Dropped

The 77 trips that did not make the board, each with the reason it was cut — thin scenery, a price the model could not stand behind, or a length that made the flying the trip. Kept visible so the board’s selection can be argued with.

Search names and reasons

77 of 77 cut trips

Savannah Way & the Gulf: Atherton Tablelands, Undara & Karumba
Fails on pricing: the only trip in its region with no activities figure at all, while the Undara lava tubes - its entire reason to exist - are a paid guided tour.
Alsace Wine Route & the Vosges (Route des Crêtes)
550 km over 10 days is the lowest driving density in the atlas, and the survey concedes the region needs a cross-border day trip to avoid repetition. A superb 5-6 day trip, not this one.
The Ardennes: Belgium, Luxembourg & Champagne-Ardenne
700 km over 10 days at tier 2 - the survey calls it 'the weakest-scenery pick in this whole region', included for novelty rather than drama.
Amalfi Coast & Cilento
The SS163 Amalfitana runs odd/even plate restrictions all of August AND all of September daily, plus Easter, late April and summer weekends - which removes the shoulder-season window the trip is built on. Lodging was also understated by 30%+.
Castile's Walled Towns & the Sierra de Gredos
Not a 10-12 day drive: the signature route is ~550 km, padded to a claimed 950. Segovia to Toledo is a comfortable five days.
Serra da Estrela & the Historic Villages of Portugal
900 km over 10 days with Coimbra and Guarda only 160 km apart; the survey's own notes admit it does not fill 10-12 days.
Orkney — Neolithic Scotland by island-hop
300 km of island driving; Skara Brae, Brodgar, Stenness and the Churchill Barriers are 3-4 days. The rest is padding on tarmac the Scotland row already covers.
The Samoeng Loop by scooter — Mae Sa, the 1269 switchbacks and Mae Kampong
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Phu Thap Boek and the Khao Kho ridge — Route 12's waterfall road to a 1,794 m cabbage mountain
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Doi Mae Salong and Phu Chi Fa by motorbike — the 1130 tea ridge and the 1155 border road
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
The Death Highway to Umphang — 1,219 curves and Thi Lo Su
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Ha Giang Loop by Motorbike
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Northwest Vietnam: Mai Chau to O Quy Ho
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Hai Van Pass and the Son Tra Headland
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Phong Nha and the Ho Chi Minh Road West
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
North Lombok: Pusuk, Senaru and the Sembalun Rim
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Shimanami Kaido — the Seto Inland Sea island crossing
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
The Death Railway: Thonburi to Nam Tok
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Sleeper to the Ping Valley: Chiang Mai, Lamphun and Lampang
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
KTM ETS Spine: Penang, Ipoh, Kuala Lumpur, Malacca and Singapore
Confirmed duplicate. The row's non-duplication case rests on 'Ipoh, Kuala Lumpur, Tampin/Malacca and Gemas are absent from all 278 rows'. Ipoh is in two existing rows, one of which does the Butterworth-Ipoh cave-temple day BY RAIL ('the Ipoh day by rail, Butterworth-Ipoh 1h34-1h46 each way, 7 trains a day, for the cave temples, Concubine Lane and white coffee') - the same corridor by the same mode
Koh Tao by Night Bus and Catamaran
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Koh Phangan's East Coast Without the Full Moon
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Southern Laos by Bus: Wat Phou, the Bolaven and Four Thousand Islands
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Lampang and Phrae: the Teak Towns by Overnight Coach
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Khiriwong and Thale Noi: the Khao Luang Foothills to the Lotus Lake
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Koh Lipe and the Tarutao Archipelago
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Koh Chang to Koh Kood by Inter-Island Boat
TRIPS_TO_DROP, confirmed duplicate: the row sleeps on Koh Chang AND Koh Kood, the exact two islands already held by 'Koh Chang with the Car on the Ferry' and 'Koh Kood — Drive and Ferry to the Last Quiet Island'. The brief's mode exception covers a road famous for a mode, not a seasonal passenger speedboat used as a transfer. Verifier: 'Do not score as written.' (Its other findings — the 2,420 THB
The Gulf Chain: Koh Tao, Koh Phangan, Ang Thong and Samui
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Similan and Surin Liveaboard: Richelieu Rock
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Raja Ampat: Waigeo, Piaynemo and the Dampier Strait
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Khao Luang from Khiriwong
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Sapa Homestay Trek, Muong Hoa Valley
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Low's Peak: the Mount Kinabalu Summit Climb
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Certify on Koh Tao
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Vietnam end to end on the Reunification Express
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Satun: Koh Lipe, the Adang-Rawi group and Tarutao
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Bukit Lawang and Tangkahan: the east door of Gunung Leuser
Confirmed duplicate — 'DROP OUTRIGHT. Hard duplicate of Lake Toba and the Karo Highlands, whose signature_route opens Medan -> Bukit Lawang (orangutans) -> Berastagi...'. Bukit Lawang is not adjacent to the incumbent, it is the incumbent's first stop, and the days_max 7 extension adds Berastagi's Sipiso-piso and Sibayak, which is verbatim the incumbent's middle section. Neither exception applies:
Koh Tao: learn to dive from Chumphon
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Mongolia: Ulaanbaatar, Terelj and the South Gobi
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Southern Laos by motorbike: the Bolaven Plateau and Si Phan Don
Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Angkor Overland: Driving to Cambodia's Temple Plain
Dead premise per the auditor's DROPS finding, not a price error. Thailand closed every land crossing with Cambodia in late June 2025 after the border-temple conflict; the 27 Dec 2025 ceasefire broke down in early Jan 2026 and all official crossings including Aranyaprathet/Poipet remained closed through 2026 with no announced reopening date, cross-border bus services suspended. The trip's entire and only stated value proposition - the only trip in the band with zero airport days - is therefore gone, and the survey's own fallback (fly BKK-Siem Reap, rent locally) is a different trip that the survey itself rates the scenically weakest row in the band at tier 2. The printed route also runs to Preah Vihear, one of the specific temple areas the fighting has been over, so the safety position is untenable as well. Consistent with the brief's verified ground truth (FCDO, 19 March 2026).
Kashmir to Ladakh Himalayan Crossing
Dead premise on safety, not length. FCDO advises against ALL travel to the Union Territory of Jammu and Kashmir, naming Sonamarg, Srinagar city and the Jammu-Srinagar national highway explicitly, with the only exception being air travel to and movement within Jammu city — the route's entire first three days (fly Srinagar -> Sonamarg -> Kargil) sit inside that zone. FCDO also advises against all travel within 10 km of the Line of Control with Pakistan and the Line of Actual Control with China, which catches four more named stops: Turtuk (~10 km from the LoC), Pangong Tso (the LAC runs through the lake) and Hanle (~20 km from the LAC). Compounding: 14 printed vs 16 real days, and the Leh taxi-union rule already breaks the self-drive premise. The auditor notes a Ladakh-only fly-in version (BKK-DEL-Leh, acclimatise 3 nights, Nubra, Pangong, fly out) is a legitimate 12-day trip but is a different trip and must be re-surveyed under its own name. Source: https://www.gov.uk/foreign-travel-advice/india/regional-risks
Jeju Island Loop
Auditor (drops): the non-stop the entire 5-day case rests on is not verifiable and may no longer exist - FlightConnections returns 'no direct flights' in BOTH directions (BKK-CJU and CJU-BKK), Trip.com routes everything via Busan/Incheon/Gimpo, and only FlightsFrom still shows a non-stop, at 2 flights a week all departing 23:20 (a red-eye, not the daytime arrival the itinerary is written around). The survey's 'roughly 17 flights a month' is contradicted by its own cited source. The fallback via Seoul is 7-8h each way, which is the same arithmetic the band used to exclude Guilin and Goa - so the 5-day premise is dead as written. The auditor's separate K-ETA verdict also states outright that 'the Jeju row it supports should be dropped on flight grounds'. Reinstate only if the non-stop is re-verified directly on jejuair.net, and then as a red-eye trip with three usable days (a rebuild, not this row).
Ladakh: Leh and Nubra (season-limited)
Auditor's explicit drop for this band: 'drop from the 6-day band (keep it at 8-10)... because the risk stack is not survivable at this length'. Taken as a drop rather than a day-range flag because the decisive element is a safety position the row does not carry, not the day count: an active, recurring civil-unrest flashpoint omitted from safety_notes entirely (lethal statehood protests in Leh 24 Sept 2025 - 4 dead, 150+ injured - a 22-day curfew with mobile-internet suspension and mass cancellations, and a 'Ladakh Peace and Justice March' set to reach Leh on 24 Sept 2026, in peak season), stacked on a single morning-only DEL-IXL window with no ground alternative, a four-month season, and a foreigner-only agent-arranged PAP gating day 4. At 6 days there is no slack to absorb any one of them. The row survives at 8-10 days in its own band; this drop removes it only from d6-asia.
Jordan: Amman to Wadi Rum on the King's Highway
Auditor's explicit instruction: 'Drop from the 8-day band; revisit only after a fresh check.' The safety position, not the price, is untenable — IRGC ballistic missiles hit Muwaffaq Salti Air Base near Azraq on 31 Aug 2026 and again on 8 Sep 2026 (four days before the survey); Safe Airspace rates Jordan 'Level Three – Caution'; France NOTAM LFFF F1813/26 tells French operators not to enter Jordanian airspace, Germany NOTAM EDWW B0678/26 recommends avoiding FIR Amman until 22 Oct 2026, and EASA CZIB 2026-08-R1 is open on Jordan. The row's own recommended Doha/Abu Dhabi routings compound it (repeated Qatari airspace closures, 2,000+ cancellations at Hamad, EASA advising avoidance of Qatari/Bahraini/Kuwaiti/UAE airspace at all altitudes), and an 8-day trip with two hard transit days and no slack cannot absorb a Gulf-hub diversion. The row's safety_notes present a June 2026 easing as the current state.
Pakistan: Ratti Gali Lake and the Neelum Valley Trek
Not a trek, and not landscape tier 4 as a walking route - it fails the survey's own exclusion standard. The operator selling it describes reaching the lake by vehicle: Day 2 travellers go 'traversing difficult terrain and winding jeep paths to reach the lake's gorgeous shoreline', and Day 3 opens 'Board jeeps to return to Dowarian'. The inclusion list is a road-tour list - 'Guided sightseeing tour according to itinerary above', 'SUV/mini-van/car (depending on number of people)', 'Driver', 'All entry/admission tickets', 'Bottled water in car', 'Toll & parking fees' - with accommodation and all meals explicitly excluded, and the product is filed under Islamabad city tours ('xmlhi-holiday-info-islamabad-tours'). This is exactly the profile the survey used to exclude Kaghan Valley / Lake Saiful Muluk ('not tier 4-5 as a trek: the lake itself is jeep/short-walk accessible') and Kel-Suu ('not a trek: reached by private 4x4'). The row's own drawbacks field already concedes 'the trekking component is genuinely short (one day); most of the trip length is overland road travel, not walking'. Its trek block (trek_days 2, trek_nights 1, guide_day_thb 417, permits 60 THB, lodging 240, food 180) is also the only itemised Pakistani trek block in the survey and is built on an informal PKR guide-fee range rather than any bookable product. Drop it rather than carry a jeep excursion in a catalogue of breathtaking treks.
Tanzania: Kilimanjaro via the Lemosho Route
Duplicate of a filed atlas row - 'Tanzania: Kilimanjaro summit via the Machame Route' is already in the atlas. Wikipedia lists Lemosho (56 km, 6-8 days) and Machame (49 km, 6-7 days) as two of seven official routes on the same mountain, both characterised as a 'forested traverse to Barafu', converging at Barafu Camp for an identical final push to Uhuru Peak (5,895 m). Same park, same gateway (JRO/Moshi), same TANAPA permit regime, same mandatory-guide rule, same summit, roughly five of eight days on shared ground - the row itself concedes it converges 'from Barranco onward'. Under the brief's own test ('the Three Passes is different from EBC; a 10-day EBC is not'), this is a route variant of a filed trek, not a genuinely different one. If the atlas nonetheless wants Kilimanjaro route coverage, the honest form is a variant note on the existing Machame row (extra Shira Plateau days, camping-only, higher park-fee load) rather than a second tier-5 row. Note also that its package price of 102,300 THB was an unsourced estimate - the operator page cited, ultimatekilimanjaro.com, 404'd for me as it did for the surveyor.
Thailand: Khao Chang Phueak savanna-summit trek, Thung Yai Naresuan/Umphang Wildlife Sanctuary
The trek described does not exist. Khao Chang Phueak is the highest peak of THONG PHA PHUM National Park in KANCHANABURI at 1,249 m (the highest point in Kanchanaburi), confirmed on both English and Thai Wikipedia - it is not in Thung Yai Naresuan or Umphang Wildlife Sanctuary, it is not 1,955 m, and it is nowhere near Mae Sot or Umphang. Consequently the gateway (BKK-Mae Sot flights, 2,200-4,800 THB), the 3,000 THB Mae Sot-Umphang-trailhead 4WD extras, the 'sanctuary trekking-quota permit with fixed departure dates', the Nov-Jan sanctuary season and the Myanmar-border malaria/safety note are all attached to the wrong protected area and would feed pure fiction into the cost model. The row is unfixable by patching - every logistics field is wrong. It is also duplicative as framed, since the atlas already holds 'Umphang and Thi Lo Su - the Death Highway'. A corrected Thong Pha Phum/Kanchanaburi row (road access from Bangkok, no flight, DNP booking) could be re-surveyed from scratch, but a 1,249 m two-day grass-ridge walk should be re-tested against the tier-4 bar before it is re-admitted.
Thailand: Doi Pha Hom Pok – Thailand's second-highest peak, Fang hill-tribe trek
Not a trek, by the survey's own exclusion rule. The elevation claim is confirmed (2,285 m, Thailand's second-highest), but the trek is not: the summit is reached by hiring a 4WD to Kiew Lom Campsite and then hiking 3.5 km, with an overnight at that vehicle-accessible campsite to catch the sunrise. That is a drive-up campsite plus a short walk - materially the same product the survey itself excluded as 'not a trek: a drive-up sea-of-mist viewpoint and campsite' (Kaeng Krachan/Phanoen Thung). The row's 3-day/2-night demanding ascent, Lisu/Lahu homestay night, summit camping night and 5-7 hours per day are unsourced and contradicted. On the atlas scale (tier 4 = Dolomites hut circuits, Kumano Kodo, Drakensberg) a 3.5 km walk from a road-head does not clear the bar. If a genuine multi-day hill-tribe approach route from Fang exists, it needs to be evidenced before the row returns.
India: Amarnath Yatra (Pahalgam route, Kashmir Himalaya)
Duplicate. The atlas already holds 'India: the Amarnath Yatra cave pilgrimage, Pahalgam route' (days_recommended 10, feasibility 'permit') in trips/roadtrip-atlas-2027/data/atlas-trek-world.json and costed.json - same cave, same Pahalgam route, same gateway. The JKSASB 2026 findings verified here (registration from 15 Apr, RFID mandatory, CHC on/after 8 Apr 2026, all routes a No Flying Zone from 1 Jul 2026 so no helicopters, a formal foreign-national procedure, and the 2026 season suspended from 9 Aug) should be folded into that existing row instead.
India: Kedarnath and Tungnath-Chandrashila Himalayan pilgrim treks (Garhwal, Uttarakhand)
Duplicate. The atlas already holds 'India: Kedarnath and the Char Dham Pilgrim Walk (Gaurikund to Kedarnath, with Vasuki Tal)' (days_recommended 8, feasibility 'permit'). The only addition here is the 5km-each-way Chopta-Tungnath-Chandrashila side walk, which does not justify a second row. Carry the verified correction to the existing row: Chandrashila is 3,690m, not ~4,000m, and the Chopta walk is 3.5km to Tungnath plus 1.5km to the summit (10km round trip, not 7km).
China: Emei Shan Golden Summit pilgrim stairs (Sichuan)
Duplicate. The atlas already holds 'China: Mount Emei - the Pilgrim Stairway to the Golden Summit' (days_recommended 6, feasibility 'open'), already present in costed.json. The verified 2026 prices (entrance CNY160/2 days, CNY110 15 Dec-15 Jan; cable car CNY60 up, CNY55 down; Leidongping bus CNY90 return; summit 3,099m; CNY 1 = 4.958 THB on 2026-09-15) should be applied to that row - and note the proposed extras figure of 743 THB silently dropped the CNY55 descent leg; bus plus both cable-car legs is CNY205 = 1,016 THB.
Turkey: Lycian Way full traverse (Fethiye to Antalya)
Duplicate of an already-filed trail, not a new landscape. The atlas holds 'Turkiye: Lycian Way, Fethiye to Kas' (10 days, 'open') and 'Lycian Coast: Antalya, Olympos, Kas, Fethiye' (10 days, route_km 850), with plan files under trips/roadtrip-atlas-2027/plans/. A 30-day thru-hike of the same waymarked route is a length variant. The route length is also contested (Wikivoyage 540km; Wikipedia's lead now says ~760km, its infobox 520km), so the day estimate is unsound as well. Keep the one verified gain for the existing row: the trail's high point is genuinely ~1,800m (Incegeris Hill 1,811m), passing just east of Tahtali Dag's 2,300m summit.
France: Vallee Blanche off-piste glacier route and Mer de Glace (TECHNICAL - not a non-technical walk)
Not a trek. The Vallee Blanche is a lift-served off-piste ski descent or roped mountaineering objective from the Aiguille du Midi, not a walking route - the only source the row cites (Wikipedia's Mer de Glace article) describes the Vallee Blanche purely as 'broad icefields that feed into the Glacier du Tacul', a glacier system, and does not characterise it as any kind of route. The row's own feasibility_notes concede it 'does not fit the non-technical framing of this batch and should be filtered accordingly', and its trek block is trek_days 1 / trek_nights 0 / lodging_style hotel / difficulty demanding: a single-day technical day activity from a Chamonix hotel. It also carries no sourced pricing whatsoever - every Compagnie des Guides and cable-car fetch failed - and its 'hundreds of extra stairs each year' drawback is unsupported by the cited source. Nothing in it is trek-shaped, priced, or verified. Note that the Mont Blanc massif is already represented in the atlas by TMB full, Haute Route and Monte Rosa, so nothing is lost.
Malaysia: Danum Valley rainforest (Borneo Rainforest Lodge)
NOT A TREK. The row self-refutes: 'THIS IS A LODGE STAY WITH GUIDED WALKS, not multi-day point-to-point trekking — 2-3 hr guided outings from a fixed lodge, returning to the same bed each night'. Wikivoyage confirms the only walking is on 'poorly labeled trails ... for 1-7 hr of walking time' radiating from a fixed base, most requiring ranger accompaniment, with no point-to-point route. Access is confirmed restricted to BRL or the Field Centre with two enforced checkpoints, which further precludes any independent multi-day route. Pricing also failed verification across two passes: the only figure retrievable is a January-2017 DVFC dorm rate (RM96) that Wikivoyage itself calls 'heavily overpriced' and which is not representative of BRL; no 2026 rate card exists at any reachable URL. Lowland dipterocarp forest viewed from a canopy walkway is also a weak tier-4 claim against Kumano Kodo and the Drakensberg.
Kenya: Loita Hills Maasai forest walk
NOT TIER 4, and the product cannot be shown to exist. The row self-refutes on landscape: 'Scenically modest compared with every other row in this batch' — which is a direct admission that it fails the tier-4 bar set by the Dolomites' hut circuits, Kumano Kodo and the Drakensberg. It also self-refutes on bookability: 'almost entirely undocumented for foreign booking this session — real risk the trek as imagined doesn't exist as a packaged product at all'. Independent verification confirms the worst case: the Wikipedia article is a stub with no elevation, no area, and 'no mention of organized walking safaris, trekking products, or tourism activities', recording only that 'the Loita Community Wildlife Conservancy operates in the Loita Hills'. No operator, no price, no permit rule, no season and no group minimum could be evidenced from any source. A row that is admittedly unremarkable to look at and cannot be shown to be bookable fails both drop tests at once.
Kenya: Chyulu Hills walking safari (ol Donyo Lodge)
NOT A TREK, and not tier 5. The row self-refutes twice in capitals: 'THIS IS A LODGE STAY WITH WALKS, not multi-day trekking ... guests sleep in the same suite each night', with daily_hours of '1-4 hrs of walking depending on which walk is booked'. The USD 1,590 pp/night rate is confirmed accurate against Great Plains Conservation's own page and is fully inclusive — but a two-night stay at 52,470 THB/night consumes 104,940 THB of a 120-140k THB whole-trip budget for two nights of optional strolls. The Kilimanjaro view is genuine; a view is not a trek, and rolling volcanic hills topping out at 2,188 m are not tier 5. Dropped on modality and tier, not on price.
Tanzania: Serengeti walking safari (lodge/mobile camp with short walks)
NOT A TREK, and not tier 5. The row self-refutes in its own comfort_note and drawbacks: 'LODGE/MOBILE-CAMP STAY WITH SHORT WALKS ONLY (1-3 hrs, escorted, usually near camp), not multi-day trekking ... no point-to-point multi-day walking product in the core Serengeti', with daily_hours of '1-3 hrs walking per day (rest of the day is vehicle-based)'. Independent verification found only that 'Night drives and guided walks are also offered at some camps and lodges' — no multi-day walking itinerary is documented anywhere. This is a vehicle-based game-drive safari with an optional stroll, filed in a batch whose modality is walking. Landscape tier 5 is also inflated: max 1,800 m of grassland against tier-5 anchors of Everest, Torres del Paine, Iceland's interior and Concordia. The migration is a wildlife spectacle, not a trek — if genuine multi-day walking is wanted in Tanzania, the Nyerere/Selous row already covers it and is confirmed to offer real walking plus boat safaris.
Botswana: Okavango Delta mokoro-and-walk
NOT A TREK, and not tier 5. The row self-refutes: 'LODGE/CAMP STAY WITH MOKORO POLING AND SHORT ISLAND WALKS (1-3 hrs), not multi-day point-to-point trekking'. Wikivoyage confirms the delta is experienced 'either by boat, airplane or mokoro', that lodge packages consist of 'two wildlife rides a day', and explicitly does not mention any established multi-day walking trail through the delta. The one concrete priced product found is a 2-hour mokoro pole at BWP 500, seasonal February to August only — which also contradicts the row's own best_months of May-September. This is a water-based wildlife product; the walking is incidental. Botswana's e-visa requirement for Thai nationals is confirmed and should be carried into the ground layer, but the row itself does not belong in a trek catalogue.
Sweden: Kungsleden in winter (Abisko–Nikkaluokta ski/snowshoe touring)
Duplicate. 'Kungsleden' is named in the do-not-re-survey list and this row covers the same Abisko-Alesjaure-Tjaktja-Salka-Kebnekaise-Nikkaluokta line already filed, differing only by season and by travelling it on skis with a pulk. The STF winter hut-season window (mid-Feb to end April) and the winter hut rate belong as a seasonal note on the existing Kungsleden row, not as a second row; neither could be verified live this session in any case (swedishtouristassociation.com returned only a booking filter interface, so the SEK900/night figure behind trek_lodging_thb_pp_night 3,064 is unsourced).
Norway: Lofoten winter walking with aurora
Duplicate and not a trek. 'Lofoten' is named in the do-not-re-survey list. The row is also self-refuting as a trek: its own drawbacks field says 'this is aurora-hunting plus short walks, not a multi-day backcountry trek', it files trek_days 3 / trek_nights 2 at max_altitude 300m in hotels, and the priced product is a 5-day Nordic Visitor sightseeing package spanning two separate archipelagos (Lofoten and Tromso). The package price could not be verified either — the Nordic Visitor tour URL returned HTTP 404. It also breaches the transit floor (13h each way requires days_min 9; the row files 8).
Norway: Senja winter walking with aurora
Duplicate and no verifiable product. 'Senja' is named in the do-not-re-survey list. Beyond that the row has no sourced price of any kind — its only source is the Wikipedia article on the island, it carries no package price and no trek-block cost fields, and its own drawbacks field calls it the 'weakest-priced row in this pair — no package figure to quote'. Like the Lofoten row it is hotel-based day walking (trek_days 3, max_altitude 400m) rather than a multi-day trek, and it breaches the transit floor (13h each way requires days_min 9; the row files 8).
Iceland: Landmannalaugar winter super-jeep excursion (labelled — vehicle-based, not backcountry trekking)
Not a trek. The row files trek_days 1, trek_nights 0, difficulty 'easy', guide_mandatory true and lodging_style 'hotel', and describes itself as 'an 8-12h round trip by super-jeep with short walks on-site, not a multi-day trek'. Guide to Iceland confirms the product is an 8-hour day tour running 'All year'. A vehicle day-excursion from a Reykjavik hotel base does not belong in a catalogue of treks, and the row's own text concedes the summer Laugavegur/Landmannalaugar trekking rows are already filed. The specific ISK32,990 price could not be re-verified either (adventures.is page 404'd; Guide to Iceland displayed 'From 0').
Nepal: Everest Base Camp in winter (Dec-Feb, honest conditions)
Duplicate. 'EBC' is the first entry in the batch brief's do-not-re-survey list, and this row is the same trek on the same Lukla-Namche-Gorak Shep-Kala Patthar route, differing only by season. The brief's instruction is explicit: 'If a candidate below turns out to be one of those, exclude it as a duplicate.' Nothing is lost by dropping it — the verified winter-season facts (Nepal Hiking Team USD1,525pp for 2-6 with a 1:2 porter ratio, the winter Lukla cancellation risk, the no-waiver VOA tier) are carried into ground_additions so they can be folded into the existing EBC row's notes. Note the row also files trek_days 12 where the operator states 11 trekking days.
Chile: Torres del Paine (Patagonia) in winter — CLOSED for the O/W circuit
Duplicate and not a trek. 'TdP O circuit' is named in the do-not-re-survey list, and the row itself rates feasibility 'closed' for the backcountry circuits, reducing the offer to a 3-day hotel-based luxury day-hike stay at max_altitude 400m — which is not a trek. The closure claim is additionally unverifiable: the cited torreshike.com FAQ returned no winter content, CONAF's park page 404'd, Fantastico Sur's winter page 404'd, and Las Torres Patagonia openly advertises a 'W Circuit 5 Days / 4 Nights' with no seasonal restriction stated. So the row is either a duplicate of an open filed trek or a closed non-trek; either way it should not stand as a separate catalogue entry.
USA: Grand Canyon South Rim to Phantom Ranch and back, via Bright Angel/South Kaibab
Route physically does not exist, and this is not a price call. Both verifiers independently, and an arbiter re-fetch of the NPS primary release, confirm: the 29 Aug 2026 Bright Angel Canyon flash flood destroyed nearly all footbridges on Bright Angel Creek and the Transcanyon Waterline intake, and NPS has closed until further notice, with 'currently no estimated reopening date': Phantom Ranch (cabins and canteen), Bright Angel Campground, the entire North Kaibab Trail and BOTH Colorado River crossings (Black Bridge and Silver Bridge); the Colorado is also closed to river traffic for debris. With neither bridge open there is no link from the South Kaibab descent to the Bright Angel ascent, so the row's signature loop cannot be walked, and the Xanterra lottery that functions as its permit cannot be entered for a closed facility. The only stated lower-effort fallback, the overnight mule ride, is separately cancelled through 28 Feb 2027. Drop for a 2027-28 atlas and re-survey only on an NPS bridge/Phantom Ranch reopening announcement; North America is then unrepresented, and the survey's own excluded list already names Canada (TRV, multi-week processing) as the alternative to resource properly. Source: https://www.nps.gov/grca/learn/news/significant-flash-flooding-closes-phantom-ranch-and-north-kaibab-trail.htm
Cambodia: Chi Phat - Cardamom Mountains community jungle trek
Dropped because the row's own load-bearing safety reassurance is refuted, not because of any price disagreement. The row argues Chi Phat is outside the trouble ('the Cardamom/Koh Kong sector has not itself been the site of the 2025-26 clashes'), but Koh Kong - the province Chi Phat sits in - is named BY NAME in Australia's Smartraveller advisory, in both the 'do not travel within 50 km of the Cambodia-Thailand border' band and the 'reconsider your need to travel, 50-80 km' band, with the stated reasons being armed conflict AND the ongoing presence of landmines/ERW - the same hazard the trek itself carries. Chi Phat measures roughly 65-75 km from the nearest Thai border point, i.e. inside the advisory, not outside it. The exposure is nationality-specific and therefore not transferable from a Western trip report: this is a THAI passport in a Cambodian border province during a live bilateral armed conflict with the traveller's own country - all six land crossings shut since late June 2025 with no reopening date, the 27 Dec 2025 ceasefire collapsed within days, Cambodia cut Thai nationals' visa-free stay as a retaliatory measure, Thailand has barred Cambodian nationals from leisure travel (making a reciprocal measure foreseeable), and there is no land exit if air links wobble. Both verifiers independently recommended the drop. Secondary and NOT the reason: the row's numbers are independently weak (package price unsourced and rebuilding to roughly a quarter to a third of the stated figure, boat and bus fares traceable to 2017-dated pages, PNH superseded by Techo/KTI on 8-9 Sep 2025, 'Cambodia Angkor Air' now Air Cambodia, the free mandatory e-Arrival card unmentioned, and a landmine warning attributed to 2026 guidance that one verifier could not find in any form). Nothing here is uniquely unavailable elsewhere in the region file: Nam Ha delivers the same guided rainforest-and-village product on a 30-day visa with a settled permit regime and no conflict exposure. If the atlas reinstates the row, the corrected numbers are preserved in _unapplied and it must be gated on a live Thai MFA advisory check within 30 days of ticketing.
Faroe Islands
About 600 km of drivable road - too small to fill 10-12 days - and it needs a separate visa endorsement on top of Schengen.
Norway – Arctic Highway to North Cape
8 days, short of the 10-12 day frame, and its headline fare was refuted - it sat 38% below the same dataset's BKK-Oslo floor for a city reached via Oslo.
North Coast 500
Duplicate of the North Atlantic survey's Scotland row; kept once, with the two surveys' conflicting fuel and lodging prices reconciled.
Pamir Highway
GBAO is rated 'reconsider travel' by the US, Canadian and Australian governments, the route ends at a border requiring a one-way international rental most firms refuse, and half the cost inputs could not be sourced.
Gobi Desert Traverse
Not self-drive: the survey concedes it is realistically a guided 4x4 trip, and its 'van day rate' was a per-person tour price misfiled as a vehicle cost.
Taiwan Round-Island Route
Taiwan admits foreign IDPs only from countries with a bilateral reciprocity arrangement, and Thailand is named among those NOT accepted. A Thai licence holder cannot legally hire a car. (Taroko Gorge is also still largely closed after the 2024 earthquake.)
Spiti Valley Self-Drive Circuit
A foreign passport needs a Protected Area Permit for Spiti itself, which cannot be applied for online and is checked twice on the survey's own route. Unsafe to recommend at this tier.
Tasmania Circle
Cannot be ranked: fuel, lodging, food and SUV rates were all missing. Also, Australian rental firms almost universally forbid their cars on the Spirit of Tasmania ferry.
Argentine Lake District (Ruta 40)
200 km of driving at the end of 32-36 hours of flying each way. Not a road trip at this length.
Kanchanaburi to Sangkhlaburi and Three Pagodas Pass
760 km round trip. A superb long weekend, not a 10-12 day trip.

Trip	Why it was cut
Savannah Way & the Gulf: Atherton Tablelands, Undara & Karumba	Fails on pricing: the only trip in its region with no activities figure at all, while the Undara lava tubes - its entire reason to exist - are a paid guided tour.
Alsace Wine Route & the Vosges (Route des Crêtes)	550 km over 10 days is the lowest driving density in the atlas, and the survey concedes the region needs a cross-border day trip to avoid repetition. A superb 5-6 day trip, not this one.
The Ardennes: Belgium, Luxembourg & Champagne-Ardenne	700 km over 10 days at tier 2 - the survey calls it 'the weakest-scenery pick in this whole region', included for novelty rather than drama.
Amalfi Coast & Cilento	The SS163 Amalfitana runs odd/even plate restrictions all of August AND all of September daily, plus Easter, late April and summer weekends - which removes the shoulder-season window the trip is built on. Lodging was also understated by 30%+.
Castile's Walled Towns & the Sierra de Gredos	Not a 10-12 day drive: the signature route is ~550 km, padded to a claimed 950. Segovia to Toledo is a comfortable five days.
Serra da Estrela & the Historic Villages of Portugal	900 km over 10 days with Coimbra and Guarda only 160 km apart; the survey's own notes admit it does not fill 10-12 days.
Orkney — Neolithic Scotland by island-hop	300 km of island driving; Skara Brae, Brodgar, Stenness and the Churchill Barriers are 3-4 days. The rest is padding on tarmac the Scotland row already covers.
The Samoeng Loop by scooter — Mae Sa, the 1269 switchbacks and Mae Kampong	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Phu Thap Boek and the Khao Kho ridge — Route 12's waterfall road to a 1,794 m cabbage mountain	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Doi Mae Salong and Phu Chi Fa by motorbike — the 1130 tea ridge and the 1155 border road	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
The Death Highway to Umphang — 1,219 curves and Thi Lo Su	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Ha Giang Loop by Motorbike	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Northwest Vietnam: Mai Chau to O Quy Ho	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Hai Van Pass and the Son Tra Headland	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Phong Nha and the Ho Chi Minh Road West	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
North Lombok: Pusuk, Senaru and the Sembalun Rim	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Shimanami Kaido — the Seto Inland Sea island crossing	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
The Death Railway: Thonburi to Nam Tok	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Sleeper to the Ping Valley: Chiang Mai, Lamphun and Lampang	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
KTM ETS Spine: Penang, Ipoh, Kuala Lumpur, Malacca and Singapore	Confirmed duplicate. The row's non-duplication case rests on 'Ipoh, Kuala Lumpur, Tampin/Malacca and Gemas are absent from all 278 rows'. Ipoh is in two existing rows, one of which does the Butterworth-Ipoh cave-temple day BY RAIL ('the Ipoh day by rail, Butterworth-Ipoh 1h34-1h46 each way, 7 trains a day, for the cave temples, Concubine Lane and white coffee') - the same corridor by the same mode
Koh Tao by Night Bus and Catamaran	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Koh Phangan's East Coast Without the Full Moon	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Southern Laos by Bus: Wat Phou, the Bolaven and Four Thousand Islands	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Lampang and Phrae: the Teak Towns by Overnight Coach	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Khiriwong and Thale Noi: the Khao Luang Foothills to the Lotus Lake	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Koh Lipe and the Tarutao Archipelago	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Koh Chang to Koh Kood by Inter-Island Boat	TRIPS_TO_DROP, confirmed duplicate: the row sleeps on Koh Chang AND Koh Kood, the exact two islands already held by 'Koh Chang with the Car on the Ferry' and 'Koh Kood — Drive and Ferry to the Last Quiet Island'. The brief's mode exception covers a road famous for a mode, not a seasonal passenger speedboat used as a transfer. Verifier: 'Do not score as written.' (Its other findings — the 2,420 THB
The Gulf Chain: Koh Tao, Koh Phangan, Ang Thong and Samui	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Similan and Surin Liveaboard: Richelieu Rock	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Raja Ampat: Waigeo, Piaynemo and the Dampier Strait	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Khao Luang from Khiriwong	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Sapa Homestay Trek, Muong Hoa Valley	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Low's Peak: the Mount Kinabalu Summit Climb	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Certify on Koh Tao	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Vietnam end to end on the Reunification Express	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Satun: Koh Lipe, the Adang-Rawi group and Tarutao	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Bukit Lawang and Tangkahan: the east door of Gunung Leuser	Confirmed duplicate — 'DROP OUTRIGHT. Hard duplicate of Lake Toba and the Karo Highlands, whose signature_route opens Medan -> Bukit Lawang (orangutans) -> Berastagi...'. Bukit Lawang is not adjacent to the incumbent, it is the incumbent's first stop, and the days_max 7 extension adds Berastagi's Sipiso-piso and Sibayak, which is verbatim the incumbent's middle section. Neither exception applies:
Koh Tao: learn to dive from Chumphon	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Mongolia: Ulaanbaatar, Terelj and the South Gobi	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Southern Laos by motorbike: the Bolaven Plateau and Si Phan Don	Cross-hole arbitration: duplicate of a kept row, or dead premise. See data/gapfill-arbitration.json.
Angkor Overland: Driving to Cambodia's Temple Plain	Dead premise per the auditor's DROPS finding, not a price error. Thailand closed every land crossing with Cambodia in late June 2025 after the border-temple conflict; the 27 Dec 2025 ceasefire broke down in early Jan 2026 and all official crossings including Aranyaprathet/Poipet remained closed through 2026 with no announced reopening date, cross-border bus services suspended. The trip's entire and only stated value proposition - the only trip in the band with zero airport days - is therefore gone, and the survey's own fallback (fly BKK-Siem Reap, rent locally) is a different trip that the survey itself rates the scenically weakest row in the band at tier 2. The printed route also runs to Preah Vihear, one of the specific temple areas the fighting has been over, so the safety position is untenable as well. Consistent with the brief's verified ground truth (FCDO, 19 March 2026).
Kashmir to Ladakh Himalayan Crossing	Dead premise on safety, not length. FCDO advises against ALL travel to the Union Territory of Jammu and Kashmir, naming Sonamarg, Srinagar city and the Jammu-Srinagar national highway explicitly, with the only exception being air travel to and movement within Jammu city — the route's entire first three days (fly Srinagar -> Sonamarg -> Kargil) sit inside that zone. FCDO also advises against all travel within 10 km of the Line of Control with Pakistan and the Line of Actual Control with China, which catches four more named stops: Turtuk (~10 km from the LoC), Pangong Tso (the LAC runs through the lake) and Hanle (~20 km from the LAC). Compounding: 14 printed vs 16 real days, and the Leh taxi-union rule already breaks the self-drive premise. The auditor notes a Ladakh-only fly-in version (BKK-DEL-Leh, acclimatise 3 nights, Nubra, Pangong, fly out) is a legitimate 12-day trip but is a different trip and must be re-surveyed under its own name. Source: https://www.gov.uk/foreign-travel-advice/india/regional-risks
Jeju Island Loop	Auditor (drops): the non-stop the entire 5-day case rests on is not verifiable and may no longer exist - FlightConnections returns 'no direct flights' in BOTH directions (BKK-CJU and CJU-BKK), Trip.com routes everything via Busan/Incheon/Gimpo, and only FlightsFrom still shows a non-stop, at 2 flights a week all departing 23:20 (a red-eye, not the daytime arrival the itinerary is written around). The survey's 'roughly 17 flights a month' is contradicted by its own cited source. The fallback via Seoul is 7-8h each way, which is the same arithmetic the band used to exclude Guilin and Goa - so the 5-day premise is dead as written. The auditor's separate K-ETA verdict also states outright that 'the Jeju row it supports should be dropped on flight grounds'. Reinstate only if the non-stop is re-verified directly on jejuair.net, and then as a red-eye trip with three usable days (a rebuild, not this row).
Ladakh: Leh and Nubra (season-limited)	Auditor's explicit drop for this band: 'drop from the 6-day band (keep it at 8-10)... because the risk stack is not survivable at this length'. Taken as a drop rather than a day-range flag because the decisive element is a safety position the row does not carry, not the day count: an active, recurring civil-unrest flashpoint omitted from safety_notes entirely (lethal statehood protests in Leh 24 Sept 2025 - 4 dead, 150+ injured - a 22-day curfew with mobile-internet suspension and mass cancellations, and a 'Ladakh Peace and Justice March' set to reach Leh on 24 Sept 2026, in peak season), stacked on a single morning-only DEL-IXL window with no ground alternative, a four-month season, and a foreigner-only agent-arranged PAP gating day 4. At 6 days there is no slack to absorb any one of them. The row survives at 8-10 days in its own band; this drop removes it only from d6-asia.
Jordan: Amman to Wadi Rum on the King's Highway	Auditor's explicit instruction: 'Drop from the 8-day band; revisit only after a fresh check.' The safety position, not the price, is untenable — IRGC ballistic missiles hit Muwaffaq Salti Air Base near Azraq on 31 Aug 2026 and again on 8 Sep 2026 (four days before the survey); Safe Airspace rates Jordan 'Level Three – Caution'; France NOTAM LFFF F1813/26 tells French operators not to enter Jordanian airspace, Germany NOTAM EDWW B0678/26 recommends avoiding FIR Amman until 22 Oct 2026, and EASA CZIB 2026-08-R1 is open on Jordan. The row's own recommended Doha/Abu Dhabi routings compound it (repeated Qatari airspace closures, 2,000+ cancellations at Hamad, EASA advising avoidance of Qatari/Bahraini/Kuwaiti/UAE airspace at all altitudes), and an 8-day trip with two hard transit days and no slack cannot absorb a Gulf-hub diversion. The row's safety_notes present a June 2026 easing as the current state.
Pakistan: Ratti Gali Lake and the Neelum Valley Trek	Not a trek, and not landscape tier 4 as a walking route - it fails the survey's own exclusion standard. The operator selling it describes reaching the lake by vehicle: Day 2 travellers go 'traversing difficult terrain and winding jeep paths to reach the lake's gorgeous shoreline', and Day 3 opens 'Board jeeps to return to Dowarian'. The inclusion list is a road-tour list - 'Guided sightseeing tour according to itinerary above', 'SUV/mini-van/car (depending on number of people)', 'Driver', 'All entry/admission tickets', 'Bottled water in car', 'Toll & parking fees' - with accommodation and all meals explicitly excluded, and the product is filed under Islamabad city tours ('xmlhi-holiday-info-islamabad-tours'). This is exactly the profile the survey used to exclude Kaghan Valley / Lake Saiful Muluk ('not tier 4-5 as a trek: the lake itself is jeep/short-walk accessible') and Kel-Suu ('not a trek: reached by private 4x4'). The row's own drawbacks field already concedes 'the trekking component is genuinely short (one day); most of the trip length is overland road travel, not walking'. Its trek block (trek_days 2, trek_nights 1, guide_day_thb 417, permits 60 THB, lodging 240, food 180) is also the only itemised Pakistani trek block in the survey and is built on an informal PKR guide-fee range rather than any bookable product. Drop it rather than carry a jeep excursion in a catalogue of breathtaking treks.
Tanzania: Kilimanjaro via the Lemosho Route	Duplicate of a filed atlas row - 'Tanzania: Kilimanjaro summit via the Machame Route' is already in the atlas. Wikipedia lists Lemosho (56 km, 6-8 days) and Machame (49 km, 6-7 days) as two of seven official routes on the same mountain, both characterised as a 'forested traverse to Barafu', converging at Barafu Camp for an identical final push to Uhuru Peak (5,895 m). Same park, same gateway (JRO/Moshi), same TANAPA permit regime, same mandatory-guide rule, same summit, roughly five of eight days on shared ground - the row itself concedes it converges 'from Barranco onward'. Under the brief's own test ('the Three Passes is different from EBC; a 10-day EBC is not'), this is a route variant of a filed trek, not a genuinely different one. If the atlas nonetheless wants Kilimanjaro route coverage, the honest form is a variant note on the existing Machame row (extra Shira Plateau days, camping-only, higher park-fee load) rather than a second tier-5 row. Note also that its package price of 102,300 THB was an unsourced estimate - the operator page cited, ultimatekilimanjaro.com, 404'd for me as it did for the surveyor.
Thailand: Khao Chang Phueak savanna-summit trek, Thung Yai Naresuan/Umphang Wildlife Sanctuary	The trek described does not exist. Khao Chang Phueak is the highest peak of THONG PHA PHUM National Park in KANCHANABURI at 1,249 m (the highest point in Kanchanaburi), confirmed on both English and Thai Wikipedia - it is not in Thung Yai Naresuan or Umphang Wildlife Sanctuary, it is not 1,955 m, and it is nowhere near Mae Sot or Umphang. Consequently the gateway (BKK-Mae Sot flights, 2,200-4,800 THB), the 3,000 THB Mae Sot-Umphang-trailhead 4WD extras, the 'sanctuary trekking-quota permit with fixed departure dates', the Nov-Jan sanctuary season and the Myanmar-border malaria/safety note are all attached to the wrong protected area and would feed pure fiction into the cost model. The row is unfixable by patching - every logistics field is wrong. It is also duplicative as framed, since the atlas already holds 'Umphang and Thi Lo Su - the Death Highway'. A corrected Thong Pha Phum/Kanchanaburi row (road access from Bangkok, no flight, DNP booking) could be re-surveyed from scratch, but a 1,249 m two-day grass-ridge walk should be re-tested against the tier-4 bar before it is re-admitted.
Thailand: Doi Pha Hom Pok – Thailand's second-highest peak, Fang hill-tribe trek	Not a trek, by the survey's own exclusion rule. The elevation claim is confirmed (2,285 m, Thailand's second-highest), but the trek is not: the summit is reached by hiring a 4WD to Kiew Lom Campsite and then hiking 3.5 km, with an overnight at that vehicle-accessible campsite to catch the sunrise. That is a drive-up campsite plus a short walk - materially the same product the survey itself excluded as 'not a trek: a drive-up sea-of-mist viewpoint and campsite' (Kaeng Krachan/Phanoen Thung). The row's 3-day/2-night demanding ascent, Lisu/Lahu homestay night, summit camping night and 5-7 hours per day are unsourced and contradicted. On the atlas scale (tier 4 = Dolomites hut circuits, Kumano Kodo, Drakensberg) a 3.5 km walk from a road-head does not clear the bar. If a genuine multi-day hill-tribe approach route from Fang exists, it needs to be evidenced before the row returns.
India: Amarnath Yatra (Pahalgam route, Kashmir Himalaya)	Duplicate. The atlas already holds 'India: the Amarnath Yatra cave pilgrimage, Pahalgam route' (days_recommended 10, feasibility 'permit') in trips/roadtrip-atlas-2027/data/atlas-trek-world.json and costed.json - same cave, same Pahalgam route, same gateway. The JKSASB 2026 findings verified here (registration from 15 Apr, RFID mandatory, CHC on/after 8 Apr 2026, all routes a No Flying Zone from 1 Jul 2026 so no helicopters, a formal foreign-national procedure, and the 2026 season suspended from 9 Aug) should be folded into that existing row instead.
India: Kedarnath and Tungnath-Chandrashila Himalayan pilgrim treks (Garhwal, Uttarakhand)	Duplicate. The atlas already holds 'India: Kedarnath and the Char Dham Pilgrim Walk (Gaurikund to Kedarnath, with Vasuki Tal)' (days_recommended 8, feasibility 'permit'). The only addition here is the 5km-each-way Chopta-Tungnath-Chandrashila side walk, which does not justify a second row. Carry the verified correction to the existing row: Chandrashila is 3,690m, not ~4,000m, and the Chopta walk is 3.5km to Tungnath plus 1.5km to the summit (10km round trip, not 7km).
China: Emei Shan Golden Summit pilgrim stairs (Sichuan)	Duplicate. The atlas already holds 'China: Mount Emei - the Pilgrim Stairway to the Golden Summit' (days_recommended 6, feasibility 'open'), already present in costed.json. The verified 2026 prices (entrance CNY160/2 days, CNY110 15 Dec-15 Jan; cable car CNY60 up, CNY55 down; Leidongping bus CNY90 return; summit 3,099m; CNY 1 = 4.958 THB on 2026-09-15) should be applied to that row - and note the proposed extras figure of 743 THB silently dropped the CNY55 descent leg; bus plus both cable-car legs is CNY205 = 1,016 THB.
Turkey: Lycian Way full traverse (Fethiye to Antalya)	Duplicate of an already-filed trail, not a new landscape. The atlas holds 'Turkiye: Lycian Way, Fethiye to Kas' (10 days, 'open') and 'Lycian Coast: Antalya, Olympos, Kas, Fethiye' (10 days, route_km 850), with plan files under trips/roadtrip-atlas-2027/plans/. A 30-day thru-hike of the same waymarked route is a length variant. The route length is also contested (Wikivoyage 540km; Wikipedia's lead now says ~760km, its infobox 520km), so the day estimate is unsound as well. Keep the one verified gain for the existing row: the trail's high point is genuinely ~1,800m (Incegeris Hill 1,811m), passing just east of Tahtali Dag's 2,300m summit.
France: Vallee Blanche off-piste glacier route and Mer de Glace (TECHNICAL - not a non-technical walk)	Not a trek. The Vallee Blanche is a lift-served off-piste ski descent or roped mountaineering objective from the Aiguille du Midi, not a walking route - the only source the row cites (Wikipedia's Mer de Glace article) describes the Vallee Blanche purely as 'broad icefields that feed into the Glacier du Tacul', a glacier system, and does not characterise it as any kind of route. The row's own feasibility_notes concede it 'does not fit the non-technical framing of this batch and should be filtered accordingly', and its trek block is trek_days 1 / trek_nights 0 / lodging_style hotel / difficulty demanding: a single-day technical day activity from a Chamonix hotel. It also carries no sourced pricing whatsoever - every Compagnie des Guides and cable-car fetch failed - and its 'hundreds of extra stairs each year' drawback is unsupported by the cited source. Nothing in it is trek-shaped, priced, or verified. Note that the Mont Blanc massif is already represented in the atlas by TMB full, Haute Route and Monte Rosa, so nothing is lost.
Malaysia: Danum Valley rainforest (Borneo Rainforest Lodge)	NOT A TREK. The row self-refutes: 'THIS IS A LODGE STAY WITH GUIDED WALKS, not multi-day point-to-point trekking — 2-3 hr guided outings from a fixed lodge, returning to the same bed each night'. Wikivoyage confirms the only walking is on 'poorly labeled trails ... for 1-7 hr of walking time' radiating from a fixed base, most requiring ranger accompaniment, with no point-to-point route. Access is confirmed restricted to BRL or the Field Centre with two enforced checkpoints, which further precludes any independent multi-day route. Pricing also failed verification across two passes: the only figure retrievable is a January-2017 DVFC dorm rate (RM96) that Wikivoyage itself calls 'heavily overpriced' and which is not representative of BRL; no 2026 rate card exists at any reachable URL. Lowland dipterocarp forest viewed from a canopy walkway is also a weak tier-4 claim against Kumano Kodo and the Drakensberg.
Kenya: Loita Hills Maasai forest walk	NOT TIER 4, and the product cannot be shown to exist. The row self-refutes on landscape: 'Scenically modest compared with every other row in this batch' — which is a direct admission that it fails the tier-4 bar set by the Dolomites' hut circuits, Kumano Kodo and the Drakensberg. It also self-refutes on bookability: 'almost entirely undocumented for foreign booking this session — real risk the trek as imagined doesn't exist as a packaged product at all'. Independent verification confirms the worst case: the Wikipedia article is a stub with no elevation, no area, and 'no mention of organized walking safaris, trekking products, or tourism activities', recording only that 'the Loita Community Wildlife Conservancy operates in the Loita Hills'. No operator, no price, no permit rule, no season and no group minimum could be evidenced from any source. A row that is admittedly unremarkable to look at and cannot be shown to be bookable fails both drop tests at once.
Kenya: Chyulu Hills walking safari (ol Donyo Lodge)	NOT A TREK, and not tier 5. The row self-refutes twice in capitals: 'THIS IS A LODGE STAY WITH WALKS, not multi-day trekking ... guests sleep in the same suite each night', with daily_hours of '1-4 hrs of walking depending on which walk is booked'. The USD 1,590 pp/night rate is confirmed accurate against Great Plains Conservation's own page and is fully inclusive — but a two-night stay at 52,470 THB/night consumes 104,940 THB of a 120-140k THB whole-trip budget for two nights of optional strolls. The Kilimanjaro view is genuine; a view is not a trek, and rolling volcanic hills topping out at 2,188 m are not tier 5. Dropped on modality and tier, not on price.
Tanzania: Serengeti walking safari (lodge/mobile camp with short walks)	NOT A TREK, and not tier 5. The row self-refutes in its own comfort_note and drawbacks: 'LODGE/MOBILE-CAMP STAY WITH SHORT WALKS ONLY (1-3 hrs, escorted, usually near camp), not multi-day trekking ... no point-to-point multi-day walking product in the core Serengeti', with daily_hours of '1-3 hrs walking per day (rest of the day is vehicle-based)'. Independent verification found only that 'Night drives and guided walks are also offered at some camps and lodges' — no multi-day walking itinerary is documented anywhere. This is a vehicle-based game-drive safari with an optional stroll, filed in a batch whose modality is walking. Landscape tier 5 is also inflated: max 1,800 m of grassland against tier-5 anchors of Everest, Torres del Paine, Iceland's interior and Concordia. The migration is a wildlife spectacle, not a trek — if genuine multi-day walking is wanted in Tanzania, the Nyerere/Selous row already covers it and is confirmed to offer real walking plus boat safaris.
Botswana: Okavango Delta mokoro-and-walk	NOT A TREK, and not tier 5. The row self-refutes: 'LODGE/CAMP STAY WITH MOKORO POLING AND SHORT ISLAND WALKS (1-3 hrs), not multi-day point-to-point trekking'. Wikivoyage confirms the delta is experienced 'either by boat, airplane or mokoro', that lodge packages consist of 'two wildlife rides a day', and explicitly does not mention any established multi-day walking trail through the delta. The one concrete priced product found is a 2-hour mokoro pole at BWP 500, seasonal February to August only — which also contradicts the row's own best_months of May-September. This is a water-based wildlife product; the walking is incidental. Botswana's e-visa requirement for Thai nationals is confirmed and should be carried into the ground layer, but the row itself does not belong in a trek catalogue.
Sweden: Kungsleden in winter (Abisko–Nikkaluokta ski/snowshoe touring)	Duplicate. 'Kungsleden' is named in the do-not-re-survey list and this row covers the same Abisko-Alesjaure-Tjaktja-Salka-Kebnekaise-Nikkaluokta line already filed, differing only by season and by travelling it on skis with a pulk. The STF winter hut-season window (mid-Feb to end April) and the winter hut rate belong as a seasonal note on the existing Kungsleden row, not as a second row; neither could be verified live this session in any case (swedishtouristassociation.com returned only a booking filter interface, so the SEK900/night figure behind trek_lodging_thb_pp_night 3,064 is unsourced).
Norway: Lofoten winter walking with aurora	Duplicate and not a trek. 'Lofoten' is named in the do-not-re-survey list. The row is also self-refuting as a trek: its own drawbacks field says 'this is aurora-hunting plus short walks, not a multi-day backcountry trek', it files trek_days 3 / trek_nights 2 at max_altitude 300m in hotels, and the priced product is a 5-day Nordic Visitor sightseeing package spanning two separate archipelagos (Lofoten and Tromso). The package price could not be verified either — the Nordic Visitor tour URL returned HTTP 404. It also breaches the transit floor (13h each way requires days_min 9; the row files 8).
Norway: Senja winter walking with aurora	Duplicate and no verifiable product. 'Senja' is named in the do-not-re-survey list. Beyond that the row has no sourced price of any kind — its only source is the Wikipedia article on the island, it carries no package price and no trek-block cost fields, and its own drawbacks field calls it the 'weakest-priced row in this pair — no package figure to quote'. Like the Lofoten row it is hotel-based day walking (trek_days 3, max_altitude 400m) rather than a multi-day trek, and it breaches the transit floor (13h each way requires days_min 9; the row files 8).
Iceland: Landmannalaugar winter super-jeep excursion (labelled — vehicle-based, not backcountry trekking)	Not a trek. The row files trek_days 1, trek_nights 0, difficulty 'easy', guide_mandatory true and lodging_style 'hotel', and describes itself as 'an 8-12h round trip by super-jeep with short walks on-site, not a multi-day trek'. Guide to Iceland confirms the product is an 8-hour day tour running 'All year'. A vehicle day-excursion from a Reykjavik hotel base does not belong in a catalogue of treks, and the row's own text concedes the summer Laugavegur/Landmannalaugar trekking rows are already filed. The specific ISK32,990 price could not be re-verified either (adventures.is page 404'd; Guide to Iceland displayed 'From 0').
Nepal: Everest Base Camp in winter (Dec-Feb, honest conditions)	Duplicate. 'EBC' is the first entry in the batch brief's do-not-re-survey list, and this row is the same trek on the same Lukla-Namche-Gorak Shep-Kala Patthar route, differing only by season. The brief's instruction is explicit: 'If a candidate below turns out to be one of those, exclude it as a duplicate.' Nothing is lost by dropping it — the verified winter-season facts (Nepal Hiking Team USD1,525pp for 2-6 with a 1:2 porter ratio, the winter Lukla cancellation risk, the no-waiver VOA tier) are carried into ground_additions so they can be folded into the existing EBC row's notes. Note the row also files trek_days 12 where the operator states 11 trekking days.
Chile: Torres del Paine (Patagonia) in winter — CLOSED for the O/W circuit	Duplicate and not a trek. 'TdP O circuit' is named in the do-not-re-survey list, and the row itself rates feasibility 'closed' for the backcountry circuits, reducing the offer to a 3-day hotel-based luxury day-hike stay at max_altitude 400m — which is not a trek. The closure claim is additionally unverifiable: the cited torreshike.com FAQ returned no winter content, CONAF's park page 404'd, Fantastico Sur's winter page 404'd, and Las Torres Patagonia openly advertises a 'W Circuit 5 Days / 4 Nights' with no seasonal restriction stated. So the row is either a duplicate of an open filed trek or a closed non-trek; either way it should not stand as a separate catalogue entry.
USA: Grand Canyon South Rim to Phantom Ranch and back, via Bright Angel/South Kaibab	Route physically does not exist, and this is not a price call. Both verifiers independently, and an arbiter re-fetch of the NPS primary release, confirm: the 29 Aug 2026 Bright Angel Canyon flash flood destroyed nearly all footbridges on Bright Angel Creek and the Transcanyon Waterline intake, and NPS has closed until further notice, with 'currently no estimated reopening date': Phantom Ranch (cabins and canteen), Bright Angel Campground, the entire North Kaibab Trail and BOTH Colorado River crossings (Black Bridge and Silver Bridge); the Colorado is also closed to river traffic for debris. With neither bridge open there is no link from the South Kaibab descent to the Bright Angel ascent, so the row's signature loop cannot be walked, and the Xanterra lottery that functions as its permit cannot be entered for a closed facility. The only stated lower-effort fallback, the overnight mule ride, is separately cancelled through 28 Feb 2027. Drop for a 2027-28 atlas and re-survey only on an NPS bridge/Phantom Ranch reopening announcement; North America is then unrepresented, and the survey's own excluded list already names Canada (TRV, multi-week processing) as the alternative to resource properly. Source: https://www.nps.gov/grca/learn/news/significant-flash-flooding-closes-phantom-ranch-and-north-kaibab-trail.htm
Cambodia: Chi Phat - Cardamom Mountains community jungle trek	Dropped because the row's own load-bearing safety reassurance is refuted, not because of any price disagreement. The row argues Chi Phat is outside the trouble ('the Cardamom/Koh Kong sector has not itself been the site of the 2025-26 clashes'), but Koh Kong - the province Chi Phat sits in - is named BY NAME in Australia's Smartraveller advisory, in both the 'do not travel within 50 km of the Cambodia-Thailand border' band and the 'reconsider your need to travel, 50-80 km' band, with the stated reasons being armed conflict AND the ongoing presence of landmines/ERW - the same hazard the trek itself carries. Chi Phat measures roughly 65-75 km from the nearest Thai border point, i.e. inside the advisory, not outside it. The exposure is nationality-specific and therefore not transferable from a Western trip report: this is a THAI passport in a Cambodian border province during a live bilateral armed conflict with the traveller's own country - all six land crossings shut since late June 2025 with no reopening date, the 27 Dec 2025 ceasefire collapsed within days, Cambodia cut Thai nationals' visa-free stay as a retaliatory measure, Thailand has barred Cambodian nationals from leisure travel (making a reciprocal measure foreseeable), and there is no land exit if air links wobble. Both verifiers independently recommended the drop. Secondary and NOT the reason: the row's numbers are independently weak (package price unsourced and rebuilding to roughly a quarter to a third of the stated figure, boat and bus fares traceable to 2017-dated pages, PNH superseded by Techo/KTI on 8-9 Sep 2025, 'Cambodia Angkor Air' now Air Cambodia, the free mandatory e-Arrival card unmentioned, and a landmine warning attributed to 2026 guidance that one verifier could not find in any form). Nothing here is uniquely unavailable elsewhere in the region file: Nam Ha delivers the same guided rainforest-and-village product on a 30-day visa with a settled permit regime and no conflict exposure. If the atlas reinstates the row, the corrected numbers are preserved in _unapplied and it must be gated on a live Thai MFA advisory check within 30 days of ticketing.
Faroe Islands	About 600 km of drivable road - too small to fill 10-12 days - and it needs a separate visa endorsement on top of Schengen.
Norway – Arctic Highway to North Cape	8 days, short of the 10-12 day frame, and its headline fare was refuted - it sat 38% below the same dataset's BKK-Oslo floor for a city reached via Oslo.
North Coast 500	Duplicate of the North Atlantic survey's Scotland row; kept once, with the two surveys' conflicting fuel and lodging prices reconciled.
Pamir Highway	GBAO is rated 'reconsider travel' by the US, Canadian and Australian governments, the route ends at a border requiring a one-way international rental most firms refuse, and half the cost inputs could not be sourced.
Gobi Desert Traverse	Not self-drive: the survey concedes it is realistically a guided 4x4 trip, and its 'van day rate' was a per-person tour price misfiled as a vehicle cost.
Taiwan Round-Island Route	Taiwan admits foreign IDPs only from countries with a bilateral reciprocity arrangement, and Thailand is named among those NOT accepted. A Thai licence holder cannot legally hire a car. (Taroko Gorge is also still largely closed after the 2024 earthquake.)
Spiti Valley Self-Drive Circuit	A foreign passport needs a Protected Area Permit for Spiti itself, which cannot be applied for online and is checked twice on the survey's own route. Unsafe to recommend at this tier.
Tasmania Circle	Cannot be ranked: fuel, lodging, food and SUV rates were all missing. Also, Australian rental firms almost universally forbid their cars on the Spirit of Tasmania ferry.
Argentine Lake District (Ruta 40)	200 km of driving at the end of 32-36 hours of flying each way. Not a road trip at this length.
Kanchanaburi to Sangkhlaburi and Three Pagodas Pass	760 km round trip. A superb long weekend, not a 10-12 day trip.