Contribution of ‘human induced fires’ to forest and savanna land conversion dynamics in the Luki Biosphere Reserve landscape, western Democratic Republic of Congo

ABSTRACT

Human-induced fire is one of the most important determinants of forest cover and change in tropical and subtropical regions of the world. Yet its impact on forest cover and forest cover change remains unclear, as fires in Africa generally do not spread over very large area. This is particularly the case in the Democratic Republic of Congo (DRC), a region of the world that is still poorly investigated. Here, we propose to study the effect of human-induced fire on land use and land cover change in a protected area of the DRC, i.e. the Luki Biosphere Reserve (LBR). We investigate tree cover changes in and around the reserve between 2002 and 2019 using Landsat 7 ETM+, Landsat 8 OLI/TIRS and MODIS MCD12Q1 images and quantify human induced fires using MODIS MCD64A1 images. The study combines land use and land cover (LULC) change detection analysis of four images, two acquired in 2002 and two acquired in 2019, with multi-temporal assessment of annual burnt area acquired between 2002 and 2019 from MODIS MCD64A1 to assess the role of fire in LULC changes and the sensitivity of different LULC types to fire. The results show a dynamic conversion of primary forest to secondary forest over about 16% of the area, the evolution of savanna to secondary forest over 9.6% (Landsat image) and the replacement of secondary forest by savanna over 8.1% (MODIS image) of the total area of Luki Reserve. Of the total area undergoing land use change, 34.1% (Landsat image) and 35.7% (MODIS image) were caused by fire, which however did not cause a significant LULC change. For the LULC types that experienced fire events, the least stable type was primary forest, which had the lowest stability rate (34.2% and 23% for Landsat and MODIS image analysis, respectively) compared to others. This result illustrates the importance of fire as a driver of primary forest loss and degradation in the region. Despite the high exposure of savannas to fire events, they were not significantly destabilized by fire (stability rates of 86.3 and 97% for Landsat and MODIS analysis, respectively). Future analyses should focus on discriminating between different fire types to better understand the complex relationship between fire and ecosystem conditions.

KEYWORDS:

• Land use and land cover
• forest fire
• slash and burn agriculture
• remote sensing
• MODIS
• Landsat
• spatial correlation
• logistic regression model
• Luki Biosphere Reserve

Acknowledgments

We acknowledged the Ecole Régionale Post Universitaire d’Aménagement et de Gestion Intégrés des forêts et Territoires topicaux (ERAIFT) and the World Wild Fund for Nature for making this work possible through its logistical and technical support. We also thank INERA Luki through its head of Station Professor Tolérant LUBALEGA who welcomed us and guided us during our stay. We also thank all the project members who contributed to the data collection.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Author contributions

Conceptualization. N. C. Cirezi. J-F. Bastin and E. Tshibasu; Data processing. N. C. Cirezi; Formal analysis. N. C. Cirezi. G. B. Chuma. J-F. Bastin and Y. Mugumaarhahama; Funding acquisition. E.L. Tipi; Methodology. N. C. Cirezi. E. Tshibasu. G. B. Chuma. Y. Mugumaarhahama. K. Karume and J.-F. Bastin; E.LTipi; Supervision. J-F. Bastin. R. S. Lumbuenamo and J. Bogaert; Validation. K. R. Sambieni. K. Karume. R. S. Lumbuenamo. J-F. Bastin and J. Bogaert; Visualization. R. S. Lumbuenamo. J-F. Bastin and J. Bogaert; Writing – original draft. N. C. Cirezi and J.-F. Bastin; Writing – review and editing. E. Tshibasu. G. B. Chuma. E. L.Tipi. Y. Mugumaarhahama. K. R. Sambieni. K. Karume. J-F. Bastin.

Additional information

Funding

This research was funded by GIZ and the European Union through the project « Strengthening the Resilience to Climate Change of Local Communities in Luki and Mai-Ndombe in DRC ».