early warning and for index-based (parametric) risk products.
Running or reusing model scenarios (baseline versus intervention, different emission or weather conditions) and presenting the differences in a form that programme staff can interpret.
Web dashboards and map applications designed around the questions users actually ask, built on standard services so data can also be reused in GIS desktops and other systems.
Linux server operation, deployment, documentation and training so the system keeps running after the project ends, and so in-house staff can maintain it.
A continuously updated view of air quality for a city, country or region that combines station observations, satellite data and model forecasts, with simple what-if scenarios.
4-9 months, delivered in phases (a working data backbone and dashboard first, scenarios later) · fixed-scope project
A system that tracks weather, climate and satellite indicators against agreed thresholds or index definitions, and turns exceedances into alerts and decision-ready maps.
3-8 months depending on the number of indicators and countries · fixed-scope project
A short discovery engagement that audits your data, systems and decisions, and gives you a realistic architecture and phased roadmap before you commit to a build.
2-5 weeks · advisory
RIMES (Regional Integrated Multi-hazard Early Warning System) · 2024–2025
Built the Temperature Sensitivity Alert System (TEMPs), integrating model, satellite and ground data, and supported ADVISE, an agricultural weather decision-support system, for the World Bank CARE project in South Asia.
Countries: Nepal, Pakistan
Seed Innovations Pvt. Ltd. · 2025–present
Ongoing development of a geospatial system for an SDC-funded parametric insurance initiative: data workflows, analytical components and user-facing geospatial functionality.
Countries: Nepal
Where
Loading map…
Tell me what you're working on. A 20-minute call is free and there's no obligation.