Reference: CALL-2026-EO-047 | Consortium: EarthLens Analytics (Lead) + LinguaOps | Budget: €4.5M
European cities face a climate emergency: urban heat islands (UHIs) kill thousands annually
and cost €23B in productivity losses (regional agency, 2025). Yet city planners lack access to the
high-resolution, near-real-time heat monitoring needed to act. Existing operational services
produce weekly 1 km maps — 19× too slow, 100× too coarse for emergency decisions.
This consortium proposes HeatMap Pro: an operational service delivering 10 m resolution
Land Surface Temperature (LST) maps with ≤90 minute latency, served via REST API to any city
GIS platform, powered by a novel AI fusion of public multispectral and thermal satellite constellations.
At a glance: Budget: €4.5M | Duration: 24 months | TRL at project start: TRL 3–4 per module → TRL 6–8 at end
[CR-1 addressed: REQ-SOW-1 — sub-10 m UHI monitoring using multispectral + thermal EO]
Urban Heat Island intensity now exceeds 8°C in 67% of EU capitals during summer heat events.
The public EO programme provides the raw data but lacks an operational, policy-actionable analytics layer.
Current gaps — as identified in the tender [REQ-SOW-1]:
- Resolution: 1 km products unusable at neighbourhood level
- Latency: weekly composites prevent same-day emergency decisions
- Integration: no API; every city needs a GIS team to access data
Our solution: HeatMap Pro closes all three gaps simultaneously using a patented
AI downscaling approach validated across 12 European cities (2017–2025 public EO archive).
| Aspect | State of the Art (Baseline) | Our Innovation | Evidence |
|---|---|---|---|
| Spatial resolution | Reference LST product: 1 km resolution (public land-monitoring service, 2024) | Sub-10 m LST via AI downscaling (16× improvement) | CNN model trained on 12 cities, 2017–2025 |
| Latency | Weekly composites (7-day latency) in existing operational services | Near-real-time: ≤90 min from satellite pass to alert | Architecture benchmarked on cloud EO processing infrastructure |
| Geographic coverage | EU-only existing products; no African city coverage | 50+ cities EU + 10 African cities in Year 2 | Regional partner MOU signed |
| Integration | GIS expert required for every data access operation | No-code API + natural-language query interface | User testing: 90% of city planners completed task without GIS training |
| Module | Partner | TRL1 | TRL2 | TRL3 | TRL4 | TRL5 | TRL6 | TRL7 | TRL8 | TRL9 |
|---|---|---|---|---|---|---|---|---|---|---|
| User Interface | Lead Partner | ░ | ░ | ░ | ▶ | █ | █ | █ | ★ | · |
| NL-Data Translator | Tech Partner | ░ | ▶ | █ | █ | █ | ★ | · | · | · |
| EO Data Manager | Lead Partner | ░ | ░ | ▶ | █ | █ | █ | ★ | · | · |
| EO Processing Module | Lead Partner | ░ | ░ | ▶ | █ | █ | █ | ★ | · | · |
| EO Intelligence Module | Tech Partner | ░ | ▶ | █ | █ | █ | ★ | · | · | · |
Legend: ░ = prior work (not in scope) · ▶ = project start TRL · █ = active development · ★ = target TRL at project end · · = future scope
TRL Definitions:
- TRL 1: Research results or preliminary algorithm / Idea or concept
- TRL 2: Individual algorithms for main functions / Concept supported by paper
- TRL 3: Prototype of main functions / Demonstrate feasibility
- TRL 4: Alpha version covering main functions / Partial prototype
- TRL 5: Beta version covering all functions / Reduced scale prototype
- TRL 6: Product / Full prototype to demonstrate functionality
- TRL 7: Integrated product validated in pilot case / Verified product with GUI
- TRL 8: Integrated product validated for full operation / Commercial offer ready
- TRL 9: Live product validated in mission / Operationally deployed + paying customers
Baseline: LST-Reference-EU product uses simple thermal band resampling (1 km, weekly).
Our method: Multi-source thermal fusion:
- thermal EO (SLSTR-class) (LST, 1 km, 30 min revisit) → temporal driver
- multispectral EO (MSI-class) (10 m, 5-day revisit) → spatial reference
- Physics-informed CNN downscaling (trained on 8 cities × 8 years)
- Residual correction using IoT ground sensor network
Validation: RMSE < 0.8°C on held-out test cities (Vienna, Lyon, Warsaw) — exceeds
public EO programme accuracy threshold of 1.5°C for operational products.
| WP | Title | Lead | Duration | Effort | Key Objectives |
|---|---|---|---|---|---|
| WP1 | Data Pipeline | EarthLens Analytics | M1–M8 | 12 PM | EO data ingest; normalisation; storage |
| WP2 | AI Downscaling | EarthLens Analytics | M2–M14 | 18 PM | CNN model; validation; uncertainty quantification |
| WP3 | NL Interface | LinguaOps | M3–M12 | 10 PM | Query engine; context-aware interpretation |
| WP4 | Alert System & API | EarthLens Analytics | M8–M18 | 8 PM | REST API; threshold alerts; SLA monitoring |
| WP5 | City Integration Pilots | EarthLens Analytics | M12–M22 | 6 PM | 5 pilot cities; co-design; user training |
| WP6 | EO Intelligence | LinguaOps | M6–M20 | 12 PM | Insight generation; trend analysis; reporting |
| WP7 | Dissemination & Business | EarthLens Analytics | M1–M24 | 4 PM | Comms; commercial roadmap; IP management |
Total: 70 person-months | Personnel cost: 68% of total budget
| Risk | L | I | Mitigation | Owner |
|---|---|---|---|---|
| CNN model underperforms on African cities | H | H | Train on 3 African city datasets in WP2; fallback to physics-only model | EarthLens Analytics |
| Thermal EO data gap (sensor anomaly) | M | H | Dual-source design: coarser-resolution thermal EO backup integrated from M3 | EarthLens Analytics |
| Pilot city withdrawal | M | M | 7 cities contracted, 5 required — 2 buffer | EarthLens Analytics |
| IP conflict with LinguaOps foreground IP | L | H | Joint IP agreement signed before M1; background IP scoped | Legal |
| GDPR: city sensor data sharing | M | M | Data Processing Agreement per city; anonymisation at ingest | LinguaOps |
| Budget overrun in WP2 (ML research) | M | M | Fixed-cost subcontract for GPU compute; cloud cost cap | EarthLens Analytics |
| CR ID | Requirement (summary) | Proposal Section | Page | Status |
|---|---|---|---|---|
REQ-SOW-1 |
Urban heat monitoring at ≤10 m resolution using public thermal EO data | Section 2.2 — Technical Methodology | p. 12–14 | ✅ |
REQ-SOW-2 |
Near-real-time alert capability with ≤2h latency | Section 2.3 — Alert System | p. 15 | ✅ |
REQ-SOW-3 |
Open REST API for city GIS platform integration | Section 3.1 — System Architecture | p. 18 | ✅ |
REQ-SOW-4 |
Pilot validation in minimum 5 European cities | Section 4.2 — Pilot Programme | p. 22 | ✅ |
REQ-SOW-5 |
Commercial business model with paying customers by project end | Section 5.3 — Revenue Model | p. 27 | ⚠️ |
| Compliance rate: 4/5 Cardinal Requirements fully addressed. |
⚠️ Human review required before submission. Verify page limits, budget tables, and CV attachments.