Forest change detection and mapping in Gatineau Park, Quėbec, 1987 to 2010 using Landsat imagery

Forest management seeks to balance sustainable economic yields with maintenance and enhancement of a diversity of ecosystem services. Forest change from natural and anthropogenic factors needs to be studied and monitored over time periods appropriate to the time scale of change, whether abrupt or gradual. This paper presents analysis of forest change in Gatineau Park, Quebec, by integrating ground-based measurements with Landsat imagery. Thirty-three 90 x 90m field plots were surveyed with respect to vegetation quantity (e.g., LAI, canopy openness, DBH, basal area), health and age during the 2010 growth season, and thirteen near-anniversary Landsat Thematic Mapper (TM) 5 scenes from 1987 to 2010 were relatively calibrated and assembled into an image time-series. Regression of the 2010 Landsat data and various vegetation indices against the field data showed that Tasseled Cap Wetness (TCW) was the most robust predictor and it differentiated between coniferous, mixed and deciduous forests. While several image variables detected abrupt forest change, TCW applied to the image time-series also detected subtle and gradual change using Mann-Kendall trend analysis (p<0.05). Landsat time-series and the change analysis techniques evaluated in this research are applicable to other large forested areas. The results of this study will aid the revision of the Gatineau Park Master Plan, which is currently underway.