Realizing the potential of restoration science

Abstract

Restoration science is growing fast. The restoration of habitats is increasingly part of the discussion over how to tackle the challenges of climate change, biodiversity loss and rural development. With this increasing role and attendant visibility, restoration science has seen increasing controversy. Here I describe six aspects of robust restoration science that should be kept in mind to help realize its potential: do data-driven studies; focus on robust results; improve reproducibility; contextualize the results; give attention to economics; consider the wider goals of restoration. Realizing the potential of restoration science, via robust scientific studies, will provide society with the knowledge and tools to make better choices about which habitats to restore and where.

This article is part of the theme issue ‘Understanding forest landscape restoration: reinforcing scientific foundations for the UN Decade on Ecosystem Restoration’.

Restoration science is rapidly expanding and restoration scientists are in demand. The UN Decade on Ecosystem Restoration (UN Resolution 73/284), the potential to remove carbon dioxide from the atmosphere [1], and the realization that habitat connection helps species move as the climate rapidly changes [2], have all shifted research and funding priorities towards restoration science. Furthermore, pervasive biodiversity loss [3] and escalating climate change impacts [4] have led many to conclude that stopping the destruction of habitats is not enough. Habitat restoration is also required to resolve these deep-seated problems [5,6].

However, with increasing visibility, restoration science has seen controversy. Some are concerned that the potential of habitat restoration has led to a focus on tree planting rather than protecting existing forests [7]. Others worry that restoration projects are being used as ‘greenwash’ by countries and companies to avoid reducing their fossil fuel emissions [8,9]. Indeed, while some argue that a focus on ecosystem restoration is harnessing the power of nature to tackle societal problems, others claim that it is merely another tool being used to extend today's unsustainable and unjust status quo [10]. Whatever your view, restoration is certainly playing an increasingly important role in climate, biodiversity and development agendas.

The new funding and opportunities available to restoration scientists, as well as increased scrutiny, requires increased attention on scientific rigour and care in how scientific results are presented. Conflicts of interest, or perceived conflicts of interest, will need to be very clearly advertised, particularly given that more science is being published by restoration scientists working for large non-governmental organizations (NGOs) who stand to benefit from funding for restoration, and by academic scientists who receive grants from those who may benefit from restoration [11].

Restoration scientists can rise to these challenges, as seen from the engaging studies in this themed issue of Philosophical Transactions of the Royal Society of London: Biological Sciences, summarized by Marshall et al. [12]. These include utilising science to advance the goals of the UN Decade on Ecosystem Restoration [13], investigations of physical factors that may alter restoration outcomes [14], from species selection choices [15,16], to soil conditions [17,18], fires [19,20], cyclones [21], the role of animals [22] and the impacts of invasive species [23]. The theme issue also assesses differing interventions in restoration practice, from spatial priority setting [24] to specific restoration interventions [25] and socio-economic analyses [26–28]. Building on the lessons implicitly embedded in these studies, I suggest that the potential of restoration science will be better realized if the following six aspects are considered.

Do data-driven studies. As restoration itself is typically within the realm of practitioners, very few restoration projects include robust well-designed scientific monitoring of their interventions, so little is known about what works and what does not work. Thus, many restoration science studies can appear opportunistic, with the literature containing many well-meaning opinion-type articles relative to the number of data-driven studies (some might say like this one!). What restoration science needs is more data and more evidence, including more well-designed experiments and robust case studies. In this themed issue, Chazdon et al. [29] show hard-won long-term data that can inform on how restoration may occur, while tackling the issue of sample sizes. Good case studies are also powerful: in a rare assessment of the impacts of tree planting over the long term and on a large scale, a study of one state in India failed to show positive effects. Tree planting did not increase forest cover, and planted areas did not support use by local people [30]. Rigorous syntheses of published studies can also move the field forward, but authors should follow guidelines for such systematic reviews, whether the evidence is quantitative [31] or qualitative [32] studies, and researchers should follow PRISMA guidance (http://prisma-statement.org/).

Focus on robust and meaningful results. Some researchers face pressures, or feel the need, to focus on getting a ‘big number’ published, which then generates widespread publicity. These pressures include publication in high-impact journals, gaining access to decision-makers and generating future funding. One highly influential study led with the claim that ‘natural climate solutions', including ecosystem restoration, could sequester or avoid emitting 23.8 petagrams of CO₂ equivalent (1 petagram = 1 billion tonnes) annually [33]. This number is big, compared to total global greenhouse gas emissions of 55.6 Pg CO₂ equivalent in 2018 [34], and is 30% higher than previous estimates [33]. It is, however, a theoretical maximum potential, with no economic constraints, rather than something that could be implemented (see below). Similarly, another ‘big number’ study on restoration [35] reported unconstrained carbon uptake values in the abstract and figures, while in the study various constraints showed a 29–49% reduction in the potential carbon sequestration [36]. Leading with such figures can easily mislead policymakers and the public. Giving primary prominence to best-estimate results will not give such a big number, but it will ensure a more nuanced understanding of what restoration can, and cannot, do.

Improve reproducibility. There is evidence of a reproducibility crisis in science [37] although others have countered this [38]. Restoration science has not escaped publishing hard-to-reproduce studies. A very high-profile estimate of the global tree restoration potential, another ‘big number’ study, suffered reproducibility and other problems [39]. Multiple groups of researchers challenged the conclusion that tree restoration could remove 204.7 Pg C from the atmosphere (750 Pg CO₂) on methodological grounds [40–42]. Bastin et al. [43] replied that the originally published methods were not correct, but the 204.7 Pg C result itself was correct, and provided new methods to calculate this. A subsequent correction showed that the new methods needed to be combined with a series of coding and transcription errors to obtain the original 204.7 Pg C result [44], making the study hard to reproduce. This case highlights that easily reproducible results are needed to save researchers time, and so ensure more rapid progress in restoration science. Checking your written methods allows the reproduction of your results, and publishing your code and dataset with your paper will allow researchers to reproduce results efficiently.

Contextualize the results. Restoration science is difficult because if often spans disciplines. One outcome of this is that it can be challenging to place results in the context of the available literature. For example, results from climate science, rather than ecology, are sometimes missing from reports on the carbon or climate impacts of restoration. For example, the ‘big number’ studies of both Bastin et al. [39] and Strassburg et al. [35] overestimated the impact of restoration on atmospheric carbon dioxide, by a factor of two, due to misunderstanding the science of the global carbon cycle, requiring published corrections [44,45]. Knowledge, for example, of the literature modelling habitat restoration and its likely impact on the global carbon cycle (e.g. [46]), through to restoration experiments within Earth System Models (e.g. [47,48]) and the reports of the Intergovernmental Panel on Climate Change can guard against such mistakes. Similar arguments can be made for contextualizing the governance system(s) in the location(s) of the restoration study, and its social and economic history. Restoration science will collectively make faster progress by improving our transdisciplinarity by incorporating relevant research from other disciplines.

Give attention to economics. Restoration is a real-world activity, and so economics matter. As restoration rises up the political agenda and more restoration is planned, analyses of the costs of restoration, including the opportunity costs, will become increasingly important [12]. The restoration potential with no economic constraints in land-use decisions is clearly larger than the potential with economic constraints. Thus figures such as 750 Pg CO₂ total sequestration [39], or 23.8 Pg CO₂-equivalent annually [33] are classed as technical potentials, and not those with economic constraints. The difference when including economic costs can be very large. Unconstrained by economics, it has been suggested that 11 Pg CO₂ yr⁻¹ can be sequestered by forest restoration, whereas just 3 Pg CO₂ yr⁻¹ is sequestered by forest restoration costing <US$100 per Mg CO₂ [33]. While an economic analysis is not always necessary, avoiding leading with the ‘big number’ and clearly highlighting that only a modest fraction of an unconstrained technical potential is likely to be possible in the real world may be enough. However, studies including economic constraints show this can be done convincingly [35,49].

Consider the social and environmental impacts and goals of restoration. Restoration is about land use and therefore must contend with the power structures that determine land-use decisions [12,50]. The power relations of different groups mediate the uses of land, and even contest what is considered restoration. For example, when analysing country pledges for restoration under the Bonn Challenge to restore 350 million hectares globally by 2030, it was discovered that approximately 45% of the area pledged by governments were for mono-culture plantations, rather than what is typically regarded as habitat restoration [51]. Furthermore, in many parts of the world, what outsiders consider marginal or under-used lands are usually essential resources for the benefit and even survival of often marginalized groups. One ‘big number’ study, while carefully attending to the economic cost of not farming land to assist with mapping restoration priorities [35], it was noted that this ‘opportunity cost’ is probably not the true cost to marginalized people who use the land, which may lead to their displacement and further impoverishment [52]. Further research is needed on how to incorporate such perverse outcomes into studies [53]. More generally, restoration science needs to better consider who restoration is for, and who is it done by [54]. Including these aspects as central to restoration science will maximize the benefits of restoration projects, and likely their longevity, as local people will need to see benefits for restoration projects to succeed in the long term.

This opinion piece makes some criticisms of past papers. I hope they can be taken in the spirit of using experience to improve restoration science. The potential of restoration science is to provide society with the knowledge and tools to make better choices about which habitats to restore and where. Producing as robust science as possible will help this potential be realized.

Data accessibility

This article has no additional data.

Conflict of interest declaration

I declare I have no competing interests.

Funding

I received no funding for this study.