Structural parameters are key indicators for the resilience and diversity of restored forest ecosystems. However, there is a lack of studies to illustrate the recovery of the structural parameters due to the absence of long-term monitoring after restoration. This study used Worldview-2 satellite data and developed an Evolutionary Algorithm-Neural Network Model to estimate coverage, biomass, and spatial structure in a semi-arid mine dump. The results show that the spectral and textural information from Worldview-2 image could effectively estimate the three structural parameters with determinant coefficients of 0.91, 0.86 and 0.62, respectively. With increase of restoration age, the structural parameters of the restored forest increased. After 23 years, the average coverage, biomass, and spatial structure reached 0.80, 27.5 kg/m² and 0.45, which were 266.6%, 245.5%, 300.0% higher than the reference sites. Among different reforestation patterns, the single forest of Pinus tabuliformis Carr. or Hippophae rhamnoides Linn. have highest coverage (0.98) and biomass (45.8 kg/m²) but the lowest spatial structure 0.22. The mixed forest (Populus L. and Pinus tabuliformis Carr.) has the highest coverage, biomass and spatial structure of 0.90, 34.9 kg/m² and 0.78. These results suggest that restoration interventions could effectively restore the forest structure in a semi-arid mine dump. However, there exist trade-offs among coverage, biomass, and spatial structure. The pattern of mixed forest is beneficial for forest regeneration. This study also demonstrates that the satellite imagery-based model and data have potential advantages for monitoring recovery and guiding the improvement of restored forest and possibly other restoration sites.
Audio/Video, Conference Presentation, SER2019
Pre-approved for CECs under SER's CERP program
Society for Ecological Restoration