The Tropical Cyclone Forecast Verification Dataset covers tropical cyclone (TC) track and intensity forecast error over the western North Pacific, with forecasting busts (judgment method is shown in Table 1) being identified. The basin is to the north of the equator and to the west of 180°E, and includes the South China Sea.

Starting from 2024, the verification samples include samples for each tropical cyclone that reach the tropical storm category or higher, while those that do not reach the tropical storm category and those after the National Meteorological Center of CMA issues the final TC bulletin are not included in the verification. The verification is based on the CMA Tropical Cyclone Best Track Dataset.

Please indicate that the Tropical Cyclone Forecast Verification Dataset was obtained from tcdata.typhoon.org.cn, and refer to the following paper in any written work using the dataset:

• YANG Mengqi, CHEN Guomin, ZHANG Xiping, TANG Lichun, BAI Lina, GUO Rong, 2024. Verification on Typhoon Forecasts over the Western North Pacific and the South China Sea in 2022[J]. Meteor Mon, 50(5): 630-641 (in Chinese). doi: 10.7519/j.issn.1000-0526.2024.031404

YYYY_tc_tra_err.csv ——

YYYY: Four-digit calendar year.

_tc_tra_err: Tropical Cyclone Track Forecast Verification Dataset.

YYYY_tc_int_err.csv ——

YYYY: Four-digit calendar year.

_tc_int_err: Tropical Cyclone Intensity Forecast Verification Dataset.

The data are saved in CSV text files.

Sample data for the Tropical Cyclone Track Forecast Verification Dataset is as follows:

Sample data for the Tropical Cyclone Intensity Forecast Verification Dataset is as follows:

Table 1 Column description

Abbreviation	Data Type	Meaning and Format Description
fcst_method	Character	Name of the forecast method.Refer to the Forecast Method Description for details.
TC_ID	Integer	TC Chinese ID.
int_time	Integer	Initial forecast time (UTC).Formatted as YYYYMMDDHH, where YYYY denotes the year, MM denotes the month, DD denotes the day, and HH denotes the hour.
fcst_lead_time	Integer	Forecast lead time (unit: hour).
fcst_lon	Real	Forecast longitude (unit: °E).
fcst_lat	Real	Forecast latitude (unit: °N).
fcst_vmax	Real	Forecast intensity (unit: m/s).
tra_err	Real	Position error (unit: km).Represents the distance between the forecast position and the actual position of the TC center.
int_err	Real	Intensity error (unit: meters/second).Represents the difference between the forecast intensity and the actual intensity of TC.
int_err_abs	Real	Absolute intensity error (unit: meters/second). Represents the absolute value of the difference between the forecast intensity and the actual intensity of TC.
Is it a forecasting bust ? 0: no, 1: yes	Integer	Identify whether the position error (tra_err) or absolute intensity error (int_err_abs) is a forecasting bust case. If tra_err or int_err_abs is greater than or equal to the 95th percentile value of all error samples in its group (grouped by year and fcst_method), it is marked as 1. If tra_err or int_err_abs is less than the 95th percentile value, it is marked as 0.

Tropical Cyclone Track Forecast Verification Dataset

2024_tc_tra_err.csv

Tropical Cyclone Intensity Forecast Verification Dataset

2024_tc_int_err.csv

Category	fcst_method	TC forecast agency or forecast method	Source
Official Guide	babj	China Meteorological Administration (CMA)	CMA
	jawt	Japan Meteorological Agency (JMA)	JMA
	pgtw	Joint Typhoon Warning Center (JTWC)	JTWC
	rksl	Korea Meteorological Administration (KMA)	KMA
	vhhh	Hong Kong Observatory (HKO)	HKO
Global Numerical Weather Prediction Model	ggfs	CMA Global Forecast System (CMA-GFS)	CMA
	ecmf	ECMWF Integrated Forecasting System (ECMWF-IFS)	ECMWF
	japn	JMA Global Spectral Model (JMA-GSM)	JMA
	avno	NCEP Global Forecast System (NCEP-GFS)	NCEP
	egrr	Met Office Unified Model (UKMO-MetUM)	UKMO
Regional Numerical Weather Prediction Model	gtym	CMA Regional Typhoon forecasting Model (CMA-TYM)	CMA
	gztm	CMA Tropical Regional Atmospheric Model System of the South China Sea (CMA-TRAMS)	ITMM/CMA
	shtm	Shanghai Weather and Risk Model System-Typhoon Model (SWARMS-TC)	STI/CMA
	hwrf	Atmosphere-ocean Coupled Hurricane Weather Research and Forecast Modeling System (HWRF)	EMC/NCEP
Artificial Intelligence-Based Weather Prediction Model	fengqing	Artificial Intelligence Global Short and Medium Range Forecasting System (Fengqing)	CMA, Tsinghua University
	fengwu	Artificial Intelligence Global Weather Forecasting Model (FengWu)	Shanghai AI Laboratory
	fuxi	Machine Learning–Based Global Weather Forecasting Model (FuXi)	Fudan University
	pangu	Artificial Intelligence Global Weather Forecasting Model (Pangu)	Huawei Cloud
	aifs	Artificial Intelligence Forecasting System (AIFS)	ECMWF
	fourcastnet	Fourier ForeCasting Neural Network Model (FourCastNet)	NVIDIA
	graphcast	Machine Learning–Based Global Weather Forecasting Model (GraphCast)	Google DeepMind