Potential of different governance mechanisms for achieving Global Biodiversity Framework goals

doi:10.21203/rs.3.rs-4170734/v1

Download PDF

Article

Potential of different governance mechanisms for achieving Global Biodiversity Framework goals

https://doi.org/10.21203/rs.3.rs-4170734/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

The Kunming-Montreal Global Biodiversity Framework includes a target of 30% of land protected by 2030 and refers to other effective area based conservation measures (OECMs) as complementary to PAs, but robust evaluations of the effectiveness of governance mechanisms that could act as OECMs in preventing forest loss and carbon emissions remain sparse. Here we assessed the impact of PAs and two potential OECMS: Indigenous Lands (ILs), and Non-Timber Forest products Concessions (NTCs) on forest loss and its associated carbon emissions in the Peruvian Amazon from 2000 to 2021. We also assessed two governance mechanisms with a commercial extractive use, Logging (LCs) and Mining Concessions (MCs). We used a robust before–after control intervention study design, with statistical matching, to account for the non-random spatial distribution of deforestation pressure and the governance mechanisms analysed. PAs were the most effective, having avoided 88% of the expected forest loss, followed by NTCs (64%) and ILs (44%). LCs also reduced expected forest loss by 29%, while MCs increased expected forest loss by 24%, showing that extractive governance mechanisms can have marked differences in their impact to forest cover. Our study provides evidence of long-term positive impacts of potential OECMs and other mechanisms at preventing forest loss and reducing carbon emission. This information is key to more effectively achieve targets from the Kunming-Montreal Global Biodiversity Framework and the UN Framework Convention on Climate Change.

Earth and environmental sciences/Environmental sciences/Environmental impact

Earth and environmental sciences/Climate sciences/Climate change/Climate-change mitigation

Earth and environmental sciences/Ecology/Conservation biology

Earth and environmental sciences/Environmental social sciences/Climate-change impacts/Governance

Other area based conservation interventions

Deforestation

Peru

Amazon

Indigenous land

forest concessions

mining concessions

carbon markets

selective logging

Tropical forests are amongst the most biodiverse ecosystems globally ^1,2. They also provide multiple crucial services to humanity at local and global scales ^3,4, including storing and sequestering large amounts of atmospheric carbon, which is key to directly regulate climate ⁵. Despite this, the loss and degradation of tropical forest has increased over time ^6,7, and it has been shown to contribute to up to 20% of the world’s greenhouse gas emissions ^5,8.

Protected areas (PAs) have always been regarded as a fundamental tool for the protection of forest ecosystems and the ecosystem services they provide. Now more than 15% of global land is under a certain type of protection regime ⁹. Despite this, evidence shows that PAs are not always effective at preventing deforestation. One third of PAs globally are under intense human pressure ¹⁰, half of protected forests have low or medium integrity ¹¹ and 3% of the forest inside PAs was lost in the first decade of the 20th century ¹². Additionally, socioeconomic conflicts may arise where PAs are established, due to restrictions in access to natural resources and territory, and those socioeconomic aspects must be taken into account when practical conservation decisions are taken ¹³. Furthermore, in some circumstances, strict protection might not be appropriate or feasible, and other conservation actions can be more suitable ¹⁴.

Due to the socioeconomic conflicts that might derive from PAs establishment under certain circumstances and that PAs alone are not enough to meet the commitments of the Kunming-Montreal Global Biodiversity Framework ¹⁵, calls have mounted for the identification of other effective area-based conservation measures (OECMs) that can deliver biodiversity outcomes to complement those provided by PAs ^9,16. Indigenous management of land, for example represents a form of governance that has proven to have positive impacts on biodiversity. There is growing evidence showing that indigenous lands (ILs) can be effective at reducing deforestation under certain management conditions ^17–19. However, there is a need for a better understanding of the interactions between deforestation trends and ILs management in different regional and local contexts ^18,20. Forest concessions, a widely used contractual arrangement, in which a government temporarily provides public forested land access, management and exclusion rights to non-government actors, is another governance mechanism that can provide positive outcomes for forest conservation while generating a potential source of income for local communities ²¹. There are at least 122 million hectares of Forest Concessions in West and Central Africa, Latin America and Southeast Asia ²¹, with different uses ranging from industrial extraction, including timber extraction or logging (logging concessions, LCs) ^22–24, to more sustainable uses, like agroforestry ²⁵, tourism and conservation ²⁶, hereafter referred to as Non-Timber Forest Product Concessions (NTCs). Their effectiveness in reducing deforestation varies depending on the type of concession and the political characteristics in the country they are established, but there is prevalent support for positive outcomes under adequate management and supportive political circumstances ^22,26,27.

Rigorous evaluations of the effectiveness of potential OECMs and other alternative governance mechanisms in preventing forest loss, particularly relative to PAs, remain relatively sparse ^28–30. For OECMs the main limiting factor is that it has not yet been possible to track their extent systematically ⁹, and some countries are still in the process of identifying OECMs within their territory ²⁹. On the other hand, studies assessing the impact of PAs preventing deforestation, have often failed to account for their non-random location. Some studies, for example, simply compare areas that are protected to all the unprotected area present in the landscape ^12,31, or to neighbouring buffers ³². This can cause bias in the estimated impact of PAs as they are non-randomly distributed across space tending in general to be located towards land that, if unprotected, is less likely than average to be cleared ^14,33,34. Identifying governance mechanisms that can be potential OECMs is needed for efficient conservation planning ^9,29,35. The Kunming-Montreal Global Biodiversity Framework includes a target of 30% of land protected by 2030 ¹⁵. On the other hand, under the UN Framework Convention on Climate Change ³⁶, countries revise their nationally determined contributions to reduce emissions every five years. Both conventions cover commitments to reduce deforestation and its associated emissions through protection of natural systems. It is then crucial to understand the contribution that PAs and potential OECMs have towards achieving these commitments and how they can complement each other to have the best conservation outcomes ²⁹.

Here we assessed the impact of government-managed PAs and other governance mechanisms on forest loss and carbon emissions using a robust before–after control intervention study design, with statistical matching. We used the Peruvian Amazon as our study area (Fig. 1). The country is megadiverse ³⁷, has multiple land tenure types ³⁰ and its Amazon region accounts for 60% of the country’s land mass and for 90% of its forest ^38,39. Additionally, the Peruvian Amazon is experiencing high rates of forest loss and degradation ^30,31,40. We used statistical matching as this methodology evaluates the effectiveness of conservation interventions by accounting for the non-random spatial distribution of those interventions and of deforestation risks ^18,41–43, comparing treatment sites to matched control ones (i.e., areas with no intervention with characteristics similar to treatment sites) ^33,44,45.

A previous study in the Peruvian Amazon found that between 2006 and 2011 PAs, ILs and Conservation Concessions avoided deforestation and degradation compared to analogous areas in the unprotected landscape ³⁰, and another study, found a positive association between titling of ILs and the reduction of clearing and forest disturbance between 2002 and 2005 ⁴⁶. While these studies suggest a positive impact of PAs, Conservation Concessions and ILs on the reduction of forest loss, they did not assess the impact of other governance regimes with conservation potential, like agroforestry, reforestation or tourism concessions, or the effect of forest loss on carbon emissions. Furthermore, the timeframe of their analysis is too short to provide an overview of the longer-term impacts of these governance mechanisms on forest loss.

Particularly in this study we assessed the impact of government-managed PAs, and two potential OECMS: Indigenous Lands (ILs), and Non-Timber Forest Products Concessions (NTCs) on forest loss and its associated carbon emissions in the Peruvian Amazon from 2000 to 2021. We also assessed two other governance mechanisms with a mostly commercial extractive use, Logging Concessions (LCs) and Mining Concessions (MCs), as a point of contrast to assess how different governance mechanisms impact forest loss. Mining concessions in particular, have been strongly associated with forest ^30,47,48 and biodiversity loss ⁴⁹ in the tropics.

Study Area and governance mechanisms analysed:

We used the Peruvian Ministry of Environment delimitation of the Peruvian Amazon ⁵⁰, which includes low elevation areas but also Amazon Andes foothills. We used the information on spatial distribution and establishment dates of government-managed PAs defined by the Peruvian Protected Areas’ Service - SERNANP ⁵¹. For the spatial distribution of ILs we made a request of information to the Ministry of Culture ⁵². Details on ILs creation dates were obtained from the database of Indigenous and Native Peoples ⁵³. For the spatial distribution of Forest Concessions we used information from the National Forest and Wildlife Service ⁵⁴ and from a request of information to the regional governments of Madre de Dios, Loreto and Ucayali. When there were discrepancies with SERFOR, information from regional governments was used, as these entities have a more detailed record of their jurisdiction than national entities. Details on Forest Concessions establishment dates were obtained from the Forestry and Wildlife Resources Supervision Agency ⁵⁵. Information on MCs spatial distribution, start and closure dates were obtained from the mining and geological geoportal from the Ministry of Energy and Mines ⁵⁶.

We divided the Forest Concessions into two groups taking into account the purpose of the concession and the definition of potential OECMs ^57,58. We aggregated the concessions that in principle do not use timber products as a source of income, including conservation, ecotourism, reforestation, restoration and agroforestry into a single category of Non-Timber Forest Product Concessions (NTCs). These concessions meet the criteria of the definition of potential OECMs ^57,58. Logging concessions (LCs) on the other hand were included in a separate group as their main objective is commercial extraction of wood. This does not align with criterion six required for their identification (Jonas et al 2023) and criterion C for their recognition and reporting as OECMs ⁵⁷. These two criteria specify the impossibility of sites that are subject to environmentally damaging industrial-scale activities to be classified as OECMs.

Mapping forest loss and carbon emissions:

We used the map of Amazon forest cover for 2000 and the one of Amazon forest loss from 2000 to 2021 developed by the Peruvian Ministry of the Environment and the Ministry of Agriculture and Irrigation ⁵⁹. These maps have a resolution of 30 by 30 meters where each pixel is either forest or no forest in the map of forest cover for the year 2000 and have a value from 1 to 21 in the map of forest loss. Each value represents the year when the forest cover on that pixel was lost. The classification of a pixel into either forest or no forest is based on different Landsat images of the Peruvian Amazon for each year, which are then processed at a pixel level. The accuracy of these maps was 92.2% for forest loss and 99.5% for no change ^40,60. We then quantified the extent and rate of forest loss for each year between 2000 and 2021 and the total cumulative loss between 2000 and 2021. This was done for the whole basin and individually for each governance mechanism assessed.

We used the high-resolution carbon per hectare density estimations for Peru, generated by Asner et al. (2014) ⁶¹, to calculate the metric tons (Mg) of carbon by forest pixel. We assumed that if a pixel was deforested, all of its carbon was emitted into the atmosphere. Then we summed the amount and proportion of carbon emissions by year and cumulative between 2000 and 2021 for the whole landscape, and for each governance mechanism as well as for the selected control and treatment pixels after the matching analysis. In this study, “carbon emissions” refers to emissions “committed” at the time of disturbance or clearing, noting that there may be a time lag until they are “realized”.

Cofounding factors:

To account for the determinants of deforestation pressure and the location of the different governance mechanisms assessed, we used a suite of biophysical and socioeconomic cofounding factors that have been found to be associated with the non-random spatial distribution of these interventions and of the distribution of deforestation risk in the Peruvian Amazon and other tropical landscapes ^30,42,43,62. These cofounding variables were estimated at the pixel level and included in the matching analysis. These included measures of forest accessibility (i.e. distance to roads, rivers and previous deforestation), anthropic pressure (i.e. distance to settlement, travel time to cities, human population density), suitability of conversion to agricultural land (i.e. precipitation, precipitation seasonality, ecoregion, elevation and slope) and the socio-political environment (i.e. administrative region) (Table 1). Spatial information that was at a different spatial resolution was resampled to the forest pixels’ resolution of 30 by 30 meters.

Statistical Matching:

We used forest loss between two time periods as alternative outcome variables for our matching analysis in order to assess the robustness of our results. Forest loss by pixel between 2000 and 2021 and forest loss by pixel between 2005 and 2021. We excluded governance mechanisms that were established after 2005. Areas of overlap, and other types of governance mechanisms with conservation potential that were region-specific were also excluded. Also, concessions that stopped operating before 2015 were excluded. To avoid the potential spill-over effects, particularly leakage ^44,63, we excluded the pixels within a buffer of 1 km (NTCs, LCs, MCs and ILs) and 5 km (PAs) around treatment areas ³⁰. In order to avoid autocorrelation we generated a set of random samples at least 500 m apart from each other for each governance mechanism and for the potential control units ^42,64. Control units were defined as pixels which were not assigned to any of the governance mechanisms assessed in this study.

To assess the impacts of the governance mechanisms (treatments) on forest loss we used a before–after control intervention study design, with matching analysis, creating artificial control groups for each treatment unit (pixel) while controlling for differences in observed covariates ^43,65. For each pixel, we extracted the values of the cofounding variables associated with deforestation and the governance mechanisms’ location. This was done both for the treatment group (each governance mechanism pixel) and the potential control group (pixels that where not classified under any of the assessed governance mechanism). To determine the best minimum set of cofounding variables, we first ran a binomial logistic regression model using each governance mechanism as the response variable and the cofounding factors as predictor variables. Then we ran a multicollinearity test to the coefficients of the covariates of each governance mechanism model. Through this analysis, we identified groups of variables with a collinearity higher than 0.65. The variable with the lowest Variance Inflation Factor (VIF) for each group was retained in the model ^30,66. We then used 'backward' and 'forward' techniques in an iterative stepwise process for model selection, to identify the optimal model. The model that achieved the greatest reduction in the Akaike’s Information Criterion (AIC) was selected ³⁰ (Supplementary Information 1). For this we used the leaps package ⁶⁷ in R (version 3.5.1) software ⁶⁸. The variables from the best model were used in the subsequent analyses and for the matching procedure for each governance mechanism.

We then used the MatchIt package ⁶⁹ in the R statistical software to pair treatment and control pixels with similar values for each of the covariates ⁶⁸. We used Propensity Score Matching to create a statistically balanced counterfactual sample to evaluate the effect of each governance mechanism on forest loss. Despite its limitations, Propensity Score Matching remains the most widely used matching approach ^30,42,43,65. We matched each treatment observation to a unique control observation to ensure each governance mechanism pixel was paired with only one control pixel that had not been matched previously. Pixels could be reselected as control samples for each independent governance mechanism analysis. Matched pairs of pixels had to be within 0.25 standard deviation of the propensity scores ^43,70. As part of the diagnostic statistics we calculated the index of covariate imbalance for every matching analysis for each governance mechanism to assess the robustness of the matching analysis reducing bias in the control units (Supplementary Information 2a). Absolute scores > 25% are considered an indication of a possible imbalance for that specific variable ^43,71.

Deforestation in the Peruvian Amazon and the different governance mechanisms

Annual deforestation over the last 20 years in the Peruvian Amazon almost doubled, from 840 km² in 2001 to more than 1,400 km² in 2021. Annual deforestation has been increasing, especially since 2006. There have been two notorious increases in yearly deforestation, one in 2005 and another in 2020 (Fig. 1 a & b). In relation to the governance mechanism, ILs had the highest absolute yearly forest loss followed by LCs (Fig. 2a). In proportional terms, MCs showed the highest yearly and accumulated loss since 2000 with a total of 21% of their forest cover lost in the last 20 years. LCs experienced the second highest cumulative loss with 4% of the forest lost in the last 20 years. On the other hand, PAs showed the lowest yearly and cumulative loss with 0.6% of the forest lost in the last 20 years (Fig. 2 b & c).

Carbon emissions in the Peruvian Amazon and the different governance mechanisms

Carbon emissions in the Peruvian Amazon have consistently increased in absolute and in percentage terms since 2000. This aligns with the increase in deforestation observed over the past 20 years (Fig. 1. a-c). Particularly in 2020 there was a high net increase in yearly emissions for the Peruvian Amazon (Fig. 2 a). This increase can also be seen in proportional terms (Fig. 2 b), and is also observed in ILs, LCs and MCs (Fig. 3). The governance mechanisms with the highest yearly net carbon emissions were ILs, followed by LCs (Fig. 3 a). MCs had the highest proportional carbon emissions yearly and cumulative to the stock in 2000, while PAs had the lowest (Fig. 3 b & c).

Robustness of the matching analyses

Most covariate imbalances between the treatment and control units were reduced through the matching analysis. For ILs, LCs and MCs these reductions meant that all the variables had indices of covariate imbalance below 25% post matching. For PAs, only the variable distance to settlements of more than 10 people in 2017 had an index of imbalance of 49% post matching. However, this was a reduction from 74% pre matching. For NTCs, 16 variables had indices of covariate imbalance below 25% post matching. Three had indices higher than 25% post matching, but for those variables these values implied reductions of 10% or more in covariate imbalance post matching. For two covariates, there was an increase in covariate imbalance post matching (Supplementary Information 2a). Overall, there was notable improvements reducing bias in the control units through the matching analysis (Supplementary Information 2a, 3a & 4a). Consistent results were obtained when using forest loss between 2005 and 2021 for the analysis (Supplementary Information 2b, 3b & 4b).

Impact of the different governance mechanisms on forest and carbon loss

Our results show that PAs, ILs, NTCs and LCs significantly decrease forest loss compared to the matched control areas. PAs were the most effective in preventing forest loss, avoiding 88% of expected loss, followed by NTCs (64%), ILs (44%), and LCs (29%). On the other hand, MCs showed a 24% increase in expected forest loss, compared to matched control areas, although the increase was not significant (Fig. 2d). Consistent results were found when using forest loss between 2005 and 2021 for the matching analysis (Supplementary Information 5).

The median carbon density by pixel was highest for forest in NTCs, followed by LCs. The lowest median carbon density was for MCs. The median carbon density of the pixels without any governance mechanism designation was higher than those of all the governance mechanism except for NTCs and LCs. However, none of these differences in median carbon densities were statistically significant (non-parametric Kruskal-Wallis test, p value; PAs = 0.17, ILs = 0.62, CCs = 0.35, LCs = 0.38, MCs = 0.49). There was a high dispersion in the carbon density values by pixels and an overlap in this dispersion between all of the governance mechanisms (Fig. 2e). Finally, with regards to carbon emissions represented by the loss of forest, PAs, ILs, NTCs and LCs had lower emissions than matched control areas. This decrease was 88% for PAs, 65% for NTCs, 42% for ILs and 20% for LCs. On the other hand, MCs had a 35% increase in carbon emissions compared to the matched control areas (Fig. 3d). Consistent results were found when using forest loss between 2005 and 2021 for the matching analysis (Supplementary Information 5).

In our study of the Peruvian Amazon we found that 4% of the forest has been lost in the last two decades. Putting this into perspective, until today the Brazilian Amazon has lost 20% of the historical forest cover ⁷². However, PAs and other governance mechanisms assessed here have played a significant role in avoiding an even stronger decline of forest cover in the region. PAs experienced nine times less forest loss than similar areas without protection, while NTCs experienced three times less and ILs two times less losses. These three types of governance mechanisms cover 17.3, 1.3 and 20% of the Peruvian Amazon’s area, respectively, and their impact on the region´s deforestation rate is quite remarkable. Identifying potential OECMs that are effective in preventing biodiversity loss, assessing how they perform in comparison to PAs and evaluating in which circumstances they can deliver biodiversity outcomes to complement those provided by PAs is fundamental for advancing the goal of encompassing other approaches to conservation beyond formally designated PAs ^9,28,29 and to achieve Target 3 of the Kunming-Montreal Global Biodiversity Framework ¹⁵. Our results clearly show that ILs and NTCs are governance mechanisms that achieve long term in situ conservation of biodiversity, thus meeting criterion six and seven required for their identification as OECMs ⁵⁸ and criterion B and C for their recognition and reporting ⁵⁷. These criteria specify that potential OECMs are expected to have a type of management that deliver long-term effective in situ conservation of biodiversity. The contribution of these governance mechanisms to regional and global conservation goals should be acknowledged and conservation funding should be channelled to maintain or increase their contributions to conservation as well as to have a better understanding of under which circumstances they can provide the best conservation outcomes.

The forest that is retained by any governance mechanism also stores large amounts of atmospheric carbon, which is key to climate regulation ⁷³. In our study we found that PAs avoided 88% of the carbon emissions expected in the area due to forest loss, while NTCs and ILs avoided 65% and 42% respectively. The estimation of the amount of carbon emissions avoided is a key measure to understand the role of these governance mechanisms in carbon cycles and regional climate regulation ^9,73. Additionally, robust estimates of the impact of different governance mechanisms on carbon retention is crucial for the acquisition of funding for carbon payments as it is a requirement for satisfying Monitoring, Evaluation and Reporting (MER) programs to access REDD + and other funding streams (e.g., the Global Environmental Facility and the World Bank) ^42,74.

Over the last two decades, ILs avoided 54% of the forest loss and 51% of the carbon emissions expected in those areas. These results show that ILs can provide important conservation benefits in terms of preventing forest loss and reducing carbon emissions. This aligns with previous research conducted at the Amazon level ^17,30,75 and globally ¹⁸. One fifth of the Peruvian Amazon is under Indigenous management, the highest proportion of all the governance mechanisms assessed. This in part explains why these territories had the highest absolute loss of forest. Due to the large extent of these territories and the potential to improve their effects on protection of forests, it is key to support Indigenous Communities in the development of financial mechanisms that allow them to receive income from the forest through activities that do not harm biodiversity. It is also important to support Indigenous Communities to further improve the governance of their communal land, especially in reference to activities potentially harmful to biodiversity, such as illegal mining and crops cultivation and deforestation ^76–78. As the influence that Indigenous land management has on the regional trends of deforestation and carbon emissions is substantial, it is crucial to support the claims by Indigenous Peoples in the Amazon and globally, who frequently advocate for more recognition on their contributions to conservation and active participation in environmental and forestry policy ^18,76. Acknowledging their contribution through stronger participation in environmental policy-making and through the provision of funding to support their manifold approaches to management is necessary. Also, the identification of the specific management practices that are effective to prevent unintended deforestation is crucial so that they can be scaled.

We found that LCs avoided 29% of the deforestation and 20% of the expected carbon emissions in the last twenty years. These concessions constitute a considerable proportion (8.1%) of the Peruvian Amazon area. While LCs do not meet all the criteria to be defined as OECMs, our results show that they could contribute to other conservation goals like target 10 of the Kunming-Montreal Global Biodiversity Framework ¹⁵, which aims to ensure the sustainable management of forestry, in particular through the sustainable use of biodiversity. While our results may be counterintuitive, as the main purpose of these concessions is to extract wood, they support findings from previous research showing that these concessions can have a positive impact on reducing forest loss, specially the certified ones ^30,79–81. The establishment of this type of land use could prevent widespread illegal logging activities due to an incentive to defend forest assets, and timber profits, by excluding other illegal and legal actors interested in extractive activities (i.e., logging, mining, land-grabbing) ^80,81. However, it is noteworthy that despite of this benefit, logging also has a negative impact on the local ecosystem, as the extraction of trees degrades the forest ^23,82.

Logging Concessions (LCs) provide economic benefits from the extraction of wood ^79,81. However, in these concessions there can be other economic activities such as agroforestry and tourism, which could increase the economic benefits these landscapes provide to the local population, with a comparatively low impact on the ecosystems than other land uses like large-scale agriculture or mining. By turning LCs into multifunctional Forest Concessions, where the management party develops not only wood extraction but also a range of other economic activities (i.e., tourism, conservation, research), there could be a more sustainable and diversified regional economy. However, verification and monitoring has to be enforced for these schemes to have benefits for people and nature. There is evidence that shows that the effectiveness of LCs in preventing forest loss is higher in areas with higher deforestation pressure ⁸⁰. More studies on LCs are needed to understand what factors and which management conditions foster positive impacts on reducing forest loss. Furthermore, we need a better understanding of how the extraction of wood affects the overall biodiversity and carbon density ^23,79,82.

Forest loss in the Peruvian Amazon has been increasing through time, especially from 2006 to 2016. It is crucial to monitor these changes in forest loss along the years as well as the abnormal increases in yearly deforestation, to understand which factors drive them, and to identify how they can be prevented or managed. The 2005 increase in deforestation in the Peruvian Amazon coincides with a particularly dry year, which had three times more fires than the previous year and twice more fires than the posterior one ⁸³. On the other hand, the 2020 increase overlaps with the implementation of confinement measures to reduce COVID-19 in the country, which reduced the capacity for enforcement and control by the authorities and permitted the appearance of new logging roads as well as increases in illegal logging and mining activities ⁸⁴.

Our results show that there are certain years, like 2005, with high forest loss that do not show congruent patterns in relation to carbon emissions, on the other hand there are years, like 2019 and 2021, with high carbon emissions in relation to the deforestation observed that year. Factors like the construction of roads may be related to these differences. The construction of the Interoceanic Highway which connects Brazil and Peru, made primary forest more accessible for selective logging and deforestation ^85,86. As primary forest is known to be more dense than secondary or disturbed forest ^87,88, these road construction events through dense forest can increase disproportionately the carbon emissions. The factors that influence the increase or decrease of carbon emissions, apart from the loss of forest, need to be better studied as these factors can have significant impacts for carbon circles and regional and global climate regulation.

The Kunming-Montreal Global Biodiversity Framework includes a target of 30% of land protected by 2030 and refers to OECMs as a complementary conservation approach to PAs ¹⁵. Our study is one of the first to provide robust evidence of long-term positive impacts of multiple types of governance mechanisms for the conservation of biodiversity. Some of them with the potential to be classified as OECMs. While our results show that PAs are the most effective measure preventing forest loss and carbon emissions, we also found that there are several other governance mechanisms that can deliver conservation benefits at different extents and with different management characteristics. To allow nations to more effectively achieve the different targets from the Kunming-Montreal Global Biodiversity Framework and the UN Framework Convention on Climate Change ³⁶, it will be crucial to (i) better understand how different types of governance mechanisms operate and under which circumstances they can provide the most beneficial conservation outcomes, and (ii) strengthen their management through funding mechanisms specific for each governance mechanism.

Author Contributions:

P.J.N., A.V., M.S. and J.G.Z. designed the research. P.J.N., S.N. and M.Q. processed the data. P.J.N. and V.R. performed the analysis. P.J.N. wrote the initial draft. P.J.N, V.R., S.N., M.Q., A.V., G.F., T.A., M.S., J.S. and J.G.Z. reviewed and edited the manuscript.

Acknowledgements:

We would like to acknowledge the people from the Asociación para la Conservación de la Cuenca Amazónica (ACCA) and from the Land Systems and Sustainability Transformations research team for helpful discussion on the interpretation of the results. This study contributes to the strategy and goals of the Global Land Programme (www.glp.earth).

Pillay, R. et al. Tropical forests are home to over half of the world’s vertebrate species. Frontiers in Ecology and the Environment 20, 10–15 (2022).
Lewis, S. L., Edwards, D. P. & Galbraith, D. Increasing human dominance of tropical forests. Science 349, 827–832 (2015).
Mori, A. S., Lertzman, K. P. & Gustafsson, L. Biodiversity and ecosystem services in forest ecosystems: a research agenda for applied forest ecology. Journal of Applied Ecology 54, 12–27 (2017).
Strand, J. et al. Spatially explicit valuation of the Brazilian Amazon Forest’s Ecosystem Services. Nature Sustainability 1, 657–664 (2018).
Maxwell, S. L. et al. Degradation and forgone removals increase the carbon impact of intact forest loss by 626%. Science Advances 5, (2019).
Albert, J. S. et al. Human impacts outpace natural processes in the Amazon. Science (2023) doi:10.1126/science.abo5003.
Lapola, D. M. et al. The drivers and impacts of Amazon forest degradation. Science 349, (2023).
Achard, F. et al. Pan-tropical monitoring of deforestation. Environ. Res. Lett. 2, 045022 (2007).
Maxwell, S. L. et al. Area-based conservation in the twenty-first century. Nature 586, 217–227 (2020).
Jones, K. R. et al. One-third of global protected land is under intense human pressure. Science 360, 788–791 (2018).
Grantham, H. S. et al. Anthropogenic modification of forests means only 40% of remaining forests have high ecosystem integrity. Nature Communications 1–10 (2020) doi:10.1038/s41467-020-19493-3.
Heino, M. et al. Forest loss in protected areas and intact forest landscapes: A global analysis. PLoS ONE 10, 1–21 (2015).
Schleicher, J. et al. Protecting half of the planet could directly affect over one billion people. Nature Sustainability 2, 1094–1096 (2019).
Negret, P. J., Venegas, R., Sonter, L. J., Possingham, H. P. & Maron, M. Conservation planning for retention, not just protection. Global Change Biology 30, e17211 (2024).
CBD. Kunming-Montreal Global Biodiversity Framework. https://www.cbd.int/doc/decisions/cop-15/cop-15-dec-04-en.pdf (2022).
Oldekop, J. A., Holmes, G., Harris, W. E. & Evans, K. L. A global assessment of the social and conservation outcomes of protected areas. Conservation Biology 30, 133–141 (2016).
Alejo, C. et al. Are indigenous territories effective natural climate solutions? A neotropical analysis using matching methods and geographic discontinuity designs. PLoS ONE 16, (2021).
Sze, J. S., Carrasco, L. R., Childs, D. & Edwards, D. P. Reduced deforestation and degradation in Indigenous Lands pan-tropically. Nature Sustainability 5, 123–130 (2022).
Vélez, M. A., Robalino, J., Cardenas, J. C., Paz, A. & Pacay, E. Is collective titling enough to protect forests? Evidence from Afro-descendant communities in the Colombian Pacific region. World Development 128, 2000–2005 (2020).
Corrigan, C., J. Robinson, C., Burgess, N. D., Kingston, N. & Hockings, M. Global Review of Social Indicators used in Protected Area Management Evaluation. Conservation Letters 11, 1–9 (2018).
Tegegne, Y. T., Cramm, M., Van Brusselen, J. & Linhares-Juvenal, T. Forest concessions and the United Nations sustainable development goals: Potentials, challenges and ways forward. Forests 10, 1–21 (2019).
Azevedo Ramos, C., Macedo Silva, J. N. & Merry, F. The evolution of Brazilian forest concessions. Elementa 1–8 (2015) doi:10.12952/journal.elementa.000048.
Clark, C. J., Poulsen, J. R., Malonga, R. & Elkan, P. W. Logging concessions can extend the conservation estate for central African tropical forests. Conservation Biology 23, 1281–1293 (2009).
Karsenty, A. & Ferron, C. Recent evolutions of forest concessions status and dynamics in Central Africa. International Forestry Review 19, 10–26 (2017).
Nunes, F. et al. Economic benefits of forest conservation: Assessing the potential rents from Brazil nut concessions in Madre de Dios, Peru, to channel REDD+ investments. Environmental Conservation 39, 132–143 (2012).
Schleicher, J. The environmental and social impacts of protected areas and conservation concessions in South America. Current Opinion in Environmental Sustainability 32, 1–8 (2018).
Radachowsky, J., Ramos, V. H., McNab, R., Baur, E. H. & Kazakov, N. Forest concessions in the Maya Biosphere Reserve, Guatemala: A decade later. Forest Ecology and Management 268, 18–28 (2012).
Donald, P. F. et al. The prevalence, characteristics and effectiveness of Aichi Target 11′s “other effective area‐based conservation measures” (OECMs) in Key Biodiversity Areas. CONSERVATION LETTERS 12, (2019).
Dudley, N. et al. The essential role of other effective area-based conservation measures in achieving big bold conservation targets. Global Ecology and Conservation 15, e00424 (2018).
Schleicher, J., Peres, C. A., Amano, T., Llactayo, W. & Leader-Williams, N. Conservation performance of different conservation governance regimes in the Peruvian Amazon. Scientific Reports 7, 1–10 (2017).
Móstiga, M., Armenteras, D., Vayreda, J. & Retana, J. Two decades of accelerated deforestation in Peruvian forests: a national and regional analysis (2000–2020). Reg Environ Change 24, 42 (2024).
Joppa, L. & Alexander, P. Reassessing the forest impacts of protection. Annals of the New York Academy of Sciences 1185, 135–149 (2010).
Joppa, L. N. & Pfaff, A. Global protected area impacts. Proceedings of the Royal Society B: Biological Sciences 278, 1633–1638 (2011).
Joppa, L. N. & Pfaff, A. High and far: Biases in the location of protected areas. PLoS ONE 4, 1–6 (2009).
Pressey, R. L., Cabeza, M., Watts, M. E., Cowling, R. M. & Wilson, K. A. Conservation planning in a changing world. Trends in Ecology and Evolution 22, 583–592 (2007).
UNFCCC. Adoption of the Paris Agreement. Proposal by the President Draft Decision -/CP.21. (2015).
Mittermeier, R. A. & Mittermeier, C. G. Megadiversity: Earth’s Biologically Wealthiest Nations. (1997).
Galarza, Elsa; La Serna, K. Las concesiones forestales en el Perú: ¿cómo hacerlas sostenibles? La Politica forestal en la Amazonía andina. Estudio de casos:Bolivia, Ecuador y Perú 445–600 (2005).
Miranda, J. J., Corral, L., Blackman, A., Asner, G. & Lima, E. Effects of Protected Areas on Forest Cover Change and Local Communities: Evidence from the Peruvian Amazon. World Development 78, 288–307 (2016).
Potapov, P. V. et al. National satellite-based humid tropical forest change assessment in Peru in support of REDD+ implementation. Environmental Research Letters 9, 2000–2010 (2014).
Geldmann, J., Manica, A., Burgess, N. D., Coad, L. & Balmford, A. A global-level assessment of the effectiveness of protected areas at resisting anthropogenic pressures. Proceedings of the National Academy of Sciences of the United States of America 1–7 (2019) doi:10.1073/pnas.1908221116.
Graham, V. et al. Southeast Asian protected areas are effective in conserving forest cover and forest carbon stocks compared to unprotected areas. Scientific Reports 11, 1–12 (2021).
Negret, P. J. et al. Spatial autocorrelation and sampling design affect estimates of protected area effectiveness. Conservation Biology 34, 1452–1462 (2020).
Andam, K. S., Ferraro, P. J., Pfaff, A., Sanchez-Azofeifa, G. A. & Robalino, J. A. Measuring the effectiveness of protected area networks in reducing deforestation - Supporting Information. Pnas 5–7 (2008) doi:10.1073/pnas.0800437105.
Morgan, S. L. & Winship, C. Counterfactuals and Causal Inference: Methods and Principles for Social Research. (Cambridge University Press, 2015).
Blackman, A., Corral, L., Lima, E. S. & Asner, G. P. Titling indigenous communities protects forests in the Peruvian Amazon. Proceedings of the National Academy of Sciences of the United States of America 114, 4123–4128 (2017).
Negret, P. J. et al. Emerging evidence that armed conflict and coca cultivation influence deforestation patterns. Biological Conservation 108176 (2019) doi:10.1016/j.biocon.2019.07.021.
Sonter, L. J., Moran, C. J., Barrett, D. J. & Soares-Filho, B. S. Processes of land use change in mining regions. Journal of Cleaner Production 84, 494–501 (2014).
Sonter, L. J., Ali, S. H. & Watson, J. E. M. Mining and biodiversity: Key issues and research needs in conservation science. Proceedings of the Royal Society B: Biological Sciences 285, (2018).
MINAM. Propuesta Tecnica Del Limite de La Amazonia. (2016).
Peruvian Ministry of the Environment. National Service of Natural Areas Protected by the State. SERNAP https://www.gob.pe/institucion/sernanp/organizacion (2023).
MINCUL. Ministerio de Cultura. https://www.gob.pe/cultura (2023).
MINCUL. Base de datos de pueblos indigenas y originarios. https://bdpi.cultura.gob.pe/buscador-de-localidades-de-pueblos-indigenas (2023).
SERFOR. Servicio Nacional Forestal y de Fauna Silvestre. https://www.gob.pe/serfor (2023).
OSINFOR. Organismo de Supervisión de los Recursos Forestales y de Fauna Silvestre. https://www.gob.pe/osinfor (2023).
MINEM. GEOCATMIN. https://geocatmin.ingemmet.gob.pe/geocatmin/ (2023).
IUCN WCPA Task Force on OECMs. Recognising and Reporting Other Effective Area-Based Conservation Measures. (IUCN, International Union for Conservation of Nature, 2019). doi:10.2305/IUCN.CH.2019.PATRS.3.en.
Jonas, H. D., MacKinnon, K., Marnewick, D. & Wood, P. Site-Level Tool for Identifying Other Effective Area-Based Conservation Measures (OECMs). (IUCN, International Union for Conservation of Nature, 2023). doi:10.2305/WZJH1425.
MINAM & MIDAGRI. Bosque – No Bosque y Pérdida de Bosque Húmedo Amazónico. (2021).
MINAM. Protocolo Metodológico Para La Detección de La Pérdida de Bosque. (2021).
Asner, G. P. et al. Targeted carbon conservation at national scales with high-resolution monitoring. Proceedings of the National Academy of Sciences of the United States of America 111, E5016–E5022 (2014).
Móstiga, M., Armenteras, D., Vayreda, J. & Retana, J. Decoding the drivers and effects of deforestation in Peru: a national and regional analysis. Environ Dev Sustain (2024) doi:10.1007/s10668-024-04638-x.
Ford, S. A. et al. Deforestation leakage undermines conservation value of tropical and subtropical forest protected areas. Global Ecology and Biogeography 29, 2014–2024 (2020).
Zhao, H. et al. Individual-level performance of nature reserves in forest protection and the effects of management level and establishment age. Biological Conservation 233, 23–30 (2019).
Schleicher, J. et al. Statistical matching for conservation science. Conservation Biology 00, 1–12 (2019).
Dormann, C. F. et al. Collinearity: a review of methods to deal with it and a simulation study evaluating their performance. Ecography 36, 27–46 (2013).
Lumley, Thomas. Package ‘leaps’: Regression Subset Selection. (2022).
R Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing. Preprint at (2022).
Ho, D., Imai, K., King, G., Stuart, E. & Alex, W. MatchIt: Nonparametric Preprocessing for Parametric Causal Inference. Preprint at (2018).
Blackman, A., Pfaff, A. & Robalino, J. Paper park performance: Mexico’s natural protected areas in the 1990s. Global Environmental Change 31, 50–61 (2015).
Olmos, A. & Govindasamy, P. Propensity Scores: A Practical Introduction Using R. Journal of MultiDisciplinary Evaluation 11, 1556–8180 (2015).
Aguiar, A. P. D. et al. Land use change emission scenarios: anticipating a forest transition process in the Brazilian Amazon. Global Change Biology 22, 1821–1840 (2016).
Duncanson, L. et al. The effectiveness of global protected areas for climate change mitigation. Nat Commun 14, 2908 (2023).
West, T. A. P. et al. Action Needed to Make Carbon Offsets from Tropical Forest Conservation Work for Climate Change Mitigation Thales. (2023).
Duarte, D. P. & Walker, B. Reducing natural vegetation loss in Amazonia critically depends on the formal recognition of indigenous territories. 279, (2023).
Camino, M. et al. Indigenous Lands with secure land-tenure can reduce forest-loss in deforestation hotspots. Global Environmental Change 81, 102678 (2023).
Dos Santos, A. M., Da Silva, C. F. A., De Melo, S. N., De Almeida Junior, P. M. & Bueno, L. F. Influence of deforestation inside and outside indigenous lands in the Brazilian Amazon Biome. Reg Environ Change 22, 77 (2022).
Siqueira-Gay, J., Soares-Filho, B., Sanchez, L. E., Oviedo, A. & Sonter, L. J. Proposed Legislation to Mine Brazil’s Indigenous Lands Will Threaten Amazon Forests and Their Valuable Ecosystem Services. One Earth 3, 356–362 (2020).
Gaveau, D. L. A. et al. Reconciling Forest Conservation and Logging in Indonesian Borneo. PLoS ONE 8, e69887 (2013).
Panlasigui, S., Rico-Straffon, J., Pfaff, A., Swenson, J. & Loucks, C. Impacts of certification, uncertified concessions, and protected areas on forest loss in Cameroon, 2000 to 2013. Biological Conservation 227, 160–166 (2018).
Rico-Straffon, J. et al. Forest concessions and eco-certifications in the Peruvian Amazon: Deforestation impacts of logging rights and logging restrictions. Journal of Environmental Economics and Management 118, 102780 (2023).
Pearson, T. R. H., Brown, S. & Casarim, F. M. Carbon emissions from tropical forest degradation caused by logging. Environ. Res. Lett. 9, 034017 (2014).
World Research Institute. Global forest watch. https://www.globalforestwatch.org/dashboards/country/PER (2023).
Céspedes, J. et al. Has global deforestation accelerated due to the COVID-19 pandemic? J. For. Res. 34, 1153–1165 (2023).
Perz, S. G. et al. Trans-boundary infrastructure and land cover change: Highway paving and community-level deforestation in a tri-national frontier in the Amazon. Land Use Policy 34, 27–41 (2013).
Song, X.-P., Huang, C., Saatchi, S. S., Hansen, M. C. & Townshend, J. R. Annual Carbon Emissions from Deforestation in the Amazon Basin between 2000 and 2010. PLoS ONE 10, e0126754 (2015).
Longo, M. et al. Aboveground biomass variability across intact and degraded forests in the Brazilian Amazon. Global Biogeochemical Cycles 30, 1639–1660 (2016).
Watson, J. E. M. et al. The exceptional value of intact forest ecosystems. Nature Ecology and Evolution2, 599–610 (2018).

Table 1

Variables used for the matching analysis.
Variable Description	Variable Name	Year	Resolution	Source
Annual Precipitation	Annual_Prec	historical	30 m	https://www.worldclim.org/
Precipitation Seasonality	Prec_Seas	historical	30 m	https://www.worldclim.org/
Distance to nearest previous deforested area	Dis_Def	2000	30 m	Ministerio del Ambiente (MINAM) & Ministerio de Agricultura y Riego (MIDAGRI)
Distance to nearest navigable river	Dis_Rivers	2021	30 m	Sistema Nacional de Información de Recursos Hídricos (SNIRH) from the Autoridad Nacional del Agua (ANA)
Distance to nearest district road	District_R	2018	30 m	Ministerio de Transportes y Comunicaciones (MTC)
Distance to nearest department road	Departme_R	2018	30 m	Ministerio de Transportes y Comunicaciones (MTC)
Distance to nearest national road	National_R	2018	30 m	Ministerio de Transportes y Comunicaciones (MTC)
Distance to nearest settlement of more than 10 people	D7Set10	2007	30 m	Instituto Nacional de Estadística e Informática (INEI)
Distance to nearest settlement of more than 1000 people	D7Set1000	2007	30 m	Instituto Nacional de Estadística e Informática (INEI)
Distance to nearest settlement of more than 5000 people	D7Set5000	2007	30 m	Instituto Nacional de Estadística e Informática (INEI)
Distance to nearest settlement of more than 10000 people	D7Set10000	2007	30 m	Instituto Nacional de Estadística e Informática (INEI)
Distance to nearest settlement of more than 10 people	D17Set10	2017	30 m	Instituto Nacional de Estadística e Informática (INEI)
Distance to nearest settlement of more than 1000 people	D17Set1000	2017	30 m	Instituto Nacional de Estadística e Informática (INEI)
Distance to nearest settlement of more than 5000 people	D17Set5000	2017	30 m	Instituto Nacional de Estadística e Informática (INEI)
Distance to nearest settlement of more than 10000 people	D17Se10000	2017	30 m	Instituto Nacional de Estadística e Informática (INEI)
Ecoregions	Ecoregions	-	-	Ministerio del Ambiente (MINAM)
Departments of the Peruvian Amazon	Department	-	-	Plataforma Nacional de Datos Abiertos
Elevation	Elevation	-	30 m	Nasa shuttle radar topography mission
Slope	Slope	-	30 m	Nasa shuttle radar topography mission
Population density	Pop2000	2000	1 km	https://www.worldpop.org/
Population density	Pop2020	2020	1 km	https://www.worldpop.org/
Travel time to major cities	Tra_Time00	2000	1 km	https://forobs.jrc.ec.europa.eu/gam
Travel time to major cities	Tra_Time15	2015	1 km	https://www.nature.com/articles/s41597-019-0265-5

There is NO Competing Interest.

SupplementaryInformation.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Potential of different governance mechanisms for achieving Global Biodiversity Framework goals

Status:

Version 1

Abstract

Figures

Introduction

Methods

Results

Discussion

Declarations

Author Contributions:

Acknowledgements:

References

Tables

Additional Declarations

Supplementary Files

Status:

Version 1