Received: 26 March 2021 Revised: 14 October 2021 Accepted: 16 October 2021 DOI: 10.1002/ldr.4134 RESEARCH ARTICLE Water-saving techniques for restoring desertified lands: Some lessons from the field Vicenç Carabassa1,2 | Daniela Alba-Patiño3 | Sergio García1 | Julián Campo4 | 5 6 3,7 Harrie Lovenstein | Gertruud Van Leijen | Antonio J. Castro | Francisco González | Gustavo Viera | Dimitrios-Sotirios Kourkoumpas10 | 8 9 Argyro Aliki Zioga10 | Christos Emmanouel Papadelis10 | Vicente Andreu4 | Eugènia Gimeno4 | Sven Kallen11 | Josep Maria Alcañiz1,2 1 CREAF, Univ Autònoma de Barcelona, Cerdanyola del Vallès, Spain Abstract 2 Universitat Autònoma de Barcelona, Nature-based solutions can significantly contribute to restoration projects in areas Cerdanyola del Vallès, Spain affected by desertification processes, where they are necessary for reversing land 3 Department of Biology and Geology, Andalusian Center for the Assessment and degradation. Currently, one innovative solution is The Cocoon™, which has been Monitoring of Global Change (CAESCG), designed as a new ecotechnology for improving seedling establishment. The Cocoon Universidad de Almería, Almería, Spain 4 consists of a doughnut-shaped container made of recycled cardboard that provides Centro de Investigaciones sobre n (CIDE-CSIC), Valencia, Spain Desertificacio water and shelter at least during the first year of a seedling, which is the most critical 5 Land Life Company, Amsterdam, The for plant establishment. To determine the effectiveness of this ecotechnology under Netherlands 6 different conditions, the Cocoon was tested on a variety of soils, climates, vegetation, Van Leijen Srl, Rome, Italy 7 and land uses. Six planting trials were performed in Spain and Greece, which covered Department of Biological Sciences, Idaho State University, Pocatello, Idaho, USA a range from humid to arid climates. With the objective of studying its functionality, 8 Consejería de Medio Ambiente, Cabildo de the survival of the seedlings, their vigor, and growth were monitored for 2 years. Gran Canaria, Las Palmas de Gran Canaria, Spain Compared with conventional planting systems, the Cocoon has effectively increased 9 n y Planeamiento Territorial y Gestio seedling survival, especially under dry growing conditions (low rainfall, soils with low Medioambiental, Gobierno de Canarias, Las water holding capacity). The Cocoon also allowed for higher growth of some species Palmas de Gran Canaria, Spain 10 Centre for Research and Technology Hellas/ (olive trees, holm oaks, and Aleppo pines). Moreover, a positive correlation between Chemical Process and Energy Resources the rainfall on the site and the biodegradation degree of the Cocoon device was Institute (CERTH/CPERI), 4th km Ptolemaida Mpodosakio Hospital, Athens, Greece observed. Overall, the Cocoon becomes more efficient in arid climates or adverse 11 VOLTERRA ECOSYSTEMS, Bellaterra, Spain growing conditions. Correspondence KEYWORDS Vicenç Carabassa CREAF, Soil Protection and climate change adaptation, Cocoon, drylands, irrigation, planting Restoration Group, Edifici C, Universitat Autònoma de Barcelona, E08193 Bellaterra (Cerdanyola del Vallès), Catalonia, Spain. Email: v.carabassa@creaf.uab.cat Funding information EU Life Program, Grant/Award Numbers: LIFE15 CCA/ES/000125, LIFE19 CCM/ES/001199 This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2021 The Authors. Land Degradation & Development published by John Wiley & Sons Ltd Land Degrad Dev. 2022;33:133–144. wileyonlinelibrary.com/journal/ldr 133 134 CARABASSA ET AL. 1 | I N T RO DU CT I O N experience nutrient limitations when the seedlings are transferred to the soil, which is typically poor in the Mediterranean region (Díaz- One of the most current ecological concerns is the increasing deserti- Hernández et al., 2003). Water stress increases in mine soils because fication rate as a direct impact of the climate crisis. According to the of their shallow depth and lack of soil structure. Thus, they have very United Nations Convention to Combat Desertification, areas with the pez-Marcos et al., 2020), which limit low water holding capacity (Lo highest susceptibility to desertification are dry, arid, semiarid and sub- plant establishment even in Mediterranean subhumid climate (Alday humid areas (MAP, 2019), such as large parts of the Mediterranean et al., 2016). Even if reforestation is carried out with regular irrigation, region. These 'drylands' occupy 41% of the Planet's land surface and the survival rate is at most 50% for many species, but in many cases are inhabited by ca 2 billion people (MEA, 2005). A common trait of even less, since root systems with inadequate irrigation do not pene- these areas is that the aridity index ranges between 0.05 and 0.65 trate deep enough into the soil and remain in the surface layers (MAPA, 2019). (Salem, 1989). Desertification, particularly its consequent reduction of ecosys- In this context, nature-based solutions could help solve environ- tem services, can threaten future improvements in human wellbeing mental problems and improve reforestation projects by increasing and impede progressive efforts in dryland areas affected by climatic seedling establishment. Nature-based solutions are defined as actions impacts, such as control measures for dust storms or floods. Desertifi- that protect, sustainably manage, and restore natural or modified eco- cation reduces primary production and microbial activity, modifies the systems addressing societal challenges effectively and adaptively, nutrient cycles, and increases soil degradation, which altogether result while simultaneously providing human well-being and biodiversity in the inability to capture carbon and the loss of biodiversity of benefits. Nature-based solutions are usually regarded as an umbrella affected ecosystems (MEA, 2005). The consequences of desertifica- concept that covers a range of different approaches (Cohen-Shacham tion can also be their causes, thus embodying a detrimental cycle. et al., 2016). Therefore, ecotechnologies designed for supporting res- Therefore, combating desertification becomes one of the great global toration projects and plantings could also be included in this group of environmental challenges, and its effects must also be considered solutions. globally. The Cocoon™, a new water-saving ecotechnology for supporting However, at the local level, desertification may depend on the plantings in drylands, is currently used abroad (Land Life combination of multiple factors and site-specific processes that may Company, 2021). The Cocoon resembles the buried clay pot used in aggravate the problem. These include indirect factors, such as popula- ancient times, in which water slowly seeps into the subsurface to sup- tion size pressure, political and socioeconomic scenarios, on the one port plant growth and restricts evaporation losses, as would be hand, and direct factors, such as land use and management as well as expected during conventional watering of the soil surface. Instead of climate-related processes, on the other. The main aggravating factors fragile and bulky pots, a paper pulp-based alternative was devised to at the local level include seasonal droughts with extreme rainfall vari- support early tree establishment in reforestation schemes under dry- ability and/or heavy rains, poor soils prone to erosion, steeped slopes land conditions. With preliminary tests of an early version of the that increase the energy of runoff, recurrent forest fires causing loss Cocoon as proof of principle, larger field tests have been laid out since of vegetation cover and changes in the physical, chemical, and biologi- 2016 in the Mediterranean basin and Canary Islands (VOLTERRA, cal soil properties (Campo et al., 2006, 2008), crisis of traditional agri- 2021) and in the Lower Rio Grande Valley of South Texas (Mohsin culture resulting in land abandonment, unsustainable exploitation and et al., 2021), among others. salinization of aquifers, bad agricultural and livestock practices, and The objective of this work is to analyze the results of the large- overpopulation in some areas (MAP, 2019). scale implementation of the Cocoon solution by means of the data All these factors cause direct impact and stress on vegetation and collected in different field trials carried out in restoration projects. The its growth. Most of these conditions occur or have occurred through- main parameters evaluated have been seedling survival, vigour, and out the Mediterranean basin. Specifically, more than two-thirds of the growth to verify the effectiveness of this technology in the wide Spanish territories are classified as arid, semiarid, and dry subhumid range of land uses and environmental conditions present in six study areas, and more than two-thirds of these territories present a risk of areas located in the Mediterranean region and the Canary Islands. desertification to a greater or lesser degree (MAP, 2019; WWF, 2016). The capacity of ecosystems to regenerate is limited in those areas, and therefore restoring degraded land is becoming essen- 2 | MATERIAL AND METHODS tial for restoring the integrity of impacted forests, rangelands, mine- affected areas, and numerous habitats that host valuable biodiversity 2.1 | The Cocoon device (Muñoz-Rojas et al., 2021). At present, despite the efforts made, reforestation in the Mediter- The Cocoon™ consists of a doughnut-shaped container (like a torus ranean region cannot be considered satisfactory in many cases due to geometrical figure) made of recycled cardboard. This device has a the slow growth of planted seedlings and extremely high mortality capacity for 25 litres of water and a central space to install the seed- rates (Valdecantos et al., 2014). Planting sites suffer water stress due ling. It is designed to provide water and shelter to the seedling, at least to droughts that last between 3 and 5 months. They can also during its first year, which is usually its most critical survival stage. A CARABASSA ET AL. 135 lid reduces evaporation losses from the bowl and a shelter protects was used. Thus, each control provides paired measures with its associ- the seedling against small herbivores and reduces evapotranspiration. ated Cocoon, whereby an encoding system was set that allowed data For the Cocoon installation, the soil must be prepared beforehand by coupling. first digging a hole of 20 cm depth and 50 cm diameter, where this In total, 22,301 seedlings of 31 different species or varieties were device can be introduced (Figure 1). Over time it biodegrades and planted on an entire surface of 73 ha (see Table S1), according to res- integrates itself into the ground. toration objectives, environmental conditions and climate change sce- narios. Planting was carried out in two phases: the first phase in Autumn 2016, and the second in Spring–Summer 2017. Cocoon 2.2 | Study areas installation was carried out using a 50 cm diameter drill installed in a tractor or a backhoe whenever topographic and soil conditions made The Cocoon system was used in six restoration areas located in Spain it possible. In steeped slopes and extremely stony soils, holes were and Greece (see Table S1). In Spain, five large demonstration areas were made manually (see Table S1). Once planted, the Cocoons were filled located in El Bruc (Catalonia), Jijona and Tous (Valencia), Sierra de María with 25 L of spring water, in addition to natural rain, while the con- (Almería), and Tifaracás (Canary Islands); and in Greece, one area in trols were watered with a similar amount of water, but no refilling/ Ptolemais (Western Macedonia). These areas cover a variety of soils, irrigation was performed thereafter, with the exception of third and Mediterranean mesoclimates (from humid to arid), vegetation and land fourth plantings in Tifaracás, where the Cocoons were refilled. use, where the effectiveness of the Cocoon device can be tested in dif- ferent conditions on several desertification scenarios, and in combina- tion with different nature-based solutions for forest fire vulnerability 2.4 | Monitoring parameters reduction, endangered/endemic species protection, open-pit mines res- toration or recuperation of agricultural land (CREAF, 2017a). Monitoring parameters were divided into two groups: (i) one for eval- For our case-studies, we focused on burned forest soils with rela- uating the Cocoon effects on plant vigour and growth, vegetation tively high organic matter content in El Bruc and Tous and poor soils exclusion, and Cocoon biodegradation; and ii) the other for evaluating with low organic matter content in Jijona (abandoned cropland) and the recovery by means of passive restoration of the plantation areas. Tifaracás (volcanic parent material). In Ptolemais, the soil derives from Plant vigor was evaluated according to the following semiquanti- the mining debris of a former coal mine, where the planting was car- tative scores during their normal growing period: ried out. The soil in this particular location presents very high contents 3: Healthy seedling, with more than 75% of green, no wilted of carbonates and coal particles. Table 1 presents the main character- leaves, with active growing points (apices) visible istics of the studied areas, and Table 2 presents the respective soils. 2: Affected seedling, with 25%–75% of the leaves being wilted, yellow, or brown 1: Severely affected seedling with less than 25% of the leaves 2.3 | Planting being green (i.e. the majority wilted, yellow or brown) 0: Presumably dead seedling with no leaves or only wilted leaves; The planting scheme was based on the combination of one set of however, seedlings may still recover by resprouting after rain seedlings planted directly in the soil (controls), which represent the occurs traditional way, and another set planted with the Cocoon. Each con- R: Resprouted seedling trol was surrounded by several associated Cocoons, depending on the With the use of a caliper, plant growth was assessed by measuring planting possibilities of the site. As a rule, a 1:3 control:Cocoon ratio maximum plant height, from the root crown to the shoot apex, and the F I G U R E 1 Cocoon scheme and functioning [Colour figure can be viewed at wileyonlinelibrary.com] 136 CARABASSA ET AL. stem diameter at the tree base (at the level of the Cocoon's lid, at 10 cm of soil). Vegetation exclusion was evaluated in 1 m diameter cir- Forest fire vulnerability Cropland restoration + Nature-based solution rambla* restoration Endangered/endemic species restoration Cropland restoration cle around the seedling, measuring vegetation cover in two perpendic- Passive restoration Mine restoration ular transects. Additionally, biomass was evaluated by harvesting and acceleration reduction weighing all the vegetation inside the circle (wet weight: weight at scenario field; dry weight: weight after drying at 60 C for 4 days). Cocoon bio- degradation was evaluated by ranging each Cocoon from States 1 to 4, 1 being the intact device, and 2, 3, and 4 being increased biodegra- Conglomerates dation until complete incorporation into the soil. State 2 corresponded Cretaceous Marls-lignite dolomites Cretaceous Geological to a Cocoon without lid or a partially collapsed one. substrate Volcanic marls Alluvial Passive restoration was measured by means of vegetation struc- Early ture measures and floristic inventories. Structure was evaluated by quantifying cover types and height, each having 20 cm in 25 m tran- sects, a minimum of 3 transects per ha, which are parallel and perpen- Mediterranean Dry subhumid Dry subhumid climate type dicular to the slopes (see the area of each study site in Table S1). Semiarid Semiarid Transect vertex was fixed with metal bars in order to repeat measure- Humid Arid ment at the same place. For this, each vertex was identified with UTM coordinates or a specific code (Carabassa et al, 2019). Aridity Floristic inventories were made identifying all plant species in Study areas identification and location, climatic parameters, geologic substrates, and nature-based solutions applied index 0.16 0.32 0.37 0.52 0.68 0.77 each area, which distinguished each site or subsite (according to site- specific variability). Additionally, an abundance estimation per species was performed using these pattern rankings: evapotranspiration 1: 0%–5% soil cover 2: low frequency (<25% soil cover) 3: high frequency (25%–75% soil cover)4: dominant (>75% soil cover) (mm yr 1) Potential A protocol for measuring all these parameters was specifically 1100 1388 1143 678 982 737 defined (CREAF, 2017b). Data were obtained in two field campaigns carried out in all plantations: one before Summer 2017 and another one after 2 years, in late Spring 2019. For the third and fourth plant- max and min T ( C) Averages monthly ings in Tifaracás, further monitoring was carried-out in 2020, whereby only seedlings survival was measured (Table S1). 14.3–28.6 8.9–24.7 10.6–25.5 0.8–30.3 6.7–22.1 1.8–22.5 2.5 | Statistical analysis precipitation (mm) Seedling survival and physiological state, Cocoon degradation, plant Mean annual height, stem diameter, root development, and microsite biomass and cover (vegetation exclusion) were analyzed using R Studio. The normal probability test and the Breusch–Pagan test were used to check nor- 177 445 424 350 666 570 mality and homoscedasticity. Analyses of the differences between treatments (control and Cocoon) were performed using the Mann– 28N 427779, 30N 721173, 30N 700667, 30N 571368, 31N 394519, 34N 565459, Whitney test, since only two treatments were tested. Analyses on the coordinates 3095798 4267646 4343691 4164540 4609198 4482995 (ETRS89) study sites were performed using the Kruskal–Wallis test, since more UTM than two treatments were tested. To determine significant differ- ences, a value of α = 0.05 was applied. Macedonia Note: * Temporary river bed. Barcelona Islands Western Valencia Alicante Almería Region Canary 3 | RE SU LT S 3.1 | Seedling survival and physiological state Ptolemais TABLE 1 Tifaracás María El Bruc Jijona Sierra Tous Site Analyzing the survival values together for all the study areas, plant species and contrasted differences could be observed. Seedlings that CARABASSA ET AL. 137 TABLE 2 Main soil characteristics of the study areas CEC pH water SOC stock 1 1 Site Texture CaCO3 (%) (cmol kg ) (1:2,5 w/v) EC (dS m ) SOM (%) (T ha 1) Comments a c a ab a a Tifaracás Clay 6.5 ± 3.0 43.7 ± 3.4 7.70 ± 0.23 0.95 ± 0.13 1.62 ± 0.35 12.99 ± 3.81 Steep slope, stony Jijona Loam 77.9 ± 6.8d 11.5 ± 4.9a 7.82 ± 0.20a 0.75 ± 0.30ab 1.79 ± 1.18a 27.67 ± 6.21ab Abandoned terraces a b a a a ab Tous Clay 9.4 ± 13.9 28.1 ± 5.4 7.49 ± 0.28 0.46 ± 0.11 3.39 ± 1.30 26.69 ± 9.59 Shallow and stony Sierra Clay loam 58.1 ± 5.7c 20.8 ± 4.5ab 7.72 ± 0.04a 0.78 ± 0.27ab 3.18 ± 1.30a 38.09 ± 7.23b Tilled, petrocalcic María horizon El Bruc Sandy clay 28.7 ± 3.3b 14.1 ± 2.2a 7.72 ± 0.12a 0.71 ± 0.26ab 3.28 ± 1.28a 58.03 ± 3.97c Shallow and stony cd ab a b a b Ptolemais Sandy 67.2 ± 14.6 24.4 ± 13.6 7.66 ± 0.20 1.51 ± 1.08 7.47 ± 14.39 29.19 ± 14.02 Mining debris, stony Note: Values represent mean ± standard deviation. Different letters indicate significant differences at p < 0.05 between study areas for each variable. Abbreviations: CaCO3, Calcium carbonate content; CEC, cation exchange capacity; EC, electric conductivity; SOC stock, Soil organic carbon stock; SOM, Soil organic matter planted with the Cocoon methodology showed greater survival, with turbinata ssp. canariensis Guyot (Canarian juniper), being 100% in con- a rate close to 60%, while the control ones showed lower rates of up trols and close to 80% in Cocoons. The best survival results were to 40% (p < 0.0001). Regarding the vigor of the survivors, the seed- obtained with Olea europaea L. ssp. guanchica (Canarian wild olive), lings with good health predominated in both cases (control and having a mortality rate that did not reach 25% of the specimens and a Cocoon treatments), but with a higher percentage in the Cocoon relatively high percentage of healthy seedlings, which was close to treatment, in which only few plants were severely affected or res- 50%. In respective controls, mortality and affected seedling rates were prouted. However, some particular tendencies can be observed when higher (32% and 38%, respectively). data are analyzed for each planting area (Figure 2). All the species planted in El Bruc and Jijona showed a similar In the El Bruc, Jijona, and Tous planting sites, differences trend, whereby a better physiological state in the Cocoon treatment between controls and Cocoons were observed in terms of mortal- could be observed. However, for some of them, such as Ceratonia ity and the number of healthy plants (see Figures S2–S5). In con- siliqua L. (carob tree), Olea europaea L. var. europaea (vera olive tree), trast, in Sierra María, there were no differences in seedling Prunus avium L. (cherry tree), and Prunus spinosa L. (blackthorn), the mortality between the control and Cocoon treatments, which both mortality of controls exceeded 80%. In El Bruc, the best results yielded 37% survival rates. In addition, control seedlings showed were obtained for the two subspecies of Quercus ilex L. (subsp. ilex greater vigor, with a greater number of healthy seedlings (36%), and subsp. ballota [Desf.] Samp.), Quercus faginea Lam. (Portuguese compared with those with Cocoon (26%). Survival in Ptolemais' oak), Olea europaea L. var. europaea (cornicabra olive tree), and planting sites was high on both treatments. The percentage of Juglans regia L. (walnut tree), where Cocoons had survival rates well affected seedlings was very low, and only appeared in the Cocoon above their respective controls (see Figure S3). In Jijona, also Que- treatment. In contrast, survival in Tifaracás was low in both treat- rcus ilex and Olea europaea exhibited good results with the Cocoon, ments, even though the seedlings planted with Cocoon presented exhibiting survival rates close to 100% in some cases (see a higher percentage of healthy seedlings (21%), compared with Figure S4), but Tetraclinis articulata (Vahl) Masters (Cartagena's controls (12%). cypress) also exhibited good results using both treatments, in which However, vigor results are not only dependent on location but controls reached 86% survival rates. also on plant species, for example, Rosmarinus officinalis L. (rosemary) and Prunus dulcis (Mill.). D. A. Webb (almond tree) in Sierra María showed high survival ratios in Cocoons like respective controls, while 3.2 | Seedling growth Tamarix gallica L. (French tamarisk) presented high mortality ratios in both treatments, with mortality being much higher in the controls, Seedling growth depended not only on local environmental conditions which reached a 100% mortality ratio (see Figure S1). In Ptolemais, but also on the species. At the same time, a similar trend toward sur- the highest mortality basically affected Cupressus sempervirens vival was observed (see Figures S7–S13, Annex 3). A clear tendency L. (cypress) specimens (see Figure S2). toward a better growth of seedlings planted with Cocoon was noted The overall results in Tifaracás were largely determined by local in diameter and height when survival is higher, such as in Quercus ilex harsh conditions, especially the drought in Summer 2017, which and Olea europaea (Figure 3). However, for some species, like occurred just after planting. Despite this, seedlings planted in the Rosmarinus officinalis, there were no significant differences between Cocoon had better physiological state than controls. Pistacia atlantica treatments, and for other species, controls presented higher growth Desf. (mastic tree) had high mortality rates in both treatments, but than Cocoons, such as almond trees in Sierra María (Figure 3). Regard- survival was higher in Cocoons, albeit with a high percentage of ing root development, differences were only observed in the severely affected seedlings. Mortality was also high in Juniperus cornicabra olive tree (Figure S13, Annex 3). 138 CARABASSA ET AL. F I G U R E 2 Physiological state of all seedlings (without discriminating by plant species) after 2 years of planting with cocoon technology and without (control) in each experimental site. Different letters show statistically significant differences (α = 0.05%) [Colour figure can be viewed at wileyonlinelibrary.com] 3.3 | Cocoon biodegradation Cocoon began showing signs of biodegradation, such as cracks or holes in its bowl (State 2). There were a lower percentage of Cocoon biodegradation is an important aspect to evaluate, since its completely biodegraded Cocoons (State 4). As observed, some design, which foresees using biodegradable material, aims at incorpo- Cocoons in State 1 could retain runoff and rainwater, thereby increas- ration into the soil once its watering function is completed. At a gen- ing the water availability for the respective seedlings. In fact, this eral level (Figure 4), the vast majority of Cocoons presented the bowl water retention capacity 2 years after implantation, which is longer in functional condition, but with the lid of the device sunk, damaged, than the expected useful life, had been utilized in the new Tifaracás or not present (State 1). In a quarter of the installed devices, the plantings (third and fourth plantings, see Table S1) for refilling the CARABASSA ET AL. 139 F I G U R E 3 Height and diameter growth between 2017 and 2019 of Quercus ilex ssp ilex and Olea europaea var. cornicabra (El Bruc), Rosmarinus officinalis and Prunus dulcis (Sierra María). Different letters show statistically significant differences (α = 0.05%). C, control; CO, Cocoon [Colour figure can be viewed at wileyonlinelibrary.com] 140 CARABASSA ET AL. F I G U R E 4 Cocoon's degradation State after 2–2.5 years installed at field. Stage 1: Cocoon OK: With or without shelter, but with lid; Stage 2: Lid collapsed but bowl apparently in good state (without cracks, holes); Stage 3: Bowl with signs of degradation (cracks, holes); Stage 4: Highly degraded bowl (almost incorporated into soil) [Colour figure can be viewed at wileyonlinelibrary.com] bowls during the summer, in which the retained water greatly plantations are subjected to tillage. Table S2 shows a reduction in the improved the survival ratios. cover and height of woody plants, which were accompanied by an These results differed according to the study area (Figure 4). increase in the cover of herbaceous plants. Regarding plant diversity, Although State 2 occurred most frequently (except in Calabria), the there was a net change in 12 species (14 new species appeared and differences could still be observed locally. The presence of Cocoon 26 were not found). Among these, we noted, on the one hand, the dis- residues incorporated into the soil was very scarce or not observed in appearance of abundant species in 2017 such as Hordeum murinum or most areas, except in El Bruc, Calabria, and Jijona. These three zones, Tamarix gallica, and on the other, a high frequency in the appearance together with Sierra María (a large proportion of Cocoons in State 3), of Avena fatua and Euphorbia sp. In 2019, the vegetation cover of the were the ones with the greatest global Cocoon biodegradation. The dry temporary riverbank upper zone suffered the effects of sporadic area with the least biodegradation was Tifaracás, with States 1 and torrential rains common in this area, which generated a flood that 2 occupying 96% of the Cocoons studied, followed by Tous and washed away the vegetation. Ptolemais (Figure 4). In the El Bruc area, there was an increase (2019 vs. 2017 sam- pling) in both herbaceous and woody covers for the three subzones of sampling, which was accompanied by an increased average height of 3.4 | Vegetation structure and diversity both types of vegetation (Table S2). This increased plant cover was also accompanied by increased species richness. With respect to the The structure and floristic biodiversity data are presented in Tables S2 inventories of 2017, in stony and shallow soils, some Asteraceae and S3 of Annex 4, respectively. All uncropped areas in the Iberian appeared abundant (Centaurea scabiosa ssp. scabiosa, Helichrysum Peninsula showed a positive trend in view of herbaceous and/or stoechas, and Scorzonera angustifolia) and grasses, such as Brachy- woody vegetation cover and/or floristic composition. However, in podium phoenicoides, increased in abundance, which eventually Ptolemais and Tifaracás, we could not identify differences in the became the dominant species in this area. In general, Rosaceae plants structure or composition of the vegetation with the 2017 sampling. (Amelanchier ovalis, Rosa canina, etc.) and Fabaceae also increased. In The characterization of the natural vegetation in Sierra María was agricultural soils that are deeper and finer textured, the trend was carried out in the temporary dry riverbank (rambla), since the almond very similar, but with some differences. In these soils, Asteraceae CARABASSA ET AL. 141 showed a reduced abundance, which became testimonial species, Within the wide range of climates tested, the driest one while the Rubiaceae like Galium lucidum and Rubia peregrina appeared. (Tifaracás) was also the most challenging for the Cocoon (unless As in the previous area, the Fabaceae, in particular Dorycnium pen- rewatered), which has a survival rate of below 30%. However, in pre- thaphyllum, presented great abundance, and several species of grasses vious restoration projects carried out in nearby areas with conven- appeared albeit with low abundance. In addition, Helianthemum tional planting systems, the mortality rates were close to 100% syriacum and Rosmarinus officinalis, abundant plants in neighbouring (CREAF, 2017a, 2017b), for which the Cocoon could be considered an areas, which were absent in 2017, appeared with high frequency interesting alternative for planting in these arid climates. It is espe- in 2019. cially interesting, in this case, to analyze the balance between the In the Jijona and Tous areas, there was also a tendency of increas- increase in survival due to rewatering and the consequent increase in ing vegetation cover of both woody and herbaceous species. More- maintenance costs. Although implementation costs of the Cocoon over, in Jijona, for woody species, the trend of cover increment was technology are initially higher than conventional methods, it is never- accompanied by an increase in the average height of the plants. How- theless regarded as a viable option for reducing seedling mortality ever, for the herbaceous species, the average height scarcely without increasing maintenance costs in the long run. In the plantings increased when compared with 2017. In Tous, there is an increase in carried out in 2018 and 2019 (see Table S1), the Cocoons were both the cover of woody and herbaceous species. However, this gain refilled twice during the summer, which improved the survival rate was not accompanied by an increase in average height, which (see Annex 5). Given the results, the option of refilling the Cocoon remained stable. Regarding floristic diversity, both areas remained bowl, despite involving higher initial cost, could be an optimal solution quite stable between 2017 and 2019. for planting in the drylands of the Canary Islands. In subhumid regions, like Ptolemais, seedlings planted with Cocoon present similar survival rates as those planted with common 3.5 | Plant competition evaluation techniques. Regarding Cupressus sempervirens, differences in mortality ratios were observed among seedlings having different heights (ages) With respect to the data collected from the vegetation surrounding planted with the Cocoon: 46% in 50-cm high specimens versus 24% the seedlings in the different study sites (Table 3), we could observe in 30-cm high specimens. This outcome supports the recommendation two different tendencies. In some areas of El Bruc and Jijona, we see that seedlings planted with the Cocoon should preferably be 1 year a pattern of higher biomass weight with greater cover in controls. In old, as reported previously (Land Life Company, 2016). Cupressus the driest areas, such as Tifaracás, or even in areas experiencing simi- seedlings are sensitive to extreme weather conditions and adapt bet- lar annual rainfall like Tous, we could see a greater development of ter when they are small in size (low height) because they are able to vegetation around Cocoons. develop stronger root systems quickly. At this site, spring plantings recorded higher survival rates than autumn ones due to better weather conditions for Cupressus implantation. 4 | DISCUSSION In contrast, in areas with drier rainfall regimes like Jijona or El Bruc, the differences between control and Cocoon are significant, as Overall, significant differences were found between seedlings planted the efficacy of this device is demonstrated in adverse conditions, such with Cocoons and controls. Seedling mortality in Cocoons was close to as the prolonged drought and high temperatures of Summer 2017. 40%, while in the control group reached 60%. In addition to this moder- This is especially true in the case of El Bruc, where a 30% reduction of ate improvement in survival, surviving plants had a better physiological annual rainfall occurred (449 mm throughout 2017), especially in the state when Cocoon was used. These differences could be attributed to summer (70% reduction, 58 mm for the whole season). nutrient uptake being highly dependent on water availability in arid and Regarding soil conditions, the Cocoons could not be properly semiarid environments (Maestre et al., 2005; Powers & Reynolds, 1999). installed in shallow and stony soils, like in some parts of the Tous site. By providing water to the planting sites using the Cocoon device, the Additionally, the strong winds at this site blew out the Cocoon shel- plants would be able to overcome or reduce this limitation and make ters, particularly those that were not properly installed. As a result, better use of available nutrients, thereby increasing their survival and the affected seedlings were exposed prematurely to high irradiation growth. In fact, the Cocoon not only provides water to the plant during and desiccating winds. Therefore, the Cocoon is not recommended the first months, but it also creates a micro catchment that allows for for Leptosols or those having a petrocalcic horizon near the surface greater infiltration of rainwater and accumulation of runoff around the (IUSS Working Group WRB, 2015), such as those existing in the Sierra plant. Moreover, it not only increases the water supply, but also reduces María almond fields. Planting under these conditions means that the water losses. The plant protector reduces evapotranspiration, and the Cocoons could not perform to their full potential, which renders this lid and the bowl itself reduce competition with herbs, especially during technology less competitive compared with usual methods. the first year. In addition, the seedlings planted with the Cocoon had a Regarding the different plant species, the high mortality in tendency toward a more developed root system than controls did, arbequin olive tree plantings in El Bruc (both in control and Cocoons) which resulted in a greater development of the aerial biomass for some should be attributed to the bad quality of the seedlings, with rotting species. roots, stem scars, leaf loss, and chlorosis (CREAF, 2017a, 2017b). In 142 CARABASSA ET AL. T A B L E 3 Herbaceous cover and plant biomass in 1 m diameter circles around control (C) and Cocoon (CO) seedlings, after 2–2.5 years of planting in four areas 2 2 Site Subsite Treatment Herbaceous cover (%) Plant biomass (g m , wet weight) Plant biomass (g m , dry weight) a a a Tifaracás TI1 C 70 ± 35 77 ± 41 71 ± 36 Tifaracás TI1 CO 95 ± 5a 95 ± 15a 89 ± 14a Jijona JI1 C 19 ± 3a 258 ± 38a 128 ± 20a a a Jijona JI1 CO 12 ± 2 333 ± 141 134 ± 49a Jijona JI2 C 52 ± 11a 596 ± 217a 342 ± 55a b b Jijona JI2 CO 22 ± 9 337 ± 31 183 ± 23b Tous TO1 C 0 ± 0a 0 ± 0a 0 ± 0a b b Tous TO1 CO 8±2 98 ± 24 48 ± 12b Tous TO2 C 3 ± 1a 57 ± 14a 23 ± 12a b b Tous TO2 CO 9±1 189 ± 26 98 ± 12b El Bruc EB1 C 47 ± 3a 178 ± 85a 76 ± 36a a a El Bruc EB1 CO 33 ± 18 271 ± 111 132 ± 51a El Bruc EB2 C 93 ± 5a 850 ± 217a 227 ± 104a b a El Bruc EB2 CO 60 ± 8 612 ± 195 204 ± 25a El Bruc EB3 C 53 ± 22a 273 ± 207a 149 ± 93a a a El Bruc EB3 CO 31 ± 7 258 ± 118 133 ± 62a Ptolemais PT3 C 47 ± 7a 773 ± 8a 702 ± 5a a a Ptolemais PT3 CO 49 ± 10 912 ± 10 833 ± 6a Note: Values represent mean ± standard error. Different letters indicate significant differences at p < 0.05 between C and CO per subsite and parameter contrast, arbequin olive trees planted in Jijona had a very high survival or the vera variety, the latter being the variety historically used in the rate (almost 90%), with approximately 75% of seedlings planted with area of El Bruc, which we also found in different places in the prov- Cocoon healthy and growing, probably aided by runoff collection in  mez-Escalonilla & Vidal, 2006). ince of Barcelona and Valencia (Go the Cocoons. In general, the Cocoon yielded very good results in the Both in El Bruc and in Jijona, the cornicabra variety responded better plantings in Jijona, a site with a semiarid climate (<450 mm per year) than the arbequin variety. The cornicabra variety also adapted better and very poor soil with an extremely high carbonate content (77%). than Vera in El Bruc, with cornicabra seedlings showing higher survival The response of the holm oak (Quercus ilex) subspecies is espe- rates and vigor. Since water deficit (moisture stress) is the most persis- cially remarkable. Ballota subspecies performed very well in El Bruc, tent environmental stress on fruit crops (Petros et al., 2020), the with a survival rate greater than 60% and a statistically significant Cocoon could help in installing crops in arid and semiarid lands. higher growth with the Cocoon. This holm oak subspecies planting Based on the data available, there is still insufficient evidence could be considered as an example of assisted migration strategy for demonstrating that the Cocoon improves the growth of seedlings in adapting to climate change (IPCC, 2007; Pramova et al., 2019). This comparison to the traditional techniques. Regarding growth in length indigenous subspecies of southern Spain and northwestern Africa was and weight of the roots, significant differences were found only for planted at higher latitude, which simulates the displacement of the cornicabra olive trees in El Bruc, which were higher in the plants with distribution area that this tree could suffer from amidst climate Cocoon. However, as the available data only reflect plant growth in change by applying the assisted migration mechanism (Sansilvestri 2 years (2017–2019) in view of the slow evolution of vegetation in et al., 2016; Schwartz et al., 2012). Another plant species that drylands (Yu & Wang, 2018), it is possible to state that the positive responds well to assisted migration is Tetraclinis articulata. This small trends observed in many cases suggest that if the growth monitoring tree, which is originally an Ibero-African endemism mostly located in were repeated after some years, these differences could increase northwestern Africa and has only two small natural populations in (Shackelford et al., 2018). Europe, namely in Malta and Sierra de Cartagena (SE Spain) (TGM, The structure and biodiversity of the accompanying vegetation 2020), was planted in Jijona (outside its distribution area) with very showed different trends for the studied areas, in terms of climatic and good results. biotic factors, including anthropogenic ones. Generally, in the non- These assisted migration tests were also performed with typical extreme Mediterranean climate sites tested, an increase in vegetation agricultural tree species. The cornicabra olive tree variety was planted growth and/or plant diversity had been observed. According to the in El Bruc and in Jijona. This variety is typical of central and southern intermediate disturbance hypothesis (Connell, 1978), the increase in Spain. They are vigorous, erect bearing, and with thick canopy density. biodiversity of these communities is an indicator that they are grow- It adapts better to continental climates than the arbequin olive trees ing in complexity and maturity, as they have not reached the CARABASSA ET AL. 143 intermediate degree of disturbance (or recovery), where maximum flo- across the Mediterranean border to increase resilience; LIFE15 ristic richness would be produced. However, the elapsed time can be CCA/ES/000125) and LIFE Nieblas (Reforestation & Climate Change considered rather short for proper assessment of improvements in Mitigation: tests, evaluation and transfer of innovative methods based biodiversity. on fog collection; LIFE19 CCM/ES/001199), co-funded by the EU The Tifaracás and Sierra María sites remained stable without LIFE program. The authors would like to express their gratitude to the appreciable changes in plant biodiversity. This slow evolution could be reviewers for their efforts in strengthening the quality of this paper. due to the hard environmental conditions in these areas. The restora- tion of degraded drylands has several limitations: (1) resource (water, DATA AVAILABILITY STAT EMEN T nutrients, soil organic matter, propagules) levels are uniformly low; The data that supports the findings of this study are available in the (2) harsh microenvironmental conditions limit seedling recruitment; supplementary material of this article. and (3) animals have a greater potential for disrupting restoration efforts in arid systems (Roundy et al., 1995). The effect of animals OR CID impeding restoration dynamics could be clearly observed in Tifaracás, Vicenç Carabassa https://orcid.org/0000-0002-8728-2818 where there is a large population of wild goats. Moreover, extreme Julián Campo https://orcid.org/0000-0003-0009-138X events are also a limitation in arid land restoration (Olsson et al., 2019). The slight changes observed in Sierra María, with a RE FE RE NCE S reduction of woody plants, are probably due to the 2018 flood that Alday, J. G., Zaldívar, P., Torroba-Balmori, P., Fernández-Santos, B., & affected the restored area. Martínez-Ruiz, C. (2016). Natural forest expansion on reclaimed coal Cocoon biodegradation is also affected by rainfall regime, increas- mines in Northern Spain: The role of native shrubs as suitable micro- ing in areas with higher rainfall values. As the lid is usually the most sites. Environmental Science and Pollution Research, 23, 13606–13616. https://doi.org/10.1007/s11356-015-5681-2 exposed part of this device, it is easily susceptible to damages. Since Campo, J., Andreu, V., Gimeno-García, E., González, O., & Rubio, J. L. Cocoon biodegradation is slower in dry conditions, trees growing (2006). Occurrence of soil erosion after repeated experimental fires in under such conditions can also benefit longer from extended, and still a Mediterranean environment. Geomorphology, 82(3-4), 376–387. needed Cocoon support: more water available by Cocoon refilling https://doi.org/10.1016/j.geomorph.2006.05.014 Campo, J., Gimeno-Garcia, E., Andreu, V., Gonzalez-Pelayo, O., & after rain events, reduced evaporation losses, and were resilient to Rubio, J. L. (2008). Aggregation of under canopy and bare soils in a competing for adjacent weeds. As mentioned above, this fact became Mediterranean environment affected by different fire intensities. advantageous for the planting carried out in Tifaracás in 2018, where Catena, 74(3), 212–218. https://doi.org/10.1016/j.catena.2008. Cocoons were refilled to increase their survival, being an experience 05.002 Carabassa, V., Ortiz, O., & Alcañiz, J. M. (2019). RESTOQUARRY: Indica- with very good results. Moreover, partially biodegraded Cocoons may tors for self-evaluation of ecological restoration in open-pit mines. still provide rainwater and be a shield against evaporation, implying an Ecological Indicators, 102, 437–445. https://doi.org/10.1016/j.ecolind. extended water availability to support tree growth. 2019.03.001 As a conclusion, the Cocoon technology proved useful for refor- Cohen-Shacham, E., Walters, G., Janzen, C., & Maginnis, S. (Eds.). (2016). estation in drylands. In general, conventional plantations showed Nature-based solutions to address global societal challenges (p. 97). Gland: IUCN. higher mortalities and relatively lower vigour rates than planting sites Connell, J. H. (1978). Diversity in tropical rain forest and coral reefs. Sci- using this ecotechnology. The direct and indirect water supply, the ence, 199, 1302–1310. mitigation of plant competition around the seedling, the reduction of CREAF. (2017a). Baseline characterisation report. Centre de Recerca evapotranspiration, and the microcatchment effect, create a suitable Ecològica i Aplicacions Forestals. https://thegreenlink.eu/wp-content/ uploads/2017/01/Deliverables-A1-Baseline-report.pdf set of conditions for improving the physiological state of plants, which CREAF.(2017b). Above ground, below ground and carbon stock monitoring increases their survival. However, a case-per-case evaluation is protocol. Centre de Recerca Ecològica i Aplicacions Forestals. https:// needed before deciding on this technology. Cocoons have an added thegreenlink.eu/wp-content/uploads/2017/05/A.2.-Del_Training- advantage when planting site conditions impose more drought stress course-material.pdf Díaz-Hernández, J. L., Barahona, E., & Linares, J. (2003). Organic and inor- (lower rainfall, sandy textured soils with poor water retention), and/or ganic carbon in soils of semiarid regions: A case study from the when tree species used are less adapted to drought stress in the early Guadix-Baza basin (Southeast Spain). Geoderma, 114, 65–80. https:// stages of development. However, Cocoons are less competitive than doi.org/10.1016/S0016-7061(02)00342-7 common techniques for planting in soils with high water retention Go mez-Escalonilla, M., & Vidal, J. (2006). Variedades del olivar. Madrid: Ministerio de Agricultura, Pesca y Alimentacio n. capacity, or in Mediterranean humid climates, or for planting drought- IPCC (2007). Climate change 2007: Impacts, adaptation and vulnerability. tolerant species. Small differences in survival and growth, combined Contribution of Working Group II to the Fourth Assessment Report of with higher costs of planting with Cocoons, make this ecotechnology the Intergovernmental Panel on Climate Change. Cambridge, UK: less interesting in these situations. Cambridge University Press, p. 976. https://www.ipcc.ch/site/assets/ uploads/2018/03/ar4_wg2_full_report.pdf IUSS Working Group WRB (2015). World Reference Base for Soil ACKNOWLEDGMENTS Resources 2014, update 2015. International soil classification system This study has been funded by the projects LIFE The Green Link for naming soils and creating legends for soil maps (World Soil (Restore desertified areas with an innovative tree growing method Resources Reports No. 106). Rome: FAO. 144 CARABASSA ET AL. Land Life Company. (2016). Cocoon planting instructions. Amsterdam: LLC. Salem, B. B. (1989). Arid zone forestry: A guide for field technicians. Rome: Land Life Company. (2021, May 28). Projects. https://landlifecompany. Publications Division, Food and Agriculture Organization of the United com/projects Nations.  pez-Marcos, D., Turrio Lo  n, M. B., & Martínez-Ruiz, C. (2020). Linking soil Sansilvestri, R., Frascaria-Lacoste, N., & Fernández-Manjarrés, J. (2016). variability with plant community composition along a mine-slope topo- One option, two countries, several strategies: Subjacent mechanisms graphic gradient: Implications for restoration. Ambio, 49, 337–349. of assisted migration implementation in Canada and France. Restora- https://doi.org/10.1007/s13280-019-01193-y tion Ecology, 24(4), 489–498. https://doi.org/10.1111/rec.12343 Maestre, F. T., Valladares, F., & Reynolds, J. F. (2005). Is the change of Schwartz, M. W., Jessica, J., Jason, M. M., Sax, D. F., Borevitz, J., Brennan, J., plant-plant interactions with abiotic stress predictable? A metaanalysis Camacho, A. E., Ceballos, G., Clark, J. R., Doremus, H., Early, R., of field results in arid environments. Journal of Ecology, 93, 748–757. Etterson, J. R., Fielder, D., Gill, J. L., Gonzalez, P., Green, N., Hannah, L., https://doi.org/10.1111/j.1365-2745.2005.01017.xMAP Jamieson, D. W., Javeline, D., … Zellmer, S. (2012). Managed relocation: MAPA (2019). La desertificacio n en España. Madrid, SP: Ministerio de Integrating the scientific, regulatory, and ethical challenges. Bioscience, agricultura, pesca y alimentacio n. https://www.mapa.gob.es/es/desar 62, 732–743. http://dx.doi.org/10.1525/bio.2012.62.8.6 rollo-rural/temas/politica-forestal/desertificacion-restauracion-forest Shackelford, N., Miller, B. P., & Erickson, T. E. (2018). Restoration of open- al/lucha-contra-la-desertificacion/lch_espana.aspx. cut mining in semiarid systems: A synthesis of long-term monitoring MEA. (2005). Ecosistemas y bienestar humano: Síntesis sobre Desertificacio n. data and implications for management. Land Degradation & Develop- Washington, DC: World Resources Institute. ment, 29, 994–1004. https://doi.org/10.1002/ldr.2746 Mohsin, F., Arias, M., Albrecht, C., Wahl, K., Fierro-Cabo, A., & TGM. (2020). The Gymnosperm Database. https://www.conifers.org/cu/ Christoffersen, B. (2021). Species-specific responses to restoration Tetraclinis.php interventions in a Tamaulipan thornforest. Forest Ecology and Manage- Valdecantos, A., Fuentes, D., Smanis, A., Llovet, J., Morcillo, L., & Bautista, S. ment, 491, 119154. https://doi.org/10.1016/j.foreco.2021.119154 (2014). Effectiveness of low-cost planting techniques for improving Muñoz-Rojas, M., Hueso-Gonzalez, P., Branquinho, C., & Baumgartl, T. water availability to Olea europaea seedlings in degraded drylands. Resto- (2021). Restoration and rehabilitation of degraded land in arid and ration Ecology, 22(3), 327–335. https://doi.org/10.1111/rec.12076 semiarid environments: Editorial. Land Degradation & Development, 32, VOLTERRA. (2021, May 28). Life The Green Link. Areas  del proyecto. 3–6. https://doi.org/10.1002/ldr.3640 https://thegreenlink.eu/es/areas-del-proyecto/ Olsson, L., Barbosa, H., Bhadwal, S., Cowie, A., Delusca, K., Flores- WWF. (2016). Recuperando paisajes: Un nuevo camino para la restauracio n Renteria, D., Hermans, K., Jobbagy, E., Kurz, W., Li, D., Sonwa, D. J., & ecologica. WWF. http://awsassets.wwf.es/downloads/informe_restau Stringer, L. (2019). Land degradation. In Climate change and land: An racion_ecologica_wwf_2016.pdf IPCC special report on climate change, desertification, land degradation, Yu, K., & Wang, G. (2018). Long-term impacts of shrub plantations in a sustainable land management, food security, and greenhouse gas fluxes in desert–oasis ecotone: Accumulation of soil nutrients, salinity, and terrestrial ecosystems. IPCC. http: ipcc.ch/srccl/ development of herbaceous layer. Land Degradation & Development, Petros, W., Tesfahunegn, G. B., Berihu, M., & Meinderts, J. (2020). Water- 29, 2681–2693. https://doi.org/10.1002/ldr.300 saving techniques on growth performance of Mango (Mangifera indica L.) Seedlings in Mihitsab-Azmati Watershed, Rama Area, Northern SUPPORTING INF ORMATION Ethiopia. Agricultural Water Management, 243, 106476. https://doi. Additional supporting information may be found in the online version org/10.1016/j.agwat.2020.106476 Powers, R. F., & Reynolds, P. E. (1999). Ten year responses of ponderosa of the article at the publisher's website. pine plantations to repeated vegetation and nutrient control along an environmental gradient. Canadian Journal of Forest Research, 29, 1027–1038. https://doi.org/10.1139/x99-104 How to cite this article: Carabassa, V., Alba-Patiño, D., García, Pramova, E., Locatelli, B., Djoudi, H., Lavorel, S., Colloff, M., & Martius, C. S., Campo, J., Lovenstein, H., Van Leijen, G., Castro, A. J., (2019). Para adaptar la restauraci on de la tierra a un clima cambiante: González, F., Viera, G., Kourkoumpas, D.-S., Zioga, A. A., Aceptemos lo que sabemos y lo que no. CIFOR. https://doi.org/10. 17528/cifor/007462 Papadelis, C. E., Andreu, V., Gimeno, E., Kallen, S., & Alcañiz, J. Roundy, B. A., McArthur, E. D., Haley, J. S. & Mann, D. K. comps. (1995). M. (2022). Water-saving techniques for restoring desertified Proceedings: Wildland shrub and arid land restoration symposium; lands: Some lessons from the field. Land Degradation & 1993 October 19-21; Las Vegas, NV. Gen. Tech. Rep. INT-GTR-315. Development, 33(1), 133–144. https://doi.org/10.1002/ US Department of Agriculture, Forest Service, Intermountain Research Station. ldr.4134