ATMOS / Forecast Dashboard
⚡ ATMOS FORECAST DASHBOARD: 0.05° Topography-Refined Monsoon Rainfall Field.
DEMO DATA — NOT OPERATIONAL CYCLE: 14 AUG 2024 • 00:00 UTC RESOLUTION: 0.05° (~5 km)
HIGH-RESOLUTION FORECAST DASHBOARD

Operational Forecast Grid & Interactive Telemetry

Explore regime-corrected rainfall fields, exceedance probabilities, and clickable grid-cell telemetry across forecast lead times.

RAINFALL FORECAST MAP REGIME: DEPRESSION / LOW CONFIDENCE: 68% FORECAST: +24H RESOLUTION: 0.05° (~5 km)
LEAD TIME:
RAINFALL LAYER:
MAP BASEMAP:
RAINFALL THRESHOLD:
RAINFALL (mm):
0 mm
1–10 mm
10–25 mm
25–50 mm
50–100 mm
100–200 mm
>200 mm
Scientific Disclaimer: ATMOS outputs are decision-support information and are not official weather warnings. DEMO DATA — Not an operational forecast
GRID CELL TELEMETRY

Mahabaleshwar (Ghats)

FORECAST IMPACT TELEMETRY
Raw NWP Baseline: 98.4 mm
AI Corrected (ATMOS): 218.6 mm
Orographic Correction ΔR: +120.2 mm
REGIME INTELLIGENCE SYNOPTIC
Active Synoptic State: Depression / Low
Classification Confidence: 68%
Shannon Entropy H(p): 1.18 bits (Moderate U)
RISK INTELLIGENCE EXCEEDANCE
P(Rain > 50 mm): 84% peak
High-Risk Region: Western Ghats & Vidarbha
Brier Skill Score (BSS): +0.64
SIH26080 SYSTEM PIPELINE FLOW
1. RAW NWP
➔
2. REGIME DETECTION
➔
3. REGIME-AWARE AI CORRECTION
➔
4. ATMOS FIELD
➔
5. RISK INTELLIGENCE

Post-Processing Model Architecture: RAW NWP ➔ ATMOS

Demonstrating the post-processing refinement pipeline that enhances existing NWP models without replacing them.

NWP ENHANCEMENT
INPUT BASELINE

Raw NWP Forecast

Coarse 0.25° dynamical ensemble mean model. Exhibits systematic dry bias along windward slopes and drizzle false alarms over arid zones.

NWP forecast ➔ ATMOS correction ➔ refined forecast
CORRECTED OUTPUT

ATMOS AI Post-Processed

Regime-aware soft gating + 0.05° terrain downscaling. Reduces continuous RMSE by 43.0% and raises extreme exceedance skill by +107%.