In 2023 and 2024, Southeast Asia endured one of its most severe El Niño droughts on record. Province-level SPEI-3 data across Thailand, Vietnam, and Malaysia registered a region-mean anomaly of approximately −1.3 — the worst single-year drought signal in 25 years of satellite-era monitoring. Reservoirs in northeastern Thailand dropped below 20% capacity. Rice yields in Vietnam's central highlands fell sharply. Sabah's oil palm estates recorded stress-induced yield losses not seen since 1998.
Now, as of June 2026, the Oceanic Niño Index has returned to near-neutral and early La Niña signals are developing in the central Pacific. Monsoon rainfall across Thailand and Vietnam is tracking at or above seasonal average. The drought is ending — but the risk is not.
The transition from El Niño drought to active monsoon is Southeast Asia's most underpriced climate event, and understanding why requires looking at soil, not just rain.
The Soil Moisture Paradox
During a prolonged drought, topsoil loses its organic structure. Clay particles contract. Surface crusting develops. In agricultural land, irrigation dependency replaces rainfall absorption. The net result: a drought-hardened landscape that cannot absorb rainfall at normal rates — at least not immediately.
When monsoon rains arrive, particularly after a significant El Niño event, the first weeks produce unusually high runoff. Water that would normally percolate into soil instead sheets across the surface, overwhelming drainage channels and elevating flash flood risk in exactly the zones that were drought-stressed six months earlier.
This soil-moisture lag effect is documented across the historical record. The provinces that recorded the deepest negative SPEI-3 anomalies in late 2023 and early 2024 — northeastern Thailand's Isaan plateau, Vietnam's Central Highlands including Lâm Đồng and Đắk Lắk, and Sabah's interior — are now among the most vulnerable to flash flooding as monsoon intensity increases through July and August.
A risk model that prices these locations as "drought risk" in one quarter and "standard monsoon exposure" in the next is structurally blind to the transition.
Province-Level Divergence: Not All Transitions Are Equal
The transition risk is sharply differentiated at the province level — and this is where aggregate models fail most visibly.
Northeast Thailand (Isaan). The 16 provinces that showed statistically significant drying trends over 2000–2025 — including Nakhon Ratchasima, Ubon Ratchathani, Roi Et, and Surin — face a compounded risk this season. Their soils carry the deepest moisture deficit. Their agricultural land is dominated by rain-fed rice, which has high sensitivity to both timing and intensity of early monsoon rainfall. Floods arriving before soil recovery are especially destructive: the water cannot be used productively, causes erosion, and damages standing crops. The window between "too dry" and "flooded" in Isaan is narrowing.
Vietnam's Central Highlands. This zone showed the strongest ENSO-drought coupling in our 25-year dataset, with Pearson correlation between the Oceanic Niño Index and drought severity reaching r = −0.53 in Lâm Đồng. It is also a zone with some of the steepest topography in the region, meaning rainfall runoff is rapid and catchment response times are short. Flash flood events in 2020 and 2022 — both La Niña years — caused significant infrastructure damage in the same provinces. A La Niña pattern developing now replicates those conditions.
Sabah, Malaysia. Our dataset showed Sabah with the highest ENSO-drought coupling among Malaysian states (r = −0.57). Sabah's unique geography — long river systems draining into the Sulu Sea and South China Sea — means that upstream drought followed by heavy monsoon rainfall produces lagged downstream flooding. Agricultural concessions and palm oil logistics infrastructure along the Kinabatangan and Segama rivers carry elevated risk this season.
The Mekong Delta (Vietnam). An anomalous case in our dataset: this zone shows elevated drought risk under La Niña conditions rather than El Niño — a reversal of the dominant regional pattern, attributable to monsoon-pathway and hydrological routing effects. As La Niña develops, the Delta faces reduced freshwater inflow from the Mekong mainstem, while simultaneously receiving direct monsoon rainfall. The result is complex: internal waterlogging without adequate upstream replenishment, complicating both agriculture and transportation.
What the Insurance Market Is Missing
Standard property and agriculture insurance models in Southeast Asia are calibrated on historical average loss rates. They do not model the sequence dependency — the fact that a La Niña flood following an El Niño drought produces materially different losses than a La Niña flood following a neutral year.
Three specific gaps stand out:
- Crop insurance mis-pricing at transition. Rice and sugarcane policies in Thailand's northeast are typically renewed in April–May, before the monsoon onset. Actuarial tables used for premium-setting reflect multi-year average precipitation, not the current soil moisture state. A post-drought monsoon year should attract a materially higher expected loss ratio — but most policies do not capture this signal.
- Property underwriting blind to runoff dynamics. Flood zone maps in Thailand and Vietnam were largely developed in normal precipitation years. Properties that have never flooded may flood this season if they sit downstream of drought-degraded catchments. Relying on historical flood zone designation without updating for current soil state overestimates policy safety margins.
- Parametric trigger design using rainfall, not soil moisture. Many parametric agriculture products in SEA trigger payouts based on rainfall thresholds. In a transition year, a province can receive above-average rainfall and still experience crop damage — because the rain arrived in intensity bursts that exceeded infiltration capacity, damaging standing crops even as cumulative totals look "normal".
Reading the Signals in Real Time
The Evaporative Stress Index (ESI) — one of the three indicators in GlobMaps' Multi-Drought Index — is particularly diagnostic during a transition period. ESI measures the ratio of actual to potential evapotranspiration, derived from land surface temperature anomalies from the ALEXI satellite product. In the weeks after monsoon onset, provinces with residual ESI stress (values remaining below −0.5 despite increased rainfall) are signalling that soil moisture recovery is incomplete and runoff risk remains elevated.
Combined with the Vegetation Condition Index (VCI), which recovers faster than soil moisture as crops respond to initial rainfall, this creates a differential signal: VCI recovering while ESI remains stressed = surface greening without deep moisture recovery = continued flash flood vulnerability.
Monitoring this differential at the province level — rather than relying on aggregate regional rainfall totals — is the practical intelligence that separates reactive response from proactive risk management.
Action Points for Risk Managers
For teams managing agricultural exposure, property portfolios, or supply chain risk across Thailand, Vietnam, and Malaysia, three actions are relevant right now:
- Map your exposure against 2023/24 drought footprint. Identify which of your assets or insured properties sit in the 16 Thai provinces, the Vietnamese Central Highlands provinces, or Sabah's agricultural zones that recorded the deepest El Niño drought. These are not the same as your highest historical flood-risk zones.
- Update loss assumptions for the transition premium. A post-El Niño monsoon year carries a 15–30% higher expected agricultural loss ratio in drought-affected zones, based on historical La Niña event comparisons in our dataset. This is a knowable adjustment — apply it before July underwriting cycles close.
- Build soil moisture state into parametric triggers. If your parametric product uses rainfall as the sole trigger variable, consider adding a soil moisture proxy (ESI or SPEI-3 at the most recent month) as a multiplier on the payout threshold. A province with SPEI-3 of −1.0 at monsoon onset warrants a lower flood trigger than one entering the season at neutral moisture levels.
The drought is ending across Southeast Asia. The risk is not. It is changing form — and the transition is the most dangerous moment of all.