Fires inside Colombia’s National Parks

Linking deforestation with conflict is not new, but the end of armed conflict presents unique research opportunities. Colombia has experienced one of the longest armed conflicts in modern times and is therefore a potential laboratory for understanding how forests fare in war and peace.


Back in 2017, prior to the implementation of the Colombian peace treaty, there was much speculation on the potential positive and negative effects of the Peace on the environment. In our new paper, we present remotely sensed data on indirect impacts on protected areas in the year following the implementation of the peace accords. By uncovering a significant increase in fire incidence in guerilla-controlled areas in the year immediately following the treaty’s implementation (March 2017-March 2018), we provide the first-ever evidence that peace and decisions associated with land management likely increased the fire incidence in protected areas in regions formerly under FARC control. 

Although remote sensing for the last year is beginning to become available, official sources only report losses up to 2017. Open-access sources have to be processed and stitched together, and the combination of tropical weather and steep Colombian mountain ranges creates vast cloud gaps for urgent, short-term forest cover analyses. For these reasons, and to compare against the time immediately preceding the implementation of the peace accords, we focused on a proxy for deforestation: fires.   

In general, climate variation can influence the incidence of fires in tropical areas, but climate is only one part of the story. In western Amazonia, our earlier work had already revealed a tight link between fires in one year and forest clearings, so much so, that a model could be built that predicted the location of deforestation with great precision! This makes sense when considering how much easier and less expensive it is to burn forests during the dry season than to use only chainsaws to clear the land without burning. Those fires also burn the forest edges, making it easier to cut later on. Hence, we used fires as a proxy for forest loss because these data are much more easily detected with daily alerts than the satellite imagery collected less frequently that needs processing to detect new clearings. Thus, we were able to both compare the incidence of fires in areas under different territorial control, as well as estimate the probability of forest clearing for the three most affected protected areas: Picachos, Macarena, and Tinigua.

The results were astounding: relative fire incidence grew the most inside the protected areas where the westernmost Amazon meets the Andes, much more than outside the parks or even in the buffer zones. Intrigued, we modeled the probability of deforestation based on the fires, finding an average increase of 50% in the probability of deforestation. The models, however, could be wrong. Therefore, we also collected satellite imagery during the dry season, to determine if the spike of fires corresponded to a spike of forest clearing in the three parks. This time, we found a 69% increase in forest clearing within the parks, corresponding to the fire hotspots we had originally identified.

Not everything is bad news, though. The spike in fires and subsequent forest loss requires immediate action, but actions need to be targeted. Our study shows remote sensing to detect fires is quite promising for monitoring protected areas and frontier deforestation following the end of civil wars. Recent years have seen massive international investment in conservation in Colombia often linked to either declaring protected areas or expanding them. Our analyses highlight the equally urgent need to build the scientific infrastructure to prevent further deforestation in those protected areas. The approaches and methods we have pioneered offer multiple avenues for doing just that. This study, for instance, received no funding from recent conservation investments in Colombia, but if properly scaled could enhance those efforts at relatively low cost. We hope this paper serves as a template for future data-driven analyses of conservation policies and investments. 

Fire in the Colombian Amazon (Copyright Uriel Murcia, Instituto Sinchi).

Our paper can be found here

Go to the profile of Dolors Armenteras-Pascual

Dolors Armenteras-Pascual

Associate Professor, Universidad Nacional de Colombia

No comments yet.