# Official road blockages along the Rasuwa–Trishuli corridor

Research and automated asset assessment completed 8 September 2026 UTC (9 September Nepal/local time). This report documents two bounded searches and 14 catalogue links for **seven shared-source road records**. It does not establish 14 independent events or a complete history of road access.

## Search outcomes and counting units

| Execution | Source | Road IDs screened | Retained | Excluded | New asset links | Native requests |
|---|---|---:|---:|---:|---:|---:|
| RUN-20260908-047 | BIPAD Highway API | 25 | 7 | 18 | 7 | 9 |
| RUN-20260908-048 | Department of Roads NAVIGATE | 10 | 7 | 3 | 7 | 16 |

The searches add two candidate service families. All seven retained DoR IDs match BIPAD `roadDataId` values. Discovery totals, asset links and candidate families are separate units. All 14 links have automated metadata assessments; human resource eligibility and manuscript inclusion remain unassigned. The register now contains 48 executions: 39 completed, eight partial and one unavailable. Overall search completeness remains unset.

“Completed” means that these bounded retrieval selections reached their recorded stopping rules. A full revision log, all closures active during the interval, exhaustive web coverage and independently validated restoration times were not recovered.

## Method declared before retrieval

Both strategies were frozen at **2026-09-08T21:58:24Z**, before new source requests. Their plan hashes and execution adaptations are retained in the exports. The question was whether the seven road-alert leads from the preceding BIPAD search could be resolved to native highway and original-provider records.

The BIPAD request selected closure starts from 25 August to 8 September 2026 UTC, using documented strict date filters. Its exact lower bound is `2026-08-24T23:59:59.999999Z` to include midnight at the target start; the exclusive upper bound is `2026-09-09T00:00:00Z`. The upper bound is a predefined query boundary, not evidence that future observations were available. Retrieval was national, followed by local selection of NH17/NH42 in Rasuwa, Nuwakot, Dhading and Chitwan. No language exclusion was applied. A similarly named Bhotekoshi location in Sindhupalchok falls outside this declared corridor.

The [BIPAD Highway API](https://bipadportal.gov.np/api/v1/highway/) returned 25 records followed by an empty page. Its reported count `9223372036854775807` is a placeholder and was never used as a population total. Seven selected detail routes were fetched. The method had a 30-page/3,000-row safeguard, same-host pagination and explicit stops for an empty page, absent/repeated next link or persistent failure.

For DoR, two supplementary web queries located the [official highway-status index](https://dor.gov.np/home/page/highway-traffic-updates), its [NAVIGATE dashboard](https://navigate.dor.gov.np/app/dashboard) and a road-condition page. The official index independently links NAVIGATE. Reading its publicly served application bundle established the anonymous history and detail routes and JSON request parameters. No authenticated operations were used. Five HTML/script responses supported discovery; eleven API responses supplied history and detail data. Web results are supplementary navigation, not a claimed exhaustive search population.

The observed DoR history API accepts POST JSON with `year`, `month`, `road_refno` and `page`. Four selections used year 2026, months 8 and 9, and roads NH17 and NH42. The native month filter is not documented as a closure-start or creation-date filter; both fields were examined. August returned nine NH17 records and one NH42 record; September returned zero for both roads. Three August records start before 25 August and were excluded from new event-window assets. Seven detail records were recovered. The native totals were exhausted on the first page of each selection; a 20-page overall history safeguard remained unused.

Each anonymous source request had a 30-second timeout, 10 MiB response cap and at most one transient-error retry. No transfer cap or persistent failure affected these responses. Source bodies, hashes and exact timestamps were preserved privately. Contact and user-account fields are excluded from public metadata. No provider was contacted.

## Seven resolved records

The times below are **BIPAD's explicitly offset Nepal-time fields**. The corresponding DoR date/time strings have matching wall-clock components but do not declare a time zone. This correspondence does not independently establish the DoR time convention.

| DoR ID / BIPAD ID | Location and road | Blockage start | Status in retrieved record | Actual-end field | Interpretation |
|---|---|---|---|---|---|
| 868 / 304 | Ghatbesi–Galaudi, Gajuri-5, Dhading · NH17 | 26 Aug 00:00 | OPEN | 28 Aug 18:30 | Midnight precision unresolved; no flood-arrival inference |
| 867 / 305 | Galchhi–Baireni, Galchi-6, Dhading · NH17 | 26 Aug 11:20 | PARTIAL_OPEN | 27 Aug 18:30 | Does not establish full restoration |
| 866 / 302 | Khalte–Syafrubesi–Timure, Rasuwa · NH42 | 26 Aug 11:30 | CLOSED | 2 Sep 11:33 | Status and actual-end field conflict |
| 869 / 303 | Chalise, Gajuri-2, Dhading · NH17 | 26 Aug 12:30 | OPEN | 28 Aug 17:00 | Source-reported end; no field confirmation |
| 871 / 306 | Jawang Khola, Benighat Rorang-10, Dhading · NH17 | 26 Aug 13:30 | OPEN | 26 Aug 19:00 | Estimated end is 06:30, before start; retained as a source contradiction |
| 870 / 307 | Mowa Khola, Ichchhakamana-3, Chitwan · NH17 | 26 Aug 13:30 | OPEN | 26 Aug 19:00 | Source-reported end; no field confirmation |
| 874 / 310 | Krishna Bhir, Benighat Rorang-7, Dhading · NH17 | 30 Aug 08:10 | CLOSED | Not stated | Estimated end 10 September is an estimate, not observed completion |

Four records carry OPEN, one PARTIAL_OPEN and two CLOSED. Six actual-end fields are populated. These are source states retrieved on this date, not present-day travel advice. The source's GLOF/flood/landslide descriptions remain attributed statements and do not independently establish the cascade mechanism or causation of every closure.

DoR's detailed location text adds named sites and textual chainages. All seven structured `chainage` fields are `0+000`, and all end-chainage fields are null. These values cannot define measured blocked-road lengths. Lengths mentioned in remarks are provider descriptions, not geometric measurements. Neither BIPAD's point locations nor its demographic associations establish flood footprints or affected-population counts. No BIPAD image links were listed for these seven records.

## Relationships with earlier evidence

The seven earlier alert IDs were 45659, 45660, 45655, 45658, 45661, 45662 and 45703. Their JSON-encoded reference lists identify `realtime.highway` primary keys; these differ from DoR IDs. Resolving both identifiers avoided requesting the wrong API record.

Six of the seven current BIPAD detail records differ from the saved alert reference in at least one charted status, end-date, repair or remarks field. Those are two retrieved representations, not a complete sequence of update events. Database modification timestamps can reflect synchronization and do not establish when a road physically changed state. Alert expiry was never treated as reopening.

The already reviewed [Nepal Flooding: Road Blockages](https://data.humdata.org/dataset/nepal-flooding-road-blcokages) resource contains 22 crowdsourced image-interpretation points from the Planetary Response Network, using Planet imagery and HOT road data. It is methodologically different from the administrative DoR records. No individual point matching or content identity was established in this phase; it must not be silently merged with, or treated as a duplicate of, these seven official records.

## Asset assessment and remaining data gaps

Descriptive asset names include the specific locality, road, start date and distribution source, while retaining native titles and identifiers. Metadata records location, coordinates, administrative area, native source fields, separate start/estimated-end/actual-end dates, repair labels, timestamps, request size, source relationships, rights and interpretation limits. Every field retains an evidence state, source URL and locator.

The combined inventory now has **6,331 automated assessments**, preserving the dates of 6,317 prior assessments. There are 100 candidate families with 6,061 assets, alongside 27 reviewed resources with 270 assets. The follow-up file contains 679 assets; all 14 additions retain rights and interpretation follow-up. Assessment completion is not independent human eligibility or scientific validation.

The [local-holdings reconciliation](2026-09-08-roads-holdings.md) identifies the strongest next searches: geology and engineering baselines, elevation products, exact Sentinel source products, source-area reports/interpretations and remaining hydrology access routes. Its 94 source-table rows include aliases, repeated citations and general references; they are not 94 new datasets.

## Reproducibility exports

- [Frozen strategies and plan hashes](2026-09-08-roads-strategies.json), [execution adaptations](2026-09-08-roads-adaptations.json)
- [Actual requests, response sizes and hashes](2026-09-08-roads-requests.json), [native query results](2026-09-08-roads-queries.json), [pagination and stopping outcomes](2026-09-08-roads-pagination.json)
- [Supplementary web queries and returned-link dispositions](2026-09-08-roads-web-navigation.json)
- [Search outcomes](2026-09-08-roads-outcomes.csv), [all 35 source-specific screening dispositions](2026-09-08-roads-dispositions.csv)
- [Seven paired road records and comparisons](2026-09-08-roads-road-records.json), [source relationships](2026-09-08-roads-relationships.csv)
- [Asset titles](2026-09-08-roads-asset-titles.csv), [asset assessments](2026-09-08-roads-asset-assessments.csv), [complete assessments](2026-09-08-roads-assessments.json), [field-level metadata](2026-09-08-roads-metadata.csv)
- [Phase counts](2026-09-08-roads-summary.json), [local-source reconciliation CSV](2026-09-08-roads-holdings.csv), [local-source reconciliation JSON](2026-09-08-roads-holdings.json)

The original research phase updated local sources without building or deploying. A subsequent user-requested [release review](2026-09-09-roads-audit.md) checked source consistency, generated outputs and browser behavior and completed a production build. The phase-summary JSON retains its original research-completion timestamp and lifecycle fields; it is not a live deployment-status endpoint.