Donate | Join | Print Page | Sign In
A Tool for Assessing Ecosystem Recovery: The 5-Star Recovery System in Action

While every restoration practitioner strives to place his/her site on a secure trajectory to full ecosystem recovery relative to an appropriate reference system, full recovery can often be slow or unrealistic in the short-term. In these cases, and for all restoration projects, practitioners are encouraged to aim and monitor for continuous improvement toward ecosystem recovery. To assist managers, practitioners and others in tracking progress towards recovery goals, the new SER International Standards for the Practice of Ecological Restoration provide a tool for assessing and ranking a site’s degree of recovery over time. The 5-Star Recovery System tool utilizes a 5-star scale that represents a cumulative gradient from very low to very high similarity to a reference ecosystem. A restoration site can be assigned to one of the five recovery levels (1 to 5 stars) in an overall assessment (Table 1); or, different ecosystem attributes can be individually assigned recovery levels based on available monitoring data, which provides a more detailed overview of recovery progress (Table 2), and accounts for the fact that different attributes may have varying rates of recovery. The Recovery Wheel (Figure 1) provides a visual way in which to communicate ecological recovery progress using the 5-star system, and can be shaded in as various sub-attributes of the site achieve greater recovery over time.

Figure 1. Progress evaluation ‘recovery wheel’ depicting a hypothetical 1-year old reconstruction project on its way to a 4-star condition. See page 20 on the Standards for a full description and use.


Five-star recovery defines an ecosystem that is on a self-organizing trajectory to full recovery and is the ‘gold standard’ to which all ecological restoration projects aim. Projects whose goals do not include full ecosystem recovery based on an appropriate reference system can still use the 5-star ranking system to gauge recovery of desired functional attributes. These projects are considered rehabilitation rather than restoration, but can still benefit from the 5-star system, which may even convince managers to adaptively upgrade project goals over time.

This tool is only as rigorous and as reliable as the monitoring data that is used to inform the evaluation; therefore, practitioners should, whenever possible, develop a robust monitoring plan that includes metrics for all objectives of the restoration project. The generic indicators described in the abovementioned tables can be modified and made specific to a given site’s needs.

The following two case studies showcase the range of potential applications of the 5-star Recovery System as the authors use the tool to assess: 1) an Australian restoration site that has achieved a 4- to 5-star recovery level and has a robust monitoring program, and 2.) a large-scale rehabilitation project in China that falls in the 1- to 2-star recovery level range and has limited monitoring data available.


Snowy Adit: Restoring natural ecosystem function in a large rock dump in Kosciuszko National Park, Australia

Contributed by Gabriel Wilks, Team Leader/Environmental Officer
Landforms & Rehabilitation Team, National Parks and Wildlife Service NSW, Office of Environment and Heritage

Kosciuszko National Park in southeastern Australia contains a diversity of ecosystems and habitats including montane, sub-alpine, and alpine flora and fauna communities. It is recognized as an International Biosphere Reserve for its high number of endemic alpine plants, its specialist high altitude fauna, and geological and cultural values.

Snowy Adit Spoil Dump is one of many large remnant construction sites within Kosciuszko National Park that were a consequence of the Snowy Mountains Hydro-Electric Scheme (Australia’s largest industrial project) built from 1949 to 1974. The Landforms and Rehabilitation Team has been reducing the environmental and safety risks at these former construction sites since 2003, utilizing the teams ecological, engineering and field restoration skills.

Restoration objectives at Snowy Adit included re-instating a level of natural ecological function and reducing the risk of soil erosion, weed invasion, and impact on adjacent forest and waterways. The surrounding national park provided the ideal reference system for monitoring the effectiveness of the work. We have used a range of survey techniques during the restoration process, and recently applied the 5-Star Recovery System in the International Standards for the Practice of Ecological Restoration (McDonald 2016) to assess recovery.  

Site history

Snowy Adit Spoil Dump “Snowy Adit” sits at an altitude of 1,000 meters with a cool humid climate and about 100 rain days per year. Annual rainfall averages between 1,000 to 1,400 millimeters, falling mainly in winter. While snow can be reasonably persistent over winter on the surrounding hills, it doesn’t remain at any depth or period of time at the site. The general landscape is relatively intact and is a transitional zone between montane and sub-alpine vegetation in the Snowy River valley. Surrounding vegetation is a wet sclerophyll forest dominated by Mountain Gum (Eucalyptus dalrympleana), Ribbon Gum (E. viminalis) and occasional stands of Alpine Ash (E. delegatensis), with a tall shrubby understory of Leptospermum grandifolium, several Acacia species, and Kunzea ericoides.

During construction of aqueduct tunnels in the 1960’s, up to 950,000 cubic meters of rock spoil was dumped at Snowy Adit. Some of this material has since been removed by quarrying activities and used for roads and buildings in nearby towns and ski resorts. This ad hoc removal of rock has added to the steep faces and exposure of industrial debris that existed at the site at the start of rehabilitation. The footprint of the site is 11 hectares, with minimal natural regeneration of native species occurring over a fifty year period. Weeds established at the site included St. John’s Wort (Hypericum perforatum), Vipers Bugloss (Echium vulgare) and Bokhara Clover (Melilotus albus). The site is exposed to frequent high winds, the effects of which were exacerbated by lack of vegetation on the denuded site.

Restoration methodology

The site was progressively restored between 2008 and 2010, with some ongoing buffer planting around a section retained for long term rock sourcing. Restoration started with earthworks to reduce steep embankments, provide track and bench access across the site for revegetation works and provide for future potential water flow across the site with a series of shallow swales and pond depressions (Figure 1). Ground ripping to address the highly compacted nature of existing surface was a critical component of the earthworks, with 260 tons of waste metal from the site being collected and recycled. Coarse woody debris was sourced from logs and tree crowns removed during local trail clearing after snow and storms, and placed in windrows to provide wind shelter and thatch to hold straw and create microclimate (Figure 2). The ground was manually prepared prior to planting, with compost and water crystals added to individual planting holes. Contract labor crews planted 110,000 tubestock grown from locally collected seed and cuttings, and mulched the planting area with rice straw. Weed control was accomplished via chemical spraying, mechanical disturbance and application of mulch and woodchip.

After high initial browsing on planted seedlings by wallabies, deer and rabbits, and limited success with grazing deterrents and guards, most planting areas were fenced. 

Figure 1 (left). The initiation of restoration earthworks at Snowy Adit, 2008 (Photo credit: Gabriel Wilks). Figure 2 (right). Midstory and coarse woody debris development at Snowy Adit (Photo credit: Gabriel Wilks).


Site Monitoring Methodologies

We have used several monitoring techniques to progressively record changes to site conditions, vegetation and fauna at Snowy Adit. Regular site inspections and photo monitoring by trained staff are key components in recording site progress and identifying and responding to any management issues that may be impeding success.

Greening Australia, an independent not-for-profit conservation agency in Australia are measuring site physical and biological characteristics using Landscape Function Analysis (LFA) methodology (Tongway & Hindley, 2004). LFA is based on specific landscape characteristics, assessing the location and size of vegetation “patches”, where resources accumulate, and bare soil areas (“inter-patches”), where resources may be mobilized and lost. Soil surface indicators gauge the surface stability, infiltration capacity and nutrient cycling potential of the study area. A modified Biometric Native Vegetation Assessment Tool has been used to assess overstory, mid, and ground layer covers, including native and exotic plant species, litter, logs, rocks and bare ground.

Fauna surveyor Martin Schulz completed vertebrate fauna surveys prior to and after restoration using standard recording techniques, including observational census, capture and release in Elliott traps and harp nets, bat ultrasonic recording and remote cameras.

The data produced from these surveys and site inspection records provided essential objective evidence in assessing recovery of the site.

Application of the 5-star Recovery System

Tein McDonald and Andre Clewell raised the concept of applying the then draft National Standards for the Practice of Ecological Restoration in Australia (2016) during a site visit to Snowy Adit in early 2016. Their interest and support for the work and survey results at Snowy Adit was a motivational force for assessing the site using the new International Standards for the Practice of Ecological Restoration.

Using the Attributes Table from the International Standards’ 5-Star Recovery System and associated Recovery Wheel to assess recovery at Snowy Adit was a straightforward process, although the glossary in the International Standards document was useful as a reference throughout the exercise. Having independent and solid monitoring data was essential to validate and provide supporting evidence for selecting the correct attribute rating. Also valuable was the inclusion of several perspectives – those very familiar with the site and relevant ecosystems, plus some independent ‘fresh eyes’ to complete the site assessment. I discovered my close connection with the site tended to bias my assessment (even when using monitoring data) when applying the 5-Star Recovery System. Interestingly, this tended to be a negative bias as I was conscious of every minor detail that had been a problem during the restoration process. Consulting with Tein McDonald, who had visited the site but was not affiliated with the rehabilitation program, was invaluable in balancing this perspective. Where there was insufficient evidence to rate a given attribute, we took a precautionary approach in assigning a recovery level; for example, while we have completed vertebrate fauna surveys, there is no current data on invertebrate populations, so this influenced the recovery level rating for desirable species.

Figure 3. Snowy Adit Project Recovery Wheel, 2017.


Snowy Adit received a 4- to 5-star rating for most sub-attributes assessed (Figure 3), indicating that the site is well on its way to achieving full recovery. Ecosystem function and external exchanges are all near 5-star recovery levels, and most threats to recovery have been mitigated. In addition, more than 60% of local indigenous plants have established at the site (Figure 4).

Figure 4. Six years after the onset of restoration work, a flowering acacia (Photo A, Gabriel Wilks) and gang gang cockatoos (Photo B, Martin Schulz) are observed at Snowy Adit.



The extent of recovery at any site is determined by many factors, ecological and otherwise. Application of the 5-Star Recovery System to the Snowy Adit site is an effective way to document and visually display the recovery progress at the site. A similar systematic and visually representative approach could also be applied to non-ecological attributes such as budget, human resources, site and public safety.

The Snowy Adit case study reveals that many of the attributes are now at an advanced point of recovery. After 50 years of being highly degraded, Snowy Adit is now developing complex plant and animal relationships within the site and is not a risk to the surrounding natural environment. As time progresses, we hope that continued assessment will reveal full recovery of all ecosystem attributes at this site.



BioMetric Tool. 2005. New South Wales Office of Environment and Heritage. Available online: (

McDonald T, Gann GD, Jonson J, and Dixon KW. 2016. International standards for the practice of ecological restoration – including principles and key concepts. Society for Ecological Restoration, Washington, D.C.

Tongway, D and N Hindley. 2004. Landscape function analysis: a system for monitoring rangeland function. African Journal of Range & Forage Science 21:109–113.



China’s Grain for Green Program: large-scale reforestation and a case of ecological rehabilitation

Contributed by Fangyuan Hua1, Ph.D. and David S. Wilcove2, Ph.D. 

1Newton International Fellow, University of Cambridge
(Formerly Associate Research Scholar at Princeton University)
2Professor, Princeton University

China’s native forest suffered extensive loss over much of the 20th century. While the Chinese government has long recognized the environmental problems associated with forest loss, and indeed initiated a number of reforestation programs as early as the late 1950s, it was not until the end of the 20th century that systematic nationwide reforestation programs were put in place. The disastrous floods across much of China in 1998 – notably attributed to unchecked deforestation upstream of the nation’s major river systems – brought about an important turning point in China’s state forest policy. Painfully awakened to forests’ critical ecological functions, the Chinese government worked swiftly to initiate six large-scale programs, nationwide and regional, aimed at forest conservation and reforestation (Liu et al. 2008). The largest of these, in terms of both geographical coverage and financial input, is the Grain for Green Program (GFGP hereafter).

GFGP’s central tenet is incentivizing rural households to reforest/afforest sloped cropland, paid for with government-provided cash or food, thus replacing “grain” production with “green” forest land cover (State Council 2002). Less widespread variants of GFGP also include reestablishing forest on sloped scrubland (usually the remnant of degraded forest), or establishing shrub and grassland on sloped cropland. According to government statistics, as of 2013, GFGP had established 27.8 million hectares (ha) of forest in 26 of China’s 31 mainland provinces since its initiation in 1999 (SFA 1999-2014). China has committed to extending GFGP until at least 2020, with planned retirement of an additional 2.93 million ha of marginal cropland, much of which will be for reforestation. At this scale, GFGP is the largest reforestation program the world has ever seen.

GFGP’s focus on sloped terrain is a reflection of its guiding vision – soil and water retention – while other environmental benefits of forests are considered secondary (State Council 2002).  Indeed, as of the late 1990s, China had extensive cropland on steep terrain, causing severe soil erosion while producing only marginal yields. GFGP is primarily aimed at reforesting these heavily eroded marginal croplands for the benefit of soil and water retention (Figure 1). In the meantime, with the financial subsidy to rural households and the expectation of production from reestablished forests, the program also intends to alleviate rural poverty and facilitate rural economy restructuring toward non-farm income.

Figure 1. Example of heavily eroded, marginal cropland in China that is targeted by the Grain for Green Program (Photo credit: Fangyuan Hua).


These visions have had profound implications for GFGP’s approach and subsequent outcomes of reforestation. Overall, apart from a coarse distinction between orchard and non-orchard forest, which receive different levels of payment, all forms of forests are eligible under GFGP and are treated equally as long as they use at least one tree species from a long list of approved species (State Council 2002). The program also includes a strong incentive to use commercially valuable species and apply intensive, harvest-oriented management. Consequently, as our research team found out through an extensive literature review (Hua et al. 2016), monoculture forests dominate GFGP forests across China, supplemented by a smaller proportion of compositionally simple, production-oriented mixed forests, and there is a near complete absence of native forest established under GFGP (Figure 2). In addition, a number of non-native (or locally non-native) species are also used under GFGP, some featuring prominently, such as eucalyptus (Figure 3).

Figure 2 (left). Distribution of different types of Grain for Green Program forests across China (Hua et al. 2016). Figure 3 (right). A eucalyptus forest planted under the Grain for Green Program (Photo credit: Fangyuan Hua).


The large-scale, incentivized land cover change from cropland to mostly production forest means that GFGP will have complex, profound environmental and socioeconomic effects. Indeed, this topic has been under intensive study since GFGP’s implementation, with the strongest focus on GFGP’s soil (Lei et al. 2012), water (Feng et al. 2016), and carbon implications (Deng et al. 2014; Song et al. 2014) on the environmental side, and household income implications (Li et al. 2011) on the socioeconomic side. The vast majority of studies are based on the comparison of GFGP forests to croplands, and quite encouragingly, point to overwhelmingly positive impacts of GFGP at least in ecosystems where forest would belong. However, despite the vast variation of tree species used under GFGP, few studies made the distinction among different types of GFGP forest, and almost none compared GFGP forests to the native vegetation cover that preceded the cropland and therefore should be the more appropriate reference point. Moreover, the biodiversity implications of GFGP, a nontrivial environmental aspect of any massive land cover change, remained strikingly unexplored.

It is against this background that our research team set out to answer three biodiversity-focused questions to explore opportunities for bettering GFGP’s biodiversity outcomes, through intensive ecological and socioeconomic fieldwork in Sichuan Province (Hua et al. 2016). First, how does GFGP impact biodiversity by converting cropland to different types of forest? Second, compared with the proper reference ecosystem of native forest, what is the yet-to-be realized biodiversity potential of GFGP? Third, what are the economic cost and feasibility of achieving better biodiversity outcomes, as informed by the two previous questions, under GFGP? 

We focused on a ~8,000 square kilometer region that spans an elevation range of 315-1,715 meters in southcentral Sichuan. The study region was forested with subtropical evergreen forests historically, and then deforested extensively for agriculture through the mid-20th century before undergoing substantial GFGP reforestation over the past 15 years. Our pilot surveys showed that the study region had four dominant types of GFGP forest: monocultures of (1) eucalyptus, (2) bamboo, and (3) Japanese cedar, and (4) compositionally simple mixed forest consisting of two to five tree species (Figure 4). We used birds and bees as representative taxa for biodiversity. Birds are reasonably good indicators of the relationship between animal diversity and habitat structure. Bees rely heavily on floral resources, and thus represent a complementary component of biodiversity that responds more strongly to habitat composition; in addition, they are also important providers of pollination services, thus serving as an indicator of an important ecosystem service. We used point count and pan trapping to survey birds and bees, respectively, and used semi-structured household interviews to obtain data on the economics of forest management and production.


Figure 4. Vegetation cover categories found across the Grain for Green Program region (Images collated by Woodrow Wilson School, Princeton University).


The details of our field findings are complex because our inclusion of four types of GFGP forest, two types of reference ecosystem (cropland and native forest), and two seasons of biodiversity survey (breeding and non-breeding, only for birds) entailed a large number of comparisons. But by and large, the answers to the three above questions are quite clear. 

First, GFGP reforestation using monocultures generally results in net losses of bird diversity, while GFGP using mixed forest generally results in net gains. All current modes of GFGP reforestation result in overwhelming losses of bee diversity. 

Second, all current modes of GFGP reforestation fall well short of restoring biodiversity to levels approximating native forests. This finding is all the more salient considering that we were able to only use degraded native forests as reference systems because of the near complete absence of undisturbed native forests in the study region. There is thus considerable scope for biodiversity gains if GFGP were to incentivize the conservation and restoration of native forests over compositionally simple forests. 

Finally, even within existing modes of reforestation, GFGP can achieve biodiversity gains by promoting mixed forests over monocultures, and it can do so at little economic cost. This is because compared with monocultures, mixed forests are associated with higher bird diversity while being no worse for bee diversity. In addition, in terms of the economics of forest production, mixed forests perform just as well as monocultures, most likely because the region’s small, sloped landholdings preclude the potentially lower production costs of monocultures.  Forest production also generally contributes only a small percentage (~10%) of household income. Therefore, switching from monocultures to mixed forests is unlikely to carry opportunity costs or pose major unforeseen economic risks to rural households. We do not yet know, however, about the economics of restoring native forests.

GFGP shares commonalities with several other large-scale afforestation/reforestation efforts across the region and globe. For many of these efforts, the objectives are the recovery of specific ecosystem services such as soil stability, carbon storage, or timber production. Often, restoration is not an explicitly stated goal; however, the SER International Standards’ 5-star evaluation system can be applied to these programs to gauge their progress and potentially assist them in upgrading to actions that promote full ecosystem recovery as an end goal.

GFGP is a reforestation program of enormous geographical scale and involves a wide spectrum of forest types, so any application of the International Standards’ 5-star evaluation system should come with the caution that deviations from the general assessment may well exist. Nonetheless, the separate evaluations of our multiple study sites can be aggregated to provide a general overview of the larger program. Given the overall pattern of forest reestablishment and management, biodiversity indicators supported, and policy as well as socioeconomic context, GFGP should be assigned a 1- or 2-star ranking on the 5-star evaluation system. In its current form, GFGP should be considered an ecological rehabilitation program rather than a restoration program, because the goals of the program do not include the reinstatement of biotic characteristics that resemble an appropriate reference system (McDonald et al. 2016).

The main consideration for such a ranking is the lack of native biota present and reduced ecosystem functionality: monocultures, particularly of non-native species, support a limited number of native species. Mixed forests, particularly those consisting of native tree species, bear a closer resemblance to native reference ecosystems, but are nonetheless associated with limited native biota. Moreover, subject to periodic harvest, current GFGP forests are unlikely to develop beyond their harvest state. Overall ecosystem functionality may also be limited because of low native biodiversity, as evidenced by the depauperate bee diversity in GFGP forests we studied. An increased emphasis on incentivizing reforestation with native and mixed composition forests would help move the GFGP program to higher ranking values, and greater ecosystem recovery, should administrators wish to move the goals of this large-scale program beyond those of soil and water retention.


Deng L, Liu G, Shangguan Z. 2014. Land-use conversion and changing soil carbon stocks in China’s ‘Grain-for-Green’ Program: a synthesis. Global Change Biology 20:3544-3556.

Feng X, Fu B, Piao S, Wang S, Ciais P, Zeng Z, Lu Y, Zeng Y, Li Y, Jiang X, Wu B. 2016. Revegetation in China’s Loess Plateau is approaching sustainable water resource limits. Nature Climate Change 6:1019-1022.

Hua F, Wang X, Zheng X, Fisher B, Wang L, Zhu J, Tang Y, Yu DW, Wilcove DS. 2016. Opportunities for biodiversity gains under the world’s largest reforestation program. Nature Communications 7:12717.

Lei D, Shangguan Z, Rui L. 2012. Effects of the Grain-for-Green program on soil erosion in China. International Journal of Sediment Research 27:120-127.

Li J, Feldman MW, Li S, Daily GC. 2011. Rural household income and inequality under the Sloping Land Conversion Program in western China. Proceedings of National Academy of Sciences 108:7721-7726.

Liu J, Li S, Ouyang Z, Tam C, Chen X. 2008. Ecological and socioeconomic effects of China’s policies for ecosystem services. Proceedings of the National Academy of Sciences 105:9477-9482.


 Previous article  Next article