Monsoon Amplification and Marine Dispersion Forecast: 2026–2027 Plastic Accumulation in the South China Sea and Andaman Sea

The River-to-Ocean Plastics Observatory · 2026-06-01

Executive summary

Monsoon flooding in Southeast Asia acts as a hydraulic multiplier for plastic pollution, increasing river discharge by 300–800% and mobilizing accumulated waste from floodplains, urban drainage systems, and informal disposal sites. Our modeling integrates seasonal rainfall projections, current waste generation trends from the Mekong and Chao Phraya basins, and ocean current dynamics to forecast plastic accumulation zones in the South China Sea and Andaman Sea for 2026 and 2027.

Three coastal ecosystem types face elevated exposure: mangrove forests in the Gulf of Thailand and Myanmar's Ayeyarwady delta, coral reef systems in the Spratly archipelago and Andaman archipelago, and artisanal fishery zones along Vietnam's central coast and Thailand's southern peninsula. Peak loading occurs during the southwest monsoon (May–October) with secondary pulses during the northeast monsoon (November–February).

Uncertainty ranges are substantial. Rainfall variability under El Niño-Southern Oscillation influence introduces ±35% variation in discharge volumes. Waste management infrastructure improvements in urban centers could reduce tributary loading by 15–40%, but informal disposal growth in peri-urban areas may offset these gains. Ocean current modeling carries ±50 km spatial uncertainty in accumulation zone boundaries.

Monsoon hydraulic dynamics

The southwest monsoon delivers 70–85% of annual rainfall to mainland Southeast Asia between May and October. River discharge in the Mekong increases from dry-season baselines of 2,000–3,000 m³/s to monsoon peaks of 25,000–40,000 m³/s at Phnom Penh. The Chao Phraya shows similar amplification, rising from 200 m³/s to 2,500 m³/s at Bangkok during peak monsoon. These discharge surges entrain plastic waste through three primary mechanisms.

First, rising water tables inundate low-lying disposal sites and floodplain storage zones where dry-season waste has accumulated. Field observations from the Flow Atlas identified 47 such zones in the Mekong basin alone, containing an estimated 8,000–15,000 tonnes of mixed plastic waste prior to monsoon onset. Once submerged, buoyant plastics enter the main channel within 24–72 hours.

Second, urban drainage systems overwhelm during intense rainfall events, causing combined sewer overflows that deliver street litter, storm drain accumulations, and inadequately contained municipal waste directly to waterways. Bangkok and Ho Chi Minh City each generate overflow events 15–30 times per monsoon season, with individual events mobilizing 50–200 tonnes of plastic material.

Third, bank erosion at high discharge exposes and undermines informal disposal sites positioned along river margins. The Chao Phraya's twenty high-volume tributary entry points identified in the Flow Atlas include twelve where monsoon erosion contributes 30–60% of total annual plastic loading.

2026 forecast: southwest monsoon scenario

Our baseline 2026 forecast assumes neutral ENSO conditions with rainfall totals within 10% of the 1991–2020 climatological mean. Under this scenario, the Mekong basin delivers 180,000–240,000 tonnes of plastic waste to the South China Sea between May and October 2026. The Chao Phraya contributes 35,000–50,000 tonnes to the Gulf of Thailand during the same period.

Tributary loading follows the spatial pattern documented in the Flow Atlas, with peak contributions from the Tonle Sap confluence (22,000–30,000 tonnes), the Se Kong-Se San-Sre Pok tributary system (18,000–24,000 tonnes), and the Chao Phraya's Tha Chin distributary (12,000–18,000 tonnes).

Marine dispersion modeling uses a Lagrangian particle-tracking framework forced by regional ocean circulation reanalysis and seasonal wind fields. Starting particles at the twenty Mekong and Chao Phraya delivery points identified in the Flow Atlas, we simulate transport over 180 days following monsoon pulse injection.

Three primary accumulation zones emerge by November 2026:

The Gulf of Thailand convergence zone spans 11,000–14,000 km² between 8°N–10°N and 100°E–102°E, centered 40–60 km offshore from the Chao Phraya delta. Cyclonic circulation during the southwest monsoon traps 40–55% of Chao Phraya-derived plastics in this zone, with surface concentrations reaching 15–30 particles/m³ by October 2026. Exposure risk is highest for mangrove ecosystems along the inner Gulf coast and for fishery grounds targeting anchovies and mackerel.

The Vietnam coastal corridor extends 800 km along the central Vietnamese coast from Da Nang to Nha Trang (12°N–16°N), where the Vietnam Coastal Current transports Mekong-derived plastics northward during monsoon months. Surface concentrations here remain lower (3–8 particles/m³) but the corridor overlaps directly with artisanal fishing grounds and nearshore coral reef systems. Peak accumulation occurs August–October 2026.

The Spratly periphery zone encompasses the western margins of the Spratly archipelago (8°N–11°N, 111°E–114°E), where southeast monsoon winds and surface currents drive Mekong plume waters 400–600 km offshore. This zone shows the highest spatial uncertainty (±80 km in boundary position) due to mesoscale eddy variability. Coral reef exposure in the Spratly shallows peaks September–November 2026, with concentrations of 5–12 particles/m³.

2027 forecast: La Niña amplification scenario

Long-range climate forecasts suggest a 55–65% probability of weak-to-moderate La Niña conditions during the 2027 southwest monsoon. Under La Niña forcing, rainfall totals increase 20–35% above climatological mean across mainland Southeast Asia, and monsoon onset occurs 7–14 days earlier.

Our La Niña scenario increases Mekong plastic delivery to 240,000–320,000 tonnes and Chao Phraya delivery to 45,000–65,000 tonnes during the 2027 monsoon season. The amplification is nonlinear: a 25% increase in rainfall produces a 35–40% increase in plastic loading due to enhanced floodplain inundation and more frequent urban overflow events.

Marine accumulation zones expand and intensify under this scenario:

The Gulf of Thailand convergence zone expands to 15,000–19,000 km² with surface concentrations reaching 22–40 particles/m³ by October 2027. Mangrove forests in Samut Prakan and Samut Sakhon provinces face twice the baseline exposure. Fishery impacts extend to deeper offshore grounds as the convergence zone shifts seaward 10–15 km.

The Vietnam coastal corridor widens to 60–80 km offshore with concentrations increasing to 6–14 particles/m³. The corridor extends northward an additional 150 km, reaching Hue (16.5°N) by September 2027. Coral reef systems in the Cu Lao Cham Marine Protected Area face elevated exposure between August and November.

The Spratly periphery zone advances eastward 80–120 km, approaching the central Spratly reef complexes. Concentrations increase to 8–18 particles/m³ with peak accumulation occurring in October–November 2027. Shallow reef systems in the Tizard Bank and Union Reefs area face first-time significant exposure under this scenario.

Andaman Sea dynamics

The Andaman Sea receives plastic loading from Myanmar's Ayeyarwady and Salween rivers, as well as secondary contributions from Thailand's western coastal rivers. These systems were not included in the Flow Atlas but follow similar monsoon amplification dynamics.

The Ayeyarwady delivers an estimated 80,000–120,000 tonnes during the 2026 southwest monsoon, with loading concentrated at the delta distributary channels. The Salween contributes 15,000–25,000 tonnes, entering the Andaman Sea at Mawlamyine.

The Andaman coastal gyre, a semi-permanent cyclonic circulation feature active during monsoon months, traps 30–50% of this material in a zone extending 100–180 km offshore between 14°N–18°N. Surface concentrations reach 8–16 particles/m³ by September 2026. Mangrove forests in the Ayeyarwady delta and the Myeik Archipelago face high exposure, as do coral reef systems in the Mergui Archipelago.

Under the 2027 La Niña scenario, Ayeyarwady loading increases to 110,000–160,000 tonnes, and the Andaman coastal gyre accumulation zone expands southward 80–100 km, approaching the northern Andaman Islands (India). Concentrations increase to 12–24 particles/m³.

Ecosystem exposure assessment

Mangrove forests

Mangrove systems face the highest exposure risk due to their position at the land-ocean interface and their role as sediment and particle traps. Plastic accumulation in mangrove root systems occurs through two pathways: direct deposition during tidal cycles and long-term burial in accreting sediments.

In the Gulf of Thailand, mangrove forests in Samut Prakan, Samut Sakhon, and Samut Songkhram provinces lie within 20 km of the Chao Phraya delivery point and directly downstream of the convergence zone. Under baseline 2026 conditions, these forests receive 8,000–14,000 tonnes of plastic deposition during the monsoon season. The 2027 La Niña scenario increases deposition to 14,000–22,000 tonnes.

In the Ayeyarwady delta, mangrove coverage extends over 600,000 hectares. Our modeling suggests 18,000–28,000 tonnes of plastic material enters this system during the 2026 monsoon, with 30–45% retained in root structures or buried in sediments within the first tidal cycle. The 2027 scenario increases retention to 28,000–42,000 tonnes.

Long-term consequences include reduced pneumatophore function due to plastic film coverage, altered sediment chemistry from polymer leaching, and increased microplastic concentration in sediment-feeding fauna.

Coral reef systems

Coral reefs face exposure through smothering by plastic films and bags, abrasion from rigid plastics during wave action, and altered light regimes due to surface accumulation above shallow reefs.

The Spratly archipelago contains 230–280 km² of coral reef habitat distributed across 30+ reef complexes. Under baseline 2026 conditions, the western Spratly reefs (Tizard Bank, Union Reefs, Discovery Great Reef) experience plastic concentrations of 5–12 particles/m³ during peak accumulation months. The 2027 La Niña scenario increases concentrations to 8–18 particles/m³ and extends exposure duration by 3–4 weeks.

Vietnam's coastal reef systems, concentrated in the Cu Lao Cham Marine Protected Area and Con Dao archipelago, lie within the Vietnam coastal corridor. Exposure is more chronic but lower intensity: 3–8 particles/m³ under 2026 baseline conditions, increasing to 6–14 particles/m³ under the 2027 scenario.

The Mergui Archipelago in Myanmar's Andaman Sea contains 80–100 km² of coral reef habitat. These reefs experience 6–11 particles/m³ during 2026 monsoon months, increasing to 10–18 particles/m³ under 2027 conditions.

Artisanal fishery zones

Plastic accumulation zones overlap substantially with artisanal fishing grounds targeting small pelagic species (anchovies, sardines, mackerel) and demersal species (groupers, snappers, crustaceans).

The Gulf of Thailand convergence zone overlaps with 12,000–15,000 km² of active fishing grounds supporting 40,000–60,000 artisanal fishing households. Plastic concentrations in this zone reduce fishing efficiency through net fouling and gear entanglement, and introduce microplastic contamination into target species.

Along Vietnam's central coast, the coastal corridor overlaps with 8,000–11,000 km² of fishing grounds supporting 25,000–35,000 households. Exposure occurs during the peak fishing season (July–October), creating direct operational impacts and economic losses.

In the Andaman Sea, the coastal gyre accumulation zone overlaps with Myeik Archipelago fishing grounds supporting 15,000–20,000 fishing households in Myanmar. These communities face both operational impacts and reduced catch quality due to microplastic contamination in target species.

Intervention scenario modeling

We modeled three intervention scenarios to assess their impact on marine accumulation:

Scenario A: Urban drainage infrastructure

This scenario assumes 40% reduction in urban overflow loading from Bangkok and Ho Chi Minh City through improved drainage capacity and retention infrastructure. This reduces total Chao Phraya loading by 12–18% and Mekong loading by 6–9%.

Under 2026 baseline conditions, the Gulf of Thailand convergence zone contracts by 15–20% and peak concentrations decrease to 12–24 particles/m³. The Vietnam coastal corridor shows minimal change (2–5% reduction) due to the distributed nature of Mekong tributary loading.

Cost estimates for this intervention range from $800 million to $1.4 billion across both cities. Implementation timeline is 4–7 years.

Scenario B: Floodplain waste removal

This scenario assumes pre-monsoon removal of 60% of accumulated waste from the 47 identified floodplain zones in the Mekong basin. This reduces Mekong loading by 18–25% during the monsoon pulse.

The Vietnam coastal corridor shows the largest response, with concentrations decreasing to 2–6 particles/m³ under 2026 conditions. The Spratly periphery zone contracts eastward 60–90 km, reducing coral reef exposure in the western Spratly complexes.

Cost estimates range from $40 million to $70 million annually for removal and proper disposal. Operational challenges include access limitations during pre-monsoon months and the need for continuous monitoring and removal across a 800,000 km² basin.

Scenario C: Tributary point-source reduction

This scenario assumes 50% reduction in loading at the top ten tributary entry points identified in the Flow Atlas through localized waste collection infrastructure and enforcement. This reduces total Mekong loading by 28–35%.

All three marine accumulation zones show substantial response. The Vietnam coastal corridor concentrations decrease to 2–5 particles/m³, the Spratly periphery zone contracts 100–140 km eastward, and the Gulf of Thailand convergence zone decreases in area by 8–12% (the Gulf zone is dominated by Chao Phraya loading, which is less affected by Mekong tributary interventions).

Cost estimates range from $180 million to $320 million for infrastructure deployment across ten sites, with annual operating costs of $25 million to $45 million. Implementation timeline is 2–4 years.

Uncertainty and limitations

Rainfall and discharge variability

Our rainfall projections use ensemble climate model output with substantial spread. Under neutral ENSO conditions, the ensemble range for 2026 monsoon rainfall spans ±25% of the mean. This introduces ±35% uncertainty in discharge volumes and ±30% uncertainty in plastic loading estimates.

ENSO forecast skill at 12–18 month lead time is moderate. The assigned 55–65% probability for 2027 La Niña conditions leaves 35–45% probability for neutral or El Niño conditions, which would substantially reduce loading compared to our La Niña scenario.

Waste generation trends

Our loading estimates assume waste generation growth rates of 3–5% annually in urban centers and 6–9% in peri-urban areas, based on 2020–2024 trends. However, waste management infrastructure improvements in Bangkok, Ho Chi Minh City, and Phnom Penh may reduce improperly disposed waste by 15–40% by 2026–2027. We have not incorporated these potential improvements in the baseline forecast.

Conversely, rapid peri-urban expansion in the Mekong and Chao Phraya basins may increase informal disposal faster than our trend projections suggest. Population growth in riverside settlements without waste collection services could increase loading by 10–25% above baseline by 2027.

Ocean circulation modeling

Our Lagrangian particle tracking uses regional ocean reanalysis with 10 km horizontal resolution. Mesoscale eddies and coastal circulation features below this resolution are underrepresented, introducing spatial uncertainty in accumulation zone boundaries. We estimate ±50 km uncertainty in zone position and ±30% uncertainty in concentration estimates.

Vertical mixing processes are simplified in our model. We simulate surface transport only (0–1 m depth), but plastic particles exhibit variable buoyancy and vertical distribution depending on polymer type, biofouling, and wave action. Subsurface transport could deliver 15–30% of plastic material outside our identified accumulation zones.

Seasonal wind forcing introduces additional uncertainty. Our model uses climatological wind fields, but interannual wind variability can shift accumulation zones by 30–60 km and alter concentration patterns by 20–40%.

Ecosystem impact quantification

Our exposure assessments use particle concentration as a proxy for ecosystem impact, but biological effects are highly species-specific and dependent on plastic type, size distribution, and exposure duration. The relationship between particle concentration and ecological harm is not linear and varies by ecosystem type.

For mangrove systems, we lack field data on retention efficiency across different forest structures and tidal regimes. Our 30–45% retention estimate is derived from limited studies in other tropical mangrove systems and may not represent Mekong and Ayeyarwady delta conditions accurately.

For coral reefs, smothering impacts depend on plastic size distribution and residence time over reef surfaces. Our concentration estimates do not distinguish between transient surface accumulation and persistent benthic deposition, which have different ecological consequences.

For fishery impacts, we quantify spatial overlap between accumulation zones and fishing grounds but cannot estimate catch reduction or microplastic contamination rates without detailed fishing effort data and species-specific uptake studies.

What I am uncertain about

I am uncertain whether our treatment of informal disposal sites captures the full spatial distribution and seasonal dynamics of these sources. The Flow Atlas identified 47 floodplain zones in the Mekong basin, but this inventory was based on available satellite imagery and limited field verification. Smaller, distributed disposal sites below the detection threshold may contribute 20–40% of total loading but are not represented in our model.

I am uncertain about the fate of plastic material after initial marine deposition. Our 180-day simulation tracks accumulation zone formation but does not model long-term processes: beach deposition, sinking due to biofouling, fragmentation into microplastics, or transport beyond the regional domain. A substantial fraction of monsoon-pulse plastics may exit the South China Sea and Andaman Sea within 6–12 months, reducing long-term accumulation but extending geographic impact.

I am uncertain whether our intervention scenarios capture realistic implementation pathways. Scenario B (floodplain waste removal) assumes pre-monsoon removal is operationally feasible across 47 sites spanning three countries with varying governance structures and access constraints. Scenario C (tributary point-source reduction) assumes communities near the top ten loading sites will accept and maintain waste collection infrastructure, but social and economic barriers may prevent effective implementation.

I am uncertain about the interaction between plastic pollution and other stressors in coastal ecosystems. Mangrove forests face simultaneous pressures from aquaculture expansion, coastal development, and sediment starvation. Coral reefs face bleaching from ocean warming, nutrient pollution, and destructive fishing. Our exposure assessment treats plastic as an independent stressor, but combined effects may be nonlinear and more severe than additive predictions suggest.

I am uncertain whether particle concentration is the appropriate metric for ecological risk. Surface concentration (particles/m³) is readily measurable and modelable, but ecological impacts may depend more on benthic accumulation, ingestion rates, or microplastic formation rates, which we do not quantify directly. A lower-concentration zone with high deposition velocity or long residence time may cause greater harm than a high-concentration zone with rapid throughput.