Untitled

Understanding Deforestation Dynamics in the Amazonian Protected Areas

Authors and Affiliations

  • Katherine J. Siegel

    • Cooperative Institute for Research in Environmental Sciences, University of Colorado, Boulder, CO, USA

    • Geography Department, University of Colorado, Boulder, CO, USA

    • Environmental Data Science Innovation & Inclusion Lab, University of Colorado, Boulder, CO, USA

  • Megan Mills-Novoa

    • Department of Environmental Science, Policy, & Management, University of California, Berkeley, California, USA

    • Energy & Resources Group, University of California, Berkeley, California, USA

  • Eva Kinnebrew

    • Department of Ecosystem Science & Sustainability, Colorado State University, Fort Collins, CO, USA

  • José Ochoa-Brito

    • Landscape Analytics Team, Revalue Nature, London, UK

    • Geography Graduate Group, University of California, Davis, California, USA

  • Elizabeth Shoffner

    • Department of Geography, Dartmouth College, Hanover, NH, USA

Keywords

  • Land-use change

  • Forest loss

  • Tropical forest

  • Amazon Basin

  • Protected areas

  • Conservation

Abstract

  • Deforestation in agricultural frontiers of the Amazon, particularly in and around protected areas (PAs), necessitates a comprehensive understanding of socio-environmental factors that contribute to forest conversion. This study integrates quantitative and qualitative analyses to refine land-use change (LUC) models across three Amazonian PA complexes:

    1. Jamanxim National Forest (Brazil)

    2. Amboró National Park and Carrasco National Park (Bolivia)

    3. Tambopata National Reserve and Bahuaja-Sonene National Park (Peru)

Goals of the Research

  1. Analyze the effectiveness of conservation discourses in informing deforestation models.

  2. Identify socio-environmental factors related to deforestation at region-wide and specific PA levels.

  3. Validate the integrative methodology’s model performance against conventional LUC models.

Key Findings

  • From 2008 to 2018:

    • Forests on steeper slopes and with higher population densities showed lower probability of conversion to agriculture.

    • Proximity to agriculture and occurrence of fires correlated positively with increased deforestation rates.

    • Legal threats to Jamanxim's conservation status amplified deforestation risks.

    • Payments for ecosystem services (PES) were associated with reduced deforestation in Amboró and Carrasco.

    • Certain commonly cited factors in deforestation discourses were not upheld by quantitative data (e.g., effectiveness of REDD+ projects).

1. Introduction

  • Deforestation Impact: Conversion of forests to agriculture impacts Indigenous communities, biodiversity, and ecosystem services.

  • Agricultural Frontiers Definition: Areas undergoing active land conversion for agriculture or livestock production.

  • Policy Interventions: Certain interventions successfully mitigated forest clearing, yet conflicts with conservation priorities (e.g., PA overlaps with agricultural expansion).

  • Protected Areas: As of 2022, around 25% of the Amazon is under protection with inconsistent effects on deforestation rates.

  • Environmental Factors: Deforestation dynamics are determined by local socio-environmental and policy contexts.

2. Methods

2.1 Study Sites

  • Characteristics of Study Sites:

    • Size (12,962–13,661 km²)

    • Deforestation Rates (2008-2018): Tambopata & Bahuaja-Sonene (<1%) vs. Jamanxim (3.6%)

  • Individual Site Descriptions:

    • Jamanxim National Forest: Established in 2006, facing deforestation primarily from ranching, agriculture, and land speculation.

    • Amboró and Carrasco National Parks: Experienced deforestation due to small-scale agriculture and ranching, challenged by coca cultivation and supposed infrastructure projects.

    • Tambopata National Reserve and Bahuaja-Sonene National Park: Facing deforestation driven by agricultural expansion and informal gold mining from 2006 onwards.

2.2 Land-use Change Maps

  • Utilized land cover maps from 2008 and 2018 derived from Landsat missions with overall accuracy rates greater than 90%.

2.3 Land-use Change Modeling

  • Logistic Regression: Selected for its interpretability and effectiveness in modeling probability of land conversion.

  • Integratory Methodology Framework: Melding qualitative and quantitative data to derive model variables. Utilized prior LUC literature and qualitative discourse analysis of local factors affecting deforestation.

2.4 Model Comparisons

  • Model performance assessed using Akaike information criterion (AIC) and ANOVA, as well as McFadden's adjusted pseudo R2.

  • Generated predicted landscapes for 2018 using Monte Carlo simulations to model potential forest conversion.

3. Results

3.1 Model Performance

  • Models that included both LUC and discourse analysis variables demonstrated superior explanatory power across all case study sites.

3.2 Site-Level Factors

  • Jamanxim: Lower conversion probability exhibited with steeper slopes and distances from agricultural activities.

  • Amboró-Carrasco: Similar trends as Jamanxim with notable dependence on PES program enrollment.

  • Tambopata-Bahuaja-Sonene: Deforestation linked to proximity to agriculture, with variations observed in the relationship with unauthorized mining.

3.3 Cross-Site Comparisons

  • Commonalities across case studies: Increased deforestation likelihood linked to flatter terrain, higher fire density, and proximity to agricultural activities.

4. Discussion

4.1 Conservation Insights

  • Findings indicate significant variances and need for tailored conservation efforts considering the nuanced behaviors of land-use variables.

4.2 Examining Conservation Discourses

  • Conservation narratives often perpetuate misconceptions like the association of poverty with deforestation.

4.3 Limitations

  • Themes from discourse analysis that did not translate into quantitative proxies limited comprehensive modeling, highlighting gaps in data availability.

4.4 Conclusions

  • Integrating qualitative and quantitative approaches in land-use modeling can illuminate the multi-faceted nature of drivers influencing deforestation in complex environments like the Amazon Basin.

References

  • Inclusion of multiple studies and literature that informed the analysis and methodology, underscoring contributions to understanding deforestation dynamics and effective conservation techniques.