hydrocraft-raven

Configure, run, and validate Raven v4.1 hydrological model emulations across multiple model structures.

155|6|Updated Mar 31, 2026
One-click install
npx skills add https://github.com/lzwei196/KISS-Knowledge-Infrastructure-for-Scientific-Simulation --skill hydrocraft-raven-lzwei196
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: hydrocraft-raven
Source: https://github.com/lzwei196/KISS-Knowledge-Infrastructure-for-Scientific-Simulation/tree/main/models/Raven
Command: npx skills add https://github.com/lzwei196/KISS-Knowledge-Infrastructure-for-Scientific-Simulation --skill hydrocraft-raven-lzwei196

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires numpy, pandas, geopandas, rasterio, shapely, netCDF4, and includes scripts (resource) and references (resource) components.

What problem does it solve? Running the Raven hydrological modelling framework requires assembling five configuration files (.rvi, .rvp, .rvh, .rvt, .rvc) with exact formats, correct units, and compatible process algorithms, and Raven silently ignores units and fills missing forcing with zeros, so mistakes produce plausible-looking but wrong results. ## Core Features & Use Cases - Multi-model emulation: Run 8 hydrological model structures (GR4J, HBV-EC, HMETS, MOHYSE, SAC-SMA, HYMOD, UBC, HYPR) on the same basin with identical forcing to quantify structural uncertainty. - End-to-end pipeline: 11 validated tools cover template selection, HRU construction from shapefiles, parameter generation, CMFD/MSWX forcing conversion with unit checks, execution, output parsing, ensemble comparison, and DDS calibration. - Diagnostic recovery: 33+ diagnostic triplets map symptoms (unit traps, class-name mismatches, period-ending output lags) to validated remedies. - Use Case: Compare five model structures on a 118,358 km2 basin against VIC output, rank them by calibration-window NSE, and report held-out performance of the selected structure. ## Quick Start Ask the agent to run the Raven HBV-EC emulation on your basin shapefile with CMFD forcing from 2000 to 2010, starting with the preflight check and following the staged pipeline from template selection through output parsing.

Frequently Asked Questions about hydrocraft-raven

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run multiple hydrological model structures on the same basin?

Use the ensemble comparison tool with a list of templates such as gr4j, hbv_ec, hmets, hymod, and sac_sma. All models run on identical forcing and HRU definitions, then are ranked by NSE and KGE computed over the calibration window.

How do I convert CMFD or MSWX forcing to Raven .rvt format?

Use the forcing conversion tool, which aggregates 3-hourly precipitation to daily mm, converts Kelvin to Celsius, and rescales radiation and pressure to Raven's expected units. Raven performs no unit conversion itself, so all conversions must happen in this tool.

Which Raven emulation template should I use for a snow-dominated basin?

Use HBV-EC, HMETS, or UBC, which include snow processes. GR4J, HYMOD, and basic SAC-SMA have no snow routines, so winter precipitation becomes rain and produces wrong seasonal patterns in cold basins.

Why does my Raven run produce zero discharge without any error?

Raven silently fills missing forcing variables with zeros, so an absent PRECIP block in the .rvt file yields zero precipitation and zero discharge. Verify PRECIP is present and check the ForcingFunctions.csv output for internally generated approximations.

Why do my external NSE scores not match Raven's Diagnostics.csv?

Hydrographs.csv is period-ending, so the row stamped with a date holds the preceding timestep's average, while observation input is period-starting. Use the provided load_discharge_series reader, which applies the calendar-date correction, instead of joining on raw date labels.

What are the limitations of the current Raven knowledge infrastructure?

It does not cover data assimilation, reservoir operations, water demand management, tracer transport, glacier processes, or MODFLOW groundwater coupling, because those require upstream data infrastructure not yet available. The generic template tool also has a known parameter-placement issue worked around with hand-crafted templates.