Case Study: ICIMOD achieves precise, high‑accuracy vegetation and land‑cover mapping (revealing 12% tree‑canopy loss) with Trimble eCognition

Out-of-the-box solution for precise vegetation and land cover mapping

The International Centre for Integrated Mountain Development (ICIMOD), working with Nepal’s Ministry of Agriculture and Cooperatives under the HIMALI program, faced the challenge of assessing vegetation and forest change in the remote, steep Lorpa watershed of Jumla. Physical surveys were impractical, and authorities needed accurate, time‑series land‑cover data to understand how climate pressure and local resource use were affecting tree cover and livelihoods.

ICIMOD combined 2006 QuickBird and 2011 Ikonos imagery with SRTM elevation data, ArcGIS indices and customized Trimble eCognition rule sets to classify 10 land‑cover classes and automatically delineate individual tree crowns. The workflows ran in about 30 minutes and achieved very high accuracy (99%/97% crown detection, 93% field validation), revealing a loss of 5,432 trees (12% canopy) between 2006 and 2011—insights now guiding targeted watershed and forestry management to improve agribusiness and resilience.

ICIMOD

Kabir Uddin

GIS and Remote Sensing Analyst