Deep-sea protected areas in the Pacific mining zone are barely connected, model shows

The network of protected areas set aside to safeguard biodiversity in the Clarion–Clipperton Zone, the vast abyssal region of the Pacific Ocean slated for the world’s first commercial seabed mining, may not function as a connected system at all, according to a new modelling study published in Nature Ecology & Evolution. Using Lagrangian particle-tracking simulations driven by ocean circulation data, researchers found that fewer than 0.02% of simulated larvae released from one protected area reached another, with the gaps between reserves far exceeding what passive near-bed currents can bridge within biologically plausible larval lifespans. The finding arrives at a critical moment, with more than 1.1 million square kilometres of the zone already under exploration licence and commercial mining looming.

To protect regional biodiversity in the Clarion–Clipperton Zone, the International Seabed Authority has designated thirteen Areas of Particular Environmental Interest, each spanning roughly 200 kilometres with surrounding buffers, arranged around the licensed mining areas. The network was designed around habitat representativity, with connectivity between the reserves judged spatially infeasible at the time of their placement. Yet whether the system functions as a coherent, connected whole had never been formally assessed until now.

The research team, led by Molly Kendall James and colleagues, simulated the dispersal of more than 68 million particles per run using the HYCOM global ocean circulation model, tracking larvae confined to the bottom layer of the water column, roughly ten metres above the seafloor. They tested three planktonic larval durations of 19, 35 and 69 days, representing the lower quartile, median and upper quartile of pooled deep-sea and eurybathic species, and ran simulations for January and July of both 2019 and 2023 to capture seasonal and interannual variability. Median dispersal distances remained under 100 kilometres even at the longest duration, far short of the 200 to 800 kilometre gaps separating the protected areas.

The results were stark. Of 156 possible direct connections between the thirteen protected areas, only five showed any exchange at all, and three routes accounted for over 99% of all transfers between reserves. Retention within individual reserves was high, with the most retentive areas keeping 88% of their particles and the least retentive still holding 60%. For particles that failed to settle within the network, the median distance to the nearest non-origin reserve was 389 kilometres, suggesting that even multi-step connectivity through intermediate sites would remain limited under the modelled assumptions.

Network analysis of dispersal across the entire zone revealed a strongly modular structure, with larval exchange largely confined within four semi-isolated connectivity communities. The largest, in the northwest, spans 4.41 million square kilometres, or 36% of the region. One community contains no protected areas at all, illustrating how connectivity-defined regions can fall outside frameworks built on habitat representativity. When the researchers removed particles originating from mining areas to simulate a disturbance scenario in which mined zones no longer function as larval sources, connectivity declined across the network, with the largest community losing 23% of its outward dispersal and 24% of its particle input. Several boundary regions shifted their community affiliation under this scenario, suggesting the regional connectivity structure could reorganise if mining proceeds.

Perhaps the most striking finding concerns where the larvae that replenish mining areas actually come from. Under undisturbed baseline conditions, mining areas are predominantly self-supplying, with 76% of settlers originating within the mining zones themselves, 19% from unprotected non-mining areas and just 5% from the protected network. When self-supply is removed in the disturbance scenario, unprotected non-mining areas dominate external replenishment, contributing 81% of remaining settlers, while the protected areas account for 19%. Many of these unprotected source regions sit adjacent to existing reserves or in the central zone, a partly geometric consequence of how the reserves were originally placed around pre-existing exploration contracts rather than positioned to supply larvae to mining areas.

The team also identified the locations that consistently act as larval sources and stepping stones across both baseline and disturbance scenarios. These scenario-stable connectivity hot spots are widely distributed, but 77% of their total area falls outside existing protected areas, many occurring along the margins of mining zones. Their persistence across both scenarios suggests they are robust features of the dispersal landscape, and their lack of formal protection, the authors argue, warrants consideration within the resilience framework the reserve network was explicitly designed to deliver.

To test whether the structural deficit could be repaired, the researchers developed a spatial prioritisation framework combining an integrated connectivity value, built from larval exchange with reserves, replenishment of mining areas, scenario-stable connectivity and transitional status, with a graph-theoretic optimisation adapted from Steiner tree methods. The algorithm identified 53 grid cells of 100 by 100 kilometres that, if protected, would act as corridors between the existing reserves. The effect was transformative: while direct links between reserves remained unchanged, the additions enabled 151 indirect pathways, increasing inter-reserve dispersal 406-fold, from one in 5,556 particles to one in 14 released from protected areas reaching another protected area within the modelled network. The corridors also halved the median near-miss distance of unsuccessful dispersers, from 389 to 154 kilometres.

The benefits extended to replenishment of mining areas as well. The source-normalised contribution of the protected network to larval supply of mining zones doubled from 4% to 8%, and the proportion of mining sites receiving sustained particle input above the study’s 1,000-particle threshold rose from 71% to 85%. Under the mining disturbance scenario the improvement was most pronounced, with the protected network’s share of remaining settlement rising from 19% to 36%, nearly doubling its backstop capacity and reducing dependence on unprotected sources from 81% to 64%. Of the 53 proposed corridor cells, 29 did not overlap mining areas, and only 11% of the total proposed corridor area intersected existing contract zones, leaving substantial unmined habitat within the proposed additions.

The authors are careful to frame their results as conditional outputs of a physically grounded model rather than empirically validated rates of exchange. Direct observations of bottom currents at abyssal depths remain sparse, larval durations for the zone’s nodule fauna are essentially unknown, and vertical swimming behaviour, which could expose larvae to faster and differently directed currents higher in the water column, was not modelled. The four simulated time periods do not span the full range of El Niño–Southern Oscillation variability, and the disturbance scenario captures only the loss of larval sources, not sediment plumes or acoustic impacts that could extend harm beyond contract boundaries, making it a conservative lower bound on mining-related disruption. The analysis evaluates structural dispersal pathways rather than demographic sufficiency, and cannot determine whether connectivity is adequate to ensure population persistence. Even so, the researchers argue that in one of the most data-poor environments on Earth, a transparent, reproducible baseline is more useful for planning than a complex model built on poorly constrained biological assumptions. They advocate embedding connectivity into conservation design and refining the projections as empirical data accumulate, before commercial mining at scale proceeds.

Subject of Research: Modelled larval connectivity among deep-sea protected areas in the Clarion–Clipperton Zone and the potential of dispersal corridors

Article Title: Dispersal corridors enhance modelled near-bed larval connectivity across deep-sea protected area networks

Article References: James, M. K., Bridges, A. E. H., Jones, D. O. B., Clubley, C. H., Knights, A. M., DeJonge, D. S. W., & Howell, K. L. (2026). Dispersal corridors enhance modelled near-bed larval connectivity across deep-sea protected area networks. Nature Ecology & Evolution. https://doi.org/10.1038/s41559-026-03172-5

Image Credits: AI Generated

DOI: 10.1038/s41559-026-03172-5

Keywords: Clarion-Clipperton Zone, deep-sea mining, larval dispersal, marine protected areas, connectivity modelling, Lagrangian particle tracking, International Seabed Authority, abyssal ecology, network analysis, dispersal corridors, HYCOM, conservation planning