One of the main shortcomings of the environmental restoration projects carried out in the past decade is the lack of an adequate later following [1] . This monitoring phase is necessary to quantify the success of the environmental improvement and to determine whether it needs to be optimized in the long term. In this regard, the evaluation of the improvement effect is an essential task for the development of the ecological restoration [2] . However, carrying out an adequate following of environmental progress is a complex process, especially working at landscape scale [3] . It requires spending time collecting data in the field and, consequently, is economically expensive, and sometimes entails the involvement of specialists [4] . It also implies continuous management and follow-up to adjust this management in the case that restoration has not resulted in the expected recovery in terms of biological diversity [5] , so much so that the environmental restoration actions suffering from a lack of the subsequent monitoring are more likely to fail to meet the initials project objectives [6] .

As background for this study, it is worth mentioning that Mediterranean forests are biodiversity hotspots and being one of the most vulnerable ecosystems to harmful human activity and its consequences [7] - [9] . Hence, these are priority areas for the implementation of practices aimed at restoring forest landscapes [10] .

In environmental improvement projects, whether they involve ecological restoration or environmental compensation, revegetation is considered one of the most effective actions because it facilitates ecological functions enhancing both biodiversity and ecosystem services, as well as carbon sequestration and soil recovery [11] , and always requires the abovementioned long-term following program. Among the most frequent interventions is the implementation of linear landscape elements which facilitate the colonization process, the settlement and the survival of species in newly established habitats [12] .

In the context of the foregoing, it should be mentioned that this research is a part of the monitoring of the compensatory measures included in the Breña’s Compensation Project, linked to the construction of the Breña II damp [13] [14] These measures were devoted to offsetting the environmental disturbance caused by the flooding of part of the nature reserve Sierra de Hornachuelos Natural Park (southern Iberian Peninsula), and by the dam infrastructure itself [15] .

The revegetation plan linked to the Breña Compensation Project corresponded to the “plantation type” [16] and consisted of the establishment of three different environmental models: 1) “Forest-islands” or patches of woody vegetation corresponding to the original forest species; 2) “Hedges,” or alignments of tree, shrub, and grass species planted to increase connectivity between patches of relict forest; and 3) “River copses” or mixed plants formations consisting of deciduous trees and shrubs species typical of Mediterranean riverbanks [17] . Each model required a specific composition of plants, which were chosen agreeing to the projected ecological characteristics [18] . In the Project, the design and implementation of a medium- to long-term monitoring plan was also included to assess the functionality of these models. Full information on the overall improvement measures carried out is available [13] .

There are several elements to be considered when approaching a long-term monitoring program: 1) Adequate selection of the bioindicator group; 2) Sufficiently long time elapsed since the implementation of the improvement actions; and 3) To have available control areas where no environmental intervention has been made to be used for comparisons [19] .

Regarding the selection of the indicator taxon, it should be supported on that the ecological requirements of the selected group are indicative of the state of the environmental conditions to be assessed during the monitoring phase [20] . Due to their sensitivity to changes in habitat structure, the Orthopterans are considered valid indicators of the effectiveness of restoration actions [21] [22] . This is because land management has a significant impact on microclimate and structure of vegetation [23] and these insects maintain a close relationship with vegetation through various aspects of their biology, particularly feeding, reproduction and dispersal power. So much so that the Orthopterans respond quickly to environmental changes [24] [25] , disturbances [26] , and habitat restoration [27] [28] . As a result, they are widely accepted bioindicators for detecting environmental changes linked to the landscape interventions [14] and effectiveness of habitat restoration [21] [22] .

Concerning the time elapsed since the revegetation actions were carried out, in our case the implementation of each environmental model, planting was done in spring 2007. First assessment of short-term effects of the revegetation program on the Orthopteran’s diversity was performed in 2012. Four years later (2016) the second monitoring phase was completed.

To have control areas, as baseline which would enable comparisons of conditions after the interventions equivalent areas were selected corresponding to the same environmental models but in which no intervention had been carried out (fully description available in Moyano et al. [14] ).

Because predictive models suggest that drought is the main present and future threat in Mediterranean terrestrial ecosystems [8] and that riparian forests are particularly vulnerable to global change [29] , this research has focused on the environmental model defined as “river copses”. Moreover, riparian vegetation is more demanding from the point of view of water requirements so that vegetation associated with temporary watercourses is more difficult to implement and maintain. Consequently, its effects on animal communities should be greater and more evident.

Similar studies to the proposed here have been previously carried out also using the Orthopterans as bioindicators for habitat quality assessment, most of them focused on grasshopper inhabiting grasslands and rangelands [22] [23] [28] [30] . In addition, research on Orthopteran assemblages (concerning both Ensifera and Caelifera) from other types of habitats is also available [31] [32] , including riparian corridors [33] [34] , as is our case.

Considering that the maximum of Orthopteran diversity is estimated to be reached 3.5 - 5.5 years after restoration [34] , we hypothesized that the maximum diversity should have been achieved before this second monitoring phase.

The aim of this study was to assess the biodiversity Orthopteran assemblages settled in the river copses after nine years of following (2007-2016); concretely, from the time of revegetation was carried out until the last recording faunistic data. The results will be interpreted as an indirect measure of the success of the environmental improvement achieved in the area.

To determine the restoration success in the research area the following two attributes from those proposed by the primer provided by the Society of Ecological Restoration International (SER) [35] were considered: diversity and community structure in comparison with reference sites; and the presence of indigenous or singular species.

The faunistic results (list of species and specimens number censused in the restored copses and control areas) corresponding to the last monitoring phase are shown in Table 2 .

There is substantial evidence that, at the local scale, plant community development determines the composition and structure of Orthopteran assemblages [21] [22] [48] . This is the rationale for using Orthopteran biodiversity as indicators of environmental recovery processes in revegetated areas affected by the compensatory measures linked to the construction of the Breña II dam. The short-term assessment of this recovery program performed in 2012 (after five years restoration), found structural similarity among the Orthopteran communities in the different habitats, whether improved or not, due to the high degree of vegetal homogeneity. The authors suggest that this could be because in the first monitoring phase, the development of the new plant cover had not yet achieved the expected achievement [14] .

According to Mola et al. [49] , four years later, the establishment and development of the plant restoration should have been consolidated. Considering that the maximum of Orthopteran diversity is estimated to be reached 3.5 - 5.5 years after restoration [34] , it would be expected that the maximum in diversity would have already been reached in the first phase of monitoring. Nevertheless, the values of the community structural indexes (Dominance, Diversity and Evenness) are more balanced in the second monitoring period, suggesting that the recovery process of the Orthopteran fauna can take up for a longer period, exceeding seven years after revegetation tasks were done. In fact, this conclusion agrees with what Moyano et al. [14] pointed out in the sense that in the first phase of following, the Orthopteran assemblages found in the control areas were still best structured than in the improved zones.

The results obtained in the second monitoring phase, after nine years of restoration, indicate somewhat higher levels of abundance and diversity in restored river copses than in the control areas. These trends are observed in both study sites but acquire statistical significance only in terms of diversity in “Las Mesas”, so it could be thought that environmental restoration has been slightly more successful in this sampling site. The highest abundance is explainable by the greater presence of certain generalist Caeliferous species as P. giornae to which corresponds almost 50% of the records in both environmental models. Kenyeres et al. [50] indicated that the habitat restoration resulted in a positive effect in the structure of Orthopteran assemblages, given that the parameters richness, density and relative frequency of species significantly increased after restoration. Nevertheless, these results contrast with those of Alignan et al. [22] who found that the Orthopteran abundance was significantly lower in the restored areas than in the original grasslands.

There is information indicating that the Orthopteran communities of unmanaged areas generally harbour lower diversity than those of managed sites [31] . Our results agree with this statement, in both phases of monitoring, since the specific richness is always higher in the revegetated river copses than in the control plots.

Now analysing the effect of the time elapsed between the two monitoring periods over the Orthopteran assemblages, and firstly considering the abundance of species, an evident faunistic impoverishment is observed in both study sites and in the two environmental models (revegetated river copses and respective control zones). Significant differences uniquely refer to variations in the populations size of the most ubiquitous Caeliferous species such as P. giornae, or D. jagoi. These are pioneer species, r-strategists (sensu Price et al. [51] ) which massively colonize the revegetated areas in the early stages of the succession, but whose populations subsequently became more balanced, the population peak observed in the first phase of monitoring having been dampened, resulting in a more structurally balanced community in the longer term. These trends have been observed in the revegetated areas of the two study sites and also, but to a lesser extent, in the control areas. So, it is not attributable to the effect of the improvement revegetation carried out.

Conversely, some Ensiferous, such as Tessellana tessellata (Charpentier, 1925), increase their presence in revegetated river copses while it is becoming scarcer or disappear in the reference models (controls) after four years following. This explains the significant differences in terms of abundance, diversity and dominance. Our results agree with those of Bieringer and Zulka [52] and Marini et al. [53] in the sense that the presence of shrubs is particularly detrimental to the Celiferous, whereas the Ensiferous seems to be less affected or even favoured, as in our case. Torma et al. [54] also observed that the density of woody vegetation along ditch banks had a negative effect on the total abundance and on the abundance of the most mobile species (Caeliferous).

Apart from the above factors, certain faunal impoverishment could be also attributable to the extreme thermal conditions recorded during 2015-2016 [55] . Accordingly, early and intense desiccation of grass vegetation, as consequence of high temperatures, could have stopped or shortened the life cycles of plants and animals. This may have affected even the species with greater ecological plasticity such as T. tesellata or P. giornae and prevented the presence of those with less environmental tolerance.

When interpreting the results, it is necessary to consider that from a conservation perspective, both community structural parameters and the presence of singular species are important [56] because of the indicative value of the exclusive species to each type of environment.

No singular species have been found in this second phase of the study, nor even those that due to their hydrophilic character would be expected, such as Xya variegata (Latreille, 1809). This species, commonly colonizing the riparian vegetation growing between sand bars and the water’s edge [57] , was exclusively located in the control areas and during the initial phase of the study. This observation agrees with Fartmann et al. [31] in the sense that the more stenotopic species are often restricted to undisturbed habitats. Disturbances linked to the revegetation tasks could be so detrimental to the organisms that initially occupy the land [39] that could prevent their presence in subsequent recolonization.

Based on our results, the environmental differences between revegetated and unmanaged river copses have not led to significant changes in the composition and structure of the Orthoptera communities they host. Regarding the temporal evolution in both, revegetated river-copses and control areas, the most recent data reveal a certain positive evolution, likely by the effect of the improvement in the vegetation cover, which is beginning to become somewhat evident at other levels of the ecosystem. This recovery is expressed by the greater biodiversity, despite the population decrease is evident in all the areas sampled, but which in no case reaches the expected levels or includes the species with the greatest indicator potential.

From the results, it can be inferred that the environmental restoration carried out in the study sites has not been as successful as could be expected, and that the previous environmental alterations have not been minimized, nor have those derived from the environmental rehabilitation itself been compensated. In our opinion, it is necessary to make a greater effort in the initial planning phase of the strategies to be followed in the environmental improvement plan before their implementation in nature to ensure that their effects have a successful impact on the biodiversity of the affected territory. Among these improvement strategies, it would be considered a more precise selection of the plots to be revegetated, prioritizing soil quality, water availability, nutrient content and shade degree, to ensure a good progress of the new planted vegetation.

Given that revegetation is one of the main actions for supporting biodiversity by restoring habitats and providing food sources for wildlife as well as protecting and improving existing ecosystems, the main efforts of a restoration program should be focused on a good revegetation approach in its initial stages.

Due to the arid climate of the study area, it would have been essential to incorporate effective irrigation mechanisms and to guarantee their functioning, in order to ensure plants survival in the first stages after planting, even under the extreme summer conditions. It would be also advisable to select varieties of plants resistant to drought and to frequent pathologies in the area, as well as to replant the specimens that have not progressed successfully.

Special attention should also be paid to prevent herbivory by means of functional enclosures that allow access to Orthopteran and other small fauna but prevent entry to wild ungulates and extensive livestock. To confirm the project is on track, it’s also necessary to monitor vegetation progress continually.

Finally, it is worth mentioning that good long-term monitoring, both of the vegetation and of other bioindicators (in our case the Orthopteran), is economically costly but indispensable for good environmental management and should be intended in the total financing of the project.

This research was supported by Aguas de las Cuencas de España, S.A. (ACUAES) (Ministry of Agriculture, Food and Environment, Government of Spain) and by Junta de Andalucía (European Union - Next Generation EU), exceptional grant in the framework of the Biodiversity R + D + I Complementary Plan, financed by the Recovery and Resilience Mechanism (RRM).