Mapping of land cover in Northern California with simulated HyspIRI images

Abstract:

Land-cover maps are important science products needed for natural resource and ecosystem service management, biodiversity conservation planning, and assessing human-induced and natural drivers of land change. Most land-cover maps at regional to global scales are produced with remote sensing techniques applied to multispectral satellite imagery with 30-500 m pixel sizes (e.g., Landsat, MODIS). Hyperspectral, or imaging spectrometer, imagery measuring the visible to shortwave infrared regions (i.e., full range) of the spectrum have shown improved capacity to map plant species and coarser land-cover associations, yet techniques have not been widely tested at regional and greater spatial scales. The Hyperspectral Infrared Imager (HyspIRI) mission is a full-range hyperspectral and thermal satellite being considered for development by NASA (hyspiri.jpl.nasa.gov). A hyperspectral satellite, such as HyspIRI, will provide detailed spectral and temporal information at global scales that could greatly improve our ability to map land cover with greater class detail and spatial and temporal accuracy than possible with conventional multispectral satellites.

The broad goal of our research is to assess multi-temporal, HyspIRI-like satellite imagery for improved land cover mapping across a range of environmental and anthropogenic gradients in California. In this study, we mapped FAO Land Cover Classification System (LCCS) classes over 30,000 km² in Northern California using multi-temporal HyspIRI imagery simulated from the AVIRIS airborne sensor. The Random Forests classification was applied to predictor variables derived from the multi-temporal hyperspectral data and accuracies were compared to that from Landsat 8 OLI. Results indicate increased mapping accuracy using HyspIRI multi-temporal imagery, particularly in discriminating different forest life-form types, such as mixed conifer and broadleaf forests and open- and closed-canopy forests.