Soil health indicators for monitoring forest ecological restoration: a critical review
Author contributions: PG, FA conceived and designed the research and wrote and edited the manuscript; PR conducted an initial search for soil indicators publications; CE provided input on the research design and participated in editing the manuscript.
Abstract
Forest restoration is considered among the most affordable and effective practices to address ecosystem and biodiversity loss and mitigate the impacts of human-induced global change. Soils are intrinsically complex systems that mediate and regulate multiple processes and functions vital for forest ecosystem restoration. Although monitoring soil attributes are critical for evaluating the success of forest restoration projects, research, and development of soil function indicators are still limited. Here, we have reviewed the most commonly reported soil indicators in forest restoration research and their recovery trajectory on a global scale. We also identified and discussed less frequently used indicators that have the potential for monitoring ecosystem recovery. We found that soil indicators have considerably increased in the literature. However, research is regionally concentrated, and a significant proportion of publications neither considered reference ecosystems (41%) nor provided basic information about soil types (<21%). The most reported indicator types were chemical (76%) (e.g., soil carbon, nitrogen, and pH). A significant proportion of the studies (46%) performed long-term evaluations (>15 years) of indicators. The majority of the indicators tended to resemble the levels of the reference ecosystem in the long term, with a few exceptions (e.g., water content and bulk density). We identified several less used but more integrative indicators with great potential for monitoring forest ecosystem recovery (e.g., aggregate stability, oxidizable carbon, soil respiration, and enzyme activity). Our results emphasize the need to effectively develop standardized soil health indicators to monitor ecosystem recovery under different conditions and expand their use in underrepresented regions.
Implications for Practice
- Practitioners and researchers should embrace soil diversity and complexity. Thus, they must ensure that soil types in the degraded and reference sites are comparable. Using available soil maps complemented with local observations could aid soil identification.
- Increasing knowledge of soil biological indicators for ecosystem restoration is essential as they can be more directly related to ecological processes and functions.
- Developing integrative indicators that provide information and link different soil functions (i.e., POXC) could help reduce the number of indicators required for evaluation and may help minimize monitoring costs.
- The evaluation of integrative, chemical, and biological indicators should be prioritized during the first years, while the assessment of physical indicators could be spread over extended evaluation periods (slow long-term response).
Introduction
Landscape transformation derived from human-induced land-use change (e.g., conversion of native forests to agricultural land or urban areas), invasion of exotic species, forest fires, and climate change, among others, have led to an accumulated loss and degradation of forest ecosystems (FAO 2011; Ghazoul & Chazdon 2017). Forest loss and degradation negatively affect critical ecosystem attributes, including soil and water's physical and chemical status, and topography (Gann et al. 2019), directly impacting soil processes and functions (Lal 2015; Kooch et al. 2020). Forest loss and degradation alter biogeochemical C, nutrient, and water cycles, affecting the capacity of forests to regenerate or recover (Crovo et al. 2021a, 2021b). Forest degradation hinders the provision of essential ecosystem services with consequential adverse effects on human well-being (Scherr 2019).
Soils and their attributes are critical components for ecosystem functioning and, as such, are essential for the success of forest ecosystem restoration. However, soils are often not considered during restoration strategies (Farrell et al. 2020), and studies generally focus only on recovering vegetation composition and structure (Callaham et al. 2008). This limits the success of restoration efforts since ecosystem processes in degraded soils cannot sustain vegetation limiting ecosystem recovery (Hobbs et al. 2009).
Soil health is defined as the “capacity of soil to function as a vital living ecosystem that sustains plants, animals, and humans” (Lehmann et al. 2020). The essential soil attributes for sustaining biodiversity are (1) physical (e.g., soil texture, bulk density), (2) chemical (e.g., cation exchange capacity, pH), and (3) biological (e.g., soil enzymes, respiration) (Lal 2016). Numerous soil functions, such as elemental cycling, water retention, and carbon sequestration, must properly occur to maintain ecological balance and promote recovery.
Monitoring allows the evaluation of restoration progress and informs new decisions or adaptation strategies to meet the goals and objectives proposed by practitioners and researchers (Gatica-Saavedra et al. 2017). Monitoring of soil health recovery must be considered during the design stage of any forest restoration plan or project (Méndez-Toribio et al. 2021). The selection of appropriate soil indicators is essential to properly assess the recovery of soil health (Heneghan et al. 2008).
Proper monitoring of soil indicators should consider the site under restoration, the control or prerestoration site, and a reference ecosystem (i.e., the soil condition to be achieved). The reference ecosystem should be based on a reference model of one or more sites close to the restoration area. These sites should have little or minimal degradation and have the same vegetation, topography, and environmental variables (including soil properties) as the area to be restored (Hernandez-Santin et al. 2021). Comparing ecological indicators between sites under restoration and reference ecosystems allows a proper evaluation of ecological restoration success.
For indicators to be effective, they must meet specific criteria such as sensitivity to environmental changes and management, ease of measurement, and cost-effectiveness (Dale & Beyeler 2001). The most frequently used soil indicators focus on the physical, chemical, and biological soil properties (Allen et al. 2011). Some of these properties allow the inference of ecosystem processes and functions such as nutrient cycling, climate and water regulation, and decomposition of plant and animal residues, among others (Costantini et al. 2016). However, it is still challenging to select appropriate indicators due to the limited understanding of soil properties, the inherent variability of soils, and forest ecosystems under restoration (Muñoz-Rojas 2018). In addition, excessive use of indicators can be ineffective and expensive. Currently, there are neither generalized suggestions for appropriate soil indicator selection nor recommendations for the evaluation time. This information could help researchers and practitioners to select the best indicator(s) and identify monitoring strategies to evaluate the recovery of soil health over time. Therefore, the objectives of this study are to (1) conduct a global review and analysis of the most commonly used soil ecological indicators in forest restoration activities, (2) analyze the sensitivity and recovery time of soil indicators, and (3) discuss less commonly used indicators with potential for monitoring restoration projects. We expect this review could serve as a guideline for selecting indicators in the practice of forest restoration.
Methods
Literature Search
A literature review was performed to identify physical, chemical, and biological soil indicators used to evaluate the recovery of ecological restoration activities from 1975 to February 2022. The Advanced Search tool of Web of Science and Scopus were used for the search. Using Boolean operators, we selected search words to achieve the most relevant result possible. After testing multiple combination, the following combination of terms was selected to search in the Abstract, Title, and Keywords sections: = “soil” AND “function” OR “response” OR “indicator*” OR “property*” OR “qualit*” OR “monitor*” OR “variable” OR “evaluat*” OR “measure” OR “assess*” AND “restor*” OR “remediat*” OR “rehab*” OR “reforestat*” OR “afforestat*” OR “recov*” AND “forest” AND “native specie*” OR “indigenous specie*.” After this first search, the title and abstract of each article were reviewed and qualitatively evaluated. This review only considered original research; thus, bibliographic reviews, conferences, congresses, and books were excluded. The selection of articles considered the following inclusion criteria: (1) must include the evaluation of ecological restoration activities or any of its stages (restoration continuum: reclamation, rehabilitation, recovery) as main objective or secondary objectives; (2) consider restoration of forest or arborescent ecosystems using native species or mixed native and exotic species; and (3) have empirically performed measurements of soil health or soil function indicators.
A database was generated with the selected articles, including the following information: study location, type of forest, or arborescent formation (standardized according to World Wildlife Fund ecoregions; Olson et al. 2001), type of soil indicator measured, and kind of comparison (i.e., reference ecosystem).
We also recorded the soil types reported by each paper; however, this was limited since explicit soil classification was not included in most papers. Some articles contained general soil or site information like local lithology or parent material and surface textures, but they did not provide detailed information on soil classification or type. Other papers used national or local classifications that are difficult to homologate to Food and Agriculture Organization of the United Nations (FAO), World Reference Base for Soil Resources (WRB) or United States Department of Agriculture Soil Taxonomy. For articles where fine-scale soil mapping was available in the USA, we correlated the soil to local map units to the Soil Survey database. To simplify the analysis, we related all classified soils to the FAO WRB classification to the major group level when possible. For restoration activities carried out in reconstructed sites (e.g., mines, landfills, etc.), we assumed the soils correspond to Technosols.
Standardization of Indicators
Indicator information and levels were obtained from articles in graphs, tables, and descriptive information. Indicators were grouped according to the similarity among general characteristics of the analysis, methodology, or reported parameters. These were then regrouped according to the indicator type (physical, chemical, or biological; Table S1). In addition, the reported level of significant differences (p values) of each indicator between the degraded and the reference ecosystems was recorded. When possible, we identified and recorded the direction of change of the indicator at the time of the evaluation, positive if the indicator changed toward the reference values and negative if it moved away from the reference levels. When the analyses considered more than one soil depth, the value of the indicator of the upper soil layer was selected.
Time of Evaluation
The evaluation time was recorded for each article, considering the time elapsed from implementing the restoration activities to when the soil indicators were analyzed. The time of evaluation was grouped into three ranges: 0–6 years (short-term evaluation), 7–14 years (medium-term evaluation), and 15 or more years (long-term evaluation). In chronosequence studies involving multiple consecutive assessments, only the maximum age of evaluation was considered.
Results
Regional and Temporal Distribution of Articles
Ninety-six articles fulfilled all the search criteria detailed above, equivalent to 8% of the total articles returned after the search (without duplicates). On the continental scale, North America was the region with most publications (n = 27). A map with the regional distribution is included in Supporting Information (Fig. S1).
Tropical and Subtropical Moist Broadleaf Forests are the biomes where most publications were conducted (41%). Most publications did not report the soil type or classification (79%).
Based on the WRB, 23 soil groups were identified (Ferralsols with 9.4%, Cambisols and Acrisols, each with 6%).
The number of articles incorporating soil indicators has increased over time. Most were related to ecological forest restoration (59%), and more than half considered reference ecosystem (Fig. 1A). The remaining studies corresponded to restorative activities like revegetation, remediation, and reclamation.
The first article that mentioned reference ecosystem was published in 2003 and has use increased since 2007 (Fig. 1B).
Type of Indicators
A total of 342 indicators were found, which were coalesced into 122 simplified indicators (see Table S1). These were categorized into chemical, biological, and physical properties (62, 28, and 32 indicators, respectively; Table S1).
The frequency of use of the different types of indicators was similar over time. However, chemical indicators have increased significantly since 2009 (Fig. S2).
Most Used Indicators
The most used indicators were chemical (76% of articles), followed by physical (58%) and biological (55%). Among the chemical indicators, soil carbon (C), nitrogen (N), pH, phosphorus (P), and potassium (available or exchangeable K+) were the most frequent. C and exchangeable K+ were reported in 66–32% of the articles, respectively. The studies encompassed different methodologies and protocols to measure soil carbon fractions, which for simplicity, were combined for this analysis (i.e., organic carbon [29% of articles], total carbon [25% of articles], and soil organic matter [SOM; 22% of articles]).
The most used physical soil indicators were water content (37% of articles), bulk density (34%), and texture (21%). The three most used biological indicators were invertebrate diversity (21% of studies), microbial biomass of C (15% of studies), and soil respiration (12% of studies; Fig. 2).
See the supporting information for more details on other indicators used less frequently (Fig. S3).
Age of Evaluation
Forty-six percent of the articles were found to have evaluations more than 15 years after the restoration activities were implemented (long-term evaluations). Thirty-one percent of the articles only evaluated between 0 and 6 years (short term), and the rest evaluated recovery after 7–14 years (medium term; Fig. 2).
Among chemical indicators, C was the most frequently reported indicator in the short term and medium term, followed by N. In the long term, C and pH were the most used indicators (>15 years). Regarding physical indicators, water content and bulk density were frequently assessed at all ages. Invertebrate diversity and microbial biomass C were the most frequently reported soil biological indicators in the short and long term (Fig. 2).
Recovery Level of Ecological Indicators
Due to the substantial number of indicators (n = 122) and the limited number of articles that reported recovery trends, we only analyzed the three most used indicators for each indicator type (i.e., physical, chemical, and biological). Soil total carbon was added as a fourth indicator for the chemical type as carbon can be regarded as a cross-type indicator for chemical and biological properties. For this analysis, only the articles that reported reference forest ecosystem sites were considered, resulting in 57 articles (60% of the total number of selected papers).
We determined whether the indicator displayed non-significant differences (NS) against the reference forest and recorded the trajectory of change for each indicator. Identifying articles with NS for a given indicator allows us to infer whether that indicator changes significantly over time (i.e., the indicator levels in the intervened site are statistically similar to the values of the reference forest). The lack of significant differences did not mean that the ecosystem had recovered but rather that the indicator had changed on a trajectory that resembled the levels of the reference ecosystem.
Overall, the indicators with the highest proportion of articles reporting positive trajectories toward the reference site were: invertebrate diversity (79%), microbial biomass C (78%), and pH (74%). On the other hand, the indicators that presented the highest proportion of articles with NS were available P (54%), carbon (52%), and pH (50%).
For the short-term evaluations the indicator with the highest proportion of articles with positive trajectories was invertebrate diversity (Fig. 3G1). In contrast, available P was the indicator with the highest proportion of articles with NS P (Fig. 3D2).
For medium-term evaluation period, pH was the indicator with the highest percentage of articles reporting positive trajectories (Fig. 3C2). Available P displayed the highest proportion of articles with NS (Fig. 3D2).
In the long term, microbial biomass and carbon were the indicators with the highest proportion of articles with positive trajectories. Similarly, the indicators with the highest proportion of papers reporting NS were carbon and total N.
It relevant to consider that some indicators were not included because there was an insufficient number of articles reporting them at certain evaluation times (texture and bulk density in the short term, and soil respiration and microbial biomass C in the medium term).
Discussion
This systematic review identified the most common soil properties used as indicators during the evaluation of ecological restoration research projects. The results also highlight soil indicators “relevance in evaluating restoration activities” progress and success. Our analysis suggests a growing interest in using soil indicators for researching and monitoring forest ecosystem restoration projects. This is reflected in the increasing number and variety of soil indicators considered in articles published in recent years. Most of this interest is probably motivated by the growing recognition of the relevance of ensuring soil functionality and enhancing ecosystem services provision during restoration projects (Robinson et al. 2012; Muñoz-Rojas et al. 2016).
We found that most of the research was concentrated in two countries: U.S.A. and Brazil. Only a modest fraction of research has been conducted in European, Asian, and African countries. It is critical to build up practical skills and knowledge at the local scale to increase and expand the use of soil indicators, especially in regions historically affected by forest degradation and deforestation.
More than half of the publications do not incorporate basic soil classification information, which is fundamental for comparison or replicating the research. Most indicators were evaluated in the upper 30 cm disregarding that most tree rooting systems reach soil depths greater than 1 m. The lack of this information could significantly hinder the definition of adequate goals and objectives of a restoration plan, as well as the possibility of adequately assessing the recovery of soil functions and the proper selection of reference ecosystems. The latter may explain some of the observed variability between intervened sites and reference ecosystems. In addition, the lack of profile descriptions and basic horizon information limits the interpretation and extrapolation of results to other regions. We suggest that forest restoration research and practice should always include basic soil type and classification information.
Worldwide, chemical indicators were the most widely used. The widespread use of chemical indicators as soil quality indicators in agronomy and forestry may have motivated its use in forest restoration. Biological indicators are considered highly sensitive compared to chemical and physical indicators (Andrés & Mateos 2006), and their popularity has increased since the 2000s.
Below we discuss the most used chemical, physical, and biological indicators and their suitability for assessing ecosystem recovery.
Soil Indicators in Forest Restoration
Chemical Indicators
The popularity of carbon (C) and total nitrogen (N) as indicators may be explained by their simplicity, low cost, and accessibility. C and N are tightly linked to many ecological processes, thus allowing practitioners to infer the recovery of other ecosystem properties and functions (Lal 2016). The growing use of C is likely related to the need to estimate carbon stocks (Xu et al. 2018a).
Soil pH was also one of the most widely used indicators. Soil pH influences most soil biogeochemical processes, nutrient availability, and plant productivity (Costantini et al. 2016). In addition, pH measurements are relatively easy to perform, accessible, and inexpensive (Harris et al. 1997; Cardoso et al. 2013).
Our findings differed from previous studies that analyze soil indicators at national scales. Lozano-Baez et al. (2021), for example, find that soil sodium (Na) and phosphorus (P) are the most used soil indicators in Colombian forests. Similarly, Mendes et al. (2019) report that pH, potassium content (K), and different phosphorus fractions (P) are the most used indicators in a review of forest restoration of Atlantic Forests in Brazil (Lozano-Baez et al. 2021). The rise of phosphorus (P) as a popular indicator in this national-scale study can be explained by the fact that P is a limiting nutrient in many tropical and subtropical forest soils (Cardoso et al. 2013).
Since a substantial amount of the published research in this review comes from restoration projects in temperate forest systems in the northern hemisphere, total N is more frequently reported than P. Temperate systems are generally thought to be more limited by N than P (Perakis & Sinkhorn 2011). However, N limitation may depend strongly on the age of the forest stand (Hou et al. 2012) or the restoration implementation period (Xu et al. 2018b). Since both elements are likely colimiting in most forest ecosystems and have different cycling pathways and origins (atmospheric vs. lithogenic), we suggest they are always evaluated together.
Physical Indicators
Soil water content was the most used physical indicator, likely because of its importance in determining plant productivity and other ecological processes (Yang et al. 2014). Soil water content is relevant for all forest types, but it is essential in recovering arid and semiarid forests (Zou et al. 2008). Despite its relevance, literature reviews have found that this indicator is rarely reported and can be considered a knowledge gap in restoration projects. Consistent with Lozano-Baez et al. (2021), we suggest that it is critical to include indicators related to the hydrological component of the soil in restoration projects, especially in temperate and Mediterranean climates currently experiencing extended and severe droughts (Sousa et al. 2011).
The second most common indicator was bulk density, which is widely assessed in agricultural or rangeland systems to monitor compaction due to heavy machinery use or livestock grazing (Allen et al. 2011). Bulk density is related to soil hydrological processes such as infiltration and aeration rates (Allen et al. 2011). However, changes in this indicator depend strongly on SOM (Cardoso et al. 2013). On the other hand, heavily compacted soils can take decades to recover (Rab 2004), limiting the usability of this indicator in the short term.
Soil texture is among the most common physical indicators and is strongly linked to many pedogenic and biogeochemical processes, including soil aggregation (Zhang et al. 2021), carbon sequestration and nutrient cycling (Matus 2021), and water and gas flow (Mentges et al. 2016). In addition, soil texture is simple and a relatively low-cost standard methodology in most labs. Despite its popularity, soil texture may not be an adequate indicator because it is generally very robust to soil management (Cardoso et al. 2013). Practitioners should evaluate texture during the initial soil characterization of degraded (prerestoration) and reference sites to ensure comparability, but we discourage its use for monitoring purposes.
Biological Indicators
Invertebrate diversity and microbial biomass C were the most frequently reported biological indicators. Invertebrate diversity was commonly used due to their known sensitivity to environmental change (Paoletti et al. 1996). Invertebrates participate in numerous soil functions, such as the decomposition of organic matter and nutrient cycling (Parkhurst et al. 2022). They also control critical physical properties like soil aggregation and macroporosity (Blanchart et al. 2004). The main disadvantage of this indicator is that it is time-consuming and generally requires specialized taxonomists to identify species.
Microbial C and N were also relatively popular among biological indicators. These indicators respond rapidly to soil degradation and recovery (Paz-Ferreiro & Fu 2016). Microbial biomass is linked to vital ecosystem functions such as nutrient and carbon cycling (Orwin et al. 2016). Similarly, microbial respiration and biomass have been considered suitable indicators for assessing soil quality and health (Paz-Ferreiro & Fu 2016) and play a crucial role in maintaining soil ecosystem functions (Singh & Gupta 2018). Microbial biomass has also been described as an estimator of the productivity of restored sites (Martucci do Couto et al. 2016) and could inform about the effect of different restoration practices (Farrell et al. 2020). Intense activities during site habilitation (e.g., removal of exotic species) could decrease the microbial biomass; therefore, an increase in this biomass would be expected as restoration progresses.
Soil respiration is among the three most common biological indicators; however, it is reported much less frequently than the first two. Similarly, Parkhurst et al. (2022) find that this has rarely been reported in projects assessing soil quality in abandoned agricultural landscapes. The low popularity of this indicator may be because its highly spatially and temporally variable, time-consuming, and generally requires specialized and costly equipment. Nevertheless, respiration is considered a key soil health indicator that integrates multiple soil processes (Allen et al. 2011). Soil respiration is discussed later as a potential integrative indicator as new, more affordable technologies may promote its adoption.
Recovery Level of Ecological Indicators
We expected to find positive recovery trends of specific soil indicators over time as the restoration sites approach or resemble the reference site. The latter should translate into a higher percentage of articles displaying indicators that are not significantly different from the reference ecosystem. In other words, we used the “non-significant statistical differences” (NS) between the means of an indicator in a restored site and a reference site to infer some level of indicator and soil function recovery.
This approach was selected because the primary purpose of monitoring indicators is to assess ecosystem recovery over time (Maharjan et al. 2020). Still, only half of the studies considered a reference ecosystem. Hence, we suggest caution when interpreting our analysis as the significance of the observed response is highly dependent on the prior status of the site, the level of interventions, and the type of soil and forest ecosystem, all of which were not explicitly considered in our analysis.
Recovery Level of Chemical Indicators
As expected, soil C showed more significant differences in the short term. However, a relevant percentage of papers (56%) still showed that C recovered in the short-term. Moreover, the evaluations during medium term showed the highest percentage of articles with significant differences in soil C.
Therefore, the response of soil C seems to be inconsistent among different publications. The differential response of soil C in the short term is likely due to different original C status (e.g., reference sites with naturally low C levels) and differences in the intensity of manipulations and strategies used in the studies (planting, invasive control, soil amendments etc.). C strongly depends on (1) the initial state of the soil in the degraded site (e.g., level of degradation), (2) the type of forest (e.g., tropical vs. deciduous), and (3) intrinsic soil characteristics related to carbon stabilization (e.g., mineralogy [Crovo et al. 2021a], texture [McLauchlan 2006], and moisture regimes [Tomar & Baishya 2020]).
Total N showed the expected trend, as a higher percentage of short-term evaluations presented significant differences with the reference. Evidence shows that total N responds consistently to active restoration strategies as native vegetation establishes and productivity and diversity increase (Fazhu et al. 2015).
Soil pH complied with the expected trend in short to medium term since the number of articles reporting NS differences increased as the restoration progressed. However, the number of studies reporting significant differences was higher in the long term. The inconsistent response of soil pH in the long term is an aspect that requires further research.
Total P did not show a clear trend of recovery in the early years; the similar percentage of items between “S and NS differences” is mainly due to the level of degradation. Studies focus on mine rehabilitation showed that total P did not recover, unlike the studies which included the application of fertilizers. This is expected as P has a lithogenic origin, and it is unlikely to respond quickly to restorative activities unless fertilization is considered. Medium-term evaluations showed that both total P and available P recovered to similar levels as the reference site. However, the number of articles reporting significant differences was slightly higher in the long term. Still, available and total P tend to improve over time in chronosequence studies (Yang et al. 2010). When soils are degraded, adding P sources may confound the evaluation of success as it may show the indicator's quick but artificial recovery (Bizuti et al. 2020). In addition, P fraction levels may change substantially throughout the seasons; thus, researchers and practicians need to consider seasonality during their evaluations.
For short-term term projects with limited budgets, indicators like nitrogen and pH can be considered, while total carbon may be more suitable for the longer term. On the other hand, phosphorus should preferably be evaluated in the medium to long term as it can be significantly affected by initial fertilization.
Recovery Level of Physical Indicators
All articles reported significant differences against the reference in the short term. Water content decreases in soils where exotic species are removed or have reduced vegetation cover and only showed recovery after the tree establishment (e.g., Ruwanza & Tshililo 2019). Therefore, it is advisable to space out the evaluations of this indicator and consider it for medium to long-term evaluations.
Most articles reported significant differences in bulk density in all evaluation periods Similar results were reported by Parkhurst et al. (2022), who find that bulk density recovers after 10 years but fails to resemble the reference ecosystem even after 50 years since the restoration of agricultural lands. Still, our analysis suggests that this indicator showed recovery in the long-term, as a small percentage of articles found NS differences. Partial recovery of bulk density is also possible. The results reported by Gu et al. (2019) show that revegetation improved bulk density over time. Thus, bulk density could be considered to evaluate more advanced stages of restoration.
Soil texture is a valuable indicator for identifying main soil types of reference ecosystems, but not necessarily for monitoring restoration success as it is not expected to change substantially during the typical evaluation time of most projects. The differences in texture observed between degraded and reference sites in the short term imply a poor selection of reference ecosystems. We recommend that forest restoration studies and projects must consider soil texture specially during morphological soil descriptions to ensure soils in degraded and reference sites are comparable.
Recovery Level of Biological Indicators
All articles using invertebrate diversity showed significant differences between the restored and reference sites in the short- and long-term evaluations. Soil invertebrate diversity is thought to be a sensitive indicator. Still, our analysis suggested that its recovery probably takes longer than the timeframe most studies considered. This could be especially the case when restoration is carried out in highly disturbed environments (e.g., mine reclamation; Orabi et al. 2010).
Most articles that reported microbial biomass C showed positive changes toward the reference site in the short term. However, these fast results occurred mainly in sites that considered the incorporation of organic matter amendments. We also found a higher percentage of articles reporting similarity to the reference site in the long term. Some articles with significant differences indicated that microbial biomass C increased but not at the levels reference site (e.g., Wei et al. 2009).
A positive trend of soil respiration (i.e., recovery) was only observable in the long term. This can be expected as organic matter accumulation, root, and microbial activity, and soil respiration tend to increase as plant productivity increases (Shi et al. 2020).
On the other hand, microbial biomass C changes positively to specific restoration treatments and could be used in short and long-term evaluation periods. Since invertebrate diversity shows some similarity with the reference site in the medium and long term, it is crucial to evaluate its original state (prerestoration or control) to identify the direction of change from the short-term assessments.
Recommendations and Additional Potential Indicators for Ecological Restoration
According to the Standards and Principles of Ecological Restoration, to determine the progress and success of restoration projects, a reference site must be considered (Prach et al. 2019). For a proper selection of a reference ecosystem, the following aspects need to be considered: (1) the site should be undisturbed or self-sustained native forests, and (2) it must be located close to the area to be restored, sharing the same soil type, topographic position, exposure, and other relevant abiotic variables (Hernandez-Santin et al. 2021). Once the reference ecosystem has been selected and characterized, it is necessary to establish realistic recovery goals and objectives in the short, medium, and long term. Likewise, it is essential to consider the assessment of selected indicators in the control (i.e., degraded condition near the restoration site or prerestoration condition) to identify trends of change and to contrast with the different strategies used postrestoration (Lindenmayer 2020). We suggest measuring soil indicators at the same time every year, considering all site conditions. The frequency of evaluations should be based on the sensitivity of each indicator. If possible, extend these evaluations for longer periods (>15 years). For long-term evaluations, select protocols and technologies that are likely to be available in the long term (i.e., technological obsolescence).
The popularity of available nutrient contents is expected as they provide helpful information about soil fertility (Lal 2015). However, not all nutrients and forms may help evaluate ecosystem recovery, and researchers must be aware of their limitations. For example, available nutrients content may be difficult to interpret as they arises from intricate dynamic relationships between soils and associated biota. In addition, nutrient availability respond strongly to initial fertilization, which is a common practice in many projects. Nutrient levels also vary substantially among soil and forest types and tend to display high spatial and temporal variability even within a single forest stand (Sayer & Banin 2016). Moreover, nutrient requirements and temporal demand for native tree species are generally unknown, which limits interpretations. Total nutrient levels do not likely represent availability to plants. In fact, the bulk of total nutrients in natural systems are tightly bound to mineral and organic matter or immobilized by microbes (Spohn et al. 2021).
If nutrient content is considered a relevant indicator by practitioners, then soil sampling should be very carefully planned, designed, and carried out. Soils are intrinsically spatially and temporally anisotropic, and this complexity and variability may limit the identification of indicator recovery. We observed that most articles included a limited number of replicates per condition (i.e., seven samples on average). We suggest first to assess which nutrients may be limiting ecosystem recovery based on a comparison of the status and stoichiometric relationships of total C in the degraded and reference ecosystems. Second, a sampling time and frequency should be defined based on the local forest dynamics that most likely represent nutrient availability (e.g., biannual sampling every early spring).
Below we list six soil indicators (one chemical, two physicals, and two biologicals) with the potential for monitoring ecological restoration projects:
Permanganate oxidizable carbon is an indicator commonly used for soil health assessment in agricultural settings with great potential for monitoring forest restoration. The relevance of this indicator is that it reflects the available or labile portion of SOM and is more related to microbial community composition and aggregate stability than total C (Wade et al. 2020).
Although microbial biomass C and soil respiration appeared among the most frequent indicators, they are reported to a much lesser extent than the most popular chemical and physical indicators.
Soil respiration is the main outflux of carbon from soils and one of the most crucial components of the carbon cycle in terrestrial ecosystems. In addition, respiration is sensitive to land-use change and soil degradation (Yao et al. 2019). An increment in respiration rates has been generally observed during the rehabilitation of degraded soils and reforestation (Shi et al. 2020). Soil respiration is a good indicator of ecosystem metabolism and should be more widely used in ecological restoration projects. For the effective use of this indicator, it is suggested to measure it with enough spatial replication and frequency due to its high variability (Ryan & Law 2005). The cost of most automatic soil respiration spectroscopic-base devices may be prohibitive for most projects. Fortunately, relatively inexpensive tools have been developed [e.g., Joshi Gyawali et al. (2019) or Sainju et al. (2021)] and may be suitable for continuous measuring and monitoring of soil respiration at higher spatial and temporal resolution. We also suggest using fast respiration assays like the 24 hour CO2 burst incubation (McGowen et al. 2018) as cost-effective alternatives. Regardless the relevance of microbial biomass as a metric for soil carbon and nutrient cycling, fumigation procedures for microbial biomass may be too complicated or expensive to implement for most labs. The use of substrate utilization assays like the fast respiration test mentioned above may be a better alternative.
There are relevant gaps in the evaluation hydrological condition of soils in restoration projects (Lozano-Baez et al. 2021). Water balance is critical in forest restoration projects, especially in highly degraded soils in semiarid and arid regions (Gao et al. 2018).
We suggest increasing the use of water content and soil infiltration as indicator. The popularity of soil infiltration has decreased in recent years. We believe infiltration is a critical indicator as it is a fundamental process of the forest hydrological cycle and link to the overall soil/plant water balance (Rahmati et al. 2018), groundwater recharge, soil erosion, and surface run-off (Lozano-Baez et al. 2018).
The evaluation of traditional physical properties like bulk density against the reference site may not be very informative due to their slow posttreatment response. For this reason, we suggest testing indicators that may respond faster such as soil aggregate stability. Soil aggregate stability is commonly used to assess soil health (Almajmaie et al. 2017), but it has not been widely used in restoration projects. Aggregate stability directly relates to water movement and storage, aeration, erosion, gas and nutrient flow, and their effect on the activity and growth of living organisms (Amézketa 1999; Eldridge & Leys 2003). It has been documented that different restoration treatments like revegetation and the incorporation of organic amendments have positive effects on the stability of aggregates, so it is possible to demonstrate changes over time (Dou et al. 2020). Most studies have only evaluated this indicator in the long term (>15 years) (Burri et al. 2009; Moreno-de las Heras 2009). Thus, assessing if this indicator responds in the short and medium term is still necessary. As for their measurement and analysis, several methodologies exist and are considered easy, reproducible, and inexpensive (Amézketa 1999; Seybold & Herrick 2001).
Soil enzyme activity is considered a sensitive indicator for soil health assessment (Utobo & Tewari 2015) and could also be a good indicator for monitoring forest restoration. Enzymes are effective indicators of biological activity, energy transfer, pollutant detoxification, nutrient renewal, soil structure, etc. (Burns 1982; Paz-Ferreiro & Fu 2016). Many enzymes are susceptible to management-induced soil changes and have been used mainly as indicators of soil quality for agriculture (Samuel et al. 2017). Still, the use of enzymes in forest restoration is relatively incipient. The oldest article that uses these indicators in forest restoration was Nogueira et al. (2006), and since 2012, there has been only a slight increase in the application of this indicator.
Conclusions
The popularity and variety of soil indicators have increased over the recent decades. However, the incorporation of more integrative soil indicators related to soil health (e.g., aggregate stability, soil enzymes, microbial C) has been slow.
Few ecological forest restoration studies assess and compare levels of soil indicators against reference ecosystems. Restoration programs must consider a reference ecosystem to properly evaluate soil indicators' recovery trajectory.
A large percentage of articles that used soil indicators presented significant differences with the reference ecosystem even after 15 years. Thus, only a few restoration projects have reached the reference soil ecosystem indicator levels. This does not mean that the restoration efforts have not been successful, but most likely that the selected soil indicators' recovery is too slow to be observed.
More monitoring and evaluation are needed for soil indicators in the short term and consider a degraded control site (baseline or prerestored) in addition to the reference site. The latter will help discern earlier if the observed changes are indeed a response to the interventions or if they are evolving in the right trajectory.
The selection of soil indicators must be realistic, considering the intensity of soil degradation, the type of forest, implemented treatments, and the years of funding for monitoring and evaluation. Quick and easy-to-measure soil indicators are convenient but may not always be the most informative for restoration programs as they may not be precise or consistent.
Acknowledgment
The authors thank the financial support from Iniciativa Foresta Nativa of the Universidad de Concepción, project UdeC-Enel reforestation project (73-J-21-ER2).
