TY - JOUR

T1 - What Peace Means for Deforestation

T2 - An Analysis of Local Deforestation Dynamics in Times of Conflict and Peace in Colombia

AU - Ganzenmüller, Raphael

AU - Sylvester, Janelle M.

AU - Castro-Nunez, Augusto

N1 - Publisher Copyright: Copyright © 2022 Ganzenmüller, Sylvester and Castro-Nunez.

PY - 2022

Y1 - 2022

N2 - Using Colombia as a case study, this analysis provides insights on deforestation dynamics in times of conflict and peace and the different factors driving these dynamics. We performed time series clustering of yearly deforestation data (2001–2018) from 708 out of 1,122 mainland Colombian municipalities (accounting for 98% of the total deforestation areas in Colombia) and produced regression models using a gradient tree boosting framework (XGBoost) to identify drivers that explain varying, local-level deforestation dynamics. Municipalities were characterized by seven categories of deforestation dynamics, with the Amazon region being largely represented by only four categories and the Andes region displaying all categories of deforestation dynamics. Notably, six of the seven representative categories exhibit substantial increases in deforestation in the years following the peace agreement. The regression analysis revealed that coca cultivation area, number of cattle, and municipality area are the top three drivers of deforestation dynamics at national, regional, and category levels. However, the importance of the different variables varied according to the different spatial dimensions. Results provide further understanding on how the drivers of deforestation change not only at a regional scale, as assumed by much of the current literature about drivers of deforestation, but also at a lower scale of analysis (intraregional and intradepartmental variation in the case of Colombia). Insights from this study can be used to understand deforestation dynamics in other countries experiencing times of conflict and peace and will support decision-makers in creating programs that align actions for peacebuilding, climate change mitigation, and biodiversity conservation more effectively.

AB - Using Colombia as a case study, this analysis provides insights on deforestation dynamics in times of conflict and peace and the different factors driving these dynamics. We performed time series clustering of yearly deforestation data (2001–2018) from 708 out of 1,122 mainland Colombian municipalities (accounting for 98% of the total deforestation areas in Colombia) and produced regression models using a gradient tree boosting framework (XGBoost) to identify drivers that explain varying, local-level deforestation dynamics. Municipalities were characterized by seven categories of deforestation dynamics, with the Amazon region being largely represented by only four categories and the Andes region displaying all categories of deforestation dynamics. Notably, six of the seven representative categories exhibit substantial increases in deforestation in the years following the peace agreement. The regression analysis revealed that coca cultivation area, number of cattle, and municipality area are the top three drivers of deforestation dynamics at national, regional, and category levels. However, the importance of the different variables varied according to the different spatial dimensions. Results provide further understanding on how the drivers of deforestation change not only at a regional scale, as assumed by much of the current literature about drivers of deforestation, but also at a lower scale of analysis (intraregional and intradepartmental variation in the case of Colombia). Insights from this study can be used to understand deforestation dynamics in other countries experiencing times of conflict and peace and will support decision-makers in creating programs that align actions for peacebuilding, climate change mitigation, and biodiversity conservation more effectively.

KW - climate change

KW - Colombia

KW - conflict

KW - deforestation

KW - land grabbing

KW - peace agreement

KW - time series clustering

KW - XGBoost

U2 - 10.3389/fenvs.2022.803368

DO - 10.3389/fenvs.2022.803368

M3 - Journal article

AN - SCOPUS:85125830589

VL - 10

JO - Frontiers in Environmental Science

JF - Frontiers in Environmental Science

SN - 2296-665X

M1 - 803368

ER -