Model Context Protocol · Free Research Version

DMAP-AI MCP Server: SPI drought analysis inside Claude and ChatGPT

The DMAP-AI MCP server, built by AgriMetSoft LLC, is a Model Context Protocol server for historical meteorological drought analysis. Connect it to an AI assistant and ask for SPI tables, drought events, wavelet scalograms, or drought periodicity for any point location, computed from NASA POWER or ERA5-Land/CDS precipitation data.

What it does

Instead of answering drought questions from general knowledge, an AI assistant connected to this server calls real scientific tools on the DMAP-AI backend. Each tool downloads historical precipitation for the requested coordinates, computes the Standardized Precipitation Index (SPI), and returns structured results the assistant can explain, tabulate, and cite. The same engine powers the DMAP-AI Research Version web app.

Available tools

1. run_spi_table — SPI time series and categories

Returns a historical SPI table (precipitation totals, SPI values, and drought/wetness categories) for a selected point and period. Supports Classic SPI thresholds, USDM-style labels, percentile bands, and k-means category schemes.

ParameterRequiredDefaultNotes
latitude, longitudeYesPoint location (decimal degrees)
data_sourceNonasa_poweror era5 (ERA5-Land/CDS)
start_date, end_dateNo1981-01-01 → 2024-12-31NASA POWER supports 1981 onward
baseline_start, baseline_endNo1981 → 2024Baseline climatology years
spi_scaleNo12SPI accumulation scale
step, yearly_methodNoyearly, jan_dec_totalsTemporal aggregation
category_schemeNoclassic_spiCategory labeling method

2. run_drought_severity_events — drought event detection

Identifies SPI-threshold drought events and returns the event table: start, end, duration, minimum SPI, and magnitude for each event.

ParameterRequiredDefaultNotes
latitude, longitudeYesPoint location
severity_thresholdNo-0.99SPI values at or below this are drought steps
Plus the same data source, dates, baseline, scale, and aggregation parameters as run_spi_table.

3. run_wavelet_scalogram — time-localized drought variability

Returns a compact wavelet scalogram (power across time and period) showing when SPI variability was strong and at which time scales. Useful for spotting multi-year variability episodes in the record.

ParameterRequiredDefaultNotes
latitude, longitudeYesPoint location
max_periods, max_timesNo24, 60Downsampling caps for the returned matrix
Plus the shared data source, dates, baseline, scale, and aggregation parameters.

4. run_drought_periodicity_analysis — dominant drought periodicities

Returns the global wavelet spectrum and ranked periodicity bands across the complete historical record (for example, an approximate 3-, 5-, or 8-year variability band).

ParameterRequiredDefaultNotes
latitude, longitudeYesPoint location
Plus the shared data source, dates, baseline, scale, and aggregation parameters.

How to connect

Claude (web, desktop, mobile)

  1. Open Settings → Connectors → Add custom connector.
  2. Enter the DMAP-AI MCP endpoint:
    https://droughtanalysis.com/mcp
  3. Start a chat and ask, for example: "Run a drought severity analysis for Ames, Iowa, 1981–2024, using NASA POWER."

ChatGPT (Custom GPT)

DMAP-AI is also available inside ChatGPT as a Custom GPT connected to the same MCP tools. See the DMAP-AI ChatGPT page for the direct link and ready-to-use prompts.

Other MCP clients

Any MCP-compatible client can use the server over Streamable HTTP at https://droughtanalysis.com/mcp. Tools are listed via the standard MCP tools/list handshake.

Example prompts

Scope and limitations

  • SPI only. No SPEI, PDSI, soil-moisture, crop-yield, or economic-impact analysis.
  • Two datasets. NASA POWER (default) and ERA5-Land/CDS. No other sources.
  • Historical and point-based. No forecasts, seasonal outlooks, or climate projections.
  • Statistical events. SPI drought events are threshold events, not direct measurements of agricultural or hydrological impacts.
  • Wavelet results are descriptive. Periodicities are variability bands in the historical record, not deterministic cycles or predictions.

Frequently asked questions

What is the DMAP-AI MCP server?

A Model Context Protocol server by AgriMetSoft LLC that gives AI assistants direct access to scientific drought-analysis tools: SPI tables, drought events, wavelet scalograms, and periodicity analysis for any point location, using NASA POWER or ERA5-Land historical data.

Which drought indices are supported?

Only SPI at this time. Additional indices (SPEI, PDSI, EDI, and others) are planned for licensed research workflows.

Can it forecast future droughts?

No. All analyses are historical and descriptive.

Is it free?

Yes — the Research Version tools exposed through MCP are free to use.

How should I cite DMAP-AI?

Kolsoumi Ayask, S. (2026). DMAP-AI: Drought Monitoring and Analysis Platform – AI (Version 1.0) [Software]. AgriMetSoft LLC. https://droughtanalysis.com/

Learn more

Knowledge Center · What is SPI? · Wavelet analysis · SPI video tutorial · Open the web app