DecodeX
Home/Problem Statements/SIH26066
All 226 Statements
SIH26066SoftwareDisaster Management
₹1,50,000 Prize (₹75k Joint)
•
Deadline: 20 September 2026
•
Submitted Ideas: 0/500

OceanEmbed - Satellite Embedding-Based Deep Learning Framework for Reconstruction of Subsurface Ocean Temperature from Surface Satellite Observations.

Ministry / Organization: Ministry of Earth Sciences (MoES)
Department: Indian National Centre for Ocean Information Services (INCOIS) Ocean Valley
SIH Official Portal
Technical Complexity
8/10
Dataset Accessibility
7/10
36h Hackathon Viability
9/10
Step 0 • Zero-Jargon Layman Summary

“We are building a smart system that uses surface-level satellite data to 'see' through the ocean surface. It predicts how warm the water is at various depths, helping scientists track heatwaves and marine life without needing expensive, hard-to-reach underwater sensors everywhere.”

The DecodeX PS Decoder™

Deconstructed Problem Blueprint

Step 1

WHAT TO BUILD?

A deep learning-based predictive software framework that reconstructs 3D subsurface ocean temperature profiles from 2D surface satellite observations.

Step 2

WHO IS IT FOR?

Primary: INCOIS (Ministry of Earth Sciences). Secondary: Marine researchers, climate scientists, fisheries management, and disaster management authorities.

Step 3

WHY DOES IT EXIST?

Direct underwater sensors (ARGO floats) are sparse and expensive. We currently lack a continuous, high-resolution 3D view of the ocean, which is critical for predicting marine heatwaves and climate shifts.

Step 4

INPUT → PROCESS → OUTPUT ARCHITECTURE PIPELINE

01 • Inputs & Telemetry

Satellite data: SST, SSS, SSH/SLA, Surface Currents (U,V), Surface Winds (U,V) at 0.25-degree resolution.

02 • Processing & AI Core

Data harmonization, spatial/temporal regridding, latent feature extraction via CNNs/Transformers, and 3D temperature profile reconstruction.

03 • Actionable Deliverables

A 3D temperature grid (15 standard depth levels) at 0.25-degree resolution for the North Indian Ocean.

Step 5

CONSTRAINTS

  • Must maintain 0.25-degree spatial resolution.
  • Must cover the specific North Indian Ocean domain (5N-30N, 45E-105E).
  • Must handle multi-source data harmonization (different formats/grids).
  • Must validate against independent ARGO float data.
Step 6

JURY METRICS

  • Root Mean Square Error (RMSE) against ARGO ground truth.
  • Correlation coefficient (R-squared) between predicted and observed profiles.
  • Bias analysis at different depth levels.
  • Inference speed for real-time monitoring.
Step 7

CORE BOTTLENECK

The 'Data Gap' problem: Aligning heterogeneous satellite datasets with different temporal/spatial footprints and ensuring the model learns physical vertical mixing dynamics rather than just statistical correlations.

Why You Should Choose This PS (Pros)

  • High national priority for MoES/INCOIS.
  • Clear, well-defined scientific objective.
  • Strong potential for research publication and real-world deployment.

Challenges & Risks to Watch (Cons)

  • High computational requirement for training.
  • Steep learning curve for oceanographic data formats (NetCDF/HDF5).
  • Requires rigorous scientific validation beyond standard software testing.
Hackathon Blueprint

Recommended Tech Stack & Grand Finale Features

Frontend & Mobile
React.jsDeck.gl for geospatial visualizationMantine UI
Backend & APIs
FastAPIXarray/Dask for large-scale data processing
AI / ML & Models
PyTorchVision Transformers (ViT)ConvLSTM for spatio-temporal dynamics
Database & Cloud
PostgreSQL with PostGISZarr/NetCDF for climate data storage

Killer Features to Impress the SIH Jury

Interactive 3D depth-slice explorer using Deck.gl.
Automated 'Marine Heatwave' alert system based on reconstructed data.
Explainable AI (XAI) module showing which surface variables contributed most to a specific depth prediction.
View Official SIH 2026 Portal Raw Text
▼
•Background Subsurface ocean temperature is a fundamental variable for understanding ocean circulation,upper-ocean heat content, stratification, climate variability, air-sea interaction and marine ecosystems. Accurate representation of the vertical ocean temperature is essential for applications such as marine heatwave monitoring, fisheries, and data assimilation, etc. However, direct measurements of subsurface temperature remain sparse because they rely primarily on in-situ observing systems such as ARGO profiling floats, moored buoys, gliders, and ship observations. While these observations provide valuable vertical information, their spatial and temporal coverage is insufficient for generating continuous, basin-scale subsurface fields.In contrast, satellite observations provide continuous, large-scale monitoring of surface ocean conditions at relatively high spatial and temporal resolution. Surface variables such as Sea Surface Temperature (SST), Sea Surface Salinity (SSS), Sea Surface Height (SSH)/Sea Level Anomaly(SLA), surface currents, and surface winds contain indirect signatures of subsurface ocean processes through physical mechanisms including thermocline displacement, mesoscale eddies,vertical mixing, transport, and ocean-atmosphere coupling.Recent advances in Artificial Intelligence (AI), Deep Learning (DL), and representation learning enable the generation of satellite embeddings, where multidimensional surface observations are transformed into compact latent representations that capture hidden ocean dynamics. Such embeddings offer the potential to learn nonlinear relationships between surface observations and subsurface ocean structure more effectively than conventional machine learning approaches.
•Detailed Description The current problem statement proposes the development of a Satellite Embedding-Based Deep Learning Framework to reconstruct depth-wise subsurface temperature from daily surface satellite observations at 0.25° spatial resolution for North Indian Ocean (5°N to 30°N and 45°E to 105°E).The objective is to estimate the three-dimensional ocean temperature using only surface satellite observations.

The proposed system shall:

1.Develop a preprocessing and harmonization pipeline for multi-source satellite and ocean datasets.
2.Standardize all datasets to:
a.Spatial Resolution: 0.25° × 0.25°
b.Temporal Resolution: Daily
3.Use surface observations as input variables:
a.Sea Surface Temperature (SST)
b.Sea Surface Salinity (SSS)
c.Sea Surface Height (SSH) / Sea Level Anomaly (SLA)
d.Surface ocean currents (U, V)
e.Surface Winds (U, V)
4.Generate compact satellite embeddings using DL architectures such as:
a.Convolutional Neural Networks (CNN)
b.Vision Transformers (ViT)
c.Autoencoders
d.Graph Neural Networks (GNN)
e.Attention-based hybrid architectures
5.Train reconstruction models that learn the relationship between surface ocean state to temperature profiles.
6.Reconstruct:
a.Temperature at standard depth levels. Standard depths in meters: (0, 5, 10, 20, 30, 50, 75, 100, 125, 150, 200, 300, 500, 700, 1000)
7.Evaluate the reconstruction using independent observations and standard skill metrics like correlation, RMSE, Bias etc.(If a dataset is not available at required resolution, the team may select the openly available product and perform appropriate spatial and temporal interpolation/regridding)
•Training Input Datasets The following datasets are recommended for building the training and evaluation pipeline. (Insert table here)
•Training Target Dataset (Subsurface Temperature)

GLORYS Global Ocean Reanalysis https://doi.org/10.48670/moi-00021 Variables: Temperature In-situ Observations dataset Gridded ARGO INCOIS Live Access Server (LAS) – Gridded ARGO

•Expected Solution
•End-to-end preprocessing pipeline for satellite and ocean datasets.
•Satellite embedding engine capable of learning latent ocean representations from surface observations.
•Deep learning reconstruction model for estimating subsurface temperature.
•Standardized output at daily temporal resolution and 0.25° spatial resolution.
•Validation framework using independent ARGO observations.
•Demonstration of a working Proof-of-Concept (PoC) over the Bay of Bengal / Arabian Sea