Longhao Wang王龙浩

PhD candidate · IGSNRR, Chinese Academy of Sciences博士研究生 · 中国科学院地理科学与资源研究所Beijing · Updated November 2026北京 · 2026 年 11 月更新

LonghaoWang

I build deep-learning models that forecast the ocean, the rain and the drought.我用深度学习预测海洋、降水与干旱。

PhD candidate in physical geography at the Key Laboratory of Water Cycle & Related Land Surface Processes, IGSNRR, advised by Prof. Yongqiang Zhang. I work on sub-seasonal to seasonal (S2S) prediction, GRACE, evapotranspiration and physically constrained AI.中国科学院地理科学与资源研究所陆地水循环及地表过程院重点实验室,自然地理学博士研究生,导师张永强研究员。研究次季节—季节(S2S)预测、GRACE 重力卫星、蒸散发与物理约束 AI。

Currently近况

Forecasting预报
Live S2S outlooks for ocean and atmosphere on 在 canglong.earth 发布海洋与大气的 S2S 实时预报
Building构建
WeatherAI, 18 published weather models re-implemented in PyTorch:用 PyTorch 复现 18 个已发表的天气模型
Writing写作
《科研论文配图:从入门到发表》, Peking University Press (in preparation),北京大学出版社(撰写中)
Read the papers阅读论文
rain降水3D Transformer3D TransformerENSOENSOwarm tongue暖舌GRACE-FOGRACE-FOET蒸散发S2S forecastS2S 预测weeks周seasons季
The water cycle, as I try to forecast it.我想预测的水循环。

Published in发表于

  • Nature Geoscience
  • National Science Review
  • Advanced Science
  • Remote Sensing of Environment
  • Hydrology and Earth System Sciences
  • Journal of Hydrology
  • Agricultural and Forest Meteorology
16peer-reviewed journal articles篇期刊论文
6as first author篇为第一作者
394GitHub stars across my repositoriesGitHub 仓库星标
18weather models re-implemented in WeatherAI个天气模型在 WeatherAI 中复现

Publications论文

The record.论文发表。

Selected from ORCID and Crossref. My name is underlined.选自 ORCID 与 Crossref,本人姓名以下划线标出。

19 entries条

IF & JCR: 2025 data, released 2026 · JCR: best subject quartile · CAS: 2025 major-category tiers. Click a metric for its source.IF / JCR:2025 年数据,2026 年发布 · JCR 取学科最高分区 · 中科院:2025 年大类分区。点击指标查看来源。

2026

  1. Improved Skills for Predicting Sub-Seasonal to Seasonal Global Sea Surface Temperature

    Longhao Wang, Xuanze Zhang, Ruby L. Leung, Francis H. S. Chiew, Amir AghaKouchak, Kairan Ying, Yongqiang Zhang

  2. Predicting root-zone soil moisture using a small sample deep learning approach

    Yalin Xie, Jing Tian, Yongqiang Zhang, Xuanze Zhang, Longhao Wang, Haoshan Wei, Xian Wang, Zixuan Tang

  3. Satellite evidence of unequal basin-scale recovery from severe droughts

    Zhenwu Xu, Yongqiang Zhang, Günter Blöschl, Thorsten Wagener, Amir AghaKouchak, Xuanze Zhang, Longhao Wang, Jing Tian, Ning Ma, Xiaojie Li, Congcong Li, Qi Huang, Zixuan Tang, Xuening Yang, Jun Xia, Changming Liu

  4. EarthArXiv (preprint)

    Satellite Embedding: A Review

    Longhao Wang

    First author第一作者Preprint / abstract预印本 / 摘要doi:10.31223/X54R20
  5. A physics-driven hybrid transformer model for hydrologic simulation under nonstationary environmental conditions

    Hongwei Zhang, Zexing Tao, Kaiwen Wang, Gang Zhao, Jiewei Chen, Duanyang Xu, Ronggao Liu, Longhao Wang, Lei Wang, Quansheng Ge

  6. EGU General Assembly 2026 (abstract)

    Refining global water cycle components under a changing climate

    Yongqiang Zhang, Günter Blöschl, Haoshan Wei, Dongdong Kong, Ning Ma, Thorsten Wagener, Jing Tian, Jun Xia, Congcong Li, Longhao Wang, Francis H.S. Chiew, L. Ruby Leung, Xingcai Liu, Hongxing Zheng, Xuanze Zhang, Changming Liu

    Preprint / abstract预印本 / 摘要doi:10.5194/egusphere-egu26-4378
  7. Overestimation of past and future increases in global river flow by Earth system models

    Yongqiang Zhang, Günter Blöschl, Haoshan Wei, Dongdong Kong, Ning Ma, Thorsten Wagener, Jing Tian, Jun Xia, Congcong Li, Longhao Wang, Francis H. S. Chiew, L. Ruby Leung, Xingcai Liu, Hongxing Zheng, Xuanze Zhang, Changming Liu

2025

  1. Hydrology and Earth System SciencesJCR Q1CAS Tier 1中科院 1区IF 7.0

    Evapotranspiration stress intensifies with enhanced sensitivity to soil moisture deficits in a rapidly greening China

    Yuan Liu, Yong Wang, Yong Zhao, Shouzhi Chen, Longhao Wang, Wenjing Yang, Xing Li, Xinxi Li, Huimin Lei, Huanyu Chang, Jiaqi Zhai, Yongnan Zhu, Qingming Wang, Ting Ye

    25 citations · Crossref次引用 · Crossrefdoi:10.5194/hess-29-3379-2025
  2. EGU General Assembly 2024 (abstract)

    Skilful prediction of mid-term sea surface temperature using 3D self-attention-based neural network

    Longhao Wang, Yongqiang Zhang, Xuanze Zhang

    First author第一作者Preprint / abstract预印本 / 摘要doi:10.5194/egusphere-egu24-9446

2024

  1. Remote Sensing of EnvironmentJCR Q1CAS Tier 1中科院 1区IF 12.3

    Filling GRACE data gap using an innovative transformer-based deep learning approach

    Longhao Wang, Yongqiang Zhang

    First author第一作者36 citations · Crossref次引用 · Crossrefdoi:10.1016/j.rse.2024.114465
  2. Using hydrological modeling and satellite observations to elucidate subsurface and surface hydrological responses to the extreme drought

    Zixuan Tang, Yongqiang Zhang, Jing Tian, Ning Ma, Xiaojie Li, Dongdong Kong, Yijing Cao, Xuening Yang, Longhao Wang, Xuanze Zhang, Yuyin Chen

    8 citations · Crossref次引用 · Crossrefdoi:10.1016/j.jhydrol.2024.132174
  3. Satellite Reveals a Coupling between Forest Displacement and Landscape Fragmentation across the Economic Corridor of the Eurasia Continent

    Ying Wang, Li’nan Dong, Longhao Wang, Jiaxin Jin

  4. Understanding the Propagation of Meteorological Drought to Groundwater Drought: A Case Study of the North China Plain

    Yuyin Chen, Yongqiang Zhang, Jing Tian, Zixuan Tang, Longhao Wang, Xuening Yang

    26 citations · Crossref次引用 · Crossrefdoi:10.3390/w16030501

2023

  1. Geo-spatial Information ScienceJCR Q1CAS Tier 2中科院 2区IF 5.4

    A novel fuzzy inference method for urban incomplete road weight assignment

    Longhao Wang, Xiaoping Rui

    First author第一作者doi:10.1080/10095020.2023.2261768
  2. Non-Ignorable Differences in NIRv-Based Estimations of Gross Primary Productivity Considering Land Cover Change and Discrepancies in Multisource Products

    Jiaxin Jin, Weiye Hou, Longhao Wang, Songhan Wang, Ying Wang, Qiuan Zhu, Xiuqin Fang, Liliang Ren

  3. Agricultural and Forest MeteorologyJCR Q1CAS Tier 1中科院 1区IF 6.6

    Diverse responses of canopy conductance to heatwaves

    Longhao Wang, Yongqiang Zhang, Ning Ma, Peilin Song, Jing Tian, Xuanze Zhang, Zhenwu Xu

    First author第一作者27 citations · Crossref次引用 · Crossrefdoi:10.1016/j.agrformet.2023.109453
  4. Journal of System Simulation 系统仿真学报No JCR / CAS record无 JCR / 中科院分区

    Efficient HMM Map Matching Method Using R-tree and Trajectory Segmentation (采用R树和轨迹分段的HMM高效地图匹配方法)

2021

  1. A Weight Assignment Algorithm for Incomplete Traffic Information Road Based on Fuzzy Random Forest Method

    Longhao Wang, Jing Wu, Rui Li, Yanjiao Song, Jiayue Zhou, Xiaoping Rui, Hanwei Xu

    First author第一作者9 citations · Crossref次引用 · Crossrefdoi:10.3390/sym13091588
  2. Uncertainty Analysis of Multisource Land Cover Products in China

    Longhao Wang, Jiaxin Jin

    First author第一作者9 citations · Crossref次引用 · Crossrefdoi:10.3390/su13168857

Full list on 完整列表见 Google Scholar · ORCID

CV履历

Experience and recognition.履历与荣誉。

Education教育经历

  1. 2022.09 — present至今PhD, Physical Geography自然地理学 博士研究生Institute of Geographic Sciences and Natural Resources Research (IGSNRR), CAS中国科学院地理科学与资源研究所Advisor: Prof. Yongqiang Zhang导师:张永强研究员
  2. 2018.09 — 2022.06BSc, Geographic Information Science地理信息科学 理学学士Hohai University河海大学Advisor: Prof. Jiaxin Jin导师:金佳鑫教授

Talks学术报告

  1. EGU24 · Vienna · 2024.04 · Oral口头报告HS3.4 Deep learning in hydrology: Skilful prediction of mid-term sea surface temperature using 3D self-attention-based neural network
  2. GEWEX 2024 · Hokkaido · 2024.07A novel 3D Transformer model for predicting global SST at S2S scales
  3. AOGS 2025 · Singapore · 2025.07Skillful sub-seasonal to seasonal precipitation forecasting using the CAS-Canglong AI model
  4. AGU26 · San Francisco · 2026.12CAS-Canglong: A Physics-Guided AI Model for Global Subseasonal-to-Seasonal Prediction

Recognition & service荣誉与服务

  • 20262026 MAZU Global Innovation Partner for an AI hydrometeorological model with wind-field embedding and physical-equation constraints (with NSMC and the Department of Meteorology, Sri Lanka), 2026 MAZU Global Intelligence Innovation Application Challenge, China Meteorological Administration2026“妈祖”全球数智创新伙伴(嵌入风场并受物理方程约束的人工智能水文气象模型,与国家卫星气象中心、斯里兰卡气象局联合),2026年“妈祖”全球数智创新应用挑战活动,中国气象局
  • 2026Best Climate Solution Award (ranked 2/147) for CAS-Canglong, AI4Climate 2026 Global Youth Climate Solution Challenge, co-hosted by the DeepSeek CommunityAI4Climate 2026 全球青年气候解决方案挑战赛最佳气候解决方案奖(中科苍龙,2/147),深求社区等主办
  • 2026Young Elite Scientists Sponsorship Program (Doctoral Student Special Plan), China Association for Science and Technology (CAST)中国科协青年科技人才培育工程博士生专项计划,中国科学技术协会
  • 2025National Scholarship for Doctoral Students, Ministry of Education of the People’s Republic of China博士研究生国家奖学金,中华人民共和国教育部
  • 2022Outstanding Undergraduate Thesis of Jiangsu Province (Second Prize), Jiangsu Provincial Department of Education江苏省优秀本科毕业论文二等奖,江苏省教育厅
  • Reviewer for审稿服务Journal of HydrologyWater Resources ResearchScientific ReportsEnvironmental Science and Pollution ResearchJournal of Open Source Software
  • Software copyright · 2023SR0498336软件著作权 · 2023SR0498336Meteorological Drought Computing Platform V1.0 (气象干旱计算平台V1.0) — Y. Zhang, H. Wei, L. Wang, P. Miao, Dagula气象干旱计算平台V1.0 — 张永强、魏豪杉、王龙浩、苗平、达古拉
  • Book · in preparation专著 · 撰写中《科研论文配图:从入门到发表》 (Scientific figures for research papers: from beginner to publication), Peking University Press,北京大学出版社

Awards获奖

Two awards for CAS-Canglong in one autumn.中科苍龙,一季两奖。

In autumn 2026 the wind-field-embedded, physics-constrained AI hydrometeorological model behind CAS-Canglong took the Best Climate Solution Award at AI4Climate 2026 and was named a MAZU Global Innovation Partner by the China Meteorological Administration.2026 年秋,风场嵌入、物理方程约束的 AI 水文气象大模型“中科苍龙”获 AI4Climate 2026 最佳气候解决方案奖,并入选中国气象局 2026“妈祖”全球数智创新伙伴。

2026.10.28 · Ningbo, China中国宁波

Best Climate Solution Award最佳气候解决方案奖

AI4Climate 2026 Global Youth Climate Solutions ChallengeAI4Climate 2026 全球青年气候解决方案挑战赛

Ranked 2 / 147排名 2 / 147¥10,000 prize奖金 ¥10,000
Project项目
CAS-Canglong: a wind-field-embedded, physics-constrained AI hydrometeorological large model中科苍龙:一种风场嵌入和物理方程约束的AI水文气象大模型
Team团队
Longhao Wang; advisors Yongqiang Zhang and Xuanze Zhang; IGSNRR, CAS王龙浩;指导老师:张永强、张选泽;中国科学院地理科学与资源研究所
Hosts主办
Zhejiang Shen Kuo Science Popularization Foundation, Zhejiang Federation of Science Popularization and DeepSeek Community (DeepSeek.club), guided by WASD; part of the 2026 Sci2U Conference浙江省沈括科普基金会、浙江省科普联合会、深求社区(DeepSeek.club)主办,世界可持续发展科学院(WASD)指导;2026 Sci2U 科普大会核心活动

2026.11.07 · Shanghai, China中国上海

MAZU Global Innovation Partner2026“妈祖”全球数智创新伙伴

2026 MAZU Global Intelligence Innovation Application Challenge, China Meteorological Administration2026年“妈祖”全球数智创新应用挑战活动,中国气象局主办

CIIE Shanghai events第九届进博会上海会议活动WMO Region II · AsiaWMO 二区协 · 亚洲
Project项目
A large AI hydrometeorological model with wind field embedding and physical equation constraints一种嵌入风场并受物理方程约束的大型人工智能水文气象模型
Partners联合单位
IGSNRR, CAS; National Satellite Meteorological Centre; Department of Meteorology, Sri Lanka中国科学院地理科学与资源研究所、国家卫星气象中心、斯里兰卡气象局
Country面向国家
Sri Lanka斯里兰卡
Accepted by领奖
Longhao Wang and Xuanze Zhang王龙浩、张选泽

Contact联系方式

Let’s talk water, weather and AI.聊聊水、天气与 AI。

Open to collaboration on AI for hydrology, climate prediction and Earth system science.欢迎在水文 AI、气候预测与地球系统科学方向交流合作。

Email · primary邮箱 · 首选

wanglonghao0857@igsnrr.ac.cn
Alternate备用 · hhuwlh@163.com

WeChat official account微信公众号

地学AI实验室

Open source开源项目

Code, models, and demos.代码、模型与演示。

Everything here is public. Star counts are updated from GitHub.以下内容全部公开,星标数从 GitHub 更新。

Flagship · PyTorch weather model zoo旗舰项目 · PyTorch 天气 AI 模型库

WeatherAI

18 published weather models, one PyTorch zoo18 个已发表的天气模型,一个 PyTorch 模型库

Native PyTorch ports of published weather and climate models. Many strict-load the official weights and are checked numerically against the reference code.原生复现已发表的天气与气候模型(PyTorch),多数可严格加载官方权重,并与官方代码做了数值一致性校验。

  • Pangu
  • FuXi
  • FengWu
  • GraphCast
  • NeuralGCM
  • FuXi-ENS
  • StormCast
  • ArchesWeather
  • ACE2-ERA5
  • ORCA-DL
  • TropiCycloneNet
  • GenFocal
  • FuXi-S2S
  • UniCM
  • GenCast
  • Aardvark
  • WeatherNext Cyclones
  • Aurora
mesh GNN网格图网络cyclones台风ensembles集合预报
S2S SST modelS2S 海温模型

CAS-Canglong

The 3D Transformer behind the Advanced Science S2S sea surface temperature paper, now issuing monthly 16-month SST and ENSO outlooks.Advanced Science 次季节—季节海温论文中的 3D Transformer,现每月发布 16 个月海温与 ENSO 展望。

canglong.earth Hugging Face Space
Jupyter Notebook110
Learning hub学习资源

DataScience-master

Curated notes and code across machine learning, deep learning, Python, Linux, Pandas, Matplotlib and Git.机器学习、深度学习、Python、Linux、Pandas、Matplotlib、Git 等数据科学学习笔记与代码合集。

Notes26294
GRACE gap fillingGRACE 缺口填补

Transformer_GRACE

Filling the GRACE / GRACE-FO data gap with a Transformer (Remote Sensing of Environment, 2024).基于 Transformer 填补 GRACE / GRACE-FO 数据缺口(Remote Sensing of Environment, 2024)。

Jupyter Notebook181
Agent-ready climate viz智能体气候制图

IPCC

Source-grounded IPCC-style climate figures for Codex and Claude, with offline search, reproducible demos and traceable plotting conventions.面向 Codex 与 Claude 的 IPCC 风格气候可视化:离线检索、可复现示例、可追溯的制图规范。

Python190
Scientific plotting科研绘图

plot-function

Reusable plotting functions for publication-quality scientific figures.可复用的科研论文绘图函数。

Jupyter Notebook3617
Developer tool开发工具

skills-dashboard

A quick look at your agent skills, with easy download.一览并下载你的智能体 Skills。

TypeScript40

Models, datasets and Spaces.模型、数据集与 Spaces。

From weather-AI demos to a geoscience flux-tower LLM.从天气 AI 演示到地学通量塔大模型。

Model · QLoRA · Qwen3-14B模型 · QLoRA · Qwen3-14B

flux-gpt-qwen3-14b

A Chinese-language geoscience assistant for eddy-covariance and flux-tower questions, QLoRA-tuned on Qwen3-14B with the companion flux-sft dataset. Apache-2.0.面向涡度相关与通量塔问题的中文地学助手,基于 Qwen3-14B 用 QLoRA 微调,配套数据集 flux-sft,Apache-2.0 许可。

geoscienceeddy-covarianceflux-towerzh
Space · Gradio

CAS-Canglong live forecastCAS-Canglong 实时预报

The 16-month global SST and ENSO outlook, with the IRI plume, regional outlooks and NetCDF downloads.16 个月全球海温与 ENSO 展望,含 IRI 预测羽对比、分区展望与 NetCDF 下载。

CAS-Canglong
Space · Gradio

GraphCast demoGraphCast 演示

The WeatherAI GraphCast port, running as a Gradio Space.WeatherAI 中 GraphCast 复现的 Gradio 演示。

weatherai-graphcast-smoke
Space · Gradio

Flux-GPT demoFlux-GPT 演示

Interactive demo for the geoscience flux-tower assistant.地学通量塔问答助手的交互演示。

gpt-flux-demo
Space · Gradio

Personal assistant个人助手

A chat assistant fine-tuned with LoRA on SmolLM2.基于 SmolLM2 LoRA 微调的对话助手。

longhaowang-assistant

Certifications认证

Badges and certificates.奖章证书。

Issued by Anthropic in September 2026. Each seal links to its credential.由 Anthropic 于 2026 年 9 月颁发,点击徽章查看证书。

View on LinkedIn在 LinkedIn 查看

Research研究方向

Where deep learning meets the water cycle.让深度学习遇见水循环。

Four threads run through my work, from predicting the ocean months ahead to filling gaps in satellite records.我的研究有四条主线:从提前数月预测海洋,到填补卫星观测记录的缺口。

S2S climate predictionS2S 气候预测

3D Transformer models (CAS-Canglong) for skilful sub-seasonal to seasonal global sea surface temperature forecasting.基于 3D Transformer(CAS-Canglong)的全球海温次季节—季节高技巧预测。

Advanced Science, 2026
~220 km~220 km

GRACE & terrestrial waterGRACE 与陆地水储量

Deep learning to bridge the GRACE / GRACE-FO data gap and track groundwater and drought.用深度学习填补 GRACE / GRACE-FO 数据缺口,追踪地下水与干旱。

Remote Sensing of Environment, 2024

Evapotranspiration & ecohydrology蒸散发与生态水文

PML_V2 coupled carbon–water modelling, canopy conductance and its responses to heatwaves and CO₂.PML_V2 碳水耦合模型、冠层导度及其对热浪和 CO₂ 的响应。

Agric. For. Meteorol., 2023
dS/dt =P-ET-Q

Physically constrained AI物理约束 AI

Hybrid models that respect physical laws, aimed at extremes such as droughts and heatwaves under a changing climate.融合物理规律的混合模型,面向气候变化下的干旱、热浪等极端事件。

Journal of Hydrology, 2026

Cover story代表作

Predicting the ocean, weeks to seasons ahead.提前数周至数季,预测全球海洋。

The model behind my Advanced Science paper now runs live forecasts every week.Advanced Science 论文中的模型,如今每周运行实时预报。

60°N30°NEQ30°S60°S
SST anomaly (°C)海温距平(°C)
Fig. 1 · An illustrative sea surface temperature anomaly field with an El Niño-like warm tongue in the eastern equatorial Pacific, the kind of pattern CAS-Canglong is built to forecast. The field is synthetic and drawn in your browser; it is not a forecast. Hover to read values.图 1 · 示意性海表温度距平场:赤道东太平洋呈类厄尔尼诺暖舌,正是 CAS-Canglong 所要预测的典型型态。该场为浏览器内生成的合成数据,并非预报结果。悬停可读取数值。
Advanced Science · 2026Open access开放获取First author第一作者

Improved Skills for Predicting Sub-Seasonal to Seasonal Global Sea Surface Temperature

CAS-Canglong is a 3D Transformer that predicts global sea surface temperature at sub-seasonal to seasonal lead times. The same model powers the forecast portal canglong.earth.CAS-Canglong 是一个 3D Transformer 模型,用于全球海温的次季节—季节预测。同一模型驱动着预报门户 canglong.earth。

Live forecast portal实时预报门户canglong.earth
16 mo
Ocean SST & ENSO horizon海温与 ENSO 预报时效
6 wk
Atmospheric outlook大气次季节展望
36
Atmospheric variables大气变量
721×1440
Global ocean grid全球海洋网格