OpenSWMM is built on the EPA SWMM 5.x foundation and extended by the contributors listed below.

• Caleb Buahin @cbuahin -- Project lead, architect, and primary developer of the OpenSWMM 6.x engine rewrite (data-oriented SoA architecture, plugin system, GeoPackage I/O, C/C++ API, dynamic wave solver alignment, Python bindings).

• Corinne Wiesner -- Verification and validation infrastructure: the manufactured benchmark suite (batches 1 and 2) and its analytical test harness, the dw-ritter-drybed-strip dry-bed benchmark and diagnostic regression test, and the benchmark provenance/tolerance reconciliation pass. Also contributed the LID exfiltration and barrel-clogging fixes, the Anderson-acceleration skip guard for surcharged EXTRAN nodes and the DPS/SLOT regimes (issue #3), the storage-node conduit half-area fix (issue #2), and thread-safety verification with concurrent-engine tests. Uncertainty-quantification sidecar work (spectral ROM, spatial fields, 1D network ROM, cross-layer coupling) is in progress on a feature branch and is not yet part of a release.

• GitHub Copilot coding agent -- swmm_NODE_OUTFLOW in the getNodeValue API and both binding trees (PR #85), and Error 227 for zero channel Manning's n in the transect reader, in both the legacy and refactored engines (PR #77).

Community members whose reports drove defects fixed in the 6.0.0-alpha.2 and 6.0.0-alpha.3 cycles. Where an issue was filed in this repository on someone else's behalf, the original reporter is credited.

• @wiesnerfriedman (Jacobs Engineering) -- The dynamic-wave convergence and Anderson acceleration campaign, reported as four separately localized defects: node convergence tested on the mixed iterate rather than the Picard residual (#97), the Anderson mixing coefficient applied to the wrong operand (#98), Anderson-accepted depths bypassing the canonical node state commit (#100), and outfalls counted in the per-node nonconvergence statistics (#101). Also reported that USE HOTSTART rejected every legacy .hsf written by a model with subcatchments (#93). Each report identified the offending code path directly.

• @MitchHeineman -- Repeated monthly RDII unit-hydrograph parameters applied silently, now WARNING 13 (#43, via USEPA/Stormwater-Management-Model#18), and the rain-gage rainfall scale factor with co-gage support (#46, via EPA discussion #147).

• @karosc -- Index and unit bugs in the storage depth Newton solve, affecting FUNCTIONAL, CONICAL and PYRAMIDAL storage shapes (#94).

• @NandanaPerera -- Pollutant mass not conserved when runoff falls below the cutoff (#90).

• @adivoky (Gemini Engineering & Sciences, Inc.) -- Outfall TIMESERIES/TIDAL stage-data references resolved by position instead of by name (#92).

Two further issues in these cycles (#74, [CONTROLS] SIMULATION TIME units; #95, per-subcatchment precipitation scale factors) were raised internally.

Portions of the OpenSWMM 6.x codebase were developed with the assistance of:

• Claude (Anthropic) -- Code generation, architecture design, GeoPackage strategy and implementation, test development.