Perceived global increase in algal blooms is attributable to intensified monitoring and emerging bloom impacts

The study addresses whether harmful algal blooms (HABs) are increasing globally and, if so, what mechanisms drive observed trends. Among ~5000 marine phytoplankton species, about 200 taxa can harm human society through toxin production and high-biomass blooms that affect seafood safety, human health, fisheries, aquaculture, recreation and coastal economies. Historical records show HABs are not new, but several toxin syndromes (e.g., DSP, ASP, AZP) have only been recognized in recent decades, prompting expanded monitoring. There is an ongoing debate: some suggest an apparent global increase driven by awareness and monitoring, others posit a global epidemic linked to eutrophication, and climate assessments project increased HAB occurrence and risk with warming and rising CO2. To provide a quantitative global assessment, the authors analyse global occurrence data of HAB species (OBIS) and harmful events (HAEDAT) from 1985–2018, distinguishing organism occurrence, toxin detection in seafood, human poisoning incidents, and fish kills and marine mortalities.

The paper situates its research within a long-standing debate on HAB trends. Earlier works proposed an apparent global increase due to heightened awareness and monitoring, while others argued for a global epidemic emphasizing eutrophication as a key driver. The IPCC has projected increasing HAB occurrence and risk under climate change. Historical accounts document HABs over centuries, but several syndromes (DSP since 1978, ASP in 1987, AZP in 1996) were only recently recognized, influencing monitoring and detection capabilities. The authors note that increased monitoring frequency, improved analytical methods for taxa and toxins, and enhanced communication likely contribute to more detections, complicating inference about true trends.

Databases: The study used (1) IOC-UNESCO Taxonomic Reference List of Harmful Microalgae; (2) HABMAP-OBIS (OBIS) for geographic ranges and occurrences of harmful species; and (3) HAEDAT, the only global database of harmful algal events. HAEDAT events include water discolorations/scum/foam with socio-economic impact (toxic or non-toxic algae), seafood biotoxin exceedances over regulatory limits, precautionary closures due to harmful algae abundance, and incidents where humans, animals, or other organisms are negatively impacted. Freshwater algae and macrophyte HABs were excluded. HAEDAT lacks records for absence of events or monitoring. Not all events have equal societal impact. Spatial framework: Global HAEDAT events were partitioned into 12 OBIS provinces/regions (ECA, CCA, SAM, WCA, ANZ, SEA, NAS, IND, BENG, MED, EUR, PAC). Regional editorial teams collated species occurrences and impacts. Sampling effort proxy: To account for variable monitoring, OBIS microalgal records (5,944,392 total, including 289,668 harmful species records) were used as a proxy for sampling effort. OBIS occurrences were matched to HAEDAT regions using shapefiles and a hexagonal discrete global grid. Within each region and month, the number of grid cells with phytoplankton records was counted and summed per year to estimate annual sampling effort. Temporal scope: Analyses focused on 1985–2018 (over 7000 HAEDAT records), when consistent phytoplankton and seafood toxin monitoring commenced in many areas. Reliable starts varied by region (e.g., 1985–1987 for ECA, MED; 1990 for EUR; 2000 for NAS). Regions with sparse data (West Africa, Benguela Current, Indian Ocean, Pacific/Oceania) were excluded from trend meta-analyses. Meta-analyses (all adjusted for temporal autocorrelation using Pyper & Peterman):

1. Events vs. Year: Random-effects meta-analysis of correlations between the annual number of HAEDAT events and year across regions (after rejecting homogeneity with Q statistic), following Worm & Myers. Combined correlation significance was tested with Z for two-tailed α=0.05; confidence limits via Zar.
2. Grids with ≥1 HAEDAT event vs. Year: Meta-analysis on the annual number of geographic monitoring grids (total 623 globally) reporting ≥1 event to reduce bias from event definitions and sampling effort.
3. HAEDAT events standardized by OBIS sampling vs. Year: Using the ratio of HAEDAT events to OBIS microalgal sampling to adjust for effort and test for standardized trends.
4. HAEDAT events vs. Aquaculture production: Tested correlation between HAEDAT events and regional aquaculture production (fish, molluscs, crustaceans, aquatic plants) from FAO/Our World in Data for ANZ, SEA, MED, ECA+WCA (combined), and SAM+CCA (combined). Data points with zero OBIS observations in a year were removed. Statistical treatment: Random-effects meta-analytic estimation of mean effect sizes per region (weighted-sample unconditional variances), back-transform via inverse z-transform to correlation coefficients; adjusted effective degrees of freedom for autocorrelation to avoid inflated Type I errors.

Database composition: As of 10 Dec 2019, 9503 HAEDAT events were recorded globally: 48% seafood biotoxins, 43% high phytoplankton counts/water discolorations with socio-economic impact, 7% mass mortalities, 2% others (foam/mucilage). About 11% of records had multiple event types. Among seafood-toxin-related events: PST 35%, DST 30%, CP 9%, cyanotoxins 9%, AST 7%, others (NST, AZA, aerosols) 10%. Regional patterns varied: PST dominated in ECA, SAM, WCA, SEA, NAS; DST in MED and EUR; CP predominant in IND and PAC. Temporal trends (unadjusted): Yearly HAEDAT events generally increased across regions; 8/9 regions increased, with 6 significant (ECA, CCA, WCA, SEA, MED, EUR). Meta-analysis across regions showed a significant global increase in events vs. year (r=0.37, z=2.97, p=0.003). The number of geographic grids with ≥1 event also increased with weaker significance (r=0.27, z=2.16, p=0.031). Adjustment for sampling effort: OBIS records generally increased over time (exceptions: SAM and NAS). When standardizing HAEDAT events by OBIS microalgal sampling, trends differed by region: significant increases in 5 regions (CCA, SAM, MED, NAS, EUR), significant decreases in 2 (WCA, ANZ), and no significant change in 2 (ECA, SEA). The global combined standardized trend was not significant (r=0.35, z=1.33, p=0.18), indicating insufficient evidence for a global increase in standardized HAB events. Aquaculture linkage: Global aquaculture production increased 16-fold from 11.35 million tonnes (1985) to 178.5 million tonnes (2018), with largest increases in SEA and SAM+CCA; ECA+WCA and EUR stabilized. HAEDAT events were positively correlated with aquaculture production across regions with data, with a strong overall relationship (r=0.43, z=3.59, p=0.0003). Case studies: Global observations of causative organisms increased markedly between 1985 and 2018 (approx. 4-fold for Dinophysis spp. [DSP], 7-fold for Pseudo-nitzschia spp. [ASP], 6-fold for Alexandrium spp. [PSP]), paralleled by increases in associated HAEDAT toxin event records. However, organism occurrence is not a reliable predictor of human poisonings due to effective risk management. Reported human impacts included about 11,000 nonfatal DSP cases (mostly Europe, South America, Japan); for ASP, 150 illnesses and 3 fatalities in 1987 in Canada with no subsequent human fatalities despite notable marine mammal mortalities in the NE Pacific and Arctic; for PSP, ~3800 human cases globally (1985–2018), with the Philippines reporting 2555 cases (165 fatalities) predominantly from Pyrodinium bahamense. Ciguatera affects an estimated 10,000–50,000 people annually globally, with regionally varying trends (declining in Hawaii, stable in French Polynesia and Caribbean, emerging in Canary Islands). Fish-killing HABs cause major economic losses in aquaculture (e.g., Japan 1972 USD71M; Korea 1995 USD70M; China 2012 USD290M; Norway 2019 USD100M; Chile 2016 USD800M), and impacts depend on proximity to aquaculture operations rather than solely on species presence.

The central question—whether HABs are increasing globally—is addressed by adjusting for monitoring effort and regional variability. While raw counts of recorded events have increased, standardized analyses provide no evidence for a uniform global increase in HAB frequency or distribution from 1985–2018. Instead, trends are heterogeneous across regions and taxa, reflecting differences in bloom species, event types, monitoring intensity, and emergent societal impacts. The perceived global increase is largely attributable to intensified monitoring, expanded aquaculture (which elevates exposure and economic sensitivity to blooms), and improved detection methods. Effective management has reduced human poisonings for some syndromes (e.g., ASP, DSP) despite rising observations of causative organisms, whereas ciguatera and fish-killing HABs remain less well monitored and regulated. The findings underscore the need for species- and region-specific assessments and management approaches, recognizing that climate change and eutrophication will exert variable, species-dependent effects and that ocean regions respond at different rates. Improved, harmonized datasets are essential to refine causal links and to predict future trends with higher confidence.

There is no conclusive evidence for a uniform global increase in HAB frequency or distribution over 1985–2018 once monitoring effort is accounted for. The widespread perception of increasing HABs arises primarily from intensified observational efforts and the expanding scale and nature of societal impacts, particularly associated with aquaculture growth. Regional trends vary, with some regions and phenomena increasing, others decreasing or remaining stable. Shellfish toxin outbreaks are increasingly well managed in many developed regions, limiting human health impacts, while fish-killing HABs continue to cause substantial and prolonged economic losses in aquaculture. The study highlights the need to assess HABs on a species-by-species and site-by-site basis, improve monitoring for under-addressed hazards (e.g., ciguatera, ichthyotoxic blooms), and expand and integrate databases (HAEDAT, OBIS, and related data) to capture ecological drivers, quantify economic and health impacts, and support predictive capability. Future research should focus on linking species-specific ecology to regional environmental change, disentangling the roles of climate change, eutrophication, and human coastal use, and developing robust, harmonized global monitoring frameworks.

Data availability and quality are uneven across regions and time. HAEDAT coverage varies by region and records only events with impacts (no data for absence of events/monitoring), and not all events have equal societal impact. OBIS is used as a proxy for sampling effort, which may not perfectly represent monitoring intensity or targeted HAB surveillance; grid sizes vary geographically. Several regions (West Africa, Benguela, Indian Ocean, parts of the Pacific/Oceania) had sparse data and were excluded from trend meta-analyses. Consistent monitoring in many areas began only after 1985; pre-1985 data are sparse and unreliable, limiting inference on earlier trends. Freshwater and macrophyte HABs were excluded. Ecological mechanisms underlying regional differences and toxin-specific distributions remain insufficiently understood, limiting causal attribution and predictive power.