Tropical cyclone (TC) landfalls in the Western North Pacific (WNP) affect several nations and territories, including the Philippines, Taiwan, China (especially southern and eastern coasts), Japan, Vietnam, Korea, Hong Kong, and Micronesia. The WNP is the most active TC basin globally, but satellite-era observations (1980–2023) show nuanced patterns in landfall frequency and intensity.

Observations in this period are more reliable due to consistent satellite coverage.

Related findings include:

Global or basin-specific studies (e.g., Klotzbach et al.) have shown flat-to-declining overall TC frequency and ACE in recent decades, with mixed signals on intensity (some increase in the proportion of intense storms but not always in overall energy).

Longer-term reconstructions (e.g., Chand et al. 2022) suggest declining TC frequency over the 20th century, consistent with weakening Hadley/Walker circulations.

Climate models (CMIP6 and others) often project future decreases in global TC frequency (e.g., 2–10% by late 21st century under high-emissions scenarios), alongside potential increases in the proportion of intense storms or shifts in tracks/intensity in some basins.

Trends vary by basin (e.g., North Atlantic has shown increases in some metrics, partly linked to aerosols or other factors). Detection/attribution remains challenging due to natural variability (ENSO, etc.), data limitations pre-satellite era, and differences in metrics (frequency vs. intensity vs. landfalls vs. ACE). Individual events like landfalling major typhoons can still cause significant impacts regardless of overall trends.

There is an open-access paper by Bell et al. (2026) in Natural Hazards, which analyzes 1980–2023 satellite-era best-track data (interpolated to 30-min intervals) for TC landfall intensity (Vmax, maximum sustained winds) across WNP nations/regions.

This study adds to evidence of stable or declining activity in the WNP during the satellite era, highlighting the gap between some projections and recent observations. For the full paper, search for “Bell et al. 2026 Tropical cyclone landfall intensity western North Pacific” (published in Natural Hazards).

_____________________________________________________________________________________

Tropical cyclone landfall intensity (V_max) for western North Pacific nations: return period and trends

Bell et al. (2026) in Natural Hazards provides the most detailed recent analysis of tropical cyclone (TC) landfall intensity (Vmax, maximum sustained winds) trends and return periods for western North Pacific (WNP) nations/regions using 1980–2023 satellite-era best-track data.

Landfall Intensity and Frequency

Landfall frequency: Mostly insignificant trends across individual nations/regions after ENSO adjustment. Significant decreases appear in the northern Philippines and parts of Micronesia (including for intense storms with Vmax > 33 m/s).

Upper-tail intensity: The distribution of stronger Vmax values increased for 5 of 7 nations/regions in 2002–2023 compared to 1980–2001 (including the Philippines and Micronesia). This partially offsets the poleward shift in TC activity.

Ex-TC (post-tropical) landfalls: Some increases, notably over southwest Japan.

Basin-wide ACE: Weak decline (~–7.6 units/decade after ENSO adjustment); land-based ACE nearly flat (+0.1 units/decade).

ENSO modulation: Strong influences on tracks and landfall patterns (e.g., more landfalls/intense events in certain areas during El Niño vs. La Niña), but long-term trends often remain insignificant once adjusted.

Mean landfall Vmax trends are largely flat or insignificant in most areas, consistent with broader satellite-era observations of stable-to-declining overall activity in the WNP despite some signals of increased pre-landfall intensification in sub-regions.

Return Periods for Vmax

The study uses a modified extreme value analysis based on “peak average” winds and a reference curve to estimate return levels on a 1°×1° grid (accounting for sparse intense events). Higher return intensities concentrate around island nations and decay inland.

Table 3 from Bell et al. (2026) — Maximum Vmax return intensities (m/s) at different return periods (R) for each region (representative high-value grid points):

Region	10 yr	50 yr	100 yr	200 yr	Coordinates
Philippines	76.5	88.6	92.7	96.2	122°E, 21°N
Japan Islands	68.4	79.2	82.3	86.1	124°E, 25°N
Micronesia	65.4	75.7	79.2	82.3	146°E, 16°N
Taiwan	62.3	72.1	75.4	78.3	122°E, 23°N
Japan-main	55.7	64.5	67.5	70.0	130°E, 33°N
China	50.8	58.9	61.6	64.0	121°E, 28°N
Hong Kong	46.0	53.3	55.7	57.9	115°E, 22°N
Korea	43.8	50.8	53.1	55.2	129°E, 35°N
Vietnam	42.5	49.3	51.6	53.5	107°E, 18°N

Interpretation: 1-in-200-year Vmax values are highest over the Philippines, southern Japanese islands, Micronesia, and Taiwan (often ~50–96+ m/s in exposed coastal spots). They are relatively lower over Vietnam, Korea, and (mainland) China. Southern Japan can have higher extremes than higher-frequency areas like Hainan despite fewer crossings. Return levels decay inland from coasts.

Maps in the paper show spatial patterns, with ENSO differences (e.g., certain areas more prone to extremes in El Niño or La Niña phases).

Caveats and Context

Data limitations: Pre-1980 records are less reliable; agency differences in wind averaging and intensity estimation add uncertainty, especially for the strongest storms.

Natural variability: ENSO and other modes dominate on decadal scales; the 44-year record is relatively short.

Projections vs. observations: Models often project decreases in overall TC frequency but increases in the proportion of intense storms and possible track shifts. Observed WNP trends (declining frequency/ACE) have not fully aligned with some earlier expectations of broad increases.

Risk implications: These metrics are useful for catastrophe modeling and planning but must combine with exposure, vulnerability, and local factors. Rare events remain highly impactful.

For full details, figures, and tables, read the open-access paper: Bell et al. (2026) in Natural Hazards. This provides a robust observational baseline grounded in satellite data.

Published: Natural Hazards

Authors: Samuel S. Bell, 
Savin S. Chand, 
Philip J. Klotzbach, 
Marie Ekström, 
Valentina Koschatzsky, 
Sarvesh Kumar, 
Krishneel K. Sharma & 
Mona Hemmati 

Abstract

Human impacts, including financial losses, are closely related to the intensity and frequency of tropical cyclone (TC) landfalls in the western North Pacific (WNP) region. Here, 44 years of observational TC wind speed (V_max) and track records (1980–2023) are used to examine TC landfall intensity trends and recurrence. Analysis is conducted over individual nations and includes modulation by El Niño-Southern Oscillation (ENSO), which has recently exhibited increasing asymmetries in spatial and temporal evolution. Two modern ENSO indices are first evaluated before one is selected to classify TC seasons into canonical phases. Landfall trends over individual nations are insignificant, but significant decreases are noted for the northern Philippines and parts of Micronesia, including for intense (>  33 m s^–1) V_max. However, the distribution of upper-tail V_max values increases for 5 of 7 nations over the recent period, including the Philippines and Micronesia, confounding the impact of the much-noted poleward shift in WNP TC activity. Ex-TC landfall trends, while less robust, are also evaluated, and shown to have increased over the southwest corner of Japan. Lastly, a modified return period framework is employed to estimate the recurrence of V_max at various locations based on “peak average” speeds and the shape of a reference extreme value curve. The framework provides a consistent platform for estimating long-term return magnitudes with limited data, such as V_max over land. 1-in-200 year V_max intensities are estimated to be the highest over the Philippines, southern Japanese islands and Micronesia, and relatively lower over Vietnam, Korea and China.