Experimental evidence for recovery of mercury-contaminated fish populations

The study addresses how reductions in mercury (Hg) loading to aquatic ecosystems translate into declines in methylmercury (MeHg) in fish, a key public health concern due to MeHg’s toxicity and biomagnification in food webs. Despite international controls like the Minamata Convention aiming to reduce Hg emissions and deposition, definitive evidence linking decreased Hg loading to fish MeHg declines has been limited by ecological complexity, trophic disruptions, climate-driven changes, and the difficulty of tracing recently deposited Hg into contemporary MeHg production. The purpose of this work is to provide unambiguous, whole-ecosystem experimental evidence on the magnitude and timing of fish MeHg recovery following reduced Hg inputs, thereby informing expectations for ecosystem and human health benefits from emission controls.

Background literature highlights global anthropogenic Hg emissions and deposition to aquatic systems, human health risks from MeHg exposure, and the complexity of MeHg production and bioaccumulation in food webs. Prior work has shown that ecological factors (e.g., food-web structure, species interactions, nutrient status), physical and chemical water properties, and climate variability can alter MeHg production, biomagnification, and fish growth, complicating predictions of fish MeHg responses to emission reductions. There has been a lack of tools to directly evaluate how recently deposited Hg contributes to MeHg in biota, limiting assessments of recovery dynamics in contaminated fish populations. The present study situates itself within this context by using enriched isotopic tracers in a whole-lake experiment to directly link changes in Hg loading to MeHg levels across ecosystem compartments and trophic levels.

Design: A 15-year whole-ecosystem experiment (METALICUS) was conducted in Lake 658, Experimental Lakes Area (ELA), Canada, spanning an addition phase (2001–2007) and a recovery phase (2008–2015). Enriched mercury isotopes were used as tracers ("lake spike") to simulate enhanced wet deposition. Hg loading: Hg enriched with distinct isotopes was applied to the lake surface, adjoining wetland, and upland catchment to increase local wet deposition approximately fivefold (from ~36 to ~191 µg m−2 y−1), comparable to polluted regions. Most Hg added to watershed surfaces remained in vegetation/soils or evaded to the atmosphere, contributing <1% of Hg in runoff and <2% to fish MeHg changes, indicating that in-lake additions dominated observed biotic responses. Sampling and compartments: MeHg and/or total Hg were measured across water (n = 516), sediments (n = 1,627), invertebrates including zooplankton (n = 211 for biota cited) and prey (e.g., Chaoborus/"Charr" as referenced), and fish (n = 1,052). Mean annual concentrations were reported for the open-water season for non-fish compartments; fish were sampled each autumn. Large-bodied fish concentrations were body-length standardized (e.g., northern pike 475 mm; lake whitefish 535 mm); planktivorous yellow perch age-1 were also standardized. Fish species and trophic guilds: Planktivore: yellow perch (Perca flavescens, age 1); piscivore: northern pike (Esox lucius); benthivore: lake whitefish (Coregonus clupeaformis). Dominant prey items for biomagnification analyses: zooplankton (to perch), forage fish (to pike), and Chaoborus/"Charr" (to whitefish). Metrics: Partitioned MeHg into lake spike (labelled) and ambient (background) components. Computed biotic concentrations and biomaniplication/biomagnification factors (BMF = [MeHg_lake(spike)/MeHg_ambient]) from prey to fish for each guild. Tracked temporal dynamics of spike vs ambient MeHg during addition and recovery phases. Recovery assessment: Ceased all enriched Hg additions in 2008 (100% reduction in spike loading). Tracked declines in spike MeHg across compartments and fish populations through 2015. For apex predator recovery, used non-lethal muscle biopsies and individual mark-recapture of northern pike sampled at the end of addition (2007) and resampled during recovery to estimate body-burden trajectories; modeled decline with exponential decay starting in year 2 of recovery. Statistical analyses: Linear regressions assessed increases during addition (significant P < 0.05 for most species). Exponential decay regression for northern pike spike MeHg body burden: y = 1.7439 × e^(−0.232 t), R² = 0.95, F(1,9) = 95.5, P = 0.0002; half-life ~2 years (50% reduction) beginning in the second year of recovery.

- Enriched Hg additions increased MeHg across ecosystem compartments and fish populations, confirming rapid incorporation of newly deposited inorganic Hg via in-lake methylation and trophic transfer. - During the addition phase, fish MeHg concentrations rose above ambient; increases varied by trophic guild, with stronger responses in pelagic-linked species and smaller responses in benthic-dwelling fishes. - Biomagnification of spike MeHg reached parity with ambient MeHg within trophic pathways over time: ~2 years for perch (zooplankton to perch), and ~4 years for pike and whitefish (via forage fish and Chaoborus/"Charr"), consistent with trophic magnification expectations. - After ceasing spike loading (2008), rapid declines in spike MeHg were observed throughout the ecosystem: within 3 years, spike MeHg declined by 81% in water, 38% in sediments, 66% in zooplankton, and 67% in Chaoborus/"Charr"; by the end of recovery, fish spike MeHg decreased by 85–91%. - Eight years post-addition, lake spike Hg contributed only ~6% to MeHg in fish and invertebrate prey. - Fish recovery was rapid but species-specific: within 8 years, spike MeHg declined by 76% in northern pike and by 32% in lake whitefish; overall key fish MeHg concentrations decreased by 38–76% within eight years following loading cessation. - Apex predator (pike) spike MeHg body burden showed an exponential decline with an estimated 50% reduction in 2 years (R² = 0.95), indicating swift benefits following loading reductions. - Differences in fish recovery rates were influenced by life history and food-web pathways: older, longer-lived whitefish (median age 17 y) with benthic associations recovered more slowly than younger, shorter-lived pike (median age 3 y). Reliance on pelagic versus benthic dietary pathways and lake water residence time modulated response timing.

The experiment provides direct, causal evidence that reductions in Hg loading to lakes rapidly propagate through lower trophic levels and culminate in substantial declines in fish MeHg, addressing a core uncertainty in translating emissions controls into ecological and human health benefits. Rapid declines in spike MeHg in water and zooplankton initiated recovery in fish, with species-specific rates reflecting trophic linkage (pelagic vs benthic pathways), longevity, and population age structure. The findings imply that emission controls under the Minamata Convention can yield relatively prompt benefits to fish consumers even in lakes with longer water residence times. However, the magnitude and timing of recovery will vary among systems depending on watershed contributions, sediment coupling, and food-web structure. Importantly, the study shows that newly deposited inorganic Hg is methylated in-lake and biomagnifies similarly to ambient Hg once equilibrated, and that stopping inputs leads to swift declines in biotic spike MeHg, particularly in apex predators with shorter generation times.

This whole-ecosystem isotopic-tracer experiment demonstrates that curtailing Hg inputs to lakes leads to rapid and substantial reductions in MeHg across food webs and in fish, providing experimental evidence that Hg emission controls translate into timely benefits for fish consumers. Key contributions include quantifying the timescales of equilibration and recovery for spike versus ambient MeHg, documenting ecosystem-wide declines after loading cessation, and identifying trophic and life-history determinants of recovery rates. Future work should assess generality across lake types, climates, and watershed conditions, quantify the roles of benthic–pelagic coupling and sediment legacy pools, and integrate climate and food-web dynamics to refine predictions of fish MeHg recovery under ongoing emission reductions.

- Single-lake, boreal ecosystem study may limit generalizability across diverse lake types, climates, and watershed settings. - Watershed Hg loading may decline more slowly than atmospheric deposition, potentially moderating or delaying recovery in systems with strong watershed or sediment inputs. - Species- and food-web-specific pathways (pelagic vs benthic reliance) and life histories influence response rates, implying variable outcomes among ecosystems. - Long water residence time and sediment interactions can affect timing; results indicate rapid response despite this, but other systems may differ. - Some data gaps (noted missing data in figures) and reliance on length-standardized fish concentrations could introduce uncertainty in cross-year comparisons.