Contact Us  |  Search  
The National Academies of Sciences, Engineering and Medicine
Partnerships for Enhanced Engagement in Research
Development, Security, and Cooperation
Policy and Global Affairs
Home About Us For Applicants For Grant Recipients Funded Projects Email Updates
Cycle 5 (2015 Deadline)

Degradation of tropical forests in Colombia: impacts of fire

PI: Dolors Armenteras (, Universidad Nacional de Colombia
U.S. Partner: Jennifer K. Balch, University of Colorado, Boulder
Project Dates: December 2016 - November 2019

Project Overview:

5-331 Fires occurring in the Savannah forest in the Bita region
Fires occurring in the Savannah forest in the Bita region [Photo courtesy of Dr. Armenteras]
Tropical forests host the highest levels of biodiversity and maintain some of the greatest carbon stocks of all terrestrial ecosystems, having an essential role in global carbon (C) cycling. Colombia is no exception and hosts a diversity of tropical forests that are rich in both C and species, but also highly threatened. The international community has established mechanisms such as REDD+ (Reduced Emissions from Deforestation and Forest Degradation plus enhancement of forest carbon stocks through conservation and sustainable forest management) as a way to combat global change. Colombia is one of the REDD+ countries that has made advances in quantifying C stocks and measuring emissions from deforestation but has yet to address the drivers and consequences of forest degradation. Recent studies have identified fire as a major driver of tropical forest degradation. Fires affect landscape structure, patterns, and processes and have shaped today´s species composition and biological diversity. However, fires have also been used more recently as a tool in management practices for land clearance, crop or pasture maintenance, and slash-and burn agriculture in tropical forest systems. Though a natural mechanism in many ecosystems, human actions have altered regimes and the extent and frequency of fires have increased in many regions of the world, particularly in tropical forests. This study will be the first in Colombia to address the impact of fire on the composition of a given forest, or how fire is partially controlled by the spatio-temporal distribution of available resources to burn (fuel amount) and climatic conditions promoting combustion (fuel moisture).

This project will provide a basis of knowledge that will help to reduce threats to agriculture systems, which are becoming more pronounced in a climate change context and causing major economic losses. Such is the case of fire dynamics, which particularly can affect vulnerable crops that are composed of monocultures of highly flammable plant species (e.g., oil palm plantations). By identifying fire drivers it is possible to design safer plantations, including elements and tools that reduce fire risk and dispersal. The project should also promote low-carbon economic growth through increased investments in low-emissions development, as well as improved community resilience to changing weather patterns and protection of significant ecosystems. By developing reliable estimations of carbon emissions from forest degradation and calibrating models combining field data and fuel moisture satellite observation data, the researchers will promote joint efforts to strengthen the capacity of regional stakeholders to integrate scientific products into development and decision making for the Colombian National REDD+ policy.

Summary of Recent Activities:

5-331 Q1 2017a
 Photo courtesy of Dr. Armenteras
Major activities during 2017 have been: (1) assessment of how fuel load and fuel moisture change along forest types, and (2) advancement of understanding of the interactions between fire frequency and intensity and resulting biomass depletion and C stocks in different types of forests affected by single and repeated fires have been the main focus during 2017.
The first study  site was selected at the Bojonawi Reserve (Vichada, Colombia) at the Orinoco Basin. The site in the Andean forests is still being determined, and is likely to be around the Iguaque Sanctuary, Boyaca district.
Building and integration of databases comprising land uses, fire dynamics, biodiversity continues. The team collected existing data on biodiversity of the area and has the species sampling determined. For the Andes site, the team has begun a preliminary study of fire dynamics with MODIS to help with the selection of the potential sites for field plots, a selection that is under process and which will help filter out sites and select an area of study for the Andes site.
Establishment of plots and field sampling of fuel, biomass, and forest composition is ongoing. Fire regimes and severities have been characterized by forest type with both field-based surveys and remote-sensing analysis. Spatial patterns of past fires were used to model fire hazard at a landscape level and to select potential sites for field plots. After performing a field verification of fire frequencies and disturb recognition, three forest types were selected according to the following criteria: 1) unburned areas (at least over 10 years since the last burn), 2) recent fires (occurred in the last year) and 3) repeated fires (occurred twice in the previous 6 years). In each forest type, three 0.1 ha plots (100 x 10m) were established from the forest edge towards the forest interior, maintaining at least 400 m of distance between the plots. Adult plants with stems ≥ 10 cm in diameter at breast height (DBH) were sampled in each   plot, and structural variables, such as abundance, frequency, basal area, coverage, and density were obtained for each sample. Height was measured with a TruPulse 200 L Hipsometer from a point with good visibility of the tree crown, and an approximate distance from the tree between 15 and 25 m (Stevenson 2004). Saplings (DBH <10 cm, Total Height > 1,5 m) and seedlings (DBH <10 cm, Total Height > 30 cm at 1.5 m) were sampled in nested subplots (5x5 m and 2x2 m, respectively) located each 10 m along the main plot (Figure 1), where the diameter and total height of each individual were measured. To study plant regeneration, sprouts (DBH <10 cm, HT <30 cm) were counted in nested 1 m2 plots (Figure 1). Adult plants were marked with aluminum numbered tags at 1.3m. 
At the end of the reporting period, all  collected botanical material has been processed at the Botanical Laboratorio of the Instituto de Ciencias Naturales of the Universidad Nacional de Colombia. Collections were identified based on comparison with specimens of the National Herbarium of Colombia (COL) and the Herbario Amazónico Colombiano (COAH), as well as with collections of virtual herbariums: Jstore Global Plants, New York Botanical Garden and Field Museum (Neotropical Herbarium Specimens). Botanical experts were also consulted for identification of complex families, such as Rubiaceae, and for identification of infertile material. A data base with structural and compositional categories was organized, including information on botanical identification, diameter at breast height (DBH), height, sprouts, wood specific gravity and coordinates, for each plant individual sampled. Two independent vegetation data bases were created for adult plants and natural regeneration (seedlings, saplings and sprouts).
Functional traits were analyzed in the laboratory, including foliar area, specific foliar area, dry matter leaf content, density of stomata and trichomes, wood density, vessels grouping, diameter, and vessels density. In addition, soil samples were analyzed in the Waters and Soils Lab from the Universidad Nacional de Colombia. Soil analyses included total carbon and nitrogen -using the DUMAS method- and organic carbon using the method of Walkley-Black digestion with colorimetric quantification. A data base for soil samples was created, including the results of total nitrogen, total carbon, organic and inorganic carbon. Subsamples of all the fuel loads were collected and weighted after collection in the field. Later the subsamples were stored in zip-close bags and brought back to the Ecology laboratory of the Universidad Nacional de Colombia, to be dried in an oven at 100ºC. The 1-, 10-, 100- hour fuels, litter and duff were dried during 24 hours. Larger fuels were weighted each 24-hours until they reached equilibrium and a final mass for each piece was obtained. A data base for fuel load was created, were each fuel class contains dimensions of the collected pieces, wet and dry mass and volume. Data bases for duff/litter profile and dead standing trees were also created. Fuel load was assessed using the Fuel Load method (FL) proposed in the FIREMON system of the Department of Agriculture, Forest Service, Rocky Mountain Research Station, developed for the monitoring and inventory of the effects of wildland fire (Lutes 2006). FL was used to sample dead and down woody debris (DWD), determine depth of the duff/litter profile and estimate the proportion of litter in the profile in the areas with different fire frequency.
Presentations at the USAID local program, IDEAM, RiquezaNatural (Usaid funded bid for Colombia), UK Universities, and local NGOs were conducted and a website featuring the project was launched and is currently being updated: featuring the project. 
Potential developmental impacts 
The results of this project will address the impact of fire composition of different types of forest and demonstrate how fire is controlled by the spatio-temporal distribution of available resources to burn and climatic conditions. These findings are anticipated to be helpful in strengthening environmental resiliency and low emissions development, as uncontrolled fires, such as the ones taking place in the Orinoquia Region, are depleting forest C stocks and increasing CO2 emissions. This project is anticipated to determine which type of forest are less resilient to fire impacts, in order to prioritize regional management actions to control fires.   
Peer reviewed publications and proceedings
Armenteras, D., Barreto, J. S., Tabor, K., Molowny, R. and Retana, J. (2017). Changing patterns of fire occurrence in proximity to forest edges, roads and rivers between NW Amazonian countries. Biogeosciences 14, 2755–2765 
Armenteras, D., Gibbes, C., Anaya, L, Dávalos, L. (2017). Integrating remotely sensed fires for predicting deforestation for REDD+. Ecological Applicattions 27(4), 1294–1304 
Andrade, R. B., J. K. Balch, A. L. Parsons, D. Armenteras, R. M. Roman-Cuesta, and J. Bulkan. (2017). Scenarios in tropical forest degradation: carbon stock trajectories for REDD+. Carbon Balance and Management 12:6.

5-331 Q1 2017c 5-331 Q1 2017b  

Back to PEER Cycle 5 Grant Recipients