Since 2007, half of the world population is leaving in urban environments, and this ratio is expected to reach 60% before the end of the decade (United Nations [UN], 2019). The huge development of urban populations consumes so much resource that natural provision cannot be sustained any more (Swilling, 2011). That is particularly the case of topsoil, which is imported to the cities from rural areas for the creation of green spaces. Concomitantly, urban agriculture has flourished in developed countries, bearing more and more weight in the local urbanism plans and in public policies. The services provided by urban agriculture strengthened its potential to contribute to quality of life in cities (Zasada, 2011; Lin et al., 2015). However, the soil quality in cities is often severely degraded (Morel et al., 2015) due to harsh environmental conditions, which may limit the development of such agronomic activities. Urban soils often bear strong legacies of human land-use, road traffic and industrialization, which can be quantified by the degree of various organic and/or mineral pollutants found at different depths. For instance, a study along a gradient of urban pressure in public green spaces of the Paris region (France) showed that most of the soils were highly contaminated with trace metals such cadmium (Cd), lead (Pb), copper (Cu), and Zinc (Zn), especially in woods (Foti et al., 2017). These contaminations make soils unsuitable for gardening or food production. Urban gardeners are particularly exposed to sanitary risks, either through the consumption of their own food that may accumulate harmful substances or by contact and inhalation of the contaminated soil.

To face the problem of soil pollution and lack of available space, soil-less agriculture as hydroponics, aquaponics or other artificial substrates are sometimes considered as an alternative. However, these technical solutions are limited by their high cost and the suboptimal nutrient supply to plants (Yep and Zheng, 2019). Moreover, they cannot entirely replace essential ecosystem services that are associated to soil biodiversity, such as nutrients cycle and water purification (Millenium Ecosystem Assessment, 2005). Thus, cities stakeholders often import huge quantity of topsoils from rural area to recover or replace urban soils. In France, the quantity of soil imported for the creation of green spaces has been estimated to 3 million m³ each year (Damas and Coulon, 2016). These practices, which are dependent of the availability of natural soils, are poorly sustainable since they rely on the extension of artificial land. The ecological and economic costs of importing soils from rural areas are continuously increasing. Therefore, alternatives that are more virtuous have to be found. Paradoxically, a solution may result from the increasing activities of construction and demolition in the cities. Indeed, urban renewal sites generate huge quantities of mineral wastes. In particular, the construction of foundations for buildings, garages or subways requires the excavation of large volumes of deep soil horizons or bedrocks that have no immediate use. Treated as wastes, such materials are commonly exported and managed outside the cities, because the urban space is scarce. In Paris, more than 15 Mt of excavated deep horizons (EDH) are treated in landfill sites every year (Ile-de-France, 2019). Should a small fraction of the latter be reuse as a substitute of natural soil, it could avoid the importation from rural areas. In New York City, it has been estimated that 1.7 Mt of surplus native soil is generated each year (Walsh et al., 2019).

Creating fertile soils by combining organic and mineral anthropogenic materials may have appeared a challenge as it takes several hundred years to obtain a few centimeters of fertile soil in natural conditions. However, during the last 15 years (Lehmann, 2011), a number of studies published in this field of research showed promising results. Most of these constructed soils may be qualified as constructed Technosols. According to the World Reference Base, Technosols are defined by soils subjected to a strong human influence and containing at least 20% of artifacts (IUSS Working Group WRB, 2014). Not all engineered soils can be included in this definition. For example, human-transported natural soil material does not qualify as Technosol (Schad, 2018). Regardless of the soil classification used, engineered soils have been shown to provide a wide variety of functions and ecosystem services (Morel et al., 2015). Urban mineral wastes such as EDH have been successfully used in several urban forestry projects (Cannavo et al., 2018; Yilmaz et al., 2018; Pruvost et al., 2020), however, very few demonstrations have been made regarding their use for urban agriculture. Yet, it has been demonstrated that excavated sediments in New York city can be used as constructed soils to reduce contamination risks for urban residents (Egendorf et al., 2018).

The objective of the present study was to evaluate the effect of compost addition and earthworms inoculation to improve the agronomic properties of EDH. Laboratory experiments performed on the same materials with ryegrass showed that the combination of compost and earthworms increased both soil water retention and plant biomass productivity (Deeb et al., 2016a,b). Here we tested in situ (1) the dose effect of green waste compost, (2) the influence of earthworm introduction and (3) the interactions between these two factors on the production of green vegetables. For this study, three types of salads were tested in mono- and co-culture. The experience was repeated tree times in spring, summer and autumn. We hypothesis that the highest biomass productivity will be reached on the highest concentration of compost in presence of earthworms.

This study clearly demonstrated that compost and earthworms have strong interactive effects on soil fertility and plant growth. As mentioned before, this hypothesis was already tested in laboratory and under controlled conditions with the same materials but a different earthworm’s species (Deeb et al., 2016a,b). However, in situ demonstration of the positive interaction between organic matter addition and earthworms inoculation on soil quality and primary production are not so frequent. In a totally different context, Senapati et al. (1999) showed that a combination of organic matter inputs and earthworms (P. corethrurus), enhance the quality of a tropical soil in India and enhanced the production of tea by 80–276%. The present study confirm that earthworms inoculation can significantly improve the fertility of engineered soils. Here, arugula production was improved two to five times with earthworms in soil containing 30% compost, depending on the season. For comparison, it is estimated that the addition of earthworms in croplands improves yield by about 25% (van Groenigen et al., 2014). Inoculations of earthworms has been shown to accelerate the restoration of degraded soils (Butt, 2008; Edwards et al., 2019). Their role on ecosystem and their beneficial effects on crop systems are widely recognized (Blouin et al., 2013). Earthworms can improve soil structural stability that optimize nutrient cycling and improve plant growth and health (Bertrand et al., 2015). This positive effect on plant production could be due to the hydrostructural structural improvement of the soil as we observed an increased in the microporosity and moisture in the presence of the earthworm, which is in line with previous results obtained in laboratory with the endogenic earthworm species A. caliginosa and L. perenne plant (Deeb et al., 2016b).

In this study, we used L. terrestris, who feeds mainly with organic matter (Griffith et al., 2013) and creates a few large vertical burrows (Bastardie et al., 2003), which are well known for improving water infiltration (Edwards et al., 1990). When the organic matter is not sufficiently available, L. terrestris, tends to go into diapause (Bouché, 1977). This could explain why the effect of earthworm was lower or not significant in soils containing less than 30% of compost. Furthermore, L. terrestris burrowing volume and depth are strongly affected by soil bulk density (Ducasse et al., 2021). Since the bulk density of EDH is about twice that of compost, soils containing 30% compost are significantly less compact than the others, so burrows are probably larger and deeper, which may stimulate both roots and microbial activities, as suggested by the small but significant effect on the pH. If one of these assumptions is true, higher doses of compost may enhance the impact of this earthworm species, which is the most commercially available and best studied species for soil restauration (Steckley, 2020). The benefits of introducing earthworms and adding organic matter show the enormous potential of this approach for the restoration of degraded soils and soil engineering in general. Many more studies should be conducted to evaluate the survival and the potential effects of other earthworm species in combination with other sources of organic matter and taking into account possible toxic effects related to the presence of certain chemical compounds in the soil beneath. To date, very few studies have been carried out to test the influence of different earthworms on the functioning of Technosols (Pey et al., 2013).

A lot of research have been done to develop a quality standard for composted green wastes as amendments, especially in Europe (Carlile, 2008; Barrett et al., 2016). Contrastingly, few studies have investigated the agronomic capacity of mineral waste such as EDH with compost addition. Used alone as growing media, green waste compost provides ideal conditions for soilless agriculture (Hartz et al., 1996; Grard et al., 2018). Lettuce, arugula and spinach usually grow better in soils amended with compost and are even less vulnerable to biological diseases (Maftoun et al., 2005; Giménez et al., 2019). Cultivated during a short period, as in our study, the plants nutrient requirement is relatively low (Grahn et al., 2015) and we expected a rapid response to the compost addition as previously observed in laboratory conditions (Deeb et al., 2016a,b, 2017). That was also the case in this outdoor experiment, but only about 6 months after soils construction. For instance, going from 10 to 30% of compost in the soil mixture allows to multiply by 2 the arugula biomass and even by 4 in the presence of earthworms. Again, the effects of compost are often either enhanced or enhanced by the presence of earthworms, depending on the type of plant or the season. Here, the high amount of carbonates in the mineral materials used may also have limited the mineralization of the organic matter and delayed the nutrients feedback to the plant (Virto et al., 2018). Indeed, organic amendments are less effective in carbonate-rich soils than in non-calcareous soils (Romanyà and Rovira, 2009). In the present study, the soils experienced three different seasons and thus, several wet and dry cycles that conditioned the biological activity carried out by microorganisms and the roots. Both physical and biological phenomena are well-known factors on soil structuring (Oades, 1984; Degens and Sparling, 1995; Chenu and Cosentino, 2011). The stability of structures has probably been reinforced over time, though it is considered short in the pedogenesis scale. Nonetheless, other studies on constructed soils have already highlighted their rapid evolution, particularly those where macrofauna have colonized (Séré et al., 2010; Jangorzo et al., 2014).

Pure mineral materials have a low capacity to support plants development. Here, by simply mixing EDH with organic matter such as compost, we were able to produce an average of nearly half a kilogram of vegetables per m² across all crops and seasons, including all types of soil with and without earthworms. The constructed soils assured a high rate of germination of all leafy vegetables when climate was wet. However, in the summer when temperature and radiation were elevated, the germination rate of the lettuce was considerably more impacted than the other two plants. This is may be due to the fragility of lettuce seedlings, which may have been more affected by surface crusting quickly in the summer. Crusts formation is a common problem in these soils and the reason is probably the high amount of carbonates in the parent material (Valentin and Bresson, 1998). Moreover, the high pH of the soils may have slowed the root development of all plants (Tang and Turner, 1999), but arugula to a lesser extent, as this plant is less sensitive to soil alkalinity (Garg and Sharma, 2014). These first results encourage the testing of many other crops adapted to this type of soil, including also aromatic plants or flower crops.

In this outdoor study, we demonstrated for the first time that mixtures of excavated deep horizons of soils and green waste compost are able to support the development of green vegetables. After 6 months of soil construction and three growing seasons, we could observe a significant effect of both compost dose and earthworms inoculation on plant production, especially for arugula. These promising results encourage further researches to improve engineered soils for urban agriculture by testing different organic amendments and/or mineral wastes as well as more demanding cultures (e.g., tomatoes or eggplants). Moreover, the impact of different earthworms species on soils restoration is still poorly explored now. It would be interesting to investigate the effect of combining species belonging to different functional groups. Finally, we show in this preliminary study that it was necessary to wait few months to observe a significant effect of compost on soil fertility. Therefore, monitoring these engineered soils over long-term period is a mandatory step to assess their evolution and thus, better inform decision makers about the choice of substrate for urban agriculture.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

TZL conceived the study, directed research, performed some statistical analyses. JHRA performed the experiment under the supervision of TZL and CH. AP-B, HA-B, and TD contributed to the analyses. TL and JHRA wrote the article. All authors contributed substantially to revisions.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.