Geography
Document Type
Article
Abstract
Tropical protected areas face pressures from smallholder agriculture and illegal mining, yet many monitoring approaches reduce these dynamics to net forest loss, obscuring whether change is alternating or persistent. In the Okapi Wildlife Reserve, Democratic Republic of the Congo, where the zoning plan allows only nonperennial crops in agriculture zones and forbids mining throughout the reserve, we analyzed annual Joint Research Centre–Tropical Moist Forest maps from 2000 to 2024 at 30 m spatial resolution. We applied Trajectory Analysis and DynamicPATCH across three study zones: agriculture, mining, and hunting–conservation. Trajectory Analysis classifies each pixel’s twenty-five-year non-forest time series into categories such as Gain without Alternation and Gain with Alternation, distinguishing persistent change from alternating change. Meanwhile, DynamicPATCH decomposes annual differences into eight patch-transition types. Trajectory Analysis summarizes each zone’s change into components of Quantity, Exchange, and Alternation, expressed as an annual percentage of each zone’s unified size to enable comparison among zones of various extents. The Quantity component most clearly distinguishes the zones as it accounts for 45% of the change in the agriculture zone, 98% of the change in the mining zone, and 27% of the change in the hunting–conservation zone. The agriculture study zone is characterized by many alternating trajectories and frequent patch transitions, but after 2022 agriculture evolves toward Gain without Alternation and more expansive transitions, suggesting a move from rotational mosaics toward more persistent non-forest, partly associated with prohibited perennial crops. The mining study zone shows consistent dominance of Gain without Alternation and Expanding or Merging transitions, reflecting progressive, and recently accelerating, expansion of illegal mining fronts. The hunting–conservation study zone remains largely stable, with only sparse and low-magnitude changes. This combined approach distinguishes land-use pattern signatures, shows that full annual series are essential for capturing patch dynamics that endpoint analyses miss, and provides actionable insight into where zoning rules are being followed, stretched, or broken. © 2026 by the authors.
Publication Title
Land
Publication Date
9-2026
Volume
15
Issue
9
ISSN
2073-445X
DOI
10.3390/land15091630
Keywords
Congo Basin, DynamicPATCH, illegal mining, patch dynamics, protected areas, smallholder agriculture, Trajectory Analysis, tropical deforestation
Repository Citation
Madibi, Pacifique Mubamba; Kibambe Lubamba, Jean-Paul; Zhang, Aiyin; and Pontius, Robert Gilmore, "GIS Methods to Distinguish Between Land Changes from Smallholder Agriculture and Mining in a Time Series for a Tropical Protected Area Forest" (2026). Geography. 1069.
https://commons.clarku.edu/faculty_geography/1069
Cross Post Location
Student Publications
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.
Copyright Conditions
Madibi, P. M., Kibambe Lubamba, J. P., Zhang, A., & Pontius Jr, R. G. (2026). GIS Methods to Distinguish Between Land Changes from Smallholder Agriculture and Mining in a Time Series for a Tropical Protected Area Forest. Land, 15(9), 1630.
