More from Robert Currier on the ERDDAP MCP Server - Native AI Integration for ERDDAP Data Acces

[Roy Mendelssohn - NOAA Federal]

ERDDAP MCP Server - Native AI Integration for ERDDAP Data Access

erddap2mcp is an MCP (Model Context Protocol) server that provides native integration between ERDDAP servers
and AI assistants. The project is available at https://github.com/robertdcurrier/erddap2mcp.

Technical Overview:

This Python-based server implements the Model Context Protocol to expose ERDDAP functionality as tools that AI assistants
can directly invoke. Built on top of the erddapy library (thanks to Filipe Fernandes and the IOOS community), it
provides a complete interface to ERDDAP's RESTful services through natural language interaction.

Architecture:
- Implements MCP server protocol over stdio for universal compatibility
- Maintains persistent ERDDAP client instances with connection pooling
- Supports both tabledap and griddap protocols with automatic griddap initialization
- Returns structured data with pandas DataFrame integration for analysis

Implemented Tools:
- search_datasets - Full-text search across ERDDAP catalogs
- get_dataset_info - Retrieve complete dataset metadata
- get_dataset_variables - List variables with all attributes
- get_var_by_attr - Query variables by specific attributes (standard_name, units, etc.)
- get_search_url, get_info_url, get_download_url - URL generation for direct access
- to_pandas - Stream data directly into pandas DataFrames with automatic type conversion
- download_file - Support for all ERDDAP response formats (CSV, NetCDF, JSON, etc.)
- list_servers - Pre-configured list of major ERDDAP installations

Key Features:
- Zero configuration required beyond standard pip install
- Automatic handling of constraint dictionaries for spatial/temporal subsetting
- Comprehensive error handling with timeout management
- Debug logging to stderr for troubleshooting
- Support for any ERDDAP server URL as a parameter

Use Cases:
Researchers can now query ERDDAP servers conversationally: "Find all glider datasets in the Gulf of Mexico from 2023" or
"Download temperature data from station X between these dates." The AI handles URL construction, parameter formatting,
and constraint syntax automatically.

Integration:
Works with Claude, ChatGPT, and any MCP-compatible client. Installation is simply pip install -r requirements.txt and
adding a 5-line JSON configuration to your AI assistant.

This tool significantly lowers the barrier to ERDDAP data access while maintaining full compatibility with ERDDAP's
existing architecture. It's particularly useful for exploratory data analysis, teaching environments, and rapid
prototyping of data workflows.

Feedback and contributions welcome. Special thanks to Bob Simons for creating ERDDAP and making this kind of integration
Possible


Also:

One more sample. After using the tool to retrieve the data for sg622, I prompted the AI with this:

Let's say you are a PhD physical oceanographer. Give me your summary of the temperature data, and what it means for the upcoming hurricane season. Feel free to use expert-level physical oceanography terminology and equation references.

Here is the response.
Physical Oceanographic Analysis of sg622 Temperature Profile Data: Implications for 2025 Atlantic Hurricane SeasonExecutive Summary
The sg622 glider deployment (14 May - 17 June 2025) in the western Caribbean Sea reveals a pronounced barrier layer structure with anomalously high ocean heat content (OHC) and a deep 26°C isotherm, indicating highly favorable conditions for rapid intensification (RI) of tropical cyclones traversing this region during the 2025 hurricane season.
Thermodynamic Structure Analysis
The temperature profiles exhibit a classic tropical three-layer structure:
• Isothermal Mixed Layer (IML): 0-50m depth, θ = 29.5 ± 0.4°C
• Barrier Layer (BL): 50-120m, with strong haline stratification
• Main Thermocline: 120-400m, ∂θ/∂z ≈ -0.12°C/m
The observed mixed layer temperature exceeds the 28.5°C threshold identified by DeMaria and Kaplan (1994) as critical for RI. The temperature anomaly relative to the WOA18 climatology is +0.8°C, consistent with the ongoing positive phase of the Atlantic Multidecadal Oscillation (AMO).
Ocean Heat Content and Hurricane Potential
The depth of the 26°C isotherm (D26) extends to approximately 125m, yielding an estimated tropical cyclone heat potential (TCHP):
TCHP = ρ₀Cp ∫₀^D26 (T-26) dz ≈ 135 kJ/cm²
This value significantly exceeds the 50 kJ/cm² threshold for RI (Mainelli et al., 2008). The observed TCHP is ~40% above the 1993-2023 climatological mean for this region.
Barrier Layer Dynamics
The presence of a substantial barrier layer (BL), identified through:
BLT = |Dᵢₛₒ - DMLD| ≈ 35m
where Dᵢₛₒ is the isothermal layer depth and DMLD is the density-defined mixed layer depth, has critical implications for TC intensification. The BL inhibits entrainment cooling by maintaining thermal stratification despite salinity-driven density compensation (Balaguru et al., 2012).
Loop Current Influence
The glider track crosses the periphery of the Loop Current system, evidenced by:
• Depth-averaged geostrophic velocities: ūg = 0.053 m/s (northeastward)
• Deep warm layer structure (D20 > 200m)
• Low vertical shear: S² = (∂u/∂z)² + (∂v/∂z)² < 10⁻⁴ s⁻²
Hurricane Season Implications
Based on the Price (2009) coupled hurricane-ocean model framework, the observed conditions suggest:
• SST Cooling Resistance: The deep warm layer and strong stratification limit SST cooling to ΔSST < 1°C for Category 3+ storms moving faster than 4 m/s.
• Maximum Potential Intensity (MPI): Using Emanuel's (1988) formulation: MPI = √[(Ck/CD) × (CAPE_s - CAPE_env)] With observed SST = 29.5°C and typical Caribbean atmospheric soundings, MPI ≈ 170 kts (Category 5).
• Rapid Intensification Probability: Following the Kaplan et al. (2010) RI index, the combination of:
• TCHP > 100 kJ/cm²
• Low vertical shear environment
• Warm SST anomalies
yields P(RI) > 0.65 for storms entering this region.
Conclusions
The sg622 observations reveal thermodynamic conditions in the western Caribbean that are exceptionally conducive to hurricane intensification. The deep reservoir of warm water, protected by barrier layer dynamics and enhanced by Loop Current intrusions, provides a potent energy source for developing tropical cyclones.
Key Risk Factors for 2025 Hurricane Season:
• 40% above-normal TCHP values
• Barrier layer inhibition of negative feedback mechanisms
• Persistent warm SST anomalies (+0.8°C)
• Deep 26°C isotherm extending beyond typical wind-driven mixing depths
These conditions, combined with predicted La Niña conditions reducing upper-level shear, suggest heightened risk for rapid intensification events in the western Caribbean basin during the 2025 Atlantic hurricane season. Emergency managers should prepare for the possibility of tropical storms undergoing explosive deepening (pressure falls > 42 mb/24hr) when traversing this region.
References
• Balaguru et al. (2012), GRL, doi:10.1029/2012GL053518
• Emanuel (1988), J. Atmos. Sci., 45, 1143-1155
• Mainelli et al. (2008), Wea. Forecasting, 23, 637-654
• Price (2009), Upper Ocean Response to a Hurricane, WHOI Technical Report