Align 1.2 research documentation and report figures

.gitattributes

@@ -199,3 +199,13 @@ evaluation/released_daily/model_1_2_benchmark.png filter=lfs diff=lfs merge=lfs
 docs/trackformer_1_2_mangkhut.gif filter=lfs diff=lfs merge=lfs -text
 evaluation/released_daily/model_1_2_after_2024_benchmark.png filter=lfs diff=lfs merge=lfs -text
 evaluation/released_daily/model_1_2_before_2024_benchmark.png filter=lfs diff=lfs merge=lfs -text
+docs/report_figures/day_1_to_5.png filter=lfs diff=lfs merge=lfs -text
+docs/report_figures/fung_wong_pressure_comparison.png filter=lfs diff=lfs merge=lfs -text
+docs/report_figures/fung_wong_track_pressure.png filter=lfs diff=lfs merge=lfs -text
+docs/report_figures/fung_wong_observed_r34.png filter=lfs diff=lfs merge=lfs -text
+docs/report_figures/mangkhut_pressure_comparison.png filter=lfs diff=lfs merge=lfs -text
+docs/report_figures/mangkhut_track_pressure.png filter=lfs diff=lfs merge=lfs -text
+docs/report_figures/mangkhut_observed_r34.png filter=lfs diff=lfs merge=lfs -text
+docs/report_figures/jebi_pressure_comparison.png filter=lfs diff=lfs merge=lfs -text
+docs/report_figures/jebi_track_pressure.png filter=lfs diff=lfs merge=lfs -text
+docs/report_figures/jebi_observed_r34.png filter=lfs diff=lfs merge=lfs -text

RELEASE_NOTES_TRACKFORMER_1_2.md

@@ -1,5 +1,7 @@
 # Trackformer 1.2 — pressure-field research model
 
+**Documentation clarification (18 October 2026):** the current [English model page](README.md) and [Chinese research explanation](docs/research_report_zh.md) follow the research report's released-1.2 scope, household-computer goal, day-1–5 chart and three typhoon examples. Route uses 270 storms / 1,473 starts; the JMA pressure comparison uses 40 of those storms / 134 starts / 2,637 common valid records. The report's R34 panels are observations, not model radius forecasts, and Taiwan rainfall remains separate ongoing research. This updates documentation and existing figure copies only; weights, neural source and benchmark arrays are unchanged. Earlier release history below is retained as an archive.
+
 **[Download weights + complete inference code](https://github.com/yu314-coder/typhoon-predict/releases/download/trackformer-1.2/trackformer_1_2_field_pressure_export_v2.tar.gz)** · **[Hugging Face](https://huggingface.co/euler314/typhoon-predict)** · **[Technical paper](https://github.com/yu314-coder/typhoon-predict/blob/main/paper/trackformer.pdf)**
 
 Trackformer 1.2 forecasts Western Pacific storm tracks, central pressure and evolving sea-level-pressure fields through +120 hours. A multiscale environmental-attention network conditions the moving pressure core; track and central pressure are read from that evolving field.

docs/report_figures/day_1_to_5.json

@@ -0,0 +1,59 @@
+{
+  "lead_hours": [
+    24,
+    48,
+    72,
+    96,
+    120
+  ],
+  "route_storms": 270,
+  "route_issues": 1473,
+  "track_error_km": {
+    "1.2": [
+      164.481223771445,
+      311.4789218238006,
+      502.4908306263511,
+      740.9340851986844,
+      1031.633570368749
+    ],
+    "deepmind": [
+      208.9564305523796,
+      357.1736913330866,
+      522.4673167277228,
+      746.7391690084348,
+      1031.9091421404373
+    ]
+  },
+  "central_pressure_mae_hpa": {
+    "1.2": [
+      10.882824059440974,
+      14.65584179709299,
+      15.699234927622467,
+      14.152747888401569,
+      12.444309436541925
+    ],
+    "deepmind": [
+      7.6849854707083125,
+      10.461160570285193,
+      13.205749144191225,
+      12.40939214066142,
+      11.023561874071152
+    ]
+  },
+  "pressure_common_issues_by_day": [
+    131,
+    133,
+    134,
+    133,
+    127
+  ],
+  "pressure_common_storms_by_day": [
+    40,
+    40,
+    40,
+    40,
+    39
+  ],
+  "aggregation": "Per exact lead: mean issues within each storm, then equal-weight mean storms; common finite JMA/1.2/DeepMind support for pressure.",
+  "deepmind_route_reconstruction": "Combine saved 230-storm and 40-storm equal-storm means with 230/270 and 40/270 weights; checked against published all-lead and +120h means."
+}

docs/report_figures/provenance.json

@@ -0,0 +1,41 @@
+{
+  "schema": "trackformer-1.2-report-figures-v1",
+  "documentation_update_date": "2026-10-18",
+  "scope": "Chinese research report plus approved text and figure amendments; existing figures copied without new forecasts or training",
+  "report_snapshot_sha256": "0e3f0ecd5c44bbd24a09697d954d8b99882935f181a7faa57081c5203387ba47",
+  "source_figure_audit_sha256": "140214e416da7057a301a74cbd0c901f8a059e89dd63ab18e76fb82673d3fc3b",
+  "source_checkpoint_sha256": "f194a23d3f91ea76ad776dfad942fabd669eeae8b3fd665815463095367e9ee0",
+  "inference_weights_sha256": "db49f36e85a3766defc4c172746897a1f783705d1ce8e6f9dfb8e87ae1d902cb",
+  "forecast_members_per_case": 50,
+  "map_lead_hours": 48,
+  "map_comparison_grid_degrees": 2.5,
+  "route_and_pressure_leads_hours": [6, 12, 18, 24, 30, 36, 42, 48, 54, 60, 66, 72, 78, 84, 90, 96, 102, 108, 114, 120],
+  "reference_pressure": "NOAA NCEP analysis; source remains analysis despite simplified visible labels",
+  "reference_track_and_central_pressure": "IBTrACS / JMA exact valid times",
+  "coastline": "Natural Earth",
+  "radius_source": "IBTrACS WP v04r01 USA_R34 quadrant records; observations only, not model predictions",
+  "ibtracs_sha256": "47ba0d51dbd22d5fb3c93d50b35e11a74b911a4019c56d6bb3424a5992dbbbf0",
+  "radius_units": "km = nautical miles * 1.852; missing values remain missing",
+  "model_radius_head_available": false,
+  "model_weights_changed": false,
+  "model_or_truth_arrays_changed": false,
+  "coordinates_or_times_shifted": false,
+  "cases": [
+    {"name": "Fung-wong", "sid": "2025308N09144", "issue_utc": "2025-11-07T00:00:00Z", "forecast_backend": "saved Metal", "saved_comparison_npz_sha256": "975e87bb5d40fe8bd081bc2ec890d1917892cb830bbef6972ce20772c787f050"},
+    {"name": "Mangkhut", "sid": "2018250N12170", "issue_utc": "2018-09-11T00:00:00Z", "forecast_backend": "saved Metal", "in_fitting_years": true, "saved_comparison_npz_sha256": "28b39089b8f03750773dba36c71bb9bf6505e5156a4bbdeb274624240210ee1f"},
+    {"name": "Jebi", "sid": "2018239N11161", "issue_utc": "2018-08-31T00:00:00Z", "forecast_backend": "CPU", "in_fitting_years": true, "saved_comparison_npz_sha256": "f47385cae42b8e6c1e77e1b93d1f078b41ad2b080732a3283a4e4b62ee2044a4"}
+  ],
+  "files_sha256": {
+    "day_1_to_5.json": "57da3a3f7c0ced7509a4892dc7028683cc3cd6c44bf156219d62eda80c25bfd2",
+    "day_1_to_5.png": "fd99254582a474b9ce2472fe18a3546572182fb466ed61022a2b230e861bda74",
+    "fung_wong_pressure_comparison.png": "9d8cd5d90e0751c3a7da6b152a34213905047430d2722965cdfebff57f98a78f",
+    "fung_wong_track_pressure.png": "d9793a66f39e1cbac9d055149f79765b891844a716068613437b057a5a29b6c8",
+    "fung_wong_observed_r34.png": "1b7855065f39cb19502f8885d7744c035ee39f8064188e0f1b45b6a72d7d5125",
+    "mangkhut_pressure_comparison.png": "35d7878efd15bb656d8104c257bdbd9e65c3026ded28a001a2a840023e69a885",
+    "mangkhut_track_pressure.png": "d2b4febc480af3d82007fd1a10a40fbe4f29c6858913d4a666d86d336da99af8",
+    "mangkhut_observed_r34.png": "ad6712a85acb4e367e1c6f06bdd6a2b3ac455f8deabb54d132647745ba871c85",
+    "jebi_pressure_comparison.png": "1413c6f6a91e1a7843f6047cd167d8608d84fab70a7ed3f554afe1e1a44cc324",
+    "jebi_track_pressure.png": "d0bb7e6d170ab471a309657baeb205ab56fac2ba5ca7bfae6ed861336573a18f",
+    "jebi_observed_r34.png": "b1582f95506b62e559d7be58c24a4c307c727e8b6cd37e0f64a976ccb26fa5d7"
+  }
+}

docs/research_report_zh.md

@@ -0,0 +1,128 @@
+# TrackFormer 1.2 中文研究說明
+
+**作品名稱：訓練AI模型探討西北太平洋氣壓系統影響臺灣附近颱風的預測評估**
+
+本頁對照中文研究報告及本次核定的文字、圖表修改。只介紹發布的 **TrackFormer 1.2**；原有技術論文與歷史比較紀錄保留於各自的檔案，不取代本頁的報告範圍。
+
+[GitHub 程式與研究紀錄](https://github.com/yu314-coder/typhoon-predict) · [Hugging Face 模型](https://huggingface.co/euler314/typhoon-predict) · [即時預測頁](https://trackformer-weatherlab.rudin-euler-8253.chatgpt.site/live) · [English overview](../README.md)
+
+## 研究動機與目標
+
+- 以一般家用電腦可訓練、可執行為目標，讓個人及學生能參與天氣預測研究。
+- 建立預測西北太平洋氣壓圖的 AI 模型，再從同一預測氣壓取得颱風路徑與中心氣壓。
+- 結合不同高度的水平風，提供研究東北風、西南風與颱風周圍氣流變化的基礎。本次成果評分以路徑與氣壓為主，並非已證明季風預測改善。
+
+## 研究設備與器材
+
+| 設備或工具 | 實際用途與範圍 |
+| --- | --- |
+| Apple M4 Mac、24 GB 統一記憶體 | 本機模型執行、資料處理與製圖；透過 PyTorch MPS 使用內建 GPU 與 Metal。家用電腦訓練與執行是研究目標，不代表每一次歷史計算都只使用同一部電腦。 |
+| NVIDIA GeForce RTX 3070、CUDA | 已完成的 DeepMind 輕量版比較預測使用此設備；RTX 3070 是硬體，不是軟體。 |
+| 外接儲存裝置與網路 | 保存資料、模型、預測及圖表，取得公開資料並分享成果。 |
+| Python、PyTorch | 建立模型、計算誤差、更新參數與執行預測。 |
+| NumPy、SciPy、xarray、netCDF4 | 數值運算，依時間及經緯度整理天氣網格。 |
+| Matplotlib | 繪製氣壓、路徑、曲線與長條圖。 |
+| XeLaTeX、xeCJK | 中文報告排版。 |
+
+## 資料來源與分配
+
+| 資料 | 模型實際使用的內容 |
+| --- | --- |
+| NOAA NCEP/NCAR Reanalysis 1 | 海平面氣壓、500 百帕高度，以及 850、500、200 百帕的東西向風 u 和南北向風 v。u、v 共同提供風向與風速。 |
+| ARCO-ERA5 氣壓 | 1,000 個具有原生 0.25 度氣壓資料的案例；以起報中心向四方各取 15 度，共 121×121 格點。 |
+| NOAA IBTrACS 颱風紀錄 | 位置、先前移動、可用的中心氣壓及最大風，缺值保留有效性標記。中心氣壓比較採日本氣象廳紀錄。 |
+| 固定地理資料 | 經緯度、陸地比例與地形高度。技術輸入同時保留環境及中心附近的地理資訊，並非未來天氣。 |
+
+環境資料涵蓋北緯 0–60 度、東經 100–180 度，間距 2.5 度，共 25×33 格點。較細氣壓案例是下表各組的子集，不另加到案例總數。
+
+| 用途 | 年份 | 歷史資料片段 | 其中有原生較細氣壓資料 |
+| --- | --- | ---: | ---: |
+| 訓練與正規化 | 2000–2021 | 13,949 | 800 |
+| 驗證與模型選擇 | 2022–2023 | 1,041 | 100 |
+| 原始測試分組 | 2024–2025 | 1,195 | 100 |
+
+同一颱風保持在同一分組；跨越分組界線的資料片段不使用。資料的平均值與變動範圍只由訓練組計算。後來取得的 2026 年資料不表示發布模型重新訓練到 2026 年。[來源與檔案檢查](trackformer_1_2_training_data.md)。
+
+## 模型如何預測與訓練
+
+1. 輸入起報前兩天至起報時刻的資料，每六小時一筆，共 **9 筆**，搭配當時的颱風位置、強度、先前移動與固定地形。
+2. 分別計算周圍氣壓、水平風與颱風中心附近氣壓的變化，再一起更新六小時後的狀態。
+3. 從預測氣壓找出颱風中心，再讀取同一中心的氣壓，得到氣壓圖、路徑與中心氣壓。
+4. 把預測狀態接著往下計算，共 **20 步、+6 至 +120 小時**。
+5. 訓練時與之後的真實紀錄比較並更新參數；預測時不使用未來觀測。訓練預報長度由一天逐步增加到兩天、三天及五天；報告使用選定的第四輪模型。
+
+模型有 21,452,595 個參數，中心氣壓計算網格為 65×65、約 20 公里。這是計算與重建間距，不是新取得的 20 公里原生觀測。兩天是本次模型固定的輸入設計，尚未以相同條件比較一天、兩天及三天，不能說兩天已被證明最佳。[完整技術輸入與輸出](../models/trackformer_1_2_field/README.md)。
+
+### 為什麼有 50 組預測
+
+報告使用**同一組模型參數**，只在正規化後的歷史天氣加入 50 組不同的小幅、平滑、平均為零的雜訊，再取平均。環境與區域資料的雜訊幅度分別為 0.025、0.015，平滑範圍分別為 5、9 格；幅度是正規化單位，並非直接加上相同數字的百帕。颱風起報位置、移動、強度、地形與有效性標記不變，原始觀測和模型參數也不修改。
+
+路徑與中心氣壓分別平均；氣壓地圖則先把各組結果放到相同經緯度，再平均。這**不是訓練 50 個模型**，也未證明 50 組是最佳數量、一定優於單次預測，或其分散程度已是可靠的不確定性範圍。
+
+## 多個颱風的比較
+
+路徑使用 **270 個颱風、1,473 次起報**，比較每六小時、未來五天的預測。先平均一次起報內各時間的有效誤差，再平均同一颱風的起報，最後讓每個颱風具有相同權重。
+
+**中心氣壓比較使用270個颱風中符合共同評分條件的40個颱風，共134次起報。將本研究模型與DeepMind輕量版模型在相同預報時間的中心氣壓預測，分別與日本氣象廳紀錄比較，共有2,637筆共同有效紀錄。**
+
+這不代表其餘 230 個颱風都沒有氣壓觀測；本次共同子集沿用原評估流程的起報強度有效條件。缺值不當成零誤差。
+
+比較模型是官方 **WeatherNext Cyclones Mini `<2024`**，訓練資料到 2023 年，原生間距 1 度，使用一個預測成員及軟體 v0.3.0，並非 DeepMind 完整大型模型。TrackFormer 使用 50 組輸入擾動平均，兩者保留各自資料流程，不是等算力的模型結構實驗。
+
+![第一至第五天路徑與中心氣壓誤差](report_figures/day_1_to_5.png)
+
+第一至第五天分別指 **24、48、72、96、120 小時**，不是一天內所有時段的平均。路徑為 270 個颱風；氣壓依各時間的共同有效紀錄使用 39–40 個颱風，樣本數列於圖下。[逐日數據](report_figures/day_1_to_5.json)。
+
+| 比較項目 | TrackFormer 1.2 | DeepMind 輕量版 |
+| --- | ---: | ---: |
+| 平均路徑誤差（公里） | 471.20 | 498.37 |
+| 第五天路徑誤差（公里） | 1031.63 | 1031.91 |
+| 六小時移動方向誤差（度） | 34.96 | 39.76 |
+| 中心氣壓平均絕對誤差（百帕） | 12.84 | 10.20 |
+
+這組開發評估的平均路徑及移動方向誤差較低，中心氣壓誤差則以比較模型較低。全組包含 230 個 1980–1999 年颱風及 40 個較近期颱風；歷史年份與比較模型的訓練期間重疊。只看 2025–2026 年起報的 18 個颱風、61 次起報，DeepMind 輕量版路徑誤差為 258.7 公里，TrackFormer 為 469.4 公里。已反覆查看的案例屬開發證據，不是全新、未使用的測試，不能由全組平均推論普遍優於 DeepMind。[原始完成數據與流程](deepmind_daily_benchmark.md)。
+
+## 三個颱風展示
+
+三案皆以相同發布模型的 50 組預測平均展示。氣壓圖比較 +48 小時、相同座標及色階下的模型與 **NCEP 參考氣壓**，採 2.5 度共同網格；NCEP 仍為分析資料，不因圖中文字簡化而變成測站觀測。路徑與中心氣壓保留全部 +6 至 +120 小時，以 IBTrACS／日本氣象廳紀錄對照，不搬移預測低壓來配合實況。
+
+### 鳳凰（Fung-wong，2025）
+
+起報：**2025-11-07 00 UTC**。使用保存的 Metal 結果。
+
+![鳳凰48小時氣壓對照](report_figures/fung_wong_pressure_comparison.png)
+
+![鳳凰五日路徑與中心氣壓](report_figures/fung_wong_track_pressure.png)
+
+![鳳凰IBTrACS觀測34節風圈半徑，非模型預測](report_figures/fung_wong_observed_r34.png)
+
+### 山竹（Mangkhut，2018）
+
+起報：**2018-09-11 00 UTC**。使用保存的 Metal 結果；位於訓練年份內，為回溯展示。
+
+![山竹48小時氣壓對照](report_figures/mangkhut_pressure_comparison.png)
+
+![山竹五日路徑與中心氣壓](report_figures/mangkhut_track_pressure.png)
+
+![山竹IBTrACS觀測34節風圈半徑，非模型預測](report_figures/mangkhut_observed_r34.png)
+
+### 燕子（Jebi，2018）
+
+起報：**2018-08-31 00 UTC**。相同發布權重在 CPU 執行的 50 組預測；未重新訓練，也未使用起報以後的天氣輸入。位於訓練年份內，為回溯展示。
+
+![燕子48小時氣壓對照](report_figures/jebi_pressure_comparison.png)
+
+![燕子五日路徑與中心氣壓](report_figures/jebi_track_pressure.png)
+
+![燕子IBTrACS觀測34節風圈半徑，非模型預測](report_figures/jebi_observed_r34.png)
+
+三張風圈圖均是 **IBTrACS 的 34 節四方向觀測半徑，不是本模型半徑預測**。左為 +48 小時四方向最大範圍，右為五日紀錄；海里乘以 1.852 換成公里，缺值保留空白。扇形不是完整風場。發布模型沒有原生風圈半徑預測頭。[圖片雜湊與來源](report_figures/provenance.json)。
+
+## 結論、成果特色與後續研究
+
+- **研究結論：**以家用電腦可訓練、可執行為核心目標，完成五日氣壓、颱風路徑及中心氣壓的整合流程，並可在一般 Mac 電腦執行預測與評估。
+- **同一套預測：**地圖、路徑與中心氣壓來自同一模型結果，可相互對照。
+- **周圍天氣與颱風中心一起更新：**同時保留環境氣壓、風場及移動中的中心氣壓計算。
+- **彩色圖表與清楚資料流程：**以相同時間、座標比較三個颱風，保留資料年份、案例數、缺值與來源紀錄。
+- **應用：**提供個人及學生進行氣象學習、歷史颱風比較與地圖展示的研究工具。這不是取代官方警報的作業預報服務。
+- **持續研究：**結合臺灣島外風向、水氣輸送與地形，研究一般降雨和強降雨的分布，並以降雨觀測另行評估。**發布的 1.2 不包含降雨預測頭或水氣輸入**，本段不是已完成的降雨準確度成果。

docs/showcase_archive.md

@@ -2,6 +2,12 @@
 
 The [model announcement](../README.md) features the geographically registered moving-core Mangkhut (2018) forecast. Its 50-member arrays and replay/playback receipts are linked there. These are development illustrations, not representative or untouched-test evidence.
 
+## Current research-report figures
+
+The [Chinese report explanation](research_report_zh.md#三個颱風展示) contains the current **Fung-wong (2025), Mangkhut (2018) and Jebi (2018)** pressure-map, five-day route/central-pressure and observed-radius panels. The day-1–5 bar chart compares released 1.2 with DeepMind Mini only, using unchanged saved scores. [Figure hashes and source identities](report_figures/provenance.json).
+
+The report's radius figures are **IBTrACS observed R34 quadrant records, not model predictions**. The pressure reference remains NCEP analysis, with a common 2.5° comparison grid and no shifted forecasts. Mangkhut and Jebi are in the training years; the examples are retrospective demonstrations, not additional untouched validation. Older films and experimental radius diagnostics below keep their original provenance and should not be confused with these report figures.
+
 ## Additional films
 
 These earlier 50-member films retain a fixed issue-relative regional pressure patch. When the centre leaves it, only the saved coarse basin field covers that position. They use 4 hPa isobars and 23-second playback with a held ending; do not confuse them with the current moving-core Mangkhut rendering. Their original model outputs remain preserved.

docs/trackformer_1_2_evaluation.md

@@ -4,28 +4,37 @@
 
 The current comparison uses **1,473 daily forecasts across 270 Western Pacific storms**, with 20 six-hour leads from +6 through +120 h. Scores average leads within issues, daily issues within each storm, then storms equally. The completed results and all forecast SHA-256 checks are in [the final report](../evaluation/daily_storm_final.json); the [daily-storm protocol](daily_storm_benchmark.md) defines the frozen cohort and metrics.
 
-Trackformer 1.2 uses the mean of 50 distinct causal input perturbations, not 50 trained networks. Trackformer 1.1 uses its archived causal route pipeline. These are matched forecast outputs with different input pipelines, not a controlled architecture ablation or an untouched generalization test.
+The current [Chinese research-report view](research_report_zh.md) compares **released 1.2 with DeepMind Mini**, using unchanged saved results. TrackFormer uses the mean of 50 distinct causal input perturbations of one model, not 50 trained networks; Mini uses one seeded member. These are matched outputs with different inputs and member policies, not a controlled architecture ablation, equal-compute test or untouched generalization test.
 
-| Equal-storm metric | 1.1 | 1.2 mean of 50 |
+| Report metric | TrackFormer 1.2 | DeepMind Mini |
 | --- | ---: | ---: |
-| Mean track error | 798.4 km | 471.2 km |
-| +120 h track error | 1,646.4 km | 1,031.6 km |
-| Six-hour direction error | 51.58° | 34.96° |
-| Centred shape similarity | 0.7544 | 0.8837 |
-| Geographic path similarity | 0.5345 | 0.6560 |
+| Mean track error | 471.20 km | 498.37 km |
+| +120 h track error | 1031.63 km | 1031.91 km |
+| Six-hour direction error | 34.96° | 39.76° |
+| Central-pressure MAE · JMA | 12.84 hPa | 10.20 hPa |
 
 Direction, centred shape, geographic path and same-time position error measure different aspects of a forecast. A translated or mistimed route can have high shape similarity without overlapping the observed route. No combined weighted ranking is introduced.
 
-The original track run's 1.2-only central-pressure MAE is **12.62 hPa on 40 storms with valid labels**; area-weighted basin MSLP MAE is **2.72 hPa on 270 storms**. The completed native 1.1 follow-up now supplies a matched pressure comparison on **134 common daily starts / 40 storms** from the same frozen plan. The default pressure chart is MAE: JMA **13.53 / 12.84 hPa**, USA **12.84 / 12.55 hPa** for 1.1 / 1.2, with paired uncertainty including zero. The optional same-time pressure-curve similarity diagnostic uses **(1 + centred cosine) / 2**, with no lag optimization or time warping: JMA scores are **0.7074 / 0.7118**, while USA scores are **0.7237 / 0.6637** on 131 non-flat eligible starts. Undefined curves remain unavailable, not zero. Both paired whole-storm similarity-difference intervals also include zero. This shape metric removes level and amplitude, so inspect the actual hPa overlays and MAE separately. [Shared verified snapshot](../evaluation/released_daily/released_daily_benchmark.json) · [Pressure protocol](intensity_benchmark.md) · [Public API](https://trackformer-weatherlab.rudin-euler-8253.chatgpt.site/api/benchmarks/released).
+**Central-pressure coverage is a subset of route coverage:** **40 of the 270 storms / 134 starts / 2,637 common valid lead records**. Each model's pressure is compared separately with the JMA record at the identical forecast time. This is not a separate 270-storm pressure evaluation. The original 1.2-only pressure mean of **12.62 hPa** uses a different mask and must not replace the report's common-support **12.84 hPa**. The original basin MSLP mean of **2.72 hPa on 270 storms** is a whole-map measure, not central-pressure error. [Common scores](../evaluation/released_daily/released_daily_benchmark.json) · [Pressure protocol](intensity_benchmark.md).
 
-The other **1,339 starts lack valid frozen 1.1 issue-time intensity inputs**, so this is not 270-storm pressure coverage. The original pressure-only mask is not substituted into the new common-mask comparison. A basin-wide average does not establish cyclone-core accuracy. The frozen cohort contains 40 recent and 230 historical storms from 1980–1999; retrospective analyses, complete-five-day eligibility and prior model selection limit interpretation.
+The inherited shared-case eligibility excludes **1,339 starts with unavailable frozen legacy issue-time intensity inputs**. That does **not** mean all excluded storms lack pressure observations. Original masks remain unchanged; unsupported scores are unavailable, not zero. The cohort contains 40 recent and 230 historical storms from 1980–1999. Retrospective analyses, complete-five-day eligibility and prior model selection limit interpretation.
+
+The [report's day-1–5 chart](report_figures/day_1_to_5.png) shows **exact 24, 48, 72, 96 and 120 h** leads rather than daily-bin averages. Track covers 270 storms at each displayed lead. JMA pressure has 40, 40, 40, 40 and 39 storms, with 131, 133, 134, 133 and 127 common starts respectively. All-lead pressure totals and per-lead counts describe different aggregations. [Exact chart values](report_figures/day_1_to_5.json).
 
 ## Completed DeepMind comparison
 
-Google's official **WeatherNext Cyclones Mini `<2024`**, using **WeatherNext software v0.3.0**, completed the same **1,473 daily starts / 270 storms** on an RTX 3070 with CUDA. Mini is one seeded member, not the full-sized WeatherNext model; 1.2 remains a 50-member mean. The common-support results are **498.4 km** mean track error, **39.76°** direction error and **10.20 hPa** JMA central-pressure MAE on **134 starts / 40 storms**. Its JMA pressure-curve similarity is **0.8088**.
+Google's official **WeatherNext Cyclones Mini `<2024`**, using **WeatherNext software v0.3.0**, completed the same **1,473 daily starts / 270 storms** on an **NVIDIA GeForce RTX 3070 with CUDA**. It is a 1° model trained through 2023. Mini is one seeded member, not the full-sized WeatherNext model; 1.2 remains a 50-member mean. The common-support results are **498.37 km** mean track error, **1031.91 km** at +120 h, **39.76°** direction error and **10.20 hPa** JMA central-pressure MAE on **134 starts / 40 storms / 2,637 valid records**. Its JMA pressure-curve similarity is **0.8088**.
 
 Mini's recent-only mean track error is **288.5 km** versus **477.2 km** for 1.2. Historical fitting-year overlap and different input/member policies limit the combined-cohort comparison; no general superiority claim is justified. The [completed protocol and period breakdowns](deepmind_daily_benchmark.md), [CUDA receipt](../evaluation/deepmind_daily/verification.json) and [independent three-model publication audit](../evaluation/deepmind_daily/publication_audit.json) establish completion. The older two-model receipt remains the original Trackformer audit, not the hash receipt for the subsequently merged snapshot. No missing or unsupported score is treated as zero.
 
+That **recent-only** group includes 2024. The report-facing strict **after-2024** group is different: 2025–2026, 61 starts / 18 storms, with **258.7 km Mini versus 469.4 km 1.2**. Calendar 2024 is not part of that strict subset. Do not interchange these period scores.
+
+## Report showcases and preserved earlier records
+
+The [three report examples](research_report_zh.md#三個颱風展示) use unchanged released-1.2 weights: Fung-wong and Mangkhut are saved Metal forecasts; Jebi was run on CPU. All use 50 input-perturbed members and exact +6 to +120 h times. Mangkhut and Jebi are within the fitting years. Their R34 panels show **IBTrACS observations only**, not model radius skill. NCEP is still an analysis reference even where the figure label is simplified to “reference pressure”.
+
+Earlier release comparisons and their pressure-curve diagnostics remain in the [archived shared snapshot](../evaluation/released_daily/released_daily_benchmark.json), [full original report](../evaluation/daily_storm_final.json) and [intensity protocol](intensity_benchmark.md). They are not silently substituted for the current report's two-model table. Centred curve similarity removes level and amplitude; read it with hPa MAE, masks and original timelines rather than treating it as accurate pressure levels.
+
 ## Recent pressure-line example
 
 ### Fung-wong MP4: selected route-and-pressure example

docs/trackformer_1_2_training_data.md

@@ -13,6 +13,8 @@ The **released 1.2 model's fitting cutoff is 2021**, not 2025 or 2026. Its selec
 
 The eight weather channels are **MSLP, HGT500, U850, V850, U500, V500, U200, V200**. No satellite cloud imagery, humidity/vapour, SST or ocean-heat-content channel is present in this released input contract. Such features in other experiments must not be attributed to 1.2.
 
+The eastward/northward **u/v pairs encode wind direction and speed together**, not direction alone. These environmental winds evolve during the model rollout; their presence is not a separate claim of validated monsoon or surface-wind skill. The [Chinese research explanation](research_report_zh.md) uses these same source variables and partition counts.
+
 The selected dataset manifest identifies the **derived IBTrACS archive by SHA-256**, not by an independently recorded raw IBTrACS release/download ID. A newer downloaded IBTrACS file elsewhere is not evidence of which raw snapshot built these frozen windows. Available wind labels and the learned auxiliary head do not establish a harmonized agency averaging period or a validated surface-wind/radius forecast.
 
 ## Temporal partitions
@@ -27,6 +29,8 @@ These are loader-eligible window counts, not independent storms or the number of
 
 Every forecast history comprises nine consecutive six-hour analyses at **−48, −42, …, 0 h**. Supervised fields and storm labels at **+6, +12, …, +120 h** are targets only; they cannot be inference inputs. Missing native detail and missing intensity labels retain their masks. Validation is not gradient fitting, but it does inform checkpoint selection. Previously inspected original tests and showcases are development evidence, not a new unused holdout.
 
+The two-day span is the input format chosen for this implementation. No matched 24/48/72-hour experiment establishes it as the best history length. Likewise, the evaluation's 50 input-perturbed forecasts share one set of weights; the count is not a training-data multiplier or proof of a calibrated ensemble.
+
 ## Newer data do not change the fitting cutoff
 
 The checksum-pinned caches contain records beyond the fitting years:

evaluation/README.md

@@ -1,5 +1,11 @@
 # Evaluation assets
 
+## Research-report view
+
+The current [report-facing summary](../README.md#results-in-the-research-report) and [Chinese explanation](../docs/research_report_zh.md) compare **TrackFormer 1.2 with DeepMind Mini**. Their [day-1–5 chart](../docs/report_figures/day_1_to_5.png) uses exact 24/48/72/96/120 h leads. Route covers **270 storms / 1,473 starts**; central pressure uses **40 of those storms / 134 starts / 2,637 common valid lead records**. The pressure subset follows frozen scoring eligibility, not a claim that the remaining storms lack all pressure observations. Source arrays and the larger historical archive below are unchanged.
+
+## Full historical archive
+
 The current **completed three-model comparison** scores Trackformer 1.1, Trackformer 1.2 and Google's **WeatherNext Cyclones Mini `<2024` (software v0.3.0)** on **1,473 daily starts / 270 Western Pacific storms**, through +120 h. Mini's RTX 3070 CUDA run is complete and verified; Trackformer 1.2 is a mean of 50 members and Mini is one member.
 
 - [Announcement chart](released_daily/model_1_2_benchmark.png) and [shared three-model scores](released_daily/released_daily_benchmark.json).

models/trackformer_1_2_field/README.md

@@ -14,6 +14,8 @@ python models/trackformer_1_2_field/predict.py causal_issue_packet.npz forecast.
 
 The `.npz` packet must contain exactly these arrays, with no object/pickle content:
 
+The nine history timestamps span **−48 to 0 h**, including both endpoints. This is the release's fixed two-day input format, not evidence that two days outperforms one or three. The eight basin channels are MSLP, HGT500, U850, V850, U500, V500, U200 and V200: each u/v pair represents **both horizontal wind direction and speed**. Static coordinates, land fraction and elevation remain part of the actual input contract below. See the [plain-language Chinese report](../../docs/research_report_zh.md) for the research motivation and household-computer equipment.
+
 | Key | Shape | Meaning |
 | --- | --- | --- |
 | `global_history` | `(9, 8, 25, 33)` | Past/issue-time basin analyses in manifest channel order, normalized with manifest means/stds. |
@@ -78,10 +80,12 @@ No model weights, learned architecture or existing benchmark results changed.
 
 **Causality:** Never populate arrays from positive-lead analysis, later best track, agency forecast or future pressure field. The wrapper checks timestamp order and shapes, but cannot authenticate the upstream archive. Do not silently replace missing values with future data. This is not a raw-weather downloader or live warning service.
 
-The published 50-member mean is a separate evaluation policy: 50 deterministic seeds perturb only normalized historical basin/regional fields with smooth zero-centred noise, then average routes, central pressures and common-grid fields. `predict.py` computes one clean forecast; do not label it a 50-member mean.
+The published 50-member mean is a separate evaluation policy: 50 deterministic seeds perturb only normalized historical basin/regional fields with smooth zero-centred noise, then average routes, central pressures and common-grid fields. The weights are the same for every member; this is **not 50 separately trained models**. Issue-time centre, motion, intensity, masks and static geography are not perturbed. `predict.py` computes one clean forecast; do not label it a 50-member mean. Neither this member count nor the 48-hour history has been established as optimal by a matched ablation; spread is not automatically calibrated uncertainty.
+
+**Report's matched development evaluation:** route uses the frozen **1,473 daily starts / 270 equal-weight storms**, +6 to +120 h. Central pressure uses **40 of those 270 storms**, **134 common starts / 2,637 valid lead records**, comparing each model separately with JMA at identical valid times. This is an inherited common-scoring subset, not evidence that all other storms lack pressure observations. Missing labels and unsupported outputs are not zero-scored. TrackFormer 1.2's mean route error is **471.20 km**, +120 h error **1031.63 km**, six-hour direction error **34.96°**, and JMA central-pressure MAE **12.84 hPa**. Central-pressure error is not whole-map error. See the [report figures](../../docs/research_report_zh.md), [frozen common scores](../../evaluation/released_daily/released_daily_benchmark.json), [pressure protocol](../../docs/intensity_benchmark.md) and [detailed evaluation notes](../../docs/trackformer_1_2_evaluation.md).
 
-**Matched development evaluation:** the released 1.1 and 1.2 route comparison uses the same frozen **1,473 daily starts / 270 equal-weight storms**, +6 to +120 h. The completed native-pressure comparison has **134 common starts / 40 storms**; the other 1,339 starts lack valid frozen 1.1 intensity inputs and are not zero-scored. The Site defaults to central-pressure MAE: **13.53 / 12.84 hPa** against JMA and **12.84 / 12.55 hPa** against USA (1.1 / 1.2), slight mean improvements with paired whole-storm uncertainty including zero. The optional curve-similarity diagnostic is **(1 + centred cosine) / 2 at exact common valid times**, without time shifting or warping. JMA similarity is **0.7074 / 0.7118**; USA is **0.7237 / 0.6637** on 131 eligible non-flat curves. Removing level/amplitude makes this a shape diagnostic, not proof of correct pressure levels. Actual unshifted hPa timelines, agency masks and uncertainty remain separate. See the [shared snapshot](../../evaluation/released_daily/released_daily_benchmark.json), [original Trackformer audit](../../evaluation/released_daily/released_daily_verification.json), [pressure protocol](../../docs/intensity_benchmark.md) and [public benchmark API](https://trackformer-weatherlab.rudin-euler-8253.chatgpt.site/api/benchmarks/released).
+**Completed DeepMind comparison:** Google's official **WeatherNext Cyclones Mini `<2024`**, run with **WeatherNext software v0.3.0**, completed all **1,473 daily starts / 270 storms** on an **NVIDIA GeForce RTX 3070 using CUDA**. This is the 1° Mini checkpoint trained through 2023, not the full-sized WeatherNext model. Mini uses one seeded member; TrackFormer 1.2 uses a 50-member mean. Mini's report values are **498.37 km** mean track error, **1031.91 km** at +120 h, **39.76°** direction error and **10.20 hPa** JMA central-pressure MAE on the **134 starts / 40 storms / 2,637 common valid records**. The combined cohort favours 1.2 for position and heading, but Mini has lower pressure MAE. Mini also has lower recent-only track error: **258.7 versus 469.4 km** on the strict 2025–2026 subset (61 starts / 18 storms). Different input/member policies and prior case inspection prevent general superiority, equal-compute or untouched-holdout claims. [Completed results and model identity](../../docs/deepmind_daily_benchmark.md) · [CUDA completion receipt](../../evaluation/deepmind_daily/verification.json) · [Independent publication audit](../../evaluation/deepmind_daily/publication_audit.json). No older benchmark's DeepMind scores were substituted.
 
-**Completed DeepMind comparison:** Google's official **WeatherNext Cyclones Mini `<2024`**, run with **WeatherNext software v0.3.0**, completed all **1,473 daily starts / 270 storms** on an **RTX 3070 using CUDA**. This is the 1° Mini checkpoint trained through 2023, not the full-sized WeatherNext model. Mini uses one seeded member; Trackformer 1.2 uses a 50-member mean. On shared support, Mini's mean track error is **498.4 km**, direction error **39.76°**, and JMA central-pressure MAE **10.20 hPa** on **134 starts / 40 storms**. Mini's recent-only track error is **288.5 km**, better than 1.2's **477.2 km**; the combined historical/recent result is not a general superiority claim. Missing outputs are unavailable, never zero errors. [Completed results and model identity](../../docs/deepmind_daily_benchmark.md) · [CUDA completion receipt](../../evaluation/deepmind_daily/verification.json) · [Independent three-model publication audit](../../evaluation/deepmind_daily/publication_audit.json). No older benchmark's DeepMind scores were substituted.
+**Rainfall and radius scope:** ongoing Taiwan rainfall work is separate from these weights. Released 1.2 has no rainfall head or humidity input. The report's three R34 diagrams are **observed IBTrACS quadrant radii**, not native 1.2 radius forecasts. Experimental pressure-derived radius outputs described above remain separately labelled diagnostics.
 
 **Pressure-display coverage:** `regional_mslp_hpa` remains the original fixed issue-relative composite. Use the separately exported `core_mslp_hpa` and its per-lead coordinates/mask when the storm moves away from that fixed patch. Do not extend it by inventing a vortex or moving it onto a route. If the moving core leaves supported basin geography, the corresponding masked cells remain unavailable. For an ensemble, register each physical member field onto a common geographic grid **before** averaging; do not average moving-frame array indices. The Mangkhut film uses that separate 50-member policy. Older fixed-patch films retain their original data in [the showcase archive](../../docs/showcase_archive.md). The automatic History archive and its separate core recovery expose geography and coverage through [the public API](https://trackformer-weatherlab.rudin-euler-8253.chatgpt.site/history-api).