Using leading and lagging indicators for forest restoration

This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1111/1365-2664.13938 This article is protected by copyright. All rights reserved Received Date: 03/19/2020 Revised Date: 02/18/2021 Accepted Date: 05/11/2021 Article Type: Commentary Handling Editor: Cristina Garcia Using leading and lagging indicators for forest restoration Liz Ota1, John Herbohn1, Jennifer Firn2, Robin Chazdon1, Nestor Gregorio1, Sharif A. Mukul1, Ricardo A. G. Viani3, Claudia Romero4 1 Tropical Forests and People Research Centre, University of the Sunshine Coast, Sippy Downs-QLD 4556, Australia 2 Queensland University of Technology, Brisbane-QLD 4001, Australia 3 Federal University of São Carlos, Araras-SP, 13600-970 Brazil 4 University of Florida, Gainesville-FL 32611, United States Correspondence author: LO, lota@usc.edu.au


Introduction
"What we measure affects what we do", stated the Nobel Prize-winning economist Joseph Stiglitz (2009).
The quote referred to the way economic activity in the world has been guided to inflate countries' gross domestic product (GDP), and how inappropriate GDP is as an indicator of economic performance and social wellbeing. GDP is an indicator focused on production in the country, leaving behind many economic and wellbeing aspects, including human health, population income earnings, environmental integrity and economic sustainability (Stiglitz 2011). The United States of America and Argentina, for example, showed high GDPs that were based on unsustainable debts used for consumption boosts and not for investments (Stiglitz 2011). This example demonstrates that what we report, and hence strive for, may not be sufficient to achieve the broader goals of improved economic performance and wellbeing (Kubiszewski et al. 2013).
An analogous situation often emerges in the context of forest restoration around the world. Forest restoration is being promoted by major initiatives globally, including the Bonn Challenge launched in 2011, the New York Declaration on Forests (launched in 2014), and, more recently, the United Nations Decade on Ecosystem Restoration (2021)(2022)(2023)(2024)(2025)(2026)(2027)(2028)(2029)(2030). Many restoration targets are based on a forest and landscape restoration (FLR) approach. FLR focuses on re-establishing ecosystem functions at the landscape level balancing restoration of ecosystem services, such as land productivity and biodiversity (Sayer et al. 2013). It also aims to harmonise global and local, public and private interests (Roderick & Chavez-Tafur 2014). Landscape approaches are concerned with enhancing landscape sustainability and multi-

Accepted Article
This article is protected by copyright. All rights reserved functionality, often under polycentric governance structures (Wilson & Cagalanan 2016;Reed et al. 2017;Long et al. 2018). Given FLR's multiple goals and multi-governance structure and management, reporting on progress should encompass multidimensional concepts. Also, performance achievement visions vary according to stakeholders' needs and aspirations across time and space (Wilson & Cagalanan 2016;Chazdon et al. 2017;Reed et al. 2017). Nevertheless, FLR projects are not often planned and implemented with a long-term landscape perspective, and results are commonly measured and reported based on more immediate outcomes such as number of seedlings planted, or area reforested. Similar to GDP, these indicators often fail to document progress towards major goals (e.g., a functional forested landscape) nor do they indicate long-term sustainability of the restoration process (Chazdon et al. 2020).
Most frequently forest restoration uses only lagging indicators, while leading indicators are rarely considered. Lagging indicators represent realised outcomes of a process. Leading indicators represent likelihood of a particular outcome and are often predictors of lagging indicators. Leading indicators are widely used within business and economic sectors. They were first developed in economics following the 1929 recession in the USA to detect a recurring event, to measure its impacts, and to guide reactive actions (Moore 1983). The use of leading indicators has been more recently applied in the occupational health and safety literature and practice. In the investment sector, an example of a leading indicator is the Dow-Jones index of industrial common stock prices (Moore 1983) and in the health and safety sector are compliance on jobsite safety audits and pre-task planning meeting attendance at construction sites (Hinze, Thurman & Wehle 2013). Examples of lagging indicators from business, and health and safety sectors are sales levels (Moore 1983) and injury rates (Hinze, Thurman & Wehle 2013), respectively.
Typically, lagging indicators used in forest restoration inform on short-term outcomes rather than on long-term achievements (e.g. increased water infiltration, stable income from forest goods and services).
Simple, short-term lagging indicators are convenient because they can be used and understood by a variety of stakeholders, are resource-effective, and their documentation needs low levels of capital and training (Ruckelshaus et al. 2015;Mansourian, Dudley & Vallauri 2017). They reflect the short-term perspective with which forest restoration projects are often defined and provide a good documentation of effort. Metrics associated with forest restoration are often developed to match short policy cycles and funding timelines, and in many cases they reflect the scale of effort rather than the magnitude of impact achieved . The short period in which outcomes are expected from forest restoration is a serious concern (Chazdon et al. 2020). This short-term view has resulted in significant failures in tree planting, given the haste with which actions are implemented that leads to use of inappropriate planting materials, site selection, and overall practices (Holl & Brancalion 2020). Short-term lagging indicators provide little indication of the likelihood of the sustainability of restoration efforts and of the outcomes and impacts at later points in time. They are insufficient under the complex adaptive systems settings

Accepted Article
This article is protected by copyright. All rights reserved underlying restoration interventions, with their focus on multiple objectives that operate at different scales and involve several groups of stakeholders.
Many restoration outcomes are often not observable in the short term, such as the realisation of financial benefits from timber marketing attributable to a forest restoration intervention. Other lagging indicators can complement those simple, short-term lagging indicators. This broader set of indicators is pertinent to monitor and evaluate systems at different scales. For instance, employment generation may be measured on the short-term at the regional scale, while changes in water quality and quantity and availability of forest products may be medium-to long-term indicators of change, at regional and landscape scales, respectively.
Because FLR encompasses long-term processes based on adaptive management, broad objectives need to be articulated into more context-based short-term goals . This can be achieved through the use of leading indicators associated with lagging indicators. Secure land tenure is an example of a leading indicator at the onset of the project lifecycle, as it contributes to future impacts. Without secure tenure, communities may have little motivation to protect and tend seedlings. In some cases, indicators can be both leading and lagging, depending when changes are assessed. Seedling survival rate, for example, is an early lagging indicator and also a leading indicator of future tree growth performance.
By predicting longer-term lagging indicators using leading indicators in the near term, stakeholders committed to restore forests and landscapes can assess the likelihood of achieving a desirable long-term outcome. For example, the presence of fleshy-fruited, fast-growing trees that attract a wide range of frugivores may be a leading indicator for the species richness of trees in the future (Viani et al. 2015).
Furthermore, their early definition, identification, and measurement may guide project implementers to adapt management decisions for that long-term view. If the leading indicator points to suboptimal or negative performance, the issue related to the leading indicator can be recognised and addressed in a timely manner. The stop-light (red, yellow, or green) indicators of the Restoration Diagnostic, for example, provides an assessment of leading indicators that are key success factors for initiating planning and implementation of FLR at national or subnational scales (Hanson et al., 2015).
Leading indicators can also be used in the design of restoration projects. For example, leading indicators can be used to select communities most likely to successfully implement projects based on how 'prepared' they (Ota et al. 2020). Leading indicators can also be used to assess the status of enabling factors that are closely associated with the core principles of FLR (Chazdon et al., 2020b), and of key success and enabling factors for restoration, such as market conditions, policy alignment, stakeholder engagement, or institutional readiness (Hanson et al., 2015). In addition, connections between leading and lagging indicators can guide reporting, which is often a short-term requirement by governments and donor agencies. Within a project cycle, observed long-term values for lagging indicators are not available.

Accepted Article
This article is protected by copyright. All rights reserved Besides presenting the current state of the forest restoration process, short-term reporting could shed light on the likelihood of achieving future targets based on leading indicators that inform on the actions being taken.
Because FLR embraces both ecological and human wellbeing principles, where outcomes are measured as both short-and long-term changes and processes, we propose the complementation of lagging indicators with leading indicators in the design, planning, implementation, assessment, and adaptive management of forest restoration. We also argue for diverse leading and lagging indicators across time and space. In forest restoration, as in other natural resource management contexts, leading and lagging indicators will comprise a mix of biophysical, socio-economic and institutional indicators. These domains are very interdependent, although narrow sectoral and disciplinary approaches for their assessment are still common (Mansourian et al. 2020). We further layout and illustrate the use of these indicators using a case study from the Philippines. We purposefully avoid the use of the term indicators of success, given its vague and controversial nature that often requires specification of thresholds (Dudley et al. 2018) -a discussion beyond the scope of this paper.

Defining leading and lagging indicators for forest restoration
Lagging indicators are defined based on the objectives of the restoration initiative -which may differ among stakeholders. To define leading indicators, an extensive understanding of a restoration system might be required. A single set of indicators is unlikely to suit a broad range of contexts, although some leading indicators are likely to be relevant in most forest restoration contexts (e.g., secure land tenure, community engagement, use of high-quality seedlings). Leading indicators might relate to direct or indirect factors that affect forest restoration, or proximate and distal drivers, as these are often referred to in the literature on land use change (Rueda et al. 2019). Understanding factors and drivers that affect land use and cover change is important in the restoration process, and there are indications in the literature of potentially important drivers that could be used as the basis for identifying leading indicators.
Our initial thoughts on leading indicators were influenced by a comprehensive assessment of the factors affecting the success of reforestation projects in the Philippines (see Le, Smith & Herbohn 2014). That assessment identified multiple biophysical, socioeconomic and institutional drivers of reforestation success, including incentives, forest protection mechanisms, road infrastructure, revegetation method and forest dependence among local people -each of which could be used to develop leading indicators for reforestation success. Success of community groups also depends on various factors, from land productivity and property rights through community group governance and bridging social capital (Baynes et al. 2015).
Apart from scientific evidence, local knowledge is required for the identification of leading indicators and their connections. Factors that are likely to lead to success are often well-understood by those with

Accepted Article
This article is protected by copyright. All rights reserved experience implementing projects. The process of developing leading indicators, if done in an appropriate manner, can help to capture local experience and knowledge and incorporate them into the project design and implementation. The participation of representatives of different groups of stakeholders might ensure that hidden underlying processes are incorporated in a framework of indicators.

Examples of leading and lagging indicators and their connections
Leading and lagging indicators can be used in combination with several globally-recognised frameworks to explain relationships between components of forest restoration systems such as the IPBES Conceptual Framework (Díaz et al. 2015), the Socio-Ecological Systems framework (Ostrom 2007 This section presents an example of applying leading indicators to complement lagging indicators in forest restoration using ToC. ToC explain how activities will lead to a desired impact through specification of causal relationships and pathways based on evidence and knowledge as assumptions underlying the change process (James 2011;Brooks et al. 2014;Rogers 2014). These conceptual models provide implementation feedback throughout the monitoring process and have been largely used in development projects and programs to address complex issues (James 2011;Brooks et al. 2014;Valters 2015;Romero & Putz 2018).
ToC promote an iterative process through learning and assists decision making as part of adaptive management (Margoluis et al. 2009;Vogel 2012). Complexity of adaptive management can be simplified through multiple ToC subsystems that help understanding the system components and the context in which they operate (Gillson et al. 2019).
The Philippines Government National Greening Program (NGP) is an initiative aimed to restore over 7 million hectares of forests between 2011 and 2028 and realise biodiversity conservation, climate change mitigation and generation of livelihood opportunities (Department of Environment and Natural Resources 2018). It is largely being implemented through community forestry, with the NGP providing funding as well as rights for communities to access and manage public lands for forest restoration.
Communities must achieve 85% seedling survival rates during annual assessments to safeguard support from the government, which irrespective of successful performance, finishes after three years. Other metrics to assess the status of NGP operation are the number of seedlings and area planted, although employment generation is also reported (Gregorio et al. 2017; Commission on Audit 2019).

Accepted Article
This article is protected by copyright. All rights reserved The NGP seedling survival rate metric (i.e. number of live seedlings at time of assessment) requires no sophisticated technology and is easy to communicate to rural people in the Philippines (UNESCO 2015).
This metric also conveys a clear idea of what communities are expected to achieve including acting as a straightforward goal to monitor progress. However, this simplistic threshold has resulted in perverse behaviours alongside the forest restoration process. Some communities, for example, have replaced dead seedlings with wildlings or other seedlings without first addressing survival threats immediately before validation assessments by the Department of Environment and Natural Resources. The implication is that, even though the number of seedlings on the ground at the time of the assessment is over 85% of the number of seedlings planted, the NGP objectives (i.e. biodiversity conservation, climate change mitigation and generation of livelihood opportunities) and the long-term success of the program are unlikely to be achieved, as the seedlings planted for the sole purpose of meeting the threshold are unlikely to survive. In this case, the communities themselves cannot be blamed for striving to achieve this metric-it highlights the need for leading indicators as well that will assist with focusing on longer term goals. In this example, seedling survival would be enhanced by the use of high-quality planting stock produced through accredited nurseries -and thus the implementation of a seedling quality accreditation process could be a relevant 'leading indicator' of high seedling survival.
To illustrate potential indicators in the articulation of a ToC with specific outcomes in the context of NGP forest restoration, Figure 1 focuses on a single intermediate outcome, namely Increased forest cover. In this ToC, leading and lagging indicators on the biophysical (e.g. tree growth rates), institutional (e.g. land tenure status) and socioeconomic (e.g. technical capacity) spheres are interconnected. Level of community capacity is a leading indicator for the intermediate outcome on Sustained communal livelihoods, which in turn affects tree growth performance. One of the potential lagging indicators related to tree growth performance is mean annual increment, which is also a leading indicator for change in tree cover. Adaptive management principles are realised when examination of site and cross-site outcomes are the basis for monitoring and evaluation. both through time and space (i.e. local to landscape). As needed,

Accepted Article
This article is protected by copyright. All rights reserved management can be adapted through improved practices to secure progress towards project goals and facilitate learning. For instance, if level of community capacity is low, it is likely that communal livelihood activities will fail, potentially leading to reduced tree growth performance or complete failure of reforested areas. If the issue is detected in a timely manner, the community capacity failure can be addressed through a range of mechanisms (e.g. training, community organising, establishment of benefit sharing arrangements and conflict resolution) to increase the likelihood of a positive outcome on the longer term. However, the robustness of this assumption is fair and there is a risk that other factors (i.e., market price oscillations, competing livelihood opportunities) might compromise the sustainability of livelihood activities even if community capacity is increased.
A program like NGP could add to their reporting framework key leading indicators that relate to the longterm success of reforestation activities. The use of appropriate leading indicators could provide greater confidence that the current measures will produce the desired outcomes or alternatively allow early identification of issues that could jeopardise the success of reforestation. For instance, short-term reporting could include the number of projects with profit-sharing arrangements. At very early stages, the level of application of grazing management and weed control could be assessed as leading indicators, as well as road conditions. In a study on drivers of reforestation success in the Philippines, Le, Smith and Herbohn (2014) found that profit-sharing arrangements had a significant and positive effect on long-term maintenance of planted areas and that grazing management, weed control and good road conditions

The way forward
Because what is measured affects what is achieved, we argue for forest restoration to explicitly include leading indicators. The system presented is framed within an evidence-based agenda and focuses on enhancing the chances of optimal outcomes based on assessments at different scales for adaptive management towards forest restoration goals. Mapping leading and lagging indicators for forest restoration is both a tool and a set of processes that assist in defining desired changes with due attention to how activities' implementation is conducive to change and the nature of constraints and opportunities surrounding implementation on different contexts. We recommend this approach for forest restoration and firmly believe that leading and lagging indicators may also have an application in the management and restoration of other ecosystems and types of natural resources (e.g. community-based conservation, sustainable tourism, mangrove restoration). A recent study by Stevenson et al. (2021)

Accepted Article
This article is protected by copyright. All rights reserved application of this concept in global species extinction and also defends that the concept can be applied to different systems and scales.
Authors' contribution: JH, JF and LO conceived the ideas and led the writing of the manuscript; RC, RV, CR, SM and NG provided substantial contribution to the development of the concept and drafting of the manuscript. All authors contributed critically to the drafts and gave final approval for publication.