Using Satellite-based Ecosystem Functional Types to Map Biodiversity Patterns and Improve Ecosystem Service Models

Abstract:

Satellite-derived Ecosystem Functional Types (EFT) are a promising way to characterize the regional patterns of ecosystem functional diversity to provide standardized metrics for monitoring and modeling of biodiversity and ecosystem services to inform land-use and natural resource decision-making. Using EFT diversity rather than land use/land cover (LULC) metrics could improve modeling and understanding of diversity-driven ecosystem services, such as pollination, wildlife-based tourism, and carbon-cycling because it better captures the true heterogeneity in ecosystem functioning that exists across landscapes.

To examine this, we first characterized EFTs using both 250-m MODIS and 30-m Landsat data from 2001 to 2019 in Costa Rica. Then we examined EFT assemblages across different ecoregions, life zones, and floristic provinces to better understand how functional ecosystem classifications relate to compositional-based classifications like LULC. Subsequently, we separately used EFTs and LULC as the primary input to ecosystem service models (using InVEST, an open source software produced by The Natural Capital Project, Stanford University), and evaluated these outputs against observed data. Results show that the EFT classification represents observed patterns of diversity on the ground, and that EFT-based ecosystem service model outputs are more consistent with empirical observations of ecosystem services compared to LULC-based outputs. This study demonstrates that EFTs hold great potential for replacing LULC classification systems for ecosystem services, providing a more functional basis for variability in service provision across space and time.