The word microplastics entered the scientific vocabulary through marine science, when Thompson et al. (2004) drew attention to microscopic plastic fragments and fibres in sediments and plankton samples. For some years the subject stayed at sea. Rillig (2012) was among the first to ask whether soils, which receive plastic from many of the same sources, might be affected as well, and later estimates suggested that the quantity added to farmland each year through sewage sludge alone could rival or exceed the stock floating on the surface of the oceans (Nizzetto et al., 2016).

Microplastics are most often defined as plastic particles smaller than 5 mm, with nanoplastics forming the fraction below one micrometre, although the boundaries and the vocabulary of shapes and polymer types are still being harmonised (Hartmann et al., 2019). Research on their behaviour in terrestrial systems has expanded quickly, and several reviews now treat soil as a major reservoir (de Souza Machado et al., 2018; Hurley and Nizzetto, 2018). Most of that work concerns the topsoil and the organisms that live in it.

The question taken up here is what happens below. Shallow groundwater beneath farmland supplies wells, feeds streams and is already known to be vulnerable to nitrate and pesticides that originate at the surface. Whether plastic particles follow the same route depends on processes that soil physics and colloid science have studied for decades in other contexts. This review connects that older knowledge with the newer literature on microplastics in soil and groundwater.

The article is a narrative review of the published literature. It was not designed as a systematic search, and it makes no claim to cover every study. Sources were selected to represent four bodies of work: studies of plastic inputs to agricultural soil, experiments on the movement of plastic particles in soil, the established theory of colloid transport in porous media, and surveys of plastic particles in groundwater. Where the microplastic literature is thin, the review relies on well-tested principles from the study of other particles and says so.

Sewage sludge is the best documented input. Treatment works retain most of the particles in wastewater, which concentrates them in the solids that are then spread on land. Zubris and Richards (2005) showed that synthetic fibres could be used as a marker of sludge application because they remained detectable in soil long after spreading had ceased, and Corradini et al. (2019) found that the microplastic content of field soils rose with the number of sludge applications they had received. Composts and digestates made from municipal organic waste carry fragments of packaging that survive sorting.

Plastic mulch film is the second large source. Thin polyethylene films are difficult to recover completely after harvest, and the residues are broken up by tillage and weathering into progressively smaller pieces (Steinmetz et al., 2016). Irrigation with untreated or partly treated wastewater, littering, wear of tyres on farm tracks and deposition from the air add smaller amounts. The result is a surface layer loaded with particles of many sizes, shapes and densities, from fibres millimetres long to fragments too small to see.

Three kinds of process carry particles downward: mechanical mixing, biological transport and movement with water. They operate over different depths and select for different particle sizes.

Direct measurements in aquifers are scarce. Panno et al. (2019) reported microplastic fibres in springs and wells of karst aquifers together with pharmaceuticals and other markers of wastewater, a setting in which conduits connect the surface to groundwater with almost no filtration. Mintenig et al. (2019) examined raw groundwater and the drinking water produced from it and found very low numbers of small fragments, whose polymer types suggested that abrasion of plastic components in the wells and treatment plant was at least as likely a source as the aquifer. Re (2019) reviewed the hydrogeological reasoning and argued that fibres deserve particular attention because their small diameter allows passage through pores that would retain a fragment of the same length.

These findings fit the theory. Rapidly recharged systems, such as karst, fractured rock and coarse gravels, are those in which particles have been found, and granular aquifers appear to be well protected against all but the finest fraction. The inference must be held loosely, because the number of studies is small and the data are of uneven quality. A critical assessment of studies on microplastics in fresh and drinking water concluded that few met basic criteria for sampling, blank correction and polymer identification (Koelmans et al., 2019).

Measuring microplastics in soil is harder than in water. Particles must be separated from a matrix of mineral grains and organic matter, usually by flotation in a dense salt solution followed by chemical digestion of organic residues, and then identified by infrared or Raman microscopy or by thermal methods such as pyrolysis coupled to gas chromatography and mass spectrometry (Bläsing and Amelung, 2018). Each step loses some particles, and the losses are greatest for the smallest sizes, which are also the most mobile. Spectroscopic identification becomes impractical well above the nanometre range, so the fraction most likely to reach groundwater is the fraction about which least is known.

Contamination is the second difficulty. Groundwater is sampled through plastic well casings, pumps and tubing, and laboratory air carries textile fibres. Without field and procedural blanks a low count cannot be distinguished from background. Studies also report results in different units, counts per volume or mass per mass, over different size ranges, which prevents comparison between sites and makes mass balances between soil and aquifer impossible to close.

The largest gap is the absence of studies that follow particles from surface to water table at a single site. What is needed is a field with a known history of sludge or mulch use, sampled by depth through the unsaturated zone, with the soil structure characterised and the underlying groundwater monitored through wet and dry seasons using plastic-free equipment. Tile-drained fields offer a convenient intermediate step, since drain outflow integrates preferential flow over a known area and can be sampled during storms.

Laboratory work should move from repacked sand to undisturbed soil monoliths, and from pristine spheres to weathered fragments and fibres of the polymers found in fields. Transport models developed for colloids and pathogens can be adapted, but they require attachment and straining parameters for realistic particles. Finally, the role of plastic particles as carriers of sorbed pesticides or metals, by analogy with natural colloids, has been proposed more often than it has been tested in soil.

Agricultural soils receive plastic particles in quantity, and several mechanisms move them below the surface. Tillage and soil animals redistribute them within the upper profile. Percolating water can carry them further, but the soil matrix is an effective filter for all except the smallest particles unless structure allows water to bypass it. Shallow groundwater beneath structured or fractured materials is therefore the most exposed, and the limited survey evidence agrees. Present methods cannot quantify the fine fraction that theory identifies as most mobile. Until measurements at field sites connect loading, soil structure and groundwater quality, statements about the scale of the pathway should be made with caution in either direction.

To be declared by the author(s) before publication.

To be declared by the author(s) before publication.