# DGs surrounding district — 2026-10-30 The generic neighborhood is replaced with 26 measured neighboring building shells, mapped streets, and a scan-derived landscape. DGs remains the existing playable club, with its selected Black survivors, detailed shotgun, four zombies and four-player co-op. ## What the sources actually support - **LiDAR:** USGS 3DEP `TX_Pecos_Dallas_B3_2018`, read through public Entwine Point Tiles. 71 LAZ nodes, clipped to 1,723,588 points including 787,218 classified ground returns. This is historical airborne elevation data, not a current scan and not facade photogrammetry. The dataset name is not asserted to be the individual flight date. - **Real aerial photograph:** USDA NAIP, `m_3209609_ne_14_060_20220704`, captured July 4, 2022, nominal 0.6 m resolution. Exported as a 2048-pixel crop. Upsampling does not create extra photographic detail. - **Footprints and roads:** OpenStreetMap ways, acquired October 30, 2026. Road names include West Northwest Highway, Wire Way, Goodnight Lane and Lombardy Lane. 83 clipped road/service-drive sections and 27 measured building records; the DGs record anchors the model but does not replace the approved playable club. - **Street-level DGs photo:** the previously approved DallasFortWorthTV photo informs the yellow facade, dark stepped fascia, entrance awning, roadside sign and overhead utility line. Its image is reference-only and is not redistributed as a texture. Neighboring facade bays and materials are interpreted. ## Construction The street-facing DGs footprint wall anchors the local coordinate system. A rigid rotation aligns it to the existing game; Web Mercator distances are corrected by cosine(latitude) to local ground metres. OSM coordinates are retained in the downloadable source and derived databases. Roof elevations use the 65th percentile of elevated returns inside each footprint, after a 1.1 m inset; minimum 40 samples. Local classified ground points establish building base and height. The terrain uses a 4 m grid and blends to the existing flat combat floor, including a margin to avoid interpolation lifting its edge. Tree locations require green aerial pixels and elevated LiDAR returns; 240 trees use 720 instanced crowns. The game ships the compact derived data and one aerial texture, not the 41 MB research cache or raw LAZ tiles. The new source/derived/aerial payload is about 1 MB. There are no map-service requests during play. Building shells use measured footprints and heights; roof and ground color come from the actual aerial image. Roller shutters, small facade bays, utility hardware and nighttime lighting remain art interpretation. ## Verification - 9 geospatial checks: hashes, scan counts, roof sample thresholds, source-coordinate preservation, four independently picked aerial roof-corner comparisons, road names, flat combat terrain, payload and attribution. - The four aerial corner checks agree within 5 metres. This is a limited alignment spot-check, **not survey accuracy**. Sub-2-mm coordinate round-trip error measures numerical conversion only. - 15 environment checks preserve all 29 authoritative co-op colliders, entrance traversal, bullet blocking, doors, scanned materials and geometry budgets. - 49 combat/movement/weapon/co-op logic tests and 13 HTTP checks with four independent local sessions. - Browser QA and actual runtime captures are recorded in the linked evidence after visual inspection. A full four-human remote session, mobile performance and a complete mission playthrough are not part of this pass. ## Deliberate limits The 2022 photo and historical LiDAR can disagree with today's tenants or buildings. The club interior/shell proportions remain the approved game interpretation. Its surrounding combat floor is flattened; the new distant streets/buildings are scenery beyond the existing playable boundary, not a new explorable open-world map. Road widths use mapped lane counts where available, otherwise an explicit visual estimate. No current tenant branding or unverified neighboring business names are invented. Airborne LiDAR cannot reconstruct detailed storefronts, signs or interiors. Some far buildings outside the bounded clip remain aerial background. This is a substantially more accurate district reconstruction, not a 1:1 digital twin. ## Sources and licenses - [USGS LiDAR Explorer](https://www.usgs.gov/tools/lidarexplorer) - [Actual EPT dataset metadata](https://s3-us-west-2.amazonaws.com/usgs-lidar-public/TX_Pecos_Dallas_B3_2018/ept.json) - [USGS/USDA NAIP service](https://imagery.nationalmap.gov/arcgis/rest/services/USGSNAIPImagery/ImageServer) - [OpenStreetMap attribution and license](https://www.openstreetmap.org/copyright) - [ODbL 1.0](https://opendatacommons.org/licenses/odbl/1-0/) - [DGs street photo reference](https://www.lemon8-app.com/@dallasfortworthtv/7589408034227012127?region=us) USGS/USDA data are public-domain government datasets. The OSM source extract and derived vector database are distributed under ODbL 1.0; this notice does not relicense the game's unrelated code or characters. Source metadata, LAZ URLs/SHA256s and derived-file hashes are in `assets/neighborhood/manifest.json`; the complete extracted vector source is `osm-source.json` and the derived database is `district.json`. ## Reproduction Use an isolated Python environment with numpy, scipy, laspy, lazrs, shapely and pillow. Run `python tools/acquire-neighborhood.py `, then `python tools/derive-neighborhood.py `. The acquisition script bounds the EPT request spatially and to depth 11. The published compact data are self-contained; regeneration can differ if upstream OSM or imagery changes, so the retained source extract and hashes define this version.